The Lancet Oncology Commission

Medical imaging and nuclear medicine: a Lancet Oncology

Commission
Hedvig Hricak*, May Abdel-Wahab*, Rifat Atun*, Miriam Mikhail Lette, Diana Paez, James A Brink, Lluís Donoso-Bach, Guy Frija, Monika Hierath,
Ola Holmberg, Pek-Lan Khong, Jason S Lewis, Geraldine McGinty, Wim J G Oyen, Lawrence N Shulman, Zachary J Ward, Andrew M Scott

The diagnosis and treatment of patients with cancer requires access to imaging to ensure accurate management Lancet Oncol 2021; 22: e136–72
decisions and optimal outcomes. Our global assessment of imaging and nuclear medicine resources identified Published Online
substantial shortages in equipment and workforce, particularly in low-income and middle-income countries (LMICs). March 4, 2021
https://doi.org/10.1016/
A microsimulation model of 11 cancers showed that the scale-up of imaging would avert 3·2% (2·46 million) of all S1470-2045(20)30751-8
76·0 million deaths caused by the modelled cancers worldwide between 2020 and 2030, saving 54·92 million life-
See Comment pages 422, 423,
years. A comprehensive scale-up of imaging, treatment, and care quality would avert 9·55 million (12·5%) of all 425, 426, 427, 429, and 430
cancer deaths caused by the modelled cancers worldwide, saving 232·30 million life-years. Scale-up of imaging would *Joint first authors
cost US$6·84 billion in 2020–30 but yield lifetime productivity gains of $1·23 trillion worldwide, a net return of Department of Radiology
$179·19 per $1 invested. Combining the scale-up of imaging, treatment, and quality of care would provide a net (Prof H Hricak MD) and
benefit of $2·66 trillion and a net return of $12·43 per $1 invested. With the use of a conservative approach regarding Department of Radiology and
human capital, the scale-up of imaging alone would provide a net benefit of $209·46 billion and net return of $31·61 Molecular Pharmacology
Programme (Prof J S Lewis PhD),
per $1 invested. With comprehensive scale-up, the worldwide net benefit using the human capital approach is Memorial Sloan Kettering
$340·42 billion and the return per dollar invested is $2·46. These improved health and economic outcomes hold true Cancer Center, New York, NY,
across all geographical regions. We propose actions and investments that would enhance access to imaging USA; Department of Radiology
equipment, workforce capacity, digital technology, radiopharma­ceuticals, and research and training programmes in (Prof H Hricak), Departments of
Pharmacology and Radiology
LMICs, to produce massive health and economic benefits and reduce the burden of cancer globally. (Prof J S Lewis), and
Departments of Radiology and
Introduction imaging, is increasingly integral to cancer diagnostics and Population Science
(G McGinty MD), Weill Cornell
The global cancer burden is increasing at an alarming treatment. Although the direct effect of imaging on overall
Medical College, New York, NY,
rate. From 2012 to 2018, the estimated number of new survival is difficult to quantify because of the complexity of USA; International Atomic
cancer cases worldwide grew by more than 28%, from cancer biology and cancer care, and with there being a Energy Agency, Division of
14·1 to 18·1 million, and the estimated number of cancer paucity of data on the subject, many studies have shown Human Health, Vienna, Austria
(Prof M Abdel-Wahab MD,
deaths rose by approximately 17%, from 8·2 to 9·6 million.1,2 that the appropriate use of imaging for indications such as
M M Lette MD, D Paez MD);
By 2030, the number of new cancer cases worldwide is cancer staging or the assessment of treatment response Radiation Oncology, National
expected to reach 22·2 million and cancer deaths to reach can improve management decisions and reduce the costs Cancer Institute, Cairo
13·2 million.3,4 These statistics are all the more concerning of cancer care (eg, by obviating the need for other tests or University, Cairo, Egypt
(Prof M Abdel-Wahab);
because approximately 80% of disability-adjusted life- invasive diagnostic procedures, indicating the need for
Graduate School of Biomedical
years are lost to cancer in low-income and middle-income neoadjuvant therapy, improving surgical or radiotherapy and Health Sciences, Hiroshima
countries (LMICs), where only approximately 5% of the planning, preventing unnecessary surgery, and discon­ University, Hiroshima, Japan
global funding for cancer control and care are applied.3,5 tinuing ineffective treatments).8–16 (Prof M Abdel-Wahab);
Department of Global Health
In 2015, The Lancet Oncology published the results of Despite the ubiquity of imaging in modern cancer care
and Population
two Commissions that assessed the gaps in access to in high-income countries, the importance of imaging in (Prof R Atun FRCP) and Center
cancer surgery and radiotherapy, and proposed actions to oncology is frequently overlooked in efforts aimed at for Health Decision Science
address the growing burden of cancer in LMICs.6,7 The improving cancer care in LMICs. Many LMICs have (Z J Ward MPH), Harvard TH
Chan School of Public Health,
Commission reports provided specific recommendations severe shortages of imaging and nuclear medicine Boston, MA, USA; Department
for increasing access to these treatment modalities, and equipment and personnel. Data on the amount of of Global Health and Social
showed that doing so could prevent avoidable human imaging equipment available in LMICs have not been Medicine (Prof R Atun) and
suffering and reduce preventable deaths, and at the same gathered systematically. There are scant data on the Department of Radiology,
Massachusetts General
time also provide substantial economic benefits. Both numbers and distribution of health professionals Hospital (Prof J A Brink MD),
reports noted that cancer care is a multidisciplinary involved in providing imaging services—including Harvard Medical School,
endeavour and that the effective use of surgery and radiologists and nuclear medicine physicians, imaging Harvard University, Boston,
radiotherapy requires, among other resources, medical radiographers and techno­logists, medical physicists, and MA, USA; Department of
Medical Imaging, Hospital
imaging. radiochemists, among others. There are few reliable Clínic of Barcelona, University
In high-income countries, imaging plays an essential studies that quantify the number and combination of of Barcelona, Barcelona, Spain
role in the management of almost all cancer types. This different types of health professionals needed to operate, (L Donoso-Bach MD); Université
medical technique is used throughout the care continuum, optimally use, and maintain imaging equipment.17 de Paris, Paris, France
(Prof G Frija MD); European
from detection, diagnosis, and staging, to treatment Furthermore, even in high-income countries with ready Society of Radiology, Vienna,
planning (especially in radiation oncology), the assessment access to imaging services, there is little appreciation for Austria (M Hierath Mag Phil);
of treatment response, and in long-term follow-up. the importance of specialised training and expertise to Radiation Protection of
Moreover, interventional radiology, which relies on the optimal interpretation and reporting of cancer Patients Unit, International

www.thelancet.com/oncology Vol 22 April 2021 e136

 The Lancet Oncology Commission

Atomic Energy Agency, Vienna, imaging.17 Without data on these crucial elements, it is expansion of access to imaging for cancer, and calls for
Austria (O Holmberg PhD); not possible to appropriately plan the introduction and action toward this goal.
Department of Diagnostic
Radiology, University of
scale-up of cancer services whose effectiveness depends
Hong Kong, Hong Kong Special on effective and efficient imaging and nuclear medicine Section 1: the evolving role of cancer imaging in
Administrative Region, China services. LMICs—opportunities and obstacles
(Prof P-L Khong FRCR); At the suggestion and with the help of the International As already described, the global cancer burden is
American College of Radiology,
Reston, VA, USA (G McGinty);
Atomic Energy Agency (IAEA), The Lancet Oncology increasing rapidly—particularly in LMICs, where funding
Department of Biomedical Commission on Medical Imaging and Nuclear Medicine for cancer care is scarce and the capacity to manage this
Sciences and Humanitas was established in 2018, with the charge of examining rising burden is low.18,19 As a result, huge inequities exist
Clinical and Research Centre, global access to imaging and nuclear medicine for cancer between countries in their access to effective services for
Department of Nuclear
Medicine, Humanitas
care. This endeavour was also charged with analysing the cancer care. In addition to intercountry inequities, large
University, Milan, Italy barriers to access to imaging for cancer care, providing inequities also exist within countries, with lower amounts
(Prof W J G Oyen PhD); new evidence to show the benefits of imaging in of access for those with a lower income and lower
Department of Radiology and improving cancer care and cancer survival, and providing education compared with those with a higher income
Nuclear Medicine, Rijnstate
Hospital, Arnhem, Netherlands
recommendations on how best to introduce and scale up and higher education. Such intracountry inequities
(Prof W J G Oyen); Department imaging services to expand access to imaging and persist both in wealthy nations such as the USA and in
of Radiology and Nuclear nuclear medicine services in LMICs. To produce this LMICs, where any available highly trained personnel
Medicine, Radboud University
Commission, the health benefits of cancer imaging were and advanced health-care infrastructure—including
Medical Centre, Nijmegen,
Netherlands (Prof W J G Oyen); analysed on a global level, with the use of data from high- imaging equip­ment—might be confined largely to private
Department of Medicine, income countries and LMICs. The financial return on practices.17,20,21 These inequities in access to cancer services
Abramson Cancer Center, investment in cancer imaging was also investigated. are reflected in inequities in health outcomes. Although
University of Pennsylvania,
Finally, given the vast imbalances in cancer burden and worldwide the overall survival rates for cancer are
Philadelphia, PA, USA
(Prof L N Shulman MD); Tumour cancer control resources between LMICs and high- improving, the improve­ ment is much less evident in
Targeting Laboratory, income countries, recommendations for scaling up LMICs.17–19 Even though the incidence of cancer in LMICs
Olivia Newton-John Cancer cancer imaging resources were produced, with a specific is lower than that in high-income countries, cancer-
Research Institute, Melbourne,
focus on LMICs. related mortality rates are significantly higher in LMICs,
VIC, Australia
(Prof A M Scott MD); This Commission is organised into eight sections. especially in people aged younger than 65 years. These
Department of Molecular Section 1 discusses the evolving role of cancer imaging circumstances are at least partly due to delays in diagnosis
Imaging and Therapy, Austin in LMICs and the main challenges that resource-poor (affected by poor access to imaging and other diagnostic
Health, Melbourne, VIC,
countries should consider when tailoring the adoption tools), inadequate access to optimal local and systemic
Australia (Prof A M Scott);
School of Cancer Medicine, and use of imaging and nuclear medicine services to treatments, and greater numbers of infection-associated
La Trobe University, the continuum of cancer care resources available to cancers in LMICs.22,23
Melbourne, VIC, Australia them. Section 2 expands on the barriers to increasing It is important to recognise that cancer care is a
(Prof A M Scott); Department of
access to cancer imaging in LMICs, presenting new continuum and requires parallel investments in imaging
Medicine, University of
Melbourne, Melbourne, VIC, data on the global availability of imaging technologies and other diagnostics, as well as in treatments. The
Australia (Prof A M Scott) and human resources and identifying specific gaps that socioeconomic benefits of investments in improve­
Correspondence to: need to be addressed. Section 3 presents an analysis of ments to cancer surgery7 and radiotherapy6 infrastructure
Prof Hedvig Hricak, Department the costs, benefits, and returns on investment that have been shown, and cancer imaging is required for
of Radiology, Memorial Sloan
could be achieved by investing in the global scale-up of diagnosis, staging, and effective treatment with either
Kettering Cancer Center,
New York, NY 10065, USA imaging technologies and human resource capabilities, surgery or radiotherapy. For example, patients
hricakh@mskcc.org alone or in tandem with the improved availability of undergoing radiotherapy require imaging for treatment
treatment modalities, quality of care, or both. Section 4 planning, and quantitative imaging affects radiotherapy
discusses financing for a global scale-up of imaging outcomes and survival.24–26 Similarly, preoperative
diagnostics. Section 5 discusses the important issue of imaging bolsters the safety, appropriateness, quality, and
ensuring radiation protection and safety for patients, effectiveness of cancer surgery. Furthermore, the use of
workers, and the public, as well as quality systems imaging to guide biopsies and minimally invasive
when scaling up imaging and nuclear medicine interventions (eg, image-guided placement of central
capabilities globally. Section 6 provides an overview of venous catheters for the administration of medicines, or
innovations in digital science technologies and novel image-guided tumour ablations) is associated with
analytical tools, such as artificial intelligence and improved quality, decreased morbidity, and enhanced
machine learning, which will transform the availability affordability of these procedures27–31 Moreover, the
of and access to imaging diagnostics and aid decision absence of staging infor­mation from imaging can lead to
making. Section 7 outlines the crucial impor­tance of the inadequate or inappropriate use of medical therapies,
teaching, training, and research, to ensuring the surgery, or radiotherapy, and consequently increase
adequate capabilities and quality of imaging sites and morbidity and mortality. Selection of the most appropriate
staff in LMICs. Section 8, the conclusion, discusses the antineoplastic regimen for patients with cancer often
success factors necessary to enabling the global relies upon imaging results.32

e137 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Use of cancer imaging and its benefits: a review of the relies on access to vaccines for common infections that
literature can lead to cancer (eg, human papillomavirus and
Although imaging plays pivotal roles in cancer care, hepatitis). Additionally, the successful delivery of cancer
because of the complexity of the care process, the direct care requires the coordination of the overall health
effects of imaging on patient outcomes have historically system, including public and private health care facilities.
been difficult to quantify. Nevertheless, we reviewed the Education of the public is necessary to promote cancer
(albeit scarce) published peer-reviewed literature and awareness and encourage them to seek care. Furthermore,
reports aimed at quantifying, on a large scale, the use of the families and careers of those affected by cancer also
imaging, and its benefits, for patients with cancer. require support. Although each of these needs demands
One study from Canada, based on a survey of centres focused attention, the process of cancer control should be
providing imaging services, examined the amount of use viewed holistically and as consisting of a dynamic,
of and the reasons for imaging; the study found that interlinked, and interdependent chain of activities, where
approximately 23·1% of CT examinations, 80·2% of weak links might cause a breakdown in the system of
PET-CT examinations, and 20·8% of MRI examinations care, and in which the links should be aligned with each
were done for cancer indications.33 However, the survey other to provide value.
relied on subjective assessments of the distribution of The shortage of a well-trained health workforce and the
indications rather than a direct analysis of administrative poor availability of health technologies in LMICs require
data, and the response rate regarding this issue was low.33 the adoption of suitable approaches to diagnostics,
Although CT scans are used to image a broad spectrum including disease staging and management during
of conditions, a report for the UK National Health Service treatment, which differ from those used in high-income
suggests that approximately 95% of the CT scanners in countries. Cancer control and care in LMICs will be
the UK National Health Service are used for cancer improved by the adoption of novel approaches to the
staging in addition to their use for non-cancer indications, management of the disease, implemented by way of the
though it does not provide details into the proportion of progressive expansion of human resources, health
CT examinations done for oncological purposes.34,35 A technologies, and health care services for prevention,
study of imaging studies in the USA that used data from diagnosis, treatment, and palliative care. For example, in
the Centers for Medicare & Medicaid Services found that LMICs, women with locally advanced breast cancer
9·5% of all advanced imaging studies (ie, CT, MRI, and might undergo a staging work-up for metastatic disease,
PET studies) were done in patients with cancer.35 which includes a chest x-ray and liver ultrasonography,
Imaging tests are included in oncology clinical practice but not CT, single photon emission computed tomog­
guidelines by every major professional group, as well as raphy (SPECT), or PET-CT, which would typically be used
the US National Comprehensive Cancer Network and the in high-income countries. Although an adapted approach
UK National Institute for Health and Care Excellence; in LMICs will miss metastatic disease in some patients
and evidence-based studies being used for the justification whose disease might have been detected with more
of reimbursement decisions for imaging examinations in advanced technologies, this systematic approach will
patients with cancer show the effect of such imaging nonetheless benefit many patients. If the initiation of the
studies in clinical practice. Data from large prospective evaluation and treatment of patients was delayed until
examinations have shown how imaging can assist in more advanced imaging (and potential treatment
management decisions; for example, the US National options) were available, it would mean that in the
Oncologic PET Registry has collected data for more than interval, which might be many years, patients would go
300 000 patients since 2006, and has shown that the use without any treatment at all.
of PET leads to substantial changes in the clinical Matching the imaging technologies with the treatments
management of 30% of patients across various cancer available in LMICs is crucial. This optimisation process
types.36,37 Our literature review did not find any relevant should be done in a systematic and evidence-informed
large-scale studies from LMICs. way for a multitude of cancer types, considering
diagnostics (including pathology and imaging), surgery,
Strengthening cancer care in LMICs: the need for a systemic therapy, and radiotherapy. The specifics for each
systems approach of the imaging and treatment modalities used will differ
Cancer control and care is complex and requires for each cancer. Investment in cancer detection and
multidisciplinary teams for a successful delivery. The control should also take into account the complexity of the
pathway encompasses prevention, screening, diagnostics health-care system and ensure equitable patient access.22
(including imaging, pathology, and laboratory services), Furthermore, over time, technology improvements and
treatments (including surgery, radiotherapy, and systemic evidence-based cost–benefit assessments of imaging and
therapies), survivorship, palliative care, and end-of-life treatment modalities will result in changes in imaging
care. A good cancer programme would ideally include recom­mendations for different cancers, depending on the
services to support all these areas at the appropriate times stage of presentation. Moreover, changes in the patterns
during the patient’s journey. Optimal cancer control also of cancer incidence and presentation that are likely to

www.thelancet.com/oncology Vol 22 April 2021 e138

 The Lancet Oncology Commission

result from economic development, because of factors and facilitate the development of strategic recommen­
such as environmental exposures, lifestyle changes, and dations for the expansion and use of cancer imaging at a
ageing populations, as well as greater access to affordable global level.
screening and diagnostic services, will require the further The need for the maintenance of imaging equipment
adaptation of cancer services.38,39 should also be taken into account when planning and
When decisions are being made about which imaging budgeting for improvements in cancer imaging services.
modalities to adopt, it is also necessary to consider the For example, in settings where there might be only one
overall resources available in a country to purchase, or two CT scanners, having one scanner out of service for
install, operate, maintain, and—when needed—repair an extended period of time will have a substantial clinical
the imaging equipment. In practice, governments effect, but equipment vendors might not have in-country
allocate a proportion of their budgets to health, which is service personnel, and it can be months before
then apportioned to different areas of need, including technicians can attend to machines at some sites. The
for maternal and child health, communicable diseases, cost of repairs and maintenance can be especially
non-communicable diseases, and injuries.23 Some of the expensive in LMICs, leading to delays in service and
funds are typically allocated to cancer control and care for prolonged down-time of equipment. Many LMICs have
capital expenditures (for infrastructural needs, including facilities with non-functioning imaging equipment
clinical space and capital outlays for radiology and (along with non-functioning pathology processors, linear
nuclear medicine equipment, pathology laboratories, accelerators, etc). Unstable power grids that lead to
and operating rooms with necessary equipment) and regular interruptions in the supply of electricity, among
operational expenditures for the salaries of health-care other factors, compound this issue. Loss of electrical
providers (eg, physicians, nurses, technologists, pharma­ power and power surges are common in many locations
cists, and community health workers, as well as trained in LMICs, in both urban and rural regions.
oncology providers and appropriately trained staff in A further challenge in LMICs is the absence of a
radiation units who are needed to safely and effectively reliable supply chain for imaging diagnostics, such as
operate them, including, for example, physicists and contrast agents and radiopharmaceuticals. Gaps in the
dosimetrists). Appropriate medicines (including chemo­ availability of crucial reagents are frequent and affect the
therapy and biological therapies), technologies (eg, for functional status of the imaging modalities that depend
radiotherapy), and diagnostics (including imaging and on them. Quality management systems are essential to
pathology) should be available to balance diagnostic ensure imaging is done in a safe and effective manner. In
capabilities with subsequent treatment options. The addition to imaging equipment, the availability of a
proportion of the funds allocated to cancer care will vary workforce appropriately trained to do imaging studies is
across and within countries depending on priorities and a notable challenge in providing timely and equitable
the different levels of services available. For example, access to imaging for cancer. At present, in some LMICs,
urban centres might have a higher level of care and more clinicians might be able to get their patients scanned in a
resources available than rural settings.17 In each setting, timely manner, but a paucity of radiologists might delay
however, all aspects of care resources should be scan reporting to a degree that affects patient care.
coordinated and appropriated to ensure effective and To help address the multitude of challenges faced by
efficient budgeting. LMICs in relation to cancer imaging, comprehensive,
When allocating scarce resources, the management global mapping of medical imaging and nuclear medicine
challenges posed by the constraints of imaging capacity resources is needed to identify existing gaps and inform
should also be considered. For instance, in some settings, strategies to mitigate them. In addition, given the
only one or two CT scanners might serve large popu­ contextual differences in cancer burden and funding
lations, not just patients with cancer but also those with availability across LMICs, as well as technical and human
other conditions (eg, trauma or infection); consequently, resource capacity, to enable strategic planning for optimal
wait times for scanning might be long, reducing the cancer care in LMICs, there is a need for evidence on how
availability of CT scans for patients with cancer. For investments in the expansion of imaging could yield clear
example, if a patient with diffuse large B-cell lymphoma improvements in patient outcomes in different countries
with extensive mediastinal involvement has to wait and health systems. These gaps and needs are addressed
6 weeks for an initial staging CT, clinicians might need to in more detail, and by the provision and analysis of new
begin treatment without the aid of the CT, which might data, in the next two sections of this report.
then not be done at all. In this context, knowledge of the
appropriate number of imaging units required per Section 2: overcoming barriers to access and
million people in a population to effectively manage mapping gaps in imaging and nuclear medicine
cancer diagnosis and treatment is necessary to allow resources to facilitate a progressive expansion
resource planning at a country level. More data on the of cancer care
use of imaging and equipment in high-income countries Greater guidance is needed to progressively expand
and LMICs would clearly assist with identifying gaps access in LMICs to cost-effective, affordable technologies,

e139 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

which include diagnostic imaging and nuclear medicine,

Imaging modality
required to address the rising burden of cancer in these
countries. WHO Health Care Level 1 Level 1 does not have adequate equipment or facilities to undertake cancer
(primary health care) care; it might have a triage role to the next level up
Applying this framework to the contemporary example
WHO Health Care Level 2 Radiography with fluoroscopy
of radiotherapy, The Lancet Oncology Commission on (secondary health care) Doppler ultrasonography
expanding global access to radiotherapy6 showed that the Mammography
cost of upscaling radiotherapy from 2015 to 2035 across Angiography
CT
all LMICs is matched by “compelling evidence that Radionuclide scintigraphy, including SPECT–CT
investment in radiotherapy not only enables treatment of WHO Health Care Level 3 Magnetic resonance imaging
large numbers of cancer cases to save lives, but also brings (tertiary health care) Positron emission tomography–CT
positive economic benefits.” Similarly, The Lancet Oncology Theranostics
Commission on sustainable care for children with cancer
The Commission recommendation comes from a consensus development process that involved discussion at Lancet
has shown substantial health and economic benefits of Oncology Commission meetings, where input from imaging experts into this topic was obtained. The differences in the
scaling up high-quality cancer services and treatment for recommendations for each WHO Health Care Level44 for imaging equipment are as follows: (1) this Commission
childhood cancers.40 The study estimated net benefits of suggests explicitly that Health Care Level 1 should not be where cancer care should be done, because the full range of
imaging equipment available at this level (including CT scans as a minimum) is not adequate for appropriate diagnosis
US$2 trillion, with an average investment of $30 billion and staging, and probably cannot provide the medical expertise or services required for complete cancer care; (2) this
each year in LMICs over a 30-year period (2020–50). Both Commission recommends the inclusion of SPECT–CT (rather than SPECT alone) in Health Care Level 2, because the use
Commissions were able to show a clear investment case, of these modalities is now standard at this level; and (3) this Commission recommends the inclusion of theranostics in
Health Care Level 3, as this procedure replaces radioimmunoscintigraphy. SPECT=single photon emission CT.
with estimated returns of up to $6 for radiotherapy and $3
for childhood cancers for every dollar invested. Table 1: Imaging technologies recommended by this Commission for cancer care facilities, adapted for
Just a few decades ago, the possibility of extending the WHO Health Care Levels44
benefits of technologies such as radiotherapy to those
without access was deemed unachievable. Since then,
many LMICs have made notable progress in primary care, (2) insufficient human resources, (3) inadequate
enabling them to begin integrating such technologies into government funding for cancer care and health systems
their health-care systems. For example, the WHO Global in general, (4) few reliable data about the availability of
Action Plan for the prevention and control of non- equipment and skilled human resources needed for
communicable diseases 2013–20 includes radiotherapy imaging, (5) a paucity of studies that quantify patient
for cervical cancer and colorectal cancer.41 Improvements imaging needs (for both cancer and non-cancer
in economic evaluation methods, applied as part of health indications), (6) the absence of evidence-based guidance
technology assessment programmes, have enabled more on investments in imaging required to achieve optimal
effective and transparent priority setting by heath care patient management, (7) inadequate and insufficient
systems and paved the way for the inclusion of new health programmes for training personnel for cancer imaging,
technologies in Universal Health Coverage (UHC).42 (8) the dearth of a procurement process that is evidence-
In the gradual development of cancer imaging capacity based and step-wise, to enable the selection of the most
in LMICs, modalities including ultrasound, conventional appropriate equipment (including appropriate technical
x-ray, CT, and mammography should be given priority specifications and requirements for the maintenance
because of their role in the initial assessment of patients, and repair for the amount of services and training
as well as their effect on patient management throughout available), (9) insufficient expertise in architectural
the disease course.43 In view of the complex nature of planning for medical imaging and nuclear medicine
cancer management for some patient groups, the type of (including radiation safety), (10) inadequate systems for
imaging equipment that should be installed and appropriate patient referral and follow-up, (11) insufficient
operational at health-care facilities should be based requisite clinical resources (eg, laboratories, resources
primarily on established, prioritised recommen­dations for pathology, and supplies of consumables such as
by WHO.44 Our Commission’s composite recom­ syringes, personal protective equipment, biopsy devices,
mendations for new imaging technologies are intended catheters, contrast media, local anaesthetic, and other
to complement and support these (table 1).44 Our aim is medicines, such as radiopharmaceuticals), and (12) poor
to promote the effective and efficient delivery of multi­ provision of safe waste disposal (including biohazards
disciplinary cancer care, with resources implemented and radio­pharma­­ceuticals).45 The barriers for the
and progressively provided in a strategic manner. This implementation of imaging equipment at appropriate
approach might be challenging in LMICs with less levels of access, as well as in the provision of adequate
funding for health care, but this framework bolsters the workforce, training, and education, are similar across
capacity of countries to develop facilities in an informed, LMICs, although differences will always exist between
contem­porary, and sustainable manner. countries.
The barriers restricting access to imaging and nuclear Furthermore, the compatibility of equipment with
medicine for cancer in LMICs, many of which were local realities, such as the availability and reliability of
mentioned earlier, include: (1) not enough equipment, electricity and clean water, optimal lighting in image

www.thelancet.com/oncology Vol 22 April 2021 e140

 The Lancet Oncology Commission

interpre­tation and procedural areas, sustainable infra­

Panel 1: Data collection for IMAGINE structure (including temperature control, or equipment
The International Atomic Energy Agency (IAEA) medical imaging and nuclear medicine that functions durably without it), and digital linkages
(IMAGINE) global resources database51 was launched in 2019, and is being continuously to patient information, are issues that need to be
updated. A total of 1857 datapoints in the profiles of 211 countries, territories, and overcome to ensure access to effective and reliable
principalities have been collected, with the dominant sources depicted in figure 1. cancer imaging services.46,47 To safeguard sustainability,
it is also essential to guarantee adequate maintenance
Primary sources for the IMAGINE database were as follows: coverage, including service contracts, warranties, the
• The IAEA (from IAEA staff and experts; reports of national, regional, and interregional availability of spare parts, and an understanding of
meetings; fact-finding missions; countries’ authorities and counterparts to IAEA anticipated software updates.
projects) and UN partner organisations and agencies such as WHO, WHO regional Furthermore, relevant patient-centred processes should
offices, the International Agency for Research on Cancer, the United Nations Scientific include an assessment of patient satisfaction, adequate
Committee on the Effects of Atomic Radiation (UNSCEAR), the UN Development communication pathways (including patient access to
Programme, the World Bank, and the ministries of health of some countries, Eurostat, telephone services), and available transportation to
and the Organisation for Economic Cooperation and Development (OECD) facilities for the entire target population. Additionally,
• National, regional, and global professional organisations and societies for medical health campaigns and community engagement can
imaging and nuclear medicine, such as the Arab Society of Nuclear Medicine; the Asia increase awareness of the target patient population
Oceania Federation of Nuclear Medicine and Biology; the Asociación Latinoamericana regarding cancer care, including the role of medical
de Sociedades de Biología y Medicina Nuclear; the European Trade Association imaging.
representing the medical imaging radiotherapy, health information and Another essential requirement is to ensure the
communication technologies, and electromedical industries (COCIR); the European availability not just of affordable imaging, but also of
Association of Nuclear Medicine; the European Society of Radiology; Global Diagnostic affordable treatment after a cancer is diagnosed. In
Imaging; the Healthcare Information Technology and Radiation Therapy Trade some LMICs, current and projected estimates of patient
Organisation; the International Organisation for Medical Physics; the International resources (including the national UHC strategy) are
Society of Radiographers and Radiation Technologists; the International Society of necessary, taking into consideration financial toxicity
Radiology; RAD-AID International; the Society of Nuclear Medicine and Molecular for individuals marginalised by the overall cost of cancer
Imaging; and the World Federation of Nuclear Medicine and Biology care.48–50
• A comprehensive review of published studies and reports on medical imaging and
nuclear medicine resources in different countries, particularly from WHO, UNSCEAR, Identifying the global gaps in the availability of
OECD, and Eurostat imaging diagnostics and human resources
• A survey of individual experts to address gaps in data, including ministry of health To address the data gaps identified as part of
representatives and radiation authority experts in countries who work with the IAEA The Lancet Oncology Commission on Medical Imaging
and agreed to share data on equipment and human resources for their respective and Nuclear Medicine, we collected new data to compre­
countries hensively analyse and map the availability of medical
imaging and nuclear medicine resources globally. The
survey and analysis were led by the IAEA. The data were
IAEA 62·09% used to construct a new database, the IAEA medical
WHO 9·48% imaging and nuclear medicine (IMAGINE) global
Professional societies or non-state actors* 4·32% resources database.51 The sources of data for the IMAGINE
Scientific publications 4·16% database are included in panel 1 and summarised in
COCIR† 4·00%
figure 1; sources for, and access to, the database are also
Organisation for Economic Cooperation
2·95% discussed further in the appendix (p 1).51 IMAGINE
and Development
data were stratified into high-income, upper-middle-
Pan-American Health Organization 2·74%
Eurostat
income, lower-middle-income, and low-income countries,
2·68%
Official bodies and authorities 2·42% according to World Bank country income classifications.6
Non-IAEA scientists and experts 1·79% Data on the amount of available CT, PET, mammo­
ISR 1·63% graphy, MRI, and SPECT equipment at a country level
Congresses or conferences or both 1·21% and by the income stratification of countries are shown
Health and country articles in the press 0·37% in figures 2–6, and more detailed interactive information
Market reports 0·16% is available on the IAEA IMAGINE database website.51
Information about the numbers of x-ray and ultrasound
Percentage of data in IMAGINE database provided by each source
equipment per country could not be accurately assessed
Figure 1: Major data sources for the IMAGINE database because of the absence of available data from the broad
IAEA=International Atomic Energy Agency. IMAGINE=IAEA medical imaging and nuclear medicine global resources range of health-care facilities, including small health
database. ISR=International Society of Radiology. *COCIR and ISR have been considered separately from the
clinics, where this equipment can be installed.
professional societies or non-state actors category, because each association independently contributed more than
1% of all data in IMAGINE. †COCIR is the European trade association of medical imaging, radiotherapy, health The survey results display a substantial difference in
information technology, and electromedical industries. the numbers of scanners per million people in the

e141 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Number of CT scanners per million inhabitants

>30·0
20·0–30·0
15·0–19·9
10·0–14·9
5·0–9·9
0·0–4·9
No CT scanners
Data not available

Figure 2: Estimates of the number of CT scanners per million inhabitants

Data are from the International Atomic Energy Agency medical imaging and nuclear medicine global resources database (IMAGINE). The map was produced by the International Atomic Energy Agency
(Vienna, Austria) and is included here with permission.

Number of PET scanners per million inhabitants

>3·0
2·0–3·0
1·0–1·9
0·0–0·9
No PET or PET–CT scanners
Data not available

Figure 3: Estimates of the number of PET scanners per million inhabitants

Data are from the International Atomic Energy Agency medical imaging and nuclear medicine global resources database (IMAGINE). The map was produced by the International Atomic Energy Agency
(Vienna, Austria) and is included here with permission.

www.thelancet.com/oncology Vol 22 April 2021 e142

 The Lancet Oncology Commission

Number of mammography units per million

inhabitants
>15·0
10·0–15·0
5·0–9·9
0·0–4·9
No mammography
Data not available

Figure 4: Estimates of the number of mammography units per million inhabitants

Data are from the International Atomic Energy Agency medical imaging and nuclear medicine global resources database (IMAGINE). The map was produced by the International Atomic Energy Agency
(Vienna, Austria) and is included here with permission.

Number of MRI units per million inhabitants

>30·0
15·0–30·0
10·0–14·9
7·5–9·9
5·0–7·4
2·5–4·9
0·0–2·4
No MRI units
Data not available

Figure 5: Estimates of the number of MRI units per million inhabitants

Data are from the International Atomic Energy Agency medical imaging and nuclear medicine global resources database (IMAGINE). The map was produced by the International Atomic Energy Agency
(Vienna, Austria) and is included here with permission.

e143 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Number of SPECT units per million inhabitants

>10·0
7·5–10·0
5·0–7·4
2·5–4·9
0·0–2·4
No SPECT or SPECT–CT scanners
Data not available

Figure 6: Estimates of the number of SPECT units per million inhabitants

Data are from the International Atomic Energy Agency medical imaging and nuclear medicine global resources database (IMAGINE). The map was produced by the International Atomic Energy Agency
(Vienna, Austria) and is included here with permission. SPECT=single photon emission CT.

population between high-income countries and LMICs

CT MRI SPECT PET
(table 2).51 For example, the mean number of people
served by one CT scanner in high-income countries is High-income countries
25 000; in upper-middle-income countries, 79 000; in Range 6·3–42·3 0·0–34·3 0·0–20·5 0·0–4·3
lower-middle-income countries, 227  000; and in low- Mean (SD) 38·8 (16·0) 27·3 (10·4) 18·2 (7·5) 3·6 (3·4)
income countries, 1 694 000.51 Although no formal Median (IQR) 20·5 (14·4–32·7) 12·6 (8·5–19·2) 5·4 (2·4–9·7) 1·2 (0·6–2·5)
recommendations for numbers of scanners per million Upper-middle-income countries
population exist, the information obtained from the Range 0·0–29·8 0·0–16·0 0·0–5·2 0·0–0·7
IMAGINE database (table 2) can be used to obtain Mean (SD) 12·1 (10·1) 5·4 (4·8) 1·6 (1·8) 0·3 (0·5)
estimates of the amount of installed imaging equipment Median (IQR) 7·8 (4·8–16·2) 3·4 (1·3–7·2) 0·9 (0·0–2·5) 0·2 (0·0–0·4)
to provide a range by different country income groups, Lower-middle-income countries
enabling the projection of requirements in different Range 0·0–7·8 0·0–3·3 0·0–0·9 0·0–0·2
settings. Additionally, evidence-based tools such as a Mean (SD) 4·3 (3·2) 1·1 (1·2) 0·3 (0·3) 0·2 (0·3)
health technology assess­ ment can enable nations to Median (IQR) 1·4 (0·9–3·9) 0·4 (0·1–1·4) 0·1 (0·0–0·4) 0·0 (0·0–0·1)
rationally set their own benchmarks. One relevant Low-income countries
example of a country using a health technology Range 0·0–1·1 0·0–0·3 0·0–0·1 0·0–0·0
assessment is the Framework for the Development of Mean (SD) 0·7 (0·8) 0·2 (0·5) 0·1 (0·1) 0·0 (0·0)
PET Services in England.52 Nations might adopt and Median (IQR) 0·4 (0·2–0·9) 0·1 (0·0–0·2) 0·0 (0·0–0·0) 0·0 (0·0–0·0)
adapt such pre-existing templates from other nations to
The data source is the International Atomic Energy Agency medical imaging and nuclear medicine global resources
set benchmarks for themselves, in support of rational,
(IMAGINE) database.⁵¹ SPECT=single photon emission CT.
achievable planning.
As with the availability and coverage of imaging Table 2: Number of different types of scanners per million inhabitants by country income group
equipment, little information exists at a global level
about the numbers of radiologists and nuclear medicine
physicians in different countries. The IMAGINE in upper-middle-income and high-income countries For the IMAGINE global
database revealed susbantial differences in the numbers (table 3).51 Although in some countries nuclear medicine resources database see
https://humanhealth.iaea.org/
of trained radiologists and nuclear medicine physicians scans are read by radiologists, the survey data suggest HHW/DBStatistics/IMAGINE.html
between countries (figures 7, 8), with substantially fewer that the use of nuclear medicine scans is less in countries
See Online for appendix
trained professionals in low-income countries than where lower access to radiopharmaceuticals and trained

www.thelancet.com/oncology Vol 22 April 2021 e144

 The Lancet Oncology Commission

Number of radiologists per million inhabitants

>100·0
50·0–100·0
25·0–49·9
10·0–24·9
0·0–9·9
Data not available

Figure 7: Estimated number of radiologists per million inhabitants

Data are from the International Atomic Energy Agency medical imaging and nuclear medicine global resources database (IMAGINE). The map was produced by the International Atomic Energy Agency
(Vienna, Austria) and is included here with permission.

Number of nuclear medicine physicians per

million inhabitants
>10·0
5·1–10·0
0·1–5·0
No nuclear medicine physicians
Data not available

Figure 8: Estimated number of nuclear medicine physicians per million inhabitants

Data are from the International Atomic Energy Agency medical imaging and nuclear medicine global resources database (IMAGINE). The map was produced by the International Atomic Energy Agency
(Vienna, Austria) and is included here with permission.

e145 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

professionals are additional confounding factors in

Nuclear Radiologists
appropriate scan use. medicine
Although imaging use data in patients with cancer in physicians
LMICs are scarce, the data from the IMAGINE database High-income countries
project suggest that in many LMICs, the availability of Range 0·0–26·2 13·9–194·0
imaging for these patients is quite restricted. As such, Mean (SD) 10·9 (10·5) 97·9 (56·2)
the main effect of imaging in LMICs is likely to be on Median (IQR) 6·5 (1·8–11·8) 93·1 (51·3–129·3)
establishing accurate staging information to guide initial Upper-middle-income countries
treatment decisions. As previously noted, the absence of Range 0·0–6·5 1·5–118·0
such information can lead to the inadequate or
Mean (SD) 2·7 (3·4) 66·8 (65·3)
inappropriate use of medicines, surgery, or radiotherapy,
Median (IQR) 1·5 (0·2–3·0) 30·6 (15·6–61·0)
and increase morbidity and mortality.53 In this context,
Lower-middle-income countries
the health outcome and economic case for improving
Range 0·0–1·2 0·4–68·4
access to imaging in LMICs for patients with cancer—as
Mean (SD) 0·5 (0·9) 22·3 (36·4)
detailed in the following section—is of great practical
Median (IQR) 0·1 (0·0–0·6) 6·9 (3·0–30·9)
relevance.
Low-income countries
Range 0·0–0·1 0·1–3·9
Section 3: costs versus health and economic
Mean (SD) 0·1 (0·1) 1·9 (2·5)
benefits of scaling up diagnostic imaging for
cancer—a case for investment Median (IQR) 0·0 (0·0–0·0) 1·1 (0·5–3·3)

Section 2 of this report presents new data on the gaps in The data source is the International Atomic Energy Agency medical imaging and
the availability of imaging modalities for cancer in LMICs. nuclear medicine global resources database.51
The expansion of cancer imaging capacity could help to Table 3: Radiologists and nuclear medicine physicians per million
improve the diagnosis, treatment, and care of patients with population by country income group
cancer worldwide. However, analysis of the IMAGINE
database reveals not only a substantial shortage of imaging
modalities, but also large variation among countries within improve 5-year net survival by more than ten times in low-
and across country income groups. For example, in high- income countries, from 3·8% (95% uncertainty interval
income countries, there is a two-times variation in the [UI] 0·5–9·2) to 45·2% (40·2–52·1), and could more than
lower quartiles and upper quartiles in the availability of CT double 5-year net survival in lower-middle-income
scanners, but a four-times difference for SPECT scanners. countries, from 20·1% (7·2–31·7) to 47·1% (42·8–50·8).
The variation in availability of all imaging modalities for There was increased survival for all country income
upper-middle-income countries, lower-middle-income groups with scale-up, with traditional imaging modalities
countries, and low-income countries is larger than that (ie, traditional treatment including surgery, radiotherapy,
observed for high-income countries (table 2). and chemotherapy; and traditional imaging including
Research undertaken in conjunction with this ultrasound and x-ray) estimated to provide the largest
Commission included modelling studies that estimated increase in low-income countries, and MRI and PET
the potential effect of scaling up treatment (chemotherapy, estimated to yield the largest increase in higher-income
surgery, radiotherapy, and targeted therapy) and imaging countries. The studies showed that investing in medical
modalities (ultrasound, x-ray, CT, MRI, PET, and SPECT) imaging would be necessary for substantial survival
on cancer survival. These studies estimated the net gains.54,55
survival benefit of scaling up treatment and imaging, both However, these studies did not estimate the cost of
individually and in combination, in 200 countries and scale-up and the potential economic benefits. Therefore,
territories, to that of the mean amount of high-income to show the health and economic benefits and costs of the
countries, for 11 cancer types (cancer of the oesophagus, scale-up of imaging modalities worldwide and to ascertain
stomach, colon, rectum, anus, liver, pancreas, lung, breast, whether a worldwide scale-up would generate positive
cervix, and prostate).54,55 We modelled all cancer sites for and substantial rates of return on these investments, we
which comparable international classification of diseases developed and extended a modelling approach that was
for oncology 3 topography codes were available in both conceived initially for The Lancet Oncology Commission
the GLOBOCAN56 (to estimate incidence) and the on expanding global access to radiotherapy and developed
CONCORD-318 (to estimate survival) studies. These for The Lancet Oncology Commission on Sustainable Care
cancers account for 60% of all global diagnosed cases of for Children with Cancer.40
cancer.55 These studies revealed substantial health benefits Briefly, we extended the microsimulation model of
of scaling up imaging modalities in the management of cancer survival for 11 cancer types in 200 countries and
cancer, in that they improved 5-year net survival. The territories, described earlier,55 to include a module on
studies showed that the simultaneous expansion of lifetime survival, treatment costs, and economic
treatment, imaging modalities, and quality of care could benefits. We used observed data from the CONCORD-3

www.thelancet.com/oncology Vol 22 April 2021 e146

 The Lancet Oncology Commission

study18 to calibrate our microsimulation model and to modalities and quality of care are scaled up. We compared
estimate 5-year net survival for 200 countries. We the potential gains from scaling up all imaging modalities
provide a detailed description of the methods in the versus all treatment modalities. We also estimated the
appendix (pp 2–7). We simulated the clinical course of potential gains foregone from not including imaging as
each individual patient with cancer diagnosed between part of a comprehensive scale-up (ie, treatment and quality
2020 and 2030 over their lifetime until death (from any of care only vs comprehensive scenarios).
cause), accounting for net cancer survival and We include a variable for quality of care to control for
competing mortality risks based on country-specific health system and facility-level factors not explicitly
lifetable projections with and without scale-up. In our included in the model, which cover health service
model we did not estimate the effect of screening, but capabilities that also affect cancer survival, such as
modelled cancer cases conditionally depending on adequate laboratory and pathology diagnostics, infection
diagnosis and stage. control, nursing standards, and coordination of care
We estimated the economic benefits of improving (appendix p 4).
cancer survival using the full income approach (also We estimated the cancer deaths averted, life-years
called the value-of-life-year approach). The full income gained, cancer treatment costs, productivity gains, and
approach recognises the intrinsic societal value of a life- lifetime return on investment for the cancer cases
year. We followed the methods used in The Lancet diagnosed in 2020–30, compared with a baseline scenario
Commission on Global Health in 2035,57 which estimated or status quo of no scale-up. We computed health and
the willingness to pay for a 1-year increase in life economic benefits, costs, and return on investment for
expectancy in countries with different income levels and the 200 countries and territories included, and for world
applied a value of 2·3 times the gross domestic product regions. We discounted costs and benefits at 3% (a
(GDP) per person per year in LMICs and 1·4 times the commonly used discount rate).59 The detailed description
GDP in high-income countries. of the data sources, methods, and the approach for the
For a sensitivity analysis, we used a more conservative modelling are provided in other published papers.55,58
human capital approach. With the human capital The results show that the com­ prehensive scenario,
approach, the economic value of a life-year is based on with a scale-up of all imaging modalities, treatment
the economic contribution of an individual and is valued methods, and quality of care in 2020–30 would avert
at one times the GDP per person. We accrued productivity 9·55 million deaths worldwide, accounting for 12·5% of
benefits only to individuals aged 18–64 years in the model the projected total worldwide deaths of 76·00 million in
when using the human capital approach to reflect typical this period and 232·30 million life-years saved. The scale-
working ages. up of imaging alone would avert 2·46 million deaths,
Because the human capital approach only values accounting for 3·2% of worldwide deaths and
productivity and economic contribution and not the 54·92 million life-years saved (table 4).58
intrinsic value of health and an additional year of life, we The vast majority of the deaths averted under a
used the full-income approach as our base case, which comprehensive scale-up scenario would be in Asia
better reflects the value of an additional year to a society. (5·28 million) accounting for 11·9% of projected cancer
Cancer treatment costs were estimated with the use of deaths in Asia in 2020–30 and 133·99 million life-years
a modelled relationship between costs and per person saved. In Asia, the scale-up of imaging alone would avert
GDP based on empirical data obtained from a targeted 1·42 million deaths, accounting for 3·2% of projected
literature review. More details on the model specifications cancer deaths in Asia, and would result in 33·47 million
and assumptions, estimations of costs, projected health, life-years saved (table 4).58
and economic benefits and restrictions with the data and Similarly, there would be major health gains in Africa
model are available in a paper by Ward and colleagues58 where the comprehensive scale-up would avert
and in the appendix (pp 2–7). 2·51 million cancer deaths amounting to 35·7% of total
Using the model, we estimated the global costs and projected cancer deaths in Africa, and result in
benefits of four different packages of scale-up, in which 61·27 million life-years saved. Scale-up of imaging alone
we improved the availability of imaging or treatment would avert 207 800 cancer deaths (3·0% of the projected
modalities, or both, and quality of care to the mean value total cancer deaths in Africa) and result in 4·64 million
of high-income countries under different scenarios: life-years saved on this continent (table 4).58
(1) imaging only, a scenario in which all imaging Worldwide scale-up of imaging alone or in conjunction
modalities (ultrasound, x-ray, CT, MRI, PET, and SPECT) with treatment and improved quality of care produces
only are scaled up; (2) treatment only, in which all substantial economic benefits and return on investments
treatment modalities (chemotherapy, radiotherapy, (table 5).58
surgery, and targeted therapy) only are scaled up; Incremental costs in 2020–30 of scaling up imaging
(3) treatment and quality of care, in which all treatment alone would be $6·84 billion, but this investment would
modalities and quality of care are scaled up; and result in productivity gains of $1·23 trillion and a net
(4) comprehensive, in which all imaging and treatment benefit of $1·22 trillion, yielding a return per dollar

e147 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Deaths from cancer averted Projected life-years saved, millions (95% uncertainty
(95% uncertainty interval) interval)
Number Proportion of total Undiscounted Discounted (3% annually)
deaths
Global
Imaging only 2 463 500 (1 154 900–4 073 900) 3·2% (1·6–5·3) 54·92 (25·15–91·40) 33·17 (15·18–54·93)
Treatment only 4 095 600 (1 632 300–7 093 400) 5·4% (2·2–9·1) 103·28 (40·37–184·19) 58·36 (22·71–102·73)
Treatment and quality of care 5 369 100 (2 894 300–8 032 800) 7·0% (3·9–10·3) 134·96 (72·84–208·11) 76·13 (40·94–116·06)
Comprehensive 9 549 500 (6 677 800–12 743 800) 12·5% (9·0–16·3) 232·30 (157·29–311·30) 133·71 (91·94–179·03)
Africa
Imaging only 207 800 (78 700–579 100) 3·0% (1·1–8·3) 4·64 (1·65–13·76) 2·72 (0·99–7·89)
Treatment only 984 300 (299 900–1 926 700) 14·1% (4·3–26·9) 23·99 (7·11–47·13) 13·50 (4·06–26·43)
Treatment and quality of care 1 569 400 (925 500–2 211 400) 22·3% (14·1–30·5) 38·54 (22·47–54·77) 21·62 (12·63–30·37)
Comprehensive 2 508 100 (2 004 500–2 932 800) 35·7% (29·8–41·7) 61·27 (49·52–72·07) 34·58 (27·86–40·30)
Asia
Imaging only 1 420 600 (381 700–2 784 800) 3·2% (0·9–6·3) 33·47 (9·16–67·14) 20·12 (5·43–39·85)
Treatment only 2 509 100 (399 600–4 813 600) 5·6% (0·9–10·4) 65·74 (10·72–124·31) 36·93 (6·09–69·93)
Treatment and quality of care 3 038 000 (822 900–5 402 900) 6·8% (1·9–11·7) 79·56 (21·62–142·02) 44·64 (12·03–79·77)
Comprehensive 5 282 200 (3 203 400–7 616 800) 11·9% (7·4–16·5) 133·99 (79·09–191·59) 76·88 (45·70–110·17)
Europe
Imaging only 435 700 (158 600–769 700) 3·2% (1·1–5·6) 8·18 (2·97–14·76) 5·16 (1·90–9·13)
Treatment only 350 500 (91 800–709 800) 2·6% (0·7–5·2) 7·40 (1·98–14·62) 4·45 (1·22–8·81)
Treatment and quality of care 455 800 (116 800–971 100) 3·3% (0·9–7·0) 9·46 (2·41–19·98) 5·68 (1·44–11·98)
Comprehensive 982 400 (610 700–1 366 200) 7·2% (4·6–9·8) 19·38 (12·02–27·12) 11·95 (7·48–16·50)
Latin America and the Caribbean
Imaging only 354 900 (26 900–633 700) 7·0% (0·6–12·6) 7·64 (0·55–14·04) 4·57 (0·33–8·36)
Treatment only 210 700 (28 600–610 400) 4·1% (0·6–12·1) 5·19 (0·77–15·17) 2·93 (0·41–8·50)
Treatment and quality of care 247 600 (53 400–728 300) 4·9% (1·1–13·8) 6·08 (1·36–17·04) 3·42 (0·75–9·77)
Comprehensive 665 000 (370 300–1 039 000) 13·1% (7·5–19·5) 15·13 (8·08–24·02) 8·84 (4·81–13·85)
North America
Imaging only 29 700 (0–219 500) 0·5% (0·0–4·0) 0·67 (0·00–4·88) 0·40 (0·00–2·94)
Treatment only 15 300 (0–119 600) 0·3% (0·0–2·2) 0·35 (0·00–2·83) 0·20 (0·00–1·72)
Treatment and quality of care 21 100 (0–129 400) 0·4% (0·0–2·4) 0·47 (0·00–2·85) 0·27 (0·00–1·72)
Comprehensive 50 900 (0–235 800) 0·9% (0·0–4·3) 1·14 (0·00–5·27) 0·68 (0·00–3·15)
Oceania
Imaging only 14 700 (700–53 900) 2·7% (0·1–9·7) 0·33 (0·01–1·23) 0·19 (0·01–0·72)
Treatment only 25 700 (800–73 300) 4·7% (0·2–12·3) 0·60 (0·02–1·70) 0·34 (0·01–0·98)
Treatment and quality of care 37 300 (3000–79 800) 6·8% (0·6–14·2) 0·86 (0·07–1·87) 0·49 (0·04–1·06)
Comprehensive 61 000 (22 800–95 800) 11·1% (4·4–17·1) 1·38 (0·50–2·27) 0·80 (0·30–1·30)

Estimates are from the global cancer survival microsimulation model. The four different scenarios are: (1) imaging only, a scenario in which all imaging modalities
58

(ultrasound, x-ray, CT, MRI, PET, and SPECT) only are scaled up; (2) treatment only, in which all treatment modalities (chemotherapy, radiotherapy, surgery, and targeted
therapy) only are scaled up; (3) treatment and quality of care, in which all treatment modalities and quality of care are scaled up; and (4) comprehensive, in which all imaging
and treatment modalities and quality of care are scaled up.

Table 4: Potential health benefits for patients with cancer diagnosed between 2020 and 2030 under various scenarios of scale-up for the 11 modelled
cancer types

invested of $179·19. The large returns that could be a 6·9% increase in the current global cost of cancer
achieved from investment are because the scale-up of treatment and care. However, the benefits of this scale-up
most of the cancer imaging modalities is not costly. would be substantial, with lifetime productivity gains of
However, the absolute numbers of deaths averted with $2·89 trillion for the cancer cases diagnosed in 2020–30.
scaling up imaging alone would be modest compared This benefit would produce a net economic benefit of
with what could be achieved with the comprehensive $2·66 trillion and a return on investment of $12·43 for
scale-up scenario (table 4). every dollar invested. Scale-up of just treatment and
The estimated incremental cost of comprehensive quality of care without imaging would produce a notably
scale-up globally would be $232·88 billion, amounting to lower net economic benefit of $1·16 trillion and a return

www.thelancet.com/oncology Vol 22 April 2021 e148

 The Lancet Oncology Commission

Incremental cancer treatment costs (2020–30), Lifetime return on investment: full income approach (95% uncertainty interval)
US$ billion (95% uncertainty interval)
Difference Percentage increase Productivity gains, US$ billion Net benefit, US$ billion Return per US$ invested
Global
Imaging only 6·84 (1·77 to 15·86) 0·2% (0·1 to 0·3) 1226·21 (540·05 to 2161·80) 1219·37 (535·47 to 2157·29) 179·19 (84·71 to 625·09)
Treatment only 50·72 (14·92 to 111·88) 1·5% (0·8 to 2·4) 1183·24 (504·90 to 2206·54) 1132·51 (489·13 to 2114·69) 23·33 (12·40 to 60·40)
Treatment and quality of care 225·50 (83·87 to 408·34) 6·7% (5·7 to 7·8) 1386·07 (726·42 to 2342·19) 1160·56 (484·04 to 2053·70) 6·15 (2·66 to 16·71)
Comprehensive 232·88 (85·92 to 421·97) 6·9% (6·0 to 8·0) 2894·41 (1794·55 to 4025·16) 2661·54 (1631·20 to 3775·64) 12·43 (6·47 to 33·23)
Africa
Imaging only 0·46 (0·23 to 0·79) 1·9% (1·2 to 3·0) 27·38 (9·61 to 65·80) 26·93 (9·29 to 65·34) 59·97 (22·11 to 128·14)
Treatment only 6·85 (3·82 to 11·22) 29·4% (17·6 to 42·2) 120·97 (52·46 to 210·96) 114·12 (44·51 to 203·06) 17·67 (8·09 to 33·93)
Treatment and quality of care 11·14 (6·64 to 16·98) 47·8% (34·1 to 63·1) 164·86 (88·57 to 237·47) 153·72 (79·95 to 225·41) 14·80 (8·05 to 25·71)
Comprehensive 11·67 (7·01 to 17·70) 50·1% (36·2 to 66·4) 249·66 (187·61 to 303·31) 237·99 (177·71 to 291·80) 21·39 (14·15 to 34·34)
Asia
Imaging only 3·42 (0·66 to 9·37) 0·4% (0·1 to 0·6) 713·38 (86·71 to 1616·35) 709·96 (86·03 to 1610·45) 208·70 (77·77 to 850·18)
Treatment only 24·58 (4·35 to 69·42) 2·7% (0·5 to 6·2) 679·76 (107·85 to 1681·10) 655·17 (103·01 to 1621·55) 27·65 (12·89 to 68·97)
Treatment and quality of care 37·98 (13·16 to 86·15) 4·4% (1·9 to 8·5) 772·73 (182·13 to 1686·61) 734·75 (164·77 to 1613·12) 20·35 (8·10 to 49·52)
Comprehensive 41·59 (14·76 to 91·25) 4·7% (2·3 to 8·9) 1653·82 (828·58 to 2458·01) 1612·22 (802·55 to 2410·54) 39·76 (17·99 to 101·74)
Europe
Imaging only 1·95 (0·23 to 5·52) 0·2% (0·0 to 0·4) 281·15 (77·79 to 612·65) 279·20 (76·86 to 605·35) 144·32 (71·07 to 686·83)
Treatment only 14·73 (1·88 to 38·95) 1·2% (0·2 to 2·6) 257·18 (82·05 to 517·31) 242·45 (72·14 to 493·25) 17·46 (8·28 to 66·89)
Treatment and quality of care 171·39 (59·50 to 314·06) 14·5% (13·3 to 16·0) 301·80 (114·77 to 602·30) 130·41 (–119·56 to 444·47) 1·76 (0·49 to 6·02)
Comprehensive 173·59 (59·79 to 315·94) 14·7% (13·6 to 16·1) 618·57 (367·27 to 884·37) 444·98 (160·23 to 737·88) 3·56 (1·64 to 10·47)
Latin America and the Caribbean
Imaging only 0·52 (0·03 to 1·31) 0·6% (0·0 to 1·1) 138·85 (8·89 to 259·83) 138·33 (8·85 to 259·06) 266·38 (109·69 to 1351·47)
Treatment only 2·21 (0·20 to 7·03) 2·9% (0·3 to 7·4) 79·99 (8·78 to 241·17) 77·79 (8·54 to 237·43) 36·28 (14·10 to 152·10)
Treatment and quality of care 2·56 (0·45 to 7·42) 3·4% (0·7 to 8·0) 87·66 (9·42 to 264·11) 85·10 (8·85 to 260·56) 34·27 (12·16 to 124·16)
Comprehensive 3·08 (0·61 to 8·04) 4·1% (1·3 to 8·7) 245·96 (123·82 to 403·20) 242·88 (122·20 to 397·69) 79·77 (30·36 to 384·86)
North America
Imaging only 0·37 (0·00 to 3·26) 0·0% (0·0 to 0·2) 47·48 (0·00 to 348·01) 47·12 (0·00 to 345·16) 128·94 (64·85 to 361·54)
Treatment only 1·22 (0·00 to 11·54) 0·1% (0·0 to 0·8) 24·24 (0·00 to 202·14) 23·02 (0·00 to 181·52) 19·83 (7·95 to 72·25)
Treatment and quality of care 1·22 (0·00 to 11·54) 0·1% (0·0 to 0·8) 32·60 (0·00 to 202·14) 31·37 (0·00 to 190·39) 26·66 (8·18 to 1398·67)
Comprehensive 1·59 (0·00 to 11·58) 0·1% (0·0 to 0·8) 80·12 (0·00 to 373·7) 78·53 (0·00 to 371·43) 50·36 (8·42 to 984·28)
Oceania
Imaging only 0·13 (0·00 to 0·59) 0·1% (0·0 to 0·6) 17·96 (0·13 to 77·95) 17·83 (0·13 to 77·42) 137·36 (24·94 to 338·03)
Treatment only 1·14 (0·02 to 4·59) 1·2% (0·0 to 4·4) 21·09 (0·12 to 86·53) 19·96 (0·11 to 83·31) 18·56 (5·28 to 51·96)
Treatment and quality of care 1·21 (0·09 to 4·68) 1·3% (0·1 to 4·5) 26·42 (0·67 to 93·98) 25·21 (0·57 to 91·45) 21·77 (5·70 to 191·78)
Comprehensive 1·35 (0·13 to 4·83) 1·4% (0·2 to 4·5) 46·29 (9·13 to 112·39) 44·95 (8·92 to 109·14) 34·41 (11·48 to 244·48)

All results discounted 3% annually. Estimates are from the global cancer survival microsimulation model.58 The four different scenarios are: (1) imaging only, a scenario in which all imaging modalities
(ultrasound, x-ray, CT, MRI, PET, and SPECT) only are scaled up; (2) treatment only, in which all treatment modalities (chemotherapy, radiotherapy, surgery, and targeted therapy) only are scaled up; (3) treatment
and quality of care, in which all treatment modalities and quality of care are scaled up; and (4) comprehensive, in which all imaging and treatment modalities and quality of care are scaled up. GDP=gross
domestic product.

Table 5: Potential economic costs and benefits for patients with cancer diagnosed between 2020 and 2030 for 11 modelled cancer types

on investment of $6·15, less than half of what would be comprehensive scale-up or with the scale-up of imaging
achieved if imaging were included in the scale-up alone or in combination with treatment and quality of
(table 5).58 To provide a specific example, we compared care (table 5). Lifetime returns on investment accrued to
our model estimates to the reported costs from Ethiopia countries worldwide are shown in figure 9.
using data from Ethiopia’s national health accounts (see The estimated variation on the return on investment
the case study in panel 2).60,61 between countries is mainly because of differences in the
The net economic benefits of a comprehensive scale-up availability of imaging modalities in different countries.
would be substantial in all world regions (table 5).58 Regional differences in these estimations are largely
All regions worldwide would achieve substantial because of: (1) differences in the baseline availability
positive returns per dollar invested on investment in of surgery, radiotherapy, and medicines and imaging

e149 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

modalities; (2) differences in the quality of care;

(3) differences in income levels in countries, which Panel 2: Incremental costs and benefits of comprehensive scale-up: Ethiopia case study
influences productivity estimates; and (4) the fact that the As a specific example, we compare our model estimates to reported costs from Ethiopia,
value placed on a life-year with the use of the full income for which national health accounts data are available. According to the Ethiopian Ministry
approach varies by income group, where the value is of Health, national health expenditures were $US3·10 billion for 2016–17 (around 4·2% of
2·3-times the GDP per person per year in LMICs, and their gross domestic product), of which an estimated 1·8% ($55·8 million) was spent on
1·4-times that the GDP per person in high-income patients with cancer.60 Our model estimates that cancer treatment costs in Ethiopia for
countries. New data compiled by this Commission on the baseline scenario (no scale-up) are $90·55 million (95% uncertainty interval [UI]
coverage of imaging modalities by country and presented 64·51–124·12) per year on average between 2020 and 2030, similar to the reported
in this report (table 2) reveal substantial variation in the estimates after accounting for population growth (UN population projections estimate
availability of imaging modalities between countries at that the population of those aged older than 50 years in Ethiopia will grow by 40%
different income levels.51 The range of per person income between 2015 and 2025).61
between and without country income categories is
We estimate that with a comprehensive scale-up, cancer treatment costs would rise to
substantial. The Gross National Income per person
$171·17 million (95% UI 125·55–224·80), accounting for an additional $80·6 million
(calculated with the use of Atlas methods62 and
(95% UI 54·3–110·0) of spending per year on average—a 90% increase in cancer costs.
purchasing power parity) ranges from $280 to $1035 in
Although this estimate represents a large increase in cancer spending, it is a small
low-income counties, from $1036 to $4045 in lower-
proportion of total health expenditures, comprising approximately 2·8% of current total
middle income countries, from $4046 to $12 535 in
health expenditures. In return, we estimate that a comprehensive scale-up would yield
upper-middle income countries, and from $12 536 to
large economic benefits over the lifetime of patients who have survived cancer, yielding
more than $100 000 in high-income countries.63
an estimated return of $18·44 (95% UI 12·94–28·80) per dollar invested in Ethiopia.
We present in the appendix (p 8) a sensitivity analysis
(based on estimates of the global cancer survival
microsimulation model)58 of costs, productivity gains,
net benefits, and return on investments that use the cancer. Hence, the results establish a compelling case for
more conservative human capital approach. This investing in the worldwide comprehensive scale-up of
sensitivity analysis shows a net benefit of $209·46 billion diagnostic imaging for cancer.
(95% UI $94·96–394·72) and a return per dollar
invested of $31·61 (95% UI $15·09–110·14) for scaling Section 4: financing the global scale-up of
up imaging alone. With comprehensive scale-up, the diagnostics
worldwide net benefit is $340·42 billion (95% UI New financing will be needed to scale-up the capacity for
$99·37–592·59) and the return per dollar invested is cancer imaging diagnostics to expand access to effective
$2·46 (95% UI $1·29–6·52), because costs of and affordable services in LMICs. But where will this
comprehensive scale-up are much higher than scaling new financing come from?
up imaging alone. There are substantial net benefits In most LMICs, the largest proportion of funding will
and returns to scaling up imaging in all world regions probably come from domestic sources—namely, public
and, with the exception of Europe, considerable net financing (the government budget allocated to health)
benefits and return on investment with comprehensive and complementary financing from the private sector.
scale-up (appendix p 8). Additionally, there is the potential for funding from
The modelling, with the use of either the full income or external private companies, Overseas Development
human capital approaches, shows notable health and Assistance, or development banks that provide loans or
economic benefits, with substantial returns on invest­ invest in health infrastructure projects; for example,
ments achieved when scaling up imaging diagnostics banks that establish new diagnostic imaging facilities or
alone or as part of a comprehensive scale-up that involves upgrade existing ones. Examples of development banks
the simultaneous scale-up of treatment and quality of include the World Bank Group, a conglomerate of
care. five institutions, as well as the European Investment
Modelling suggests synergistic benefits when all of Bank, African Development Bank, InterAmerican
these aspects are scaled up simultaneously. Therefore, the Development Bank, Islamic Development Bank, and
results are not additive. Scaling up imaging without scale- Asian Development Bank.
up in treatment is not likely to lead to major improvements Donations can also come from or be facilitated by
in cancer survival, because treatment capacity is soon non-state actors or non-governmental organisations
reached and additional cases will not be adequately and UN organisations, such as WHO and the IAEA.
treated. Similarly, scaling up quality of care without For example, the IAEA allocated €5·74 million in 2019
diagnostics or improving treatment availability will for the support of nuclear medicine and diagnostic
probably have little effect on cancer survival in LMICs, imaging, including the procurement of medical
because many individuals will not be diagnosed, and even imaging equip­ment and the expansion of capacity. The
when they are diagnosed they will not receive the surgery, beneficiaries of cooperation are member state LMICs
radiotherapy, or medicines that they need to treat their (eg, Algeria).64

www.thelancet.com/oncology Vol 22 April 2021 e150

 The Lancet Oncology Commission

Return per US$ invested

$0 $25 $50 $75 $100

Figure 9: Estimated lifetime return on investment (comprehensive scale-up of imaging, treatment, and quality of care) by country for 11 cancer types
Comprehensive scale-up refers to scale-up of all imaging and treatment modalities and quality of care to the mean amount of that of high-income countries. Returns per US$ invested are estimated
for patients diagnosed with cancer in 2020–30, compared with a baseline scenario of no scale-up. Estimates are presented in US$ in 2018 and discounted at 3% annually.

The amount of public financing for any sector is help to increase government spending on health per
established by the so-called fiscal space available to the person by around 5·3% each year in upper-middle-
government, which is defined as “…the availability of income countries, 4·2% in middle-income countries,
budgetary room that allows a government to provide and 1·8% in low-income countries.69 However, notably,
resources for a desired purpose without any prejudice to these estimates are based on pre-COVID-19 economic
the sustainability of a government’s financial position”.65 variables. An investment case for imaging diagnostics is
Fiscal space depends on the sources of finances, which crucial to harness new funding for this area.
can be from: (1) economic growth that creates favourable
macroeconomic conditions for increased government Generation of revenues by strengthening tax
revenues and budget, (2) the strengthening of tax administration
administration, (3) the reprioritisation of health within In LMICs, government revenues from tax are low, being
the governments’ budget, (4) borrowing from domestic on average 15% of the GDP in low-income countries,
and international sources or Overseas Development 25% in lower-middle-income countries, 30% in upper-
Assistance to invest in health, (5) more effective and middle-income countries, and 40% in high-income
efficient allocation of available health resources, and countries.70 Modelling studies estimate that an increase
(6) innovative domestic and international financing.66,67 in tax revenue, where at least a third of newly raised
In the following paragraphs, we describe the main revenues is allocated to health, could on average increase
sources of financing that could be used to expand fiscal public expenditure on health in LMICs by 78% (95% CI
space and summarise the potential magnitude of funds 60–90%).71
and the suitability of different funding sources for Increased taxes on tobacco and alcohol are highly cost-
investing in the scale-up of imaging diagnostics and effective public policies. Egypt, the Philippines, and
cancer care. Thailand have successfully introduced tobacco taxes to
generate funding for the health sector.72 A 20% and 50%
Improved economic growth price increase in tobacco prices could generate more
The International Monetary Fund projects positive than 50 years’ worth of additional tax revenues globally,
economic growth in LMICs between 2020 and 2025.68 with a 20% price increase resulting in approximately
Other estimates suggest that in 2015–40, the continued $1987 billion (UI 1613–2297 billion) in additional tax
growth of GDP and higher government revenues could revenues over 50 years, and a 50% price increase

e151 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

generating $3625 billion (UI 2534–4599 billion) over More effective and efficient allocation of available
50 years; and in low-income countries, an average health resources in health systems
additional revenue of 0·17% of GDP each year in the With appropriate priority setting and the more efficient
50% price increase scenario.73 allocation and use of financial resources, governments
in LMICs could generate a 26% (95% CI 21–31%)
Reprioritisation of health within government budgets increase in public expenditure on health.70 For example,
Evidence for the health and economic benefits of new governments in LMICs could undertake reviews of
health investments could be made use of to persuade their health budgets to reduce the spending on
governments to reprioritise their investments. Modelling interventions and programmes that are not cost-
estimates that budget reprioritisation could potentially effective, and channel these resources to more cost-
increase the funds allocated to health in LMICs by 72% effective interventions. These governments could also
(95% CI 57–87%).71 improve the procurement of health products, by
benchmarking the prices achieved or through the use
Borrowing from domestic and international sources of pooled procurement to secure a better value for these
and Overseas Development Assistance products.
Concessional financing with low interest rates and
generous grace periods for repayments could be Innovative financing
mobilised from international development banks to Funding mobilised from non-traditional sources is
invest in the expansion of diagnostics capacity. In 2017, another potential source of financing for diagnostic
the World Bank had 45 active projects for a total sum of imaging. Innovative financing mechanisms such as the
US$470 million for medical equipment procurement.74 Global Fund to Fight AIDS, Tuberculosis and Malaria,
In 2020, the African Development Bank approved an Gavi, and Unitaid78,79 (which link different elements of
equity investment that will raise $100 million to fund the financing value chain—namely, resource mobili­
health infrastructure projects in Africa.75 sation, pooling, channelling, resource allocation, and
Investment in diagnostic imaging is particularly imple­mentation) have channelled more than $55 billion
attractive for development banks, because these are to LMICs for the health sector.
infrastructure investments that can generate an income Social or development impact bonds are promising
stream for the investors to service the loans over time innovative financing instruments that could be used to
and also provide an opportunity for public–private finance the expansion of diagnostics capability in LMICs.
partnerships or private sector investments for the A social or development impact bond is created by a
provision of public services that can be outsourced by government agency (or external funder such as a
governments. In addition to loans, guarantees provided development agency or a charitable foundation) that
by development banks can be used to encourage the aims to achieve a desired social or health outcome.80,81 The
mobilisation of private financing by mitigating invest­ government agency or external funder engages an
ment risks in LMICs for projects to establish or develop external organisation to achieve the outcome. A third-
facilities for imaging diagnostics. party investor provides upfront working capital to the
Over the past 20 years, World Bank Group guarantees external organisation as an at-risk investment. If the
have mobilised more than US$42 billion in commercial desired social outcome is accomplished, the government
capital and private investments.76 The guarantees were agency or external funder releases payment to the
structured as partial risk guarantees, partial credit external organisation, on the basis of terms specified in
guarantees, or policy-based guarantees.77 Partial risk an upfront contract, which repays its investors their
guarantees support private sector invest­ment, including principal, plus a return on the investment. If the outcome
public–private partnerships. Partial credit guarantees is not met, the government agency or external funder
enable commercial borrowing in support of public disburses no payment.
investment projects, and policy-based guarantees support The potential new funding from multiple sources to
commercial borrowing for budget financing or reform expand fiscal space (table 6)68–77,79–82 far exceeds the
programmes. Guarantees offer several benefits to the financing needed globally for the comprehensive scale-
borrowers. The reduced risk of default improves the up of interventions for cancer care. With measurable
country’s ability to borrow for investment. Guarantees performance indicators, the investment in population-
can reduce the cost of capital as a result of lower interest based health can be a tool towards a nation’s develop­
rates that the borrowing government must pay, because ment, rather than a mere byproduct of it. Medical
these rates are moderated by the guarantor’s credit imaging is a cornerstone of the strengthening of health
worthiness (the World Bank has an AAA rating). systems to address the disability-adjusted life-years lost
Guarantees also allow governments to share the risk of to cancer, a burden that falls disproportionately (80%) on
projects with the private sector. Such guarantees would LMICs, even though these nations receive only
be suited to investments in expanding the capacity for approximately 5% of current global funding for cancer
imaging diagnostics in LMICs. control.3,5

www.thelancet.com/oncology Vol 22 April 2021 e152

 The Lancet Oncology Commission

Potential additional fiscal space that Feasibility of creating additional fiscal Suitability for funding the scale-up of
could be created space imaging diagnostics for cancer
Improved economic Substantial. Could help increase Feasible. LMICs are projected to have robust Would generate sustainable general
growth government spending on health per economic growth,68 and despite the revenue income for allocation to health
person each year by approximately 5·3% COVID-19 pandemic, many have returned
in upper-middle-income countries, to positive growth trajectories82
4·2% in middle-income countries,
and 1·8% in low-income countries69
Generation of revenues Susbtantial. Allocating at least a third of Feasible. Tax revenues in LMICs are only 15– Additional revenues would need to be
by strengthening tax newly raised revenues to health could on 30% of GDP compared with 40% in high- allocated to health; however, it is a
administration average increase public expenditure on income countries, but would require sustainable funding source
health in LMICs by 78% (95% CI stronger tax collection systems, which
60–90%)71 would take time to implement70
Increased taxes on Substantial. In low-income countries, Feasible, but would require political will to Sustainable funding with additional
tobacco, alcohol, a 50% increase in tobacco prices could fight opposition. Highly cost-effective72 health and economic benefits. Could be
and sugary beverages generate on average an additional earmarked for health
revenue of 0·17% of GDP each year73
Reprioritisation of health Substantial. In LMICs, governments could Less feasible. Would require strong political Sustainable funding
within the government increase funds allocated to health by 72% capital to achieve reprioritisation
budget (95% CI 57–87%)71
Borrowing from Substantial, but underused. Could be in Feasible. Low interest rates make this an This option would encourage public-
domestic and the form of hybrid financing: a mix of attractive option. Infrastructure loans are private partnerships to reduce capital
international sources and loan and equity from public and private available from the World Bank and regional investment requirements for
Official Development sectors development banks. Export guarantees governments; and could provide a
Assistance would substantially reduce borrowing revenue stream to investors to offset
costs74–77 costs
Innovative financing Substantial, with a large potential Feasible. Social or development impact This option would encourage public-
bonds could be used to invest in scale- private partnerships to reduce capital
up.79–81 Easily measurable results with investment requirements for
investment in imaging diagnostics governments; and provides a revenue
stream to investors to offset costs

The sources for this table was an analysis synthesis of evidence68–77,79–81 and the International Monetary Fund 2020 report.82 GDP=gross domestic product. LMICs=low-income
and middle-income countries.

Table 6: Potential funding sources for expanding fiscal space for health and investment in the scale-up of imaging diagnostics and cancer care in LMICs

Section 5: radiation protection and safety and from man-made sources. Between the global surveys for
quality systems 1991–96 and 1997–2007, the total annual number of
The safe use of medical imaging in cancer care requires diagnostic medical examinations (both medical and
appropriate standards for radiation protection and safety dental) was estimated to have risen by 50%.83 However,
with regard to patients, families, workers, and the public, more recent national figures for the USA84 suggest that
irrespective of the level of economic development of a the largest contributor to radiation doses, CT scanning,
country. Responsibilities to ensure that appropriate has stabilised in numbers. The second largest contributor,
standards are met, lie at the national, institutional, and imaging with the use of nuclear medicine, has shown
individual levels. Whether the imaging modality makes similar numbers per year in the last 5 years for SPECT-CT
use of ionising or non-ionising radiation, adequate safety procedures, and continued to increase its contribution to
infrastructure, education and training of staff, appropriate radiation doses in PET-CT studies (mainly in patients
staffing amounts, and effective quality assurance systems with cancer) globally, in both high-income countries and
are all essential. LMICs.85–87 In relation to occupational radiation exposure,
according to the UN Scientific Committee on the Effects
Protecting patients and workers when ionising of Atomic Radiation,83 worldwide, the estimated number
radiation is used in medicine of health-care workers involved in the medical uses of
The latest figures published by the UN Scientific radiation is 7·4 million (estimated in 2008), which is
Committee on the Effects of Atomic Radiation83 indicate considered to be increasing with time.
that approximately 3·6 billion diagnostic radiology x-ray For the use of ionising radiation in medicine, radiation
examinations and 33 million diagnostic nuclear medicine protection for patients and workers needs to be approached
examinations are done each year worldwide. However, systematically.88 In the past century, remarkable progress
imaging frequency during cancer care is not explicitly has been made in understanding the health effects of
considered in these figures.83 Medical uses of ionising radiation. There is a need to increase awareness among
radiation (excluding therapeutic uses) constitute more the medical community about the amount of radiation
than 98% of the world population’s exposure to radiation received by patients in imaging procedures.89 However,

e153 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

there is an absence of qualified support for medical safety standards,95 it is the responsibility of the
physics, in particular in diagnostic radiology and nuclear government to ensure that a country’s diagnostic
medicine theranostics, in LMICs.90 This shortfall poses reference levels, an optimisation tool for diagnostic
notable risks for patients and health-care workers because imaging, are established through consultation between
radiation safety, quality systems, and maintenance are the relevant health authorities, professional bodies, and
insufficiently guaranteed. Furthermore, in many LMICs, the regulatory agencies. The regulatory agency has
the medical radiation devices and their use are not different means of ensuring compliance, such as the
sufficiently governed by appropriate governmental, legal, authorisation and inspection of facilities and activities,
and regulatory frameworks for safety. The rapid evolution and enforcement of regulatory requirements.94 At a
of technology for imaging involving radiation exposure national level, other organisations have an important
poses challenges for maintaining the safety of patients role for the safety of patients, workers, and the public,
and health-care workers, because this maintenance such as health authorities, professional bodies, technical
requires the education and training of health professionals standards associations, regulatory agencies involved in
and regulatory staff; moreover, the rapid evolution of this the approval of medical devices, and agencies involved
technology makes it challenging to keep regulations up to in health technology assessments.95 Many countries do
date. Regulation of the use of ionising radiation in not have adequate infrastructure for radiation safety. For
medicine differs between countries globally.91 LMICs and other countries that might need to strengthen
The radiation exposure of patients for diagnosis, this infrastructure at a national level, the IAEA has
intervention, or therapy differs from other uses of published guidance on overcoming this challenge,
radiation in that it is done for the direct benefit of the including on national policy, regulatory framework, and
individual, who also incurs the radiation risk and other technical infrastructure.96
risks of the procedure.92 The guidelines that justify the
use of a procedure should be developed by health Responsibilities at the facility and individual levels
authorities together with professional bodies and should Hospitals and other health-care institutions that do
be reviewed regularly to ensure that radiological radiological and nuclear medicine imaging procedures
procedures that are no longer justified are removed from should have appropriate equipment (with planned
guidelines and medical practice.93 The optimisation of replacement cycles), maintenance and quality systems,
radiation protection in imaging means that the amount and enough staffing to do studies in an optimal manner.
of protection and safety should be the best possible under Health professionals working in such facilities should
the prevailing circumstances, and should be implemented have appropriate training and qualifications in clinical
in all scenarios. Notably, this pertains not only to practice and adhere to relevant radiation safety
radiation doses that are excessive for the given imaging standards. The optimisation of radiation protection is
being done but also to doses that are too low to generate inadequate in facilities in many countries and can be
images of a suitable diagnostic quality for accurate improved with the use of simple and inexpensive
interpretation. This trade-off between radiation exposure techniques.97
and a suitable diagnostic quality is a challenging issue in Clinical imaging guidelines and appropriate use
cancer care, because repeated exposure to radiation over criteria are the imaging referral guidelines developed
short and long intervals is common. Dose limits apply to by international expert groups that facilitate the choice
occupational exposure and public exposure arising from of the best imaging test for a clinical scenario, and help
medical uses of ionising radiation, but not to the to strengthen the justification of exposure to radiation
exposure of patients. For some areas of medical uses of in imaging procedures.98 Justified procedures, by
ionising radiation, such as image-guided interventional definition, bring individual patients more benefit than
procedures, good radiation protection practice for staff risk. This means that the proposed overall increase of
must be followed to not exceed occupational dose limits.93 imaging with the use of ionising radiation will bring
the global population more benefit than risk, as long as
Responsibilities at a national level a generic justification of the radiological procedure has
For the safe operation of facilities and use of radiation been done by the health authority in conjunction with
sources, a country must have appropriate governmental, appropriate professional bodies, and the justification of
legal, and regulatory frameworks for safety.94 The the medical exposure for the individual patient has
government establishes laws and adopts policies relating been done by means of consultation between the
to safety as well as the responsibilities and functions of radiological medical practitioner and the referring
different governmental bodies involved in safety. The medical practitioner. Improving the appropriate use of
important responsibilities of a government include the imaging is important for the radiation protection of
establishment of an independent regulatory body with patients and for overall patient care. According to the
the necessary legal authority, competence, and resources international basic safety standards developed by the
to oversee radiation safety for the public and radiation IAEA,95 relevant national or international referral
workers. In the health sector, according to international guidelines should be taken into account when justifying

www.thelancet.com/oncology Vol 22 April 2021 e154

 The Lancet Oncology Commission

the medical exposure of an individual patient in a formulations and, therefore, they should be produced in
radiological or nuclear medicine procedure. These facilities that have appropriate quality management
guidelines are produced, maintained, and disseminated systems in place. Radiopharmaceuticals can be produced
by many international organisations,99–104 are for the use by a licensed commercial organisation, or alternatively by
of referring physicians, radiologists, and nuclear hospital-based facilities that comply with appropriate
medicine physicians, and are important for the domestic or international standards.109–111 Testing of the
radiation protection of patients. However, it should be final product and radiation safety are essential in
noted that knowledge in cancer care, especially for new ensuring safe and appropriate use.
therapeutic drugs, is evolving rapidly, which makes it Access to, and availability of, radiopharmaceuticals are
challenging to keep guidelines up to date. a major factor in the provision of nuclear medicine pro­
cedures that are clinically necessary. Barriers to accessing
Quality systems radiopharmaceuticals include an absence of coordinated
The provision of safe, high-quality imaging services supply (especially in LMICs), transportation issues,
depends on the control of several variables, including inadequate facility infrastructure, and little appropriate
infrastructure, staffing, regulatory environment, quality staff training and availability. The provision of essential
control of instruments, compliance with national nuclear medicine procedures for patients with cancer
regulations for patients’ and workers’ safety, and for the therefore requires a health system and regulatory
conduct of imaging studies according to appropriate framework that facilitates access to radiopharmaceuticals,
clinical need. This frame­work requires the identification as well as the infrastructure and trained staff needed
of quality policies and objectives, and the production of to do these procedures.110 In this context, the local
a documented system with clearly defined processes, production of radiopharmaceuticals for immediate
procedures, and responsibilities. Such a system is injection should not necessarily require facilities that
usually referred to as a quality management system, meet Current Good Manufacturing Practice standards in
and its purpose is to help direct activities to meet full, but the radiopharmaceuticals should undergo
patient and regulatory requirements and to continually appropriate quality control before administration.110
improve the effectiveness and efficiency of the imaging With regard to the radiation protection of patients and
service. Typically, a quality manage­ment system also workers, the safety of the public and of family members
provides a platform to identify areas for improvement. should also be considered.112 Many nuclear medicine
The IAEA has developed quality management audit procedures are done on an outpatient basis and the
methods for nuclear medicine (QUANUM)105,106 and exposure to the public and patients’ families after a
radiology (QUAADRIL),107 which facilitate the adoption procedure needs to be considered.31 The mitigation of
of quality policies in medical imaging departments. this risk includes educating the patient on how to reduce
The programmes cover all aspects of medical imaging, the risk of public and family exposure to the ionising
including management, radiation regulations and radiation from the radiopharmaceuticals that have been
safety, radiation protection of patients, quality control administered to the patient for the diagnostic test or for
of instruments, operations and services, diagnostic radionuclide therapy.113
clinical services, and radiopharmacy. The European
Society of Radiology has also published guidance on Protecting patients and health-care workers when
clinical audits.108 using MRI
In contrast to imaging procedures with ionising radiation,
Radiopharmaceuticals and targeted therapy there are few comprehensive data in the field of MRI. The
Radiopharmaceuticals are radiolabelled compounds that, number of workers involved in MRI worldwide is
once administered to the patient, are incorporated into unknown, although the safety of health-care workers
cells or tissues to provide diagnostic information or to involved with MRI is an important area of consideration.
trigger a therapeutic effect. These unique molecular In particular, for some types of MRI procedures, the
tools, which are indispensable for the practice of nuclear occupational exposure of health professionals to the
medicine, need to be prepared shortly before being magnetic fields can be substantive, and requires
administered to patients, because of the short physical considerable protective measures, especially in the case of
half-life of the radionuclides used. Most radiopharma­ high and very high magnetic fields. Workers’ protection
ceuticals that are used for diagnostic and therapeutic has been comprehensively addressed in the directive
purposes are dosed in subpharmacological quantities 2013/35/EU of the European Parliament114 on the health
of ligand attached to radioisotope, such as ¹⁸F-fluoro­ and safety requirements regarding the exposure of
deoxyglucose for PET imaging or ¹³1I-meta­iodo­benzyl­ workers to the risks arising from physical agents
guanidine for imaging and therapy of neuroblastoma, (electromagnetic fields) and is also mentioned in some
thereby avoiding clinically relevant drug-related side- national and professional guidelines.115
effects. According to the international pharmacopoeia, MRI safety is mostly dominated by the interaction of
radio­pharma­ ceuticals are defined as medicinal implanted devices with the different magnetic fields used

e155 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

to make the images. Therefore, it is of utmost importance patient workflow and logistics. Furthermore, the growth
to have a policy to assess the safety of medical implants of wireless technologies (mobile phones and other
and devices before MRI (eg, cardiac pacemakers, vascular wireless devices that acquire and transmit data) is opening
clips in the brain, neurostimulators, cochlear implants, new possibilities for innovation in health-care delivery.
medication patches, and delivery pumps); access to an Indeed, according to WHO, mobile health, which might
updated list of device magnetic compatibility is necessary. be defined as the application of mobile phones or other
Guidelines for the safety of patients undergoing MRI wireless devices for medical or public health purposes,
procedures are necessary at an institutional and national could potentially transform health service delivery around
level, with some countries developing standards that can the world.124 Advances in digital sciences promise to
be used by LMICs.115 reduce the cost and improve the deployment of cancer
MRI protocols should be integrated within clinical sites imaging in both high-income countries and LMICs.
that use this imaging method. Furthermore, safety culture Although digital technologies are gradually replacing
developed in the field of ionising radiation should be existing established structures in high-income countries,
expanded to the use of MRI, even if the health effects of LMICs with less developed digital infrastructures are in a
ionising radiation and MRI are fundamentally different. unique position to implement digital technologies from
Radiation regulatory bodies do not always consider MRI the start, and therefore possibly at a faster pace. For
and, in general, the safety of MRI is mostly a concern of example, in some LMICs, mobile phone systems have
labour organisations in the general context of medical already superseded communication with traditional
and non-medical magnetic fields. The establishment of a landlines for health telecommunications124 and mobile
legal and regulatory framework for magnetic fields would health is already used for cancer screening.125 Mobile
be helpful, provided medical applications are considered teleradiology, in particular, is a branch of mobile health
separately from non-medical use. The involvement of that makes use of mobile phone technology to provide
professional medical bodies in this endeavour is specialist expertise in image interpretation. Mobile
considered essential. The potential benefits of such a teleradiology refers not only to radiology and nuclear
framework for LMICs would be substantial, and ensure medicine specialists providing services remotely, but also
patient and worker safety in MRI facilities. to communication with the patient via telemedicine
Safety processes are fundamental in the daily life of visits—a strategy that has been used in high-income
MRI facilities, and mostly involve the screening of countries and has expanded markedly during the
patients for implanted devices and avoiding the missile COVID-19 crisis. In LMICs, the dissemination of
effects of ferromagnetic objects in the MRI scanner technology for telemedicine (including teleradiology)
room, which can harm both patients and staff members. would not only help with the COVID-19 crisis and future
The use of quality management systems should be pandemics, but would also help more generally to
increased and incentivised. provide country-wide care, lessening the need for travel
Specific attention should be paid to pregnant women. to medical centres. Hospital stakeholders in LMICs need
Although no harmful fetal effects of MRI on pregnant to overcome many hurdles, because they first need to
workers are known, some national authorities recom­ assess information technology infrastructure, internet
mend avoiding any magnetic exposure during pregnancy. access, and the electricity supply to establish appropriate
Staff at MRI facilities should be educated and incentivised regional goals that leverage technologies that are easily
to develop a safety-oriented culture, based on published accessible, affordable, and user-friendly, and at the same
guidelines, so that near-miss events are shared and used time guarantee patient privacy. According to a 2016 WHO
for process improvement.116 survey, only 28% of lower-middle-income countries and
30% of low-income countries had legislation for the
Section 6: the potential of advances in digital protection of eHealth data, as opposed to more than 80%
sciences and device engineering for improving of high-income countries.126 Nevertheless, progress is
cancer care in LMICs being made, at least in some eHealth areas: the
Unprecedented advances in computing, data science, implementation of e-learning, for example, has already
information technology, and engineering in the last enhanced access to self-learning modules and video
decade are affecting all aspects of health care, including conferences in many LMICs.127
radiology and nuclear medicine.117,118 For example, in In this section is a discussion of various digital
cancer imaging specifically, artificial intelligence (AI) and technologies that hold particular promise for advancing
its subfields, machine learning and natural language cancer imaging in LMICs, now or in the future. It should
processing, have been used to assist in clinical diagnosis be noted that the infrastructure required to implement
and outcome prediction in various ways, including many of these technologies includes electronic medical
tumour detection and characterisation, and for the record (EMR) systems. Although EMR systems are
identification of cohorts of patients who require vigilant widely used in high-income countries, their distribution
monitoring.119–123 Novel analytical techniques based on AI in LMICs is less pervasive. Additionally, although more
are also being implemented to tackle unmet needs in than 50% of upper-middle-income and high-income

www.thelancet.com/oncology Vol 22 April 2021 e156

 The Lancet Oncology Commission

patients with cancer. In many LMICs, hospital systems

Panel 3: Use of ultrasound in low-income settings continue to function in the analogue world, with digital
In the past two decades, the widespread adoption of smartphone technology has image data often only available in private practices.
facilitated the near-ubiquitous availability of powerful computation and high-resolution However, where hospital systems in LMICs are able to
displays in such devices. Ultrasound manufacturers have leveraged the availability of invest in high-quality digital image data, then connectivity
these technologies to create a new class of low-cost mobile health (mHealth) portable between imaging sites can assist with technical queries
devices (ie, ultrasound probes) that connect directly to consumer electronic devices and enhance the quality of acquired image data.124 The
(smartphones). New ultrasound transducer technologies that mitigate the frequency imaging systems must be installed according to protocols
limitations of piezoelectric crystals permit a single transducer to be used for several that meet the standard of care in high-income countries
clinical applications.131 In combination, these technologies have vastly increased the and local health-care professionals, including tech­
availability of medical ultrasound at the same time as reducing its cost. Medical nologists, nurses, and pharmacists, must be adequately
ultrasound is routinely available in low-income and middle-income countries (LMICs), trained in using these systems.
where its central use is for oncological diagnosis and monitoring in the female pelvis, The availability of any imaging devices in LMICs is
thyroid, liver, breast, peritoneal cavity, and kidneys, and is commonly also used for biopsy often restricted by cost; hence, innovative technologies
and tumour ablation guidance. For example, mHealth devices are facilitating a have been used to create next-generation scanners that
competency-based training programme that enables Nigerian radiologists to do breast are less expensive to purchase and operate and have
biopsies guided by ultrasonography, which are the standard of care in high-income mobile capabilities. The development of these tech­
countries and are recommended by the Breast Health Global Initiative for many LMICs.132 nologies has required collaboration between industry
This project was started in Nigeria in 2020, because it is the most populous country in and academia, and has immediate relevance for their
Africa, with the highest rate of breast cancer mortality.133 Furthermore, the Nigerian implementation in LMICs. The average hospital-grade
Government is committed to cancer control, with more than 350 available radiologists ultrasound unit can cost more than a hospital’s annual
nationwide, a number similar to that found in other African countries. This work was capital budget and often serves as the primary diagnostic
done with the African Research Group for Oncology (ARGO), a National Cancer Institute- imaging modality in many LMICs. The more than
recognised cancer consortium that aims to improve outcomes for patients with cancer in 65-times disparity factor between high-income and low-
Nigeria. In 2017, none of the ARGO radiologists were able to do an ultrasonography- income countries in the number of CT installations, as
guided breast biopsy because they had not been trained for it. indicated by the IAEA IMAGINE data and mentioned
earlier,51 is therefore unsurprising. A relevant factor in
The project’s first step was a multidisciplinary assessment of the needs of local
this context might be that most (>90%) high-income
stakeholders, which identified a need for and favourability towards an mHealth-based
countries rely chiefly on the public funding of eHealth
ultrasonography-guided biopsy training programme in Nigeria.132,133 The local
programmes, whereas in the majority of low-income and
stakeholders included surgeons, radiologists, and pathologists, because the proposed
lower-middle-income countries (70%), donor funding is
change in practice was feasible only with multidisciplinary support. The training
the dominant source of support.126 This difference in
programme approach was competency-based and included instructor-led and e-learning
commitment by governments might affect middle-term
modules, as well as simulation-based training. This approach enabled independent
and long-term strategic goals and investment decisions
learning and provided users wtih access to newly developed artificial intelligence
by stakeholders. This infrastructural deficit also greatly
applications, which helped in the successful training and clinical implementation of
restricts the use of available scanners for image-guided
ultrasonography-guided biopsies. The training programme is self-propagating and the
procedures, which is one reason why many LMICs
assessment metrics are being validated.
continue to rely on blind (non-image guided) or surgical
biopsies for cancer diagnosis. New, innovative, low-cost
countries have adopted national electronic health record solutions, such as handheld mobile health ultrasound
systems that are based on EMRs, adoption rates in lower- devices that are used at the point-of-care, now offer a
middle-income and low-income countries are much safe, simple, and sustainable solution toward building
lower, at 35% for lower-middle-income countries and capacity for cancer control in LMICs. For example, new
15% for low-income countries.126 However, open-source ultrasound transducer technologies mitigating the
EMR platforms have been used in dozens of countries in frequency limitations of piezoelectric crystals131 permit a
Africa, Asia, and Latin America,128 and as the implemen­ single low-cost, portable transducer to be used for
tation of eHealth solutions in LMICs is a key factor in multiple clinical applications (panel 3).132,133
improving health outcomes, novel approaches for Advances in the design of x-ray sources, detectors, and
providing low-cost, easily accessible electronic health reconstruction algorithms have made possible the
records are a major focus of governments, international potential for motion-free, completely solid-state CT
bodies (eg, WHO), and industry.126,129,130 scanners.134 Compared with standard CT scanners, these
scanners promise to be less expensive, and easier to
Imaging technology and image acquisition: mobile and transport, assemble, and service, owing to the
low-cost imaging equipment elimination of moving parts in the CT gantry, which will
The acquisition of high-quality digital image data is a be ideal for use in LMICs. Specialised MRI systems that
prerequisite for accurate diagnosis with any of the have been developed that use permanent magnets
imaging technologies used in the management of instead of superconducting or resistive electromagnets

e157 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Registration Streamlined imaging Radiologist review and interpretation

The entire process is automatic. Biosensors for optimal image Voice-activated point-of-care based on relevant
Medical information is analysed acquisition and standardisation clinical and pathological information;
from radio-frequency identification triage by artificial intelligence, with the use of
abnormal findings, and display of key images
with quantitative actionable information

Patient
arrives

Protocoling Postprocessing Consultations and integrated diagnostics

Protocol is automatically personalised Automated extraction of information Radiologist in person, or via telemedicine
for each individual on the basis of clinical from abnormalities to allow a complete with the use of virtual reality or augmented
indication, risk factors, and body type assessment of the disease reality (eg, holograms), discusses clinical
implications of imaging results with patient
or clinical team

Figure 10: Artificial intelligence-driven workflow for imaging in patients with cancer
An illustration of a streamlined, artificial intelligence-driven imaging workflow, in which digital technologies enable the automation, standardisation, and optimisation
of every step, from patient registration to imaging acquisition and interpretation.

will enable low-cost, portable, and point-of-care MRI quantitative imaging features are affected by the vendor-
scans.135 Although the resulting field strength (<0·3T) is specific settings and image acquisition protocols, AI-
lower than that of standard 1·5T MRI scanners, based approaches for standardised image analysis are
advances in hardware design and reconstruction currently being investigated.142 With MRI as an example,
algorithms have made the use of low-field MRI scanners figure 10 presents a vision of a streamlined, AI-driven
possible, particularly for niche applications such as workflow, in which digital technologies enable the
brain imaging.136 Such scanners promise to be automation, standardisation, and optimisation of every
lightweight, low cost, and portable, enabling them to be step, from patient registration to imaging acquisition
deployed more readily than standard MRI scanners in and interpretation.
LMICs. Similarly, other technological advances include
PET systems with scalable ring configurations, which Patient registration and protocoling: improvement of
reduce costs while maintaining diagnostic capabilities.137 patient safety with radio-frequency technology
LMICs looking to invest in these new technologies need Radio-frequency identification (RFID) technology has
to be informed about the type of regional support that is been commercially available in one form or another since
available, and partnerships between manufacturers, the 1970s, but it has only recently been introduced into
governments, and private providers in LMICs will be health care. RFID is a wireless system of communication,
required to ensure that equipment can be maintained whereby tags containing patient data transmit that data
and operational for routine patient access and avoid through radio waves, which can be picked up or read by
scenarios where longer downtime might occur. New AI- stationary or portable devices.143 Many health-care device
based approaches will reduce—or in some cases manufacturers are incorporating RFID technology into
eliminate—the need for in-person equipment services, their workflow solutions. Similarly to the way contactless
will monitor quality and safety, and will also allow more payment services that have become standard in the
information to be extracted from imaging examinations, consumer economy allow efficient, convenient, and safe
because digital imaging data could be analysed not just financial transactions, contactless patient identification
qualitatively but also quantitatively. AI-based approaches and registration by means of RFID is expected to improve
for optimising imaging include the use of biosensors the workflow, patient safety, and patient experience.144
(eg, for MRI and PET scanners) that automatically Prerequisites for the use of RFIDs are a compatible
adjust for patient bodyweight and anatomy, optimise Hospital Information System and EMR system. A key
coil positions, and analyse heartbeat and breathing advantage to the use of RFIDs and accessible EMRs is the
rhythm to correct for body motion.138 Furthermore, AI- improvement of patient safety through the prevention of
based image reconstruction algorithms are fast and can human error,145 including the failure to recognise a
suppress noise and artifacts and produce higher-quality predisposition to a contrast media reaction, the need for
images, as shown in CT,139 MRI,140 and PET.141 Because premedication, or the presence of an implantable medical

www.thelancet.com/oncology Vol 22 April 2021 e158

 The Lancet Oncology Commission

device that precludes a patient from undergoing high- science, with the main goals being automation, improved
field MRI examinations. Another advantage is that, with accuracy, and decision support.148–153
the help of RFID technology, amendments to national or Computer-aided detection systems have been applied in
global safety guidelines can be implemented automatically different cancer types and organs or tissues, most
after approval by a central health-care authority, thereby extensively for lung nodules and breast cancer.122,123,150,153–156
enabling the application of safety standards that are Although commercial solutions have been available for
uniform throughout a country. An additional important several years, widespread clinical implementation is still
benefit of modern digital technology is the potential of AI pending. This situation is likely to change as positive and
to manage, predict, and reduce patient exposure to negative predictive values improve with the amount of
ionising radiation and thus further contribute to model complexity and generalisability, as offered by novel
improved patient safety.146 AI-driven approaches that use mathematical patterns
Further advantages can be found in the use of RFIDs extracted from imaging data—the so-called radiomic
and EMR information to directly guide image acquisition features. Because the application of deep learning
to tailor imaging protocols to a particular type of cancer algorithms to cranial CT has been shown to allow for the
or clinical question, without the need for manual expert-level identification of findings that require urgent
interaction by a radiologist or a nuclear medicine attention (eg, haemorrhage and fractures),157,158 machine
physician, such as directing imaging protocols for learning algorithms could be used for the triage of
specific body areas. This approach enables the country- patients with cancer. For instance, machine learning
wide standardisation of imaging protocols that adhere to algorithms could be applied in lung cancer and breast
the latest versions of published expert guidelines, and cancer screening programmes in high-risk populations,
ensures that state-of-the-art imaging can be done in areas or in the follow-up of patients with cancer undergoing
and at institutions that do not have relevant specialists. surveillance after complete remission. In LMICs, such an
Finally, the use of RFIDs might reduce physical approach could help to address the gaps in expertise and
interaction between patients and health-care personnel, availability in rural, difficult-to-access areas where few
depending on the imaging test being done—a benefit trained radiologists are available to provide care,128,159 as
that is particularly valuable during the COVID-19 well as in situations where radiologists are overwhelmed
pandemic, with its obligatory physical distancing rules. by the volumes of images they are required to interpret.147
Notably, implementation of this type of technology is The same applies to ultrasound, which, for example, is
facilitated by a supporting legal framework, which is used extensively in LMICs to stage cervical cancer.160 The
often missing in most LMICs. As the 2016 WHO survey high operator dependence of ultrasound makes the
shows, policies or legislation to address patient safety absence of sufficiently trained experts even more
and quality of care are only in place in 10–20% of low- detrimental, so that deep learning algorithms, such as
income and lower middle-income countries, compared those which have been used to interpret thyroid, breast,
with almost 80% in high-income countries.126 and abdominal ultrasonographies,153,161,162 are expected to
have a substantial effect. For example, AI could be used as
Image analysis and interpretation: AI and machine a second reader to confirm accuracy or serve as a reference
learning to bring tertiary care image interpretation to standard. This application of AI could have immediate
community hospitals in LMICs applicability in LMICs where there are few radiologists
State-of-the-art diagnostic image analysis and inter­ and ultrasounds are often done by technicians and
pretation require digital imaging, lossless compression, nurses.147
and transfer with the use of picture archiving and Decision support represents another application of
communication system technology. Moreover, advanced computer-assisted image analysis, although this is still
workstations and screens are needed to view radiology experimental and therefore not yet in clinical use.163
and nuclear medicine images, which most facilities in On the basis of radiomic data, diagnostic confidence
LMICs do not have147 (often, a laptop serves as the could be improved for the interpretation of equivocal
diagnostic workstation and the radiology report is lesions that are difficult to characterise by human visual
handwritten and placed in the patient’s paper chart). perception. For instance, studies have suggested that
Additionally, the availability of an EMR system is highly radiomics can help to differentiate CNS lymphoma and
desirable for the effective management of imaging data, atypical glioblastoma multiforme on PET164 and MRI,165 or
but again, most LMICs do not have this system either. different types of gastric malignancies on CT.166 Notably,
Access to an open-source picture archiving and radiomic features can be extracted not only after the
communication system that is integrated with an open selection of a lesion by the radiologist, but also fully
EMR would provide crucial information for clinical automatically by AI algorithms such as the convolutional
decision making and possibly help to reduce costs. neural network U-Net, which segments lesions without
Advanced AI-based image analysis and interpretation are the need for human interaction.167 This use of AI,
among the most extensively investigated topics in however, requires powerful computing infrastructure,
radiology and nuclear medicine, as well as computer with especially powerful graphics processing units. In

e159 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

view of the reported association between molecular ensuring consistently personalised, evidence-based cancer
tumour phenotypes and radiomic features, these features management and optimised patient outcomes.
could possibly have a role as surrogate markers in LMICs
where genomic and molecular biomarkers are not readily Section 7: research and training
available and accessible.168–170 The use of radiomics to Research is essential to the formation of practices and
predict tumour phenotype is also an area of ongoing policies in cancer care; in fact, integrating research and
research, and further validation will be required before it teaching into clinical practice ultimately leads to improved
becomes part of the standard of care. care and better patient outcomes.172 Hence, research
should also be considered as essential to elevating practice
Integrated reporting and the promise of integrated standards and driving training and education in any
diagnostics institution. Although available resources, socioeconomic
An important goal in cancer imaging is the efficient issues, and health systems in high-income countries
production of integrated imaging reports, in which all differ vastly to those in LMICs, the integration of research
pertinent imaging and other patient data are accounted into clinical practice is no less important. The creation
for and combined. This process can be enhanced by AI. and support of LMIC-based research groups is a
For example, the use of natural language processing for precondition for setting research priorities that address
qualitative content extraction from routine clinical local situations, developing evidence-based practices
reports could provide radiologists and nuclear medicine uniquely suited to LMICs, and adapting evidence
physicians with relevant clinical information that can developed in high-income countries to an LMIC context.
be readily used during image interpretation.120,149 The Research requires data, and the acquisition of prospective,
automated extraction of quantitative metrics (eg, PET complete, and accurate data is a challenge in many
standardised uptake values) and derivation of changes settings. The provision of cancer care, including the
over time could also enhance and accelerate image associated imaging services, in LMICs, should be
interpretation. Radiologists and nuclear medicine continually assessed to establish patient outcomes and
physicians might then integrate all of this information gaps in care. Many of these gaps could relate to imaging,
into final reports to better assist referring clinicians with either poor availability or suboptimal quality, but
regards to patient management decisions.171 continual prospective data collection can help to design
There is an unmet need to condense the wealth of interventions to overcome these challenges. This data
medical diagnostic data produced during routine patient collection can be viewed as part of the spectrum of
tests into a form that retains and emphasises all clinically implementation research, and is crucial in these settings.
relevant information. Efforts to develop this novel, holistic
approach, termed integrated diagnostics, strive to provide Evidence-based research
a digital framework for combining imaging, pathology, Clinical trials are essential to the evolution and
laboratory, genomic, and other diagnostic and clinical data development of cancer treatment. Trials are increasingly
to give clinicians easy access to aggregated information. being done for novel radionuclide therapy, interventional
A prerequisite for integrated diagnostics is the collection radiology, and diagnostic imaging studies, and these
and aggregation of digitally structured big data118—for imaging approaches also serve to evaluate treatment
example, through the use of electronic health records. In response and disease progression as study endpoints for
practice, the first step in applying integrated diagnostics to treatment efficacy and decision making.173–176 For cancer
an individual patient would be the extraction of all the trials of solid tumours (phase 3 trials especially),
relevant types of clinical and diagnostic data from that conventional CT size measurements by Response
patient in digitised form. The second step would be Evaluation Criteria In Solid Tumours (RECIST) are used
the visualisation and integrated display of the data on a in the vast majority of evaluations, although different
single dashboard. The final step would be the use of criteria might be used for modern technologies,
computational data analytics to integrate the patient’s data including hybrid PET (eg, PET Response Criteria in Solid
in light of insights drawn from big data, and offer precise Tumours, and Deauville criteria) in some trials.177 Clinical
predictive and prognostic information on which to base trials can be extended to LMICs to evaluate LMIC-specific
clinical decisions and patient counselling. One of the pathologies or to do multicentre, multinational trials. An
substantial hurdles to the implementation of this vision of innovative approach could also be to pool data from
integrated diagnostics, even at elite institutions in high- several individual trials, including sites in LMICs, as has
income countries, is the need to be able to mine clinical been proposed for data obtained from trials in patients
notes digitally—a process for which natural language with COVID-19.178 High-income countries are working on
processing will be a key tool. However, with natural major training programmes, for example in nuclear
language processing technology quickly evolving, and medicine, to establish cooperative trial networks and site
with the growing need to streamline information resulting validation processes.179 Such programmes, extending
from the rapid increase in the complexity and volume of from high-income countries to LMICs, advance the goal
patient data, integrated diagnostics is the best hope for of population-based evidence for new indications and

www.thelancet.com/oncology Vol 22 April 2021 e160

 The Lancet Oncology Commission

data registries, which is essential for health technology and supports the development of quality-assured clinical
assessments. research in LMICs. Furthermore, the programme allows
The introduction of new health-care technology, for cross-specialty research collab­
orations (panel 4;
including imaging, should be evidence based, and figure 11).184 Other grant funding bodies include the
systematic evaluation of its effect and cost-effectiveness Medical Research Council (UK), The Bill and Melinda
should inform policies related to technology in health Gates Foundation (USA), and the Wellcome Trust (UK).
care.179 Health technology assessments can be initiated
in high-income countries and adapted for submission to Research, education, and training
LMICs with the use of local country health systems and The establishment of a research culture in imaging
cost information. Evidence-based assessments of new departments is essential, and requires institutional
imaging (and radionuclide therapy) indications arising commitment, dedicated leadership, and exemplary role
from high-income countries could arguably be made models; these aspects are highly relevant in both high-
available for regulatory approval and funding in LMICs income countries and LMICs. Research should be
to avoid duplicating trials or health technology integrated into training programmes. Research
assessments in multiple countries. Additionally, policies structures within LMICs should include a well-
that have been successful in high-income countries organised policy framework that facilitates research,
should be evaluated in the context of LMICs and subject and the provision of appropriate infrastructure for
to relevant science and research. Different approaches research. The provision of protected research time,
for the integration of imaging into cancer care might although challenging in a busy clinical practice
well be needed, parti­ cularly in the context of low- environment, should be prioritised in LMICs, where
resource settings. time constraints represent a substantial barrier to
research activities. A special priority should be placed
Global health research on implementation research, which is essential to
LMICs carry the highest burden of cancer globally.41 translate research from high-income countries to
However, most of the world’s research funding originates clinical practice in LMICs. Currently, the research
in and is distributed to high-income countries, both infrastructure in many LMICs is either weak or non-
for adult and childhood cancers.5,180,181 This situation existent. There is frequently little or no in-country
influences the development of new imaging technologies, expertise in clinical and implementation science
radio­pharmaceutical innovation, and analytic approaches research, and although increasing funding sources are
(eg, AI), which require essential infrastructure and encouraging, personnel should be hired and dedicated
expertise to generate and implement novel approaches to to cancer research to begin the process. Continuing
imaging. Global health research fosters collaboration reviews and quality assurance and audit programmes
between high-income countries and LMICs and provides should be integrated into the routine activity of imaging
opportunities to address global health disparities, departments. These endeavours can form an important
accelerating the development of therapeutics and building research activity that is often underemphasised and
research capacity in LMICs. The overarching goal is to might include assessing the accuracy and consistency of
foster independence and promote professional develop­ reports, quality and safety studies, workflow, and unique
ment in LMICs to sustainably develop resources and practices to improve the quality of imaging services and
capacity, expand access to cancer imaging, and provide cancer care in general.
affordable and high-quality cancer care. In addition, Education and training activities in LMICs can extend
global research initiatives provide an opportunity to not from country-based programmes to overseas attach­
only assess resource-sparing approaches, but also to ments, distance learning, online didactic lectures, and
implement new techniques in LMICs in a real-world workshops. With the support of digital technologies, the
research setting that is controlled to allow for an in-depth transmission of images for training in image
and unbiased assessment of these techniques. Several interpretation can also be facilitated in LMICs, and this
grant funding bodies have dedicated funds to global might be combined with practical training in local
health research; for example, the National Institutes of facilities in a blended learning approach. For example,
Health offer international research training grants that tele-ultrasound training by real-time image inter­pretation
sup­port research training programmes that develop and and guidance from experts from afar has been shown to
strengthen the scientific leadership and expertise needed be feasible and of value in training and patient
for research in LMICs. Global research from patterns of management in the LMIC setting.185 Many international
care studies to randomised phase 3 trials are funded professional imaging societies have organised outreach
and done through the IAEA coordinated research program­mes to LMICs for this purpose, including the
programme.182 The pro­ gramme facilitates research Society of Nuclear Medicine and Molecular imaging,
collaboration between high-income countries and LMICs the European Association of Nuclear Medicine, the
in medical disciplines that use radiation (eg, nuclear Radiological Society of North America, the European
medicine, radiology, radiotherapy, and medical physics) Society of Radiology, and the World Federation of

e161 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Panel 4: Research and training support for low-income and middle-income countries (LMICs)
To improve outcomes for patients with cancer, LMICs should Bangladesh, Bulgaria, Mexico, Montenegro, Morocco,
support the development of workforces suited to and Thailand, and worked with faculty from institutions in
contemporary practice in imaging and nuclear medicine. Australia, Belgium, Italy, the UK, and the USA. The related core
Many meaningful initiatives by governments and professional research projects assessed the effectiveness, applications, quality,
organisations around the world have been implemented, optimisation, and safe use of advanced imaging techniques.
with the most comprehensive global coordination of such The students learned how to do advanced clinical research and
programmes undertaken by the International Atomic Energy implement practice and quality improvement strategies.
Agency (IAEA) since 1987. A primary mission of the IAEA is to The research measurably enhanced local and national training
promote and support research on the practical applications of programmes and improved the clinical practice of advanced
atomic energy and related techniques for peaceful purposes imaging in radiology and nuclear medicine in the researchers’
worldwide, including in health care, with a particular focus on home countries.
LMIC member states. The challenges of doing such work in Another CRP aimed to improve the clinical applications of
LMICs include insufficient resources (human and PET–CT in LMICs. This project included an international study on
infrastructural), an absence of training in clinical research, the use of PET–CT for stage III non-small-cell lung cancer
and underestimation of participant countries’ own capabilities radiotherapy planning (the IAEA-PERTAIN study) that involved
to support projects. Through the IAEA Coordinated Research more than 350 patients in LMICs including Brazil, Estonia, India,
Activities platform, pertinent activities and plans to strengthen Jordan, Pakistan, Turkey, Uruguay, and Vietnam.183 Following
health systems are initiated, supported, and coordinated rigorous and comprehensive training from hands-on courses,
between LMICs and high-income countries. Through webinars, and participant feedback, knowledge and skills were
well-designed, multicentre, international research protocols, successfully transferred to study sites for the delineation of
participants are supported in their work to develop and radiotherapy target volumes, and a study on the effect of
contribute to local research and autonomously implement PET–CT in radiotherapy planning on 2-year survival rates was
quality improvements. done. Additional outcomes included the development of
So far, approximately 100 coordinated research projects (CRPs) in guidelines for PET–CT in image acquisition and target volume
the field of nuclear medicine and diagnostic imaging have been delineation, the adoption of new protocols, and changes in
initiated, with more than 1000 research institutions clinical practice. Instrumental to the success of CRPs was the
participating. These collaborative strategies aim to engage LMICs accreditation of ¹⁸F-fluorodeoxyglucose-PET–CT studies by
in well-designed, international, multicentre clinical trials, to means of quality control and quality assurance measures by the
address the most relevant scientific questions, including those European Association of Nuclear Medicine Research.
that are specific to LMICs, and to improve daily clinical practice. This accreditation was provided through the collaboration of
In nuclear medicine and diagnostic imaging, projects range from the European Association of Nuclear Medicine with imaging
workforce training for advanced imaging modalities, to scaling facilities in the target countries. Local trainers were trained, and
up the local applications of advanced imaging modalities, such as their experience and expertise were subsequently disseminated
PET, to addressing specific types of cancer prevalent in LMICs. through seminars and conferences. This CRP also fostered
The worldwide distribution of countries active in the IAEA’s CRPs multidisciplinary training and skill development on contouring
devoted to addressing health conditions is illustrated in figure 11. with the use of PET–CT for radiation oncologists and medical
CRPs also support the optimal supervision of research by imaging specialists alike. Successful CRP examples such as this
postgraduate students in LMICs. For example, a doctoral CRP in one are amenable to being applied in other LMICs and tailored
advances in medical imaging techniques linked PhD students on to their local contexts. Future programmes will address areas of
a medical physics course from LMICs with faculty supervisors unmet need, including updates on the use of diagnostic
from degree-conferring institutions in high-income countries. imaging in LMICs, the application of digital connectivity and
Students were selected from LMICs across the globe, including artificial intelligence, and theranostic techniques.

Paediatric Imaging, among others, who also provide Section 8: scaling up capacity for sustainable
online education on their websites. Furthermore, inter­ access to cancer imaging diagnostics—a call to
national organisations, including WHO and the IAEA, action
regularly reach out to LMICs to provide training and This Commission has identified several important chal­
education in radiation safety and skillsets required for lenges hindering access to effective services for cancer
establishing imaging facilities. These activities are imaging diagnostics, especially in LMICs; these
essential to ensuring that radiologists, nuclear medicine challenges include inadequate investment in imaging
physicians, and other imaging professionals gain equipment, a low workforce capacity, an absence of
practical education and training, and enhance the quality digital technology including electronic clinical data, poor
of imaging studies done in LMICs. access to radio­ pharmaceuticals, and a deficiency in

www.thelancet.com/oncology Vol 22 April 2021 e162

 The Lancet Oncology Commission

Number of coordinated research projects

>28
25–27
21–24
17–20
13–16
9–12
5–8
1–4
None

Figure 11: Active International Atomic Energy Agency coordinated research projects in human health
The map was produced by the International Atomic Energy Agency (Vienna, Austria) and is included here with permission.

research and training. We have also presented new and metrics in global health statistics and country progress
compelling evidence on the substantial health and monitoring is essential.
economic benefits of scaling up cancer imaging The second crucial success factor relates to the
diagnostics in LMICs, where they are most needed and development of a compelling case for investing in the
where the widest inequities exist in access to effective scale-up of cancer imaging diagnostics. The results of this
cancer services and in cancer outcomes. These benefits Commission show that such investments can yield
will be greatest with a com­prehensive approach to scale- substantial health and economic benefits. Now that clear
up, where the scale-up of diagnostic capacity is aligned evidence of an investment case exists, a straightforward
with treatment capacity and where there is a simultaneous narrative should communicate the benefits of investment
improvement in quality of care. for individuals, households, and countries, and the
In this section, we examine crucial success factors for potential opportunities provided by imaging diagnostics
scaling up, the roles that key stakeholders could play in for patients with cancer worldwide.
the scale-up process, and targets that will help to translate The third crucial success factor relates to alignment.
aims into actions and accomplish the vision of an effective Activities aimed at the scale-up of services for cancer
and equitable scale-up of cancer imaging diagnostics in imaging diagnostics align with global efforts to
LMICs. achieve Sustainable Development Goals. In particular,
the health-related Sustainable Development Goal 3,
Crucial success factors for scaling up cancer imaging “Ensure healthy lives and promote well-being for all at
diagnostics all ages,” has set the achievement of UHC by 2030 as
The challenges and opportunities in the global fight the target.187 Global efforts to scale-up cancer imaging
against cancer and crucial success factors for an effective diagnostics should be fully aligned and integrated with
response with comprehensive scale-up have been actions aimed at achieving UHC. The alignment of
outlined in earlier studies.6,40,186 the expansion of imaging diagnostics with UHC will
The first crucial success factor is strong and visible require a compre­hensive approach to scale-up, where
leadership, at both a global and country level. International the scale-up of diagnostic capacity is aligned with a
development agencies, global leaders, and governments scale-up in treatment capacity. This alignment will
with commensurate funding should firmly commit to optimise the use of available resources in countries,
scaling up imaging diagnostics capabilities. Additionally, help to strengthen health systems, ensure a more
the inclusion of medical imaging and nuclear medicine strategic approach to the provision of diagnostic

e163 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Panel 5: An inclusive global coalition to scale up capabilities for diagnostic cancer imaging in low-income and middle-income
countries
An inclusive coalition of partnerships and networks is essential Audit for Diagnostic Radiology Improvement and Learning
for the development of an effective global-level and country- (QUAADRIL; for radiology) have been instrumental in supporting
level response to the scale-up of cancer imaging diagnostics. quality programmes in many countries, including LMICs.105
All actors involved in the scale-up—such as governments, civil
Civil society
society, affected individuals, health professionals, professional
Civil society involvement is crucial for bringing a voice to those
associations, researchers, funders, international agencies,
affected by cancer, building awareness at the global and national
the private sector, and innovators—bring capabilities that can be
levels, and mobilising support for concerted action. Civil society
harnessed to create synergies in the scale-up process.
has an important role in articulating health rights, and
Governments influencing global actors and country-level policies to help to
Governments can use the evidence generated by this Commission include cancer imaging diagnostics as an integral part of UHC
to convene relevant stakeholders and coordinate investments in expansion. The Union for International Cancer Control, which For details on the Union for
diagnostic imaging services for patients with cancer as part of the has brought together more than 1000 non-governmental International Cancer Control
see https://www.uicc.org
efforts aimed at the expansion of universal health coverage (UHC). organisations involved in cancer, is well positioned to strengthen
Governments are needed to provide leadership and make political civil society and help to mobilise global leaders through the
and fiscal decisions to invest in health systems that generate World Cancer Summit and the World Cancer Declaration.
health and economic returns for their citizens and economies.
Professional associations
International agencies Professional associations are important for establishing
International agencies, such as WHO, can be integral in the professional standards, developing capacity, expanding access to
incorporation of cost-effective imaging diagnostics into high-quality health-care services for patients with cancer, and for
essential diagnostics lists, in that these agencies support their the appropriate use of imaging technologies (eg, the American
inclusion as part of benefits packages for UHC. The WHO Best College of Radiology, the American Society of Clinical Oncology,
Buys list for non-communicable diseases188 and the WHO priority the American Society for Radiation Oncology, the European
medical devices list43 include diagnostic imaging, and imaging is Society for Medical Oncology, the Radiological Society of North
also included in a WHO publication on providing cancer care for America, the European Society of Radiology, the International
all.189 WHO provides leadership in the establishment of Society of Radiology, the International Society for Strategic
guidelines and policies on human health, including for cancer, Studies in Radiology, the European Society for Radiotherapy and
and in the implementation of programmes aimed at improving Oncology, the Society of Nuclear Medicine & Molecular Imaging,
access to essential diagnostics and treatment to reduce the the European Association of Nuclear Medicine, the Asia Oceania
burden of disease globally, particularly in LMICs. Federation of Nuclear Medicine and Biology, and the World
Global and regional development banks have a crucial role in Federation of Nuclear Medicine and Biology). These groups could
working with governments and the private sector to develop effectively contribute to and accelerate the scale-up of the
innovative financing solutions (see section 3) to enable the capacity for imaging diagnostics and access to effective imaging
expansion of cancer imaging diagnostics in LMICs. services in LMICs by working with international and country-level
partners to expand human resource capacity through education
The International Atomic Energy Agency (IAEA), an independent, and training, by providing clinical guidelines adapted to the LMIC For details on the IAEA see
intergovernmental, and technology-based, organisation within https://www.iaea.org
setting for the optimal use of imaging resources, and by
the UN family, is an important stakeholder in the scale-up of establishing or strengthening regional collaborations in research,
cancer imaging diagnostics in LMICs. As the focal point for development, and innovation.
nuclear cooperation worldwide, the IAEA works to promote the
safe, secure, and peaceful use of nuclear technologies, including Philanthropic organisations
diagnostic imaging and nuclear medicine. This agency provides a In LMICs, philanthropic organisations have been key in mobilising
wide range of support, which encompasses the provision of donations and public funding to establish academic cancer centres
equipment, education, and training; quality and safety of clinical that provide high-quality services to some populations. Many of
practice through guidance documents; equipment calibration; these centres have twinning arrangements with cancer centres in
and support of clinical and health economics research. Working high-income countries and provide an opportunity to integrate
with WHO and its International Agency for Research on Cancer, operations with publicly funded elements of health systems to For details on the International
the IAEA has undertaken fact-finding missions and imPACT establish integrated cancer networks. Such integration will help to Agency for Research on Cancer
see https://www.iarc.fr
reviews190 in more than 100 countries to assess their cancer create synergies to optimise the exapansion of access to care for
control, from national registries to palliation, including patients with cancer. A good example is the International Cancer
diagnostic imaging. In addition, IAEA quality assurance methods Research Centre in Kyebi, Ghana, which is being constructed by the
such as Quality Management Audits in Nuclear Medicine Eugène Gasana Jr Foundation. This state-of-the-art children’s
Practices (QUANUM; for nuclear medicine) and Quality Assurance (Continues on next page)

www.thelancet.com/oncology Vol 22 April 2021 e164

 The Lancet Oncology Commission

(Panel 5 continued from previous page) scale-up imaging diagnostics and expand access to effective
services.
cancer research centre will be aligned with the University of
Ghana Medical Centre, and the medical programme will be However, the private for-profit sector for health-care providers is
designed in cooperation with Memorial Sloan Kettering Cancer not well regulated in many LMICs, and there are few data on the
quality of services provided or the outcomes achieved. The private
Center in the USA. The facility is intended to serve as a centre of
excellence in cancer care for the continent of Africa. sector is also a major funder of research and development, and
innovation for diagnostics, medicines, and health technologies
The private sector for the management of cancer, but much of this effort is similarly
In LMICs, the private for-profit sector has created substantial targeted for high-income countries. Novel collaborations of
capacity for cancer imaging diagnostics, but generally only for public–private institutions, universities, philanthropic
those who can afford to pay for the services. The private sector organisations, and international development agencies could help
can use this experience to work with governments, to harness the private sector’s capability to develop affordable
international agencies, and philanthropic organisations to imaging diagnostics solutions for cancer in LMICs.
develop innovative financing and service delivery models to

services for cancer, and help with the sustainability of diagnostics with oncologists and other health
the scale-up. professionals to ensure quality standards and the
The fourth crucial success factor is the creation of appropriate use of medical imaging and nuclear
inclusive coalitions of partnerships and networks to drive medicine in clinical care is a key driver of improved
the scale-up of cancer imaging diagnostics (panel 5).43,105,188–190 outcomes of patients with cancer.
Such coalitions should involve, among others, civil society, At present, no clear, overarching global strategy for
individuals affected by cancer, professional associations, scaling up cancer imaging diagnostics exists in many
health professionals, researchers, funders, international LMICs, and efforts are often fragmented as a result.
agencies, the private sector, and innovators. A multistakeholder coalition should develop a global
Wide-ranging initiatives have emerged over the years to strategy for scaling up imaging diagnostics to ensure
expand the capacity for cancer care in LMICs by improving alignment with and the coordination of the many short-
clinical knowledge, increasing the amount and quality of term initiatives and pilot projects, which do not sus­
cancer care, and establishing research activities. These tainably address the shortcomings in access to effective
initiatives have been underpinned by collaborations cancer imaging diagnostics.
involving multiple stakeholders from LMICs and high- The fifth crucial success factor is investment in
income countries, typically through academic institutions research, development, and innovation to develop novel
that have established twinning arrangements (ie, technological solutions and service delivery models that
partnerships). For example, St Jude Children’s Research can rapidly address any shortages in human resources,
Hospital in the USA, a pioneer of this model, has infrastructure, affordable diagnostics, care models, and
established close collaborative relationships with two financing. For example, these initiatives could involve
dozen partner sites in more than 15 countries, including the expansion of the use of new, less expensive scanner
Brazil, China, Guatemala, Haiti, Jordan, Morocco, and the technologies through the wider application of digital
Philippines.191 To be successful, such collaborations should connectivity solutions that can enable radiologists in-
involve a two-way transfer of expertise, advice, knowledge, country or internationally to interpret scans remotely,
and skills, and be characterised by mutual respect between and through the use of virtual digital learning platforms
the local stakeholders and the international partners.192 to train and support health professionals. Investment in
However, although beneficial to those institutions research, development, and innovation will also enable
involved in the collaborations and patients accessing the the better application of evidence-based solutions, best
institutions involved in these collaborations, many such practices, and transfer of knowledge. The application of
initiatives have been small-scale projects; as such, they these innovative approaches can provide opportunities
have not always produced noticeable differences in the for the rapid and more affordable scale-up of the capacity
access to cancer services for a large numbers of citizens in for imaging diagnostics and digital health solutions in
LMICs, or made cancer outcomes more equitable at a LMICs.
population level. The sixth crucial success factor is the mobilisation
The implementation of multidisciplinary teams and better use of existing resources by optimising the
including oncologists, surgeons, radiologists, nuclear use of the existing health workforce, equipment, and
medicine physicians, and pathologists is necessary to infrastructure assets in countries through networks or
ensure the provision of high-quality care for patients collaboratives for cancer imaging diagnostics. These
with cancer. The establishment of collaborative networks networks or collaboratives could be operationally aligned
in LMICs that bring together experts in cancer imaging with cancer networks and include public, private, and

e165 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Panel 6: Major actions and targets

Action 1: incorporate imaging diagnostics into essential Action 4: ensure the provision of optimal access to effective
benefits packages when expanding universal health coverage imaging diagnostics by establishing collaboratives for cancer
(UHC) in low-income and middle-income countries (LMICs) imaging diagnostics
Cancer imaging diagnostics should be incorporated into national Countries should work with stakeholder coalitions to create
essential benefits packages for diagnostics when expanding national and regional collaboratives focused on cancer imaging
UHC, with explicit targets for the scale-up of capacity in health diagnostics, or to expand them where they already exist, to
systems to expand the coverage of effective services. better use available capacity for providing packages of effective
Target cancer services. These collaborations could be enabled through
By 2030, as part of the efforts to expand UHC, at least 80% of virtual digital linkages.
LMICs should incorporate appropriate cancer imaging Target
diagnostics in their essential benefits packages to expand access By 2030, establish collaborative networks of imaging diagnostics
to effective services. in 50% of LMICs to expand the coverage of effective imaging
diagnostics services for cancer.
Action 2: incorporate costed actions into national cancer
control plans to scale-up cancer imaging diagnostics Action 5: invest in education and training to expand human
Predictable financing is essential for the scale-up of cancer resources
imaging diagnostics and to sustain these services. LMICs should The establishment of a trained workforce of radiologists,
develop national cancer plans that are fully costed that establish nuclear medicine physicians, radiographers and technologists,
how sustainable cancer care could be progressively developed nurses, physicists, and radiochemists is essential to ensure that
and funded. safe and effective imaging and nuclear medicine services can
Target be provided and that quality systems provide accurate and
By 2030, 60% of LMICs should have national cancer control plans reliable information for cancer care. Digital solutions and
that specify actions for the scale-up of cancer imaging diagnostics, virtual platforms that facilitate the development of workforce
with the necessary fiscal space for funding this expansion. planning and training could enable the rapid scale-up of
training in LMICs.
Action 3: expand access to effective services for imaging
Target
diagnostics by scaling up the current capacity of human
By 2030, 80% of LMICs should establish plans for workforce
resources and imaging equipment
development and for the use of digital platforms for workforce
The ability of LMICs to improve health outcomes for patients
training.
with cancer depends on their ability to expand the availability of
imaging equipment and a suitable trained workforce to an Action 6: invest in training, research, development, and
amount that provides appropriate access for these patients. innovation to develop affordable cancer imaging diagnostics
The quantity of imaging equipment and human resources per in LMICs
million people in the population varies substantially in countries Research funding related to cancer imaging diagnostics in LMICs
of similar and different income groups. The difference in the is small, fragmented, and largely inaccessible to researchers
average and median amounts of imaging equipment and outside high-income countries. The absence of affordable
human resources per million people in the population ranges solutions for imaging diagnostics hinders the achievement of
from three-times to ten-times between low-income and lower- improved health outcomes for patients with cancer. Investments
middle-income countries, between lower-middle-income are needed in research and innovation in LMICs to ensure the
countries and upper-middle-income countries, and between better use of available interventions and create affordable and
upper-middle-income countries and high-income countries accessible imaging solutions and new care delivery models for
(see sections 2 and 3). patients with cancer appropriate for LMICs.
Target Target
By 2040, at least 50% of low-income, lower-middle-income, and By 2025, a US$100 million innovation fund for cancer imaging
upper-middle-income countries should expand the capacity of diagnostics should be established to improve the coordination
human resources and availability of imaging equipment to reach of funding for education, training, research and development,
or exceed the median amounts per million people in the and innovation in LMICs, with a target of mobilising and
population of that currently achieved in countries of the next investing thereafter at least $25 million per year.
income group up, adjusted for cancer incidence.

philanthropic institutions. The development of such could be augmented with the strategic purchasing of
networks or collaboratives requires careful planning at imaging diagnostic services by national authorities to
both the national and subnational level to ensure appro­ produce economies of scale and the equitable allocation
priate investment to address capacity gaps. Planning of available funds.

www.thelancet.com/oncology Vol 22 April 2021 e166

 The Lancet Oncology Commission

The findings of this Commission show the substantial Group Capital, InVicro, and pHLIP Technologies; equity interests in Telix
health and economic benefits of the successful scale-up of Pharmaceuticals, Evergreen Theragnostics, and pHLIP Technologies;
preclinical research support from Eli-Lilly, Sapience Therapeutics,
the capacity for cancer imaging diagnostics in LMICs and MabVax Therapeutics, SibTech, Thermo Fisher Scientific, Ground Fluor
high-income countries. These benefits will be the greatest Pharmaceuticals, ImaginAb, Merck & Company, AbbVie, Bristol-Myers
with a comprehensive approach to scale-up, where the Squibb, Genentech; fee-for-service work from Y-mAbs and Regeneron
scale-up of diagnostic capacity is aligned with treatment Pharmaceuticals; and income from licensed intellectual property from
Summit Biomedical Imaging, CheMatech, Elucida , Theragnostics,
capacity. The pathway to scale-up and the speed of the Daiichi Sankyo, and Samus Therapeutics, outside the submitted work.
expansion of imaging diagnostics for cancer in each AMS reports trial funding from Abbvie, EMD Serono, Isotopen
country will necessarily vary, given that the political will, Technologien München, Telix, and Cyclotek; research funding from
infrastructure, the availability of radiotherapy, surgery, Medimmune, AVID Radiopharmaceuticals, Adalta, and Theramyc; and
personal fees from Life Science Pharmaceuticals and Imagion, outside
medical treatment, imaging modalities, human resources, the submitted work. All other authors declare no competing interests.
and financing will be different in each country. However,
Acknowledgments
there are a set of actions that each country could take to We thank Ada Muellner and Garon Scott, both medical editors in the
enable scale-up. Department of Radiology, Memorial Sloan Kettering Cancer Center
We propose six main actions, with targets, to achieve (New York, NY, USA) for editing portions of the manuscript. HH and JSL
the important goal of equitable access to imaging were supported by a National Insitutes of Health National Cancer
Institute Cancer Center support grant (P30 CA008748). JSL was supported
diagnostics worldwide (panel 6). by a National Insitutes for Health National Cancer Institute grant (R35
CA232130). AMS was supported by a National Health and Medical
Conclusion Research Council grant (1177837). Harvard TH Chan School of Public
Health also provided funding support for this study. We thank the
Compelling evidence exists for the substantial health
following organisations for their financial or in-kind support: the African
benefits of scaling up medical imaging and access to Association of Nuclear Medicine, the American College of Radiology,
nuclear medicine for patients with cancer. Improvements the Association of Latin American Societies of Biology and Nuclear
in science have enabled rapid developments in affordable Medicine, the Australian and New Zealand Society of Nuclear Medicine,
the Asia Oceania Federation of Nuclear Medicine and Biology, the African
imaging technologies and solutions, and flexible, low-
Organisation for Research & Training in Cancer, the American Society of
cost digital platforms for virtual training. Science and Clinical Oncology, the Arab Society of Nuclear Medicine, the African
technology are not the barriers to a worldwide equitable Society of Radiology, the American Society for Radiation Oncology, the
scale-up of effective cancer imaging diagnostics; rather, Breast Cancer Research Foundation, the European Association of Nuclear
Medicine, the European Society for Medical Oncology, the European
achieving equitable scale-up is a matter of vision and
Society of Radiology, the European Society for Radiotherapy and Oncology,
will. Successful scale-up will result from effective political the Hong Kong College of Radiologists, the International Atomic Energy
leadership, active participation from all major stake­ Agency, the International Society for Strategic Studies in Radiology, the
holders, and the alignment of country-level and global International Society of Radiology, the National Cancer Institute, the
Pan-Arab Association of Radiological Societies, the Radiological Society of
efforts to expand access to medical imaging and nuclear North America, the South African Society of Nuclear Medicine, the
medicine, leading to better outcomes for patients with Society of Nuclear Medicine & Molecular Imaging, the Union for
cancer worldwide. International Cancer Control, the World Federation of Nuclear Medicine
and Biology, and the World Molecular Imaging Society. We thank the
Contributors
following individuals for their contributions: Giles Boland (Brigham and
RA, HH, MA-W, and AMS were co-leaders of the Commission and co-
Women’s Hospital, Harvard Medical School, Boston, MA, USA),
developed and co-wrote the study design with input from co-authors.
Juan C Bucheli (International Atomic Energy Agency, Vienna, Austria),
MML, DP, and MA-W co-conceived the commission and co-led the
Juliano Cerci (Diagnóstico e Terapia, Curitiba, Brazil),
International Atomic Energy Agency Secretariat, which convened the
Maria del Rosario Perez (World Health Organisation, Geneva,
commission. HH wrote section 1 with MA-W, MH, AMS, GM, MML,
Switzerland), Francesco Giammarile (International Atomic Energy
DP, and LNS. MML wrote section 2, with DP, MA-W, HH, and AMS.
Agency), Christian J Herold (Department of Medical Imaging and Image-
MML and DP accessed and verified the data in the IMAGINE database.
Guided Therapy, Medical University of Vienna, Vienna General Hospital,
RA wrote sections 3 and 4. RA also conceived the modelling approach
Vienna, Austria), André Ilbawi (World Health Organization),
for sections 3 and 4. RA and ZJW led the modelling and analysis for
Krishna Juluru (Memorial Sloan Kettering Cancer Center, New York, NY,
section 3, with input from AMS and HH. ZJW collated the data and
USA), Peter T Kingham (Memorial Sloan Kettering Cancer Center),
built the model. OH wrote section 5, with contributions from GF, MH,
Gabriel P Krestin (Department of Radiology & Nuclear Medicine,
JSL, DP, and AMS. JAB co-wrote section 6 with HH, MH, LD-B,
Erasmus MC University Medical Center Rotterdam, Rotterdam,
and AMS. P-LK wrote section 7 with AMS, JSL, MML, WJGO, and DP.
Netherlands), Ronilda Lacson (Brigham and Women’s Hospital, Harvard
RA wrote section 8 with input from HH, AMS, MML, and MA-W. RA,
Medical School), Marius Mayerhöfer (Department of Radiology, Memorial
HH, and AMS revised all sections of the report. All authors contributed
Sloan Kettering Cancer Center), Daniel Mollura (RAD-AID International,
and approved the final version of the submitted manuscript.
Chevy Chase, MD, USA), Tetiana Okolielova (International Atomic Energy
Declaration of interests Agency), Olivier Pellet (International Atomic Energy Agency),
HH reports personal fees from Ion Beam Applications for service on its Yaroslav Pynda (International Atomic Energy Agency), Mike Sathekge
Board of Directors, outside the submitted work; and serves without (Department of Nuclear Medicine, University of Pretoria and Steve Biko
compensation on the External Advisory Board of Sidney Kimmel Academic Hospital, Pretoria, South Africa), Heinz-Peter Schlemmer
Comprehensive Cancer Center (Johns Hopkins University), the (German Cancer Research Centre, Heidelberg, Germany), Veronica
International Advisory Board of the University of Vienna, the Scientific Sichizya (University Teaching Hospitals, Lusaka, Zambia), Elizabeth
Committee of the German Cancer Research Centre, the Board of Trustees Sutton (Memorial Sloan Kettering Cancer Center), Alberto H Vargas
of the German Cancer Research Centre, and the Scientific Advisory (Memorial Sloan Kettering Cancer Center), and Cherian Varghese (World
Board of Euro-BioImaging. JAB reports personal fees from Board of Health Organization). The commissioners at large, who participated in
Directors of Accumen, outside the submitted work. JSL reports personal meetings and contributed to the planning and outline of the paper, but do
fees from Clarity Pharmaceuticals, Varian Medical Systems, Texas Pacific not meet all authorship criteria, were: Kwanele Asante (African

e167 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

Organisation for Research & Training in Cancer, Johannesburg, South 16 Vijenthira A, Chan K, Cheung MC, Prica A. Cost-effectiveness of
Africa), Jean-Yves Douillard (European Society for Medical Oncology, first-line treatment options for patients with advanced-stage Hodgkin
Nantes, France), Tarek El Diasty (Urology and Nephrology Center, lymphoma: a modelling study. Lancet Haematol 2020; 7: e146–56.
Mansoura University, Mansoura, Egypt), Mary Gospodarowicz (Princess 17 Schlemmer HP, Bittencourt LK, D’Anastasi M, et al. Global
Margaret Cancer Centre, University of Toronto, Toronto, ON, Canada), challenges for cancer imaging. J Glob Oncol 2018; 4: 1–10.
Paul Harari (American Society for Radiation Oncology, Madison, WI, 18 Allemani C, Matsuda T, Di Carlo V, et al. Global surveillance of
USA), Jun Hatazawa (Asia Oceania Federation of Nuclear Medicine and trends in cancer survival 2000–14 (CONCORD-3): analysis of
Biology, Osaka, Japan), Dina Husseiny Salama (Atomic Energy Authority, individual records for 37 513 025 patients diagnosed with one of
Cairo, Egypt), Renato Adam Mendonca (Diagnosticos da America, São 18 cancers from 322 population-based registries in 71 countries.
Lancet 2018; 391: 1023–75.
Paulo, Brazil), Satoshi Minoshima (Department of Radiology and Imaging
19 Allemani C, Weir HK, Carreira H, et al. Global surveillance of
Sciences, University of Utah, Salt Lake City, UT, USA; the Society of
cancer survival 1995-2009: analysis of individual data for
Nuclear Medicine and Molecular Imaging, Reston, VA, USA), Sharyl Nass
25 676 887 patients from 279 population-based registries in
(Health and Medicine Division, National Academies of Sciences, 67 countries (CONCORD-2). Lancet 2015; 385: 977–1010.
Engineering, and Medicine, Washington DC, USA), Paul Parizel (Royal
20 Bagley AF, Anscher MS, Choi S, et al. Association of
Perth Hospital & University of Western Australia, Perth, Australia), sociodemographic and health-related factors with receipt of
Vjay Rao (Thomas Jefferson University, PA, USA), Ferid Shannoun nondefinitive therapy among younger men with high-risk prostate
(United Nations Scientific Committee on the Effects of Atomic Radiation, cancer. JAMA Netw Open 2020; 3: e201255.
Vienna, Austria), Ben J Slotman (Amsterdam University Medical Centres, 21 Pfister DG, Rubin DM, Elkin EB, et al. Risk adjusting survival
Amsterdam, Netherlands), and Edward Trimble (US National Cancer outcomes in hospitals that treat patients with cancer without
Institute, Bethesda, MD, USA). Editorial note: the Lancet Group takes a information on cancer stage. JAMA Oncol 2015; 1: 1303–10.
neutral position with respect to territorial claims in published maps and 22 Editorial. Think globally about cancer. Nat Med 2019; 25: 351.
institutional affiliations. 23 Shah SC, Kayamba V, Peek RM Jr, Heimburger D. Cancer control in
References low- and middle-income countries: is it time to consider screening?
1 Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A. J Glob Oncol 2019; 5: 1–8.
Global cancer statistics 2018: GLOBOCAN estimates of incidence 24 Lou B, Doken S, Zhuang T, et al. An image-based deep learning
and mortality worldwide for 36 cancers in 185 countries. framework for individualizing radiotherapy dose.
CA Cancer J Clin 2018; 68: 394–424. Lancet Digit Health 2019; 1: e136–47.
2 Ferlay J, Soerjomataram I, Dikshit R, et al. Cancer incidence and 25 Mac Manus MP, Hicks RJ, Ball DL, et al. F-18 fluorodeoxyglucose
mortality worldwide: sources, methods and major patterns in positron emission tomography staging in radical radiotherapy
GLOBOCAN 2012. Int J Cancer 2015; 136: e359–86. candidates with nonsmall cell lung carcinoma: powerful correlation
3 Bray F, Jemal A, Grey N, Ferlay J, Forman D. Global cancer with survival and high impact on treatment. Cancer 2001;
transitions according to the Human Development Index (2008–30): 92: 886–95.
a population-based study. Lancet Oncol 2012; 13: 790–801. 26 Press RH, Shu HG, Shim H, et al. The use of quantitative imaging
4 Stewart BW, Wild CP, eds. World Cancer Report, 2014. Lyon: in radiation oncology: a quantitative imaging network (QIN)
International Agency for Research on Cancer, 2014. https:// perspective. Int J Radiat Oncol Biol Phys 2018; 102: 1219–35.
publications.iarc.fr/Non-Series-Publications/World-Cancer-Reports/ 27 Accordino MK, Wright JD, Buono D, Neugut AI, Hershman DL.
World-Cancer-Report-2014 (accessed Oct 6, 2020). Trends in use and safety of image-guided transthoracic needle
5 Farmer P, Frenk J, Knaul FM, et al. Expansion of cancer care and biopsies in patients with cancer. J Oncol Pract 2015; 11: e351–59.
control in countries of low and middle income: a call to action. 28 Galluzzo A, Genova C, Dioguardi S, Midiri M, Cajozzo M. Current
Lancet 2010; 376: 1186–93. role of computed tomography-guided transthoracic needle biopsy of
6 Atun R, Jaffray DA, Barton MB, et al. Expanding global access to metastatic lung lesions. Future Oncol 2015; 11 (2 suppl): 43–46.
radiotherapy. Lancet Oncol 2015; 16: 1153–86. 29 Huo J, Aloia TA, Xu Y, Chung TH, Sheu T, Tina Shih YC.
7 Sullivan R, Alatise OI, Anderson BO, et al. Global cancer surgery: Comparative effectiveness of computed tomography- versus
delivering safe, affordable, and timely cancer surgery. Lancet Oncol ultrasound-guided percutaneous radiofrequency ablation among
2015; 16: 1193–224. medicare patients 65 years of age or older with hepatocellular
carcinoma. Value Health 2019; 22: 284–92.
8 Brown G, Davies S, Williams GT, et al. Effectiveness of preoperative
staging in rectal cancer: digital rectal examination, endoluminal 30 Preim B, Botha C. Visual computing for medicine (second edition):
ultrasound or magnetic resonance imaging? Br J Cancer 2004; theory, algorithms, and applications. Amsterdam: Elsevier, 2014.
91: 23–29. 31 WHO. Communicating radiation risks in paediatric imaging:
9 Brown LC, Ahmed HU, Faria R, et al. Multiparametric MRI to information to support healthcare discussions about benefit and
improve detection of prostate cancer compared with transrectal risk. 2016. https://www.who.int/ionizing_radiation/pub_meet/
ultrasound-guided prostate biopsy alone: the PROMIS study. radiation-risks-paediatric-imaging/en/ (accessed Nov 30, 2020).
Health Technol Assess 2018; 22: 1–176. 32 Martens MH, Lambregts DMJ, Kluza E, Beets-Tan RGH. Tumor
10 Dzau VJ, Ginsburg GS, Chopra A, et al. Realizing the full potential response to treatment: prediction and assessment. Curr Radiol Rep
of precision medicine in health and health care. NAM Perspectives 2014; 2: 62.
2016; published online Sept 19. https://doi.org/10.1001/ 33 The Canadian Medical Imaging Inventory. March, 2018. https://
jama.2016.14117. www.cadth.ca/canadian-medical-imaging-inventory-2017 (accessed
11 Facey K, Bradbury I, Laking G, Payne E. Overview of the clinical Nov 30, 2020).
effectiveness of positron emission tomography imaging in selected 34 The Institute of Physics and Engineering in Medicine, The College
cancers. Health Technol Assess 2007; 11: iii–iv, xi–267. of Radiographers, The Royal College of Radiologists. CT equipment,
12 Fischer B, Lassen U, Mortensen J, et al. Preoperative staging of operations, capacity and planning in the NHS: report from the
lung cancer with combined PET–CT. N Engl J Med 2009; 361: 32–39. Clinical Imaging Board. June, 2015. https://www.rcr.ac.uk/sites/
default/files/ct_equipment_in_the_nhs_report_cib_may_2015_v2_
13 Hricak H, Powell CB, Yu KK, et al. Invasive cervical carcinoma: role final240615.pdf (accessed Nov 30, 2020).
of MR imaging in pretreatment work-up—cost minimization and
diagnostic efficacy analysis. Radiology 1996; 198: 403–09. 35 Rosenkrantz AB, Chaves Cerdas L, Hughes DR, Recht MP, Nass SJ,
Hricak H. National trends in oncologic diagnostic imaging.
14 Langer A. A systematic review of PET and PET/CT in oncology: J Am Coll Radiol 2020; 17: 1116–22.
a way to personalize cancer treatment in a cost-effective manner?
BMC Health Serv Res 2010; 10: 283. 36 Hillner BE, Siegel BA, Liu D, et al. Impact of positron emission
tomography/computed tomography and positron emission
15 Scott AM, Gunawardana DH, Kelley B, et al. Positron emission tomography (PET) alone on expected management of patients with
tomography changes management and improves prognostic cancer: initial results from the National Oncologic PET Registry.
stratification in patients with recurrent colorectal cancer: results of J Clin Oncol 2008; 26: 2155–61.
a multi-center prospective study. J Nucl Med 2008; 49: 1451–57.

www.thelancet.com/oncology Vol 22 April 2021 e168

 The Lancet Oncology Commission

37 Lindsay MJ, Siegel BA, Tunis SR, et al. The National Oncologic PET 57 Jamison DT, Summers LH, Alleyne G, et al. Global health 2035:
Registry: expanded medicare coverage for PET under coverage with a world converging within a generation. Lancet 2013; 382: 1898–955.
evidence development. AJR Am J Roentgenol 2007; 188: 1109–13. 58 Ward ZJ, Scott AM, Hricak H, Atun R. Global costs, health, and
38 Bellanger M, Zeinomar N, Tehranifar P, Terry MB. Are global economic benefits of scaling up treatment and imaging modalities
breast cancer incidence and mortality patterns related to country- for survival of 11 cancers: a simulation-based analysis. Lancet Oncol
specific economic development and prevention strategies? 2021; 22: 341–51.
J Glob Oncol 2018; 4: 1–16. 59 Neumann PJ, Sanders GD, Russell LB, Siegel JE, Ganiats TG.
39 Bray F, Jemal A, Torre LA, Forman D, Vineis P. Long-term realism Cost-effectiveness in health and medicine, 2nd edn. New York, NY:
and cost-effectiveness: primary prevention in combatting cancer Oxford University Press, 2016.
and associated inequalities worldwide. J Natl Cancer Inst 2015; 60 Ministry of Health, Federal Democratic Republic of Ethiopia.
107: djv273. Ethiopia Health Accounts 2016/17. http://www.moh.gov.et/ejcc/
40 Atun R, Bhakta N, Denburg A, et al. Sustainable care for children sites/default/files/2020-01/Ethiopia%207th%20Health%20
with cancer: a Lancet Oncology Commission. Lancet Oncol 2020; Accounts%20Report_2016-17.pdf (accessed Oct 6, 2020).
21: e185–224. 61 UN Department of Economic and Social Affairs. World population
41 WHO. Technical annex: updated appendix 3 of the WHO Global prospects 2019. https://population.un.org/wpp/Download/
NCD Action Plan 2013–2020. April 12, 2017. https://www.who.int/ Standard/CSV/ (accessed Oct 6, 2020).
ncds/governance/technical_annex.pdf?ua=1 (accessed Nov 30, 2020). 62 The World Bank. The World Bank Atlas method - detailed
42 WHO. 2015 Global survey on health technology assessment by methodology. https://datahelpdesk.worldbank.org/knowledgebase/
national authorities. 2015. https://www.who.int/health-technology- articles/378832-what-is-the-world-bank-atlas-method (accessed
assessment/MD_HTA_oct2015_final_web2.pdf?ua=1 (accessed Jan 27, 2021).
Nov 30, 2020). 63 Serajuddin U, Hamadeh N. New World Bank country classifications
43 WHO. WHO list of priority medical devices for cancer by income level: 2020–2021. World Bank Blogs. July 1, 2020. https://
management. 2017. https://apps.who.int/iris/ blogs.worldbank.org/opendata/new-world-bank-country-
handle/10665/255262?mode=full (accessed Dec 1, 2020). classifications-income-level-2020-2021 (accessed Nov 20, 2020).
44 Maksimovic R, Berument AV. Innovative technology in addressing 64 International Atomic Energy Agency. IAEA Technical Cooperation
global health issues: the WHO perspective. https://www.who.int/ Projects – nuclear medicine and diagnostic imaging section.
diagnostic_imaging/imaging_modalities/ https://www.iaea.org/projects/technical-cooperation-projects/3328
InnovativeTechAddressingGlobalHealthIssues_WHOPerspective. (accessed Nov 30, 2020).
pdf?ua=1 (accessed Nov 20,2020). 65 Heller PS. Understanding fiscal space. March, 2005. https://www.
45 RAD-AID International. RAD-AID radiology readiness survey. 2013. imf.org/external/pubs/ft/pdp/2005/pdp04.pdf (accessed
https://www.rad-aid.org/resource-center/radiology-readiness Oct 6, 2020).
(accessed Nov 30, 2020). 66 Reddy CL, Peters AW, Jumbam DT, et al. Innovative financing to
46 Dodd D, Jack A. Cancer optimism tempered by ‘financial toxicity’. fund surgical systems and expand surgical care in low-income and
Financial Times. June 1, 2018. https://www.ft.com/content/ middle-income countries. BMJ Glob Health 2020; 5: e002375.
a796face-658f-11e8-a39d-4df188287fff (accessed Oct 6, 2020). 67 Tandon A, Cashin C. Assessing public expenditure on health from a
47 IAEA, WHO. Roadmap towards a National Cancer Control fiscal space perspective. February, 2010. https://openknowledge.
Programme. Milestones for establishing nuclear medicine, worldbank.org/bitstream/handle/10986/13613/560530WP0Box341p
diagnostic imaging and radiotherapy services. 2019. https://www. enditureFiscalSpace.pdf?sequence=1&isAllowed=y (accessed Oct 6,
iaea.org/sites/default/files/19/10/milestones-document-2019.pdf 2020).
(accessed Jan 16, 2020). 68 International Monetary Fund. Real GDP growth. 2019. https://www.
48 National Cancer Institute. Financial toxicity and cancer treatment imf.org/external/datamapper/NGDP_RPCH@WEO/OEMDC/
(PDQ®)–health professional version. September, 2019. https:// ADVEC/WEOWORLD (accessed Nov 30, 2020).
www.cancer.gov/about-cancer/managing-care/track-care-costs/ 69 Global Burden of Disease Health Financing Collaborator Network.
financial-toxicity-hp-pdq (accessed Nov 30, 2020). Future and potential spending on health 2015–40: development
49 Schnipper L. Elephant in the room: financial toxicity in cancer care. assistance for health, and government, prepaid private, and out-of-
ASCO 2019 Annual Meeting; Chicago, IL; May 31–June 4, 2019. pocket health spending in 184 countries. Lancet 2019;
50 WHO. Can people afford to pay for health care? New evidence on 389: P2005–30.
financial protection in Europe: regional report. https://apps.who. 70 Glenday G, Bharali I, Wang Z. Enhancing domestic revenues:
int/iris/bitstream/handle/10665/311654/9789289054058-eng.pdf constraints and opportunities: a cross-country comparative study of
(accessed Oct 6, 2020). tax capacity, effort and gaps. April, 2019. https://
51 International Atomic Energy Agency. IMAGINE - IAEA medical centerforpolicyimpact.org/wp-content/uploads/sites/18/2019/04/
imaging and nuclear medicine global resources database. October, CPIGH-Report_Tax-report_Enhancing-Domestic-Revenues__
2019. https://humanhealth.iaea.org/HHW/DBStatistics/IMAGINE. April-2019_FINAL.pdf (accessed Oct 6, 2020).
html (accessed Nov 30, 2020). 71 Barroy H, Sparkes S, Dale E, Mathonnat J. Can low- and middle-
52 Department of Health. A framework for the development of income countries increase domestic fiscal space for health: a mixed-
positron emission tomography (PET) services in England. methods approach to assess possible sources of expansion.
Oct 10, 2005. http://www.inahta.org/wp-content/uploads/2014/09/ Health Syst Reform 2018; 4: 214–26.
PET_A_framework_for_development_of_PET_services_in_ 72 WHO. Earmarked tobacco taxes: lessons learnt from nine countries.
England.pdf ((accessed Nov 30, 2020). 2015. http://apps.who.int/iris/bitstream/10665/206007/1/
53 van Tinteren H, Hoekstra OS, Smit EF, et al. Effectiveness of 9789241510424_eng.pdf?ua=1 (accessed Nov 30, 2020).
positron emission tomography in the preoperative assessment of 73 Summan A, Stacey N, Birckmayer J, Blecher E, Chaloupka FJ,
patients with suspected non-small-cell lung cancer: the PLUS Laxminarayan R. The potential global gains in health and revenue
multicentre randomised trial. Lancet 2002; 359: 1388–93. from increased taxation of tobacco, alcohol and sugar-sweetened
54 Ward ZJ, Grover S, Scott AM, et al. The role and contribution of beverages: a modelling analysis. BMJ Glob Health 2020; 5: e002143.
imaging in global cervical cancer management – survival estimates 74 The World Bank. Industry Engagement Program. Large medical
from a simulation-based analysis. Lancet Oncol 2020; 21: 1089–98. diagnostic imaging equipment market review and action plan. 2017.
55 Ward ZJ, Scott AM, Hricak H, et al. Estimating the impact of http://pubdocs.worldbank.org/en/452061518466768717/
treatment and imaging modalities on 5-year net survival of Procurement-Market-Review-and-Action-Plan.pdf (accessed
11 cancers in 200 countries: a simulation-based analysis. Nov 30, 2020).
Lancet Oncol 2020; 21: 1077–88. 75 African Development Bank. African Development Bank Group
56 Global Cancer Observatory, International Agency for Research on approves $10 million equity in Razorite Healthcare Fund for Africa.
Cancer. Estimated number of new cases in 2020, worldwide, July 23, 2020. https://www.afdb.org/en/news-and-events/press-
both sexes, all ages, http://gco.iarc.fr/today/online-analysis-table releases/african-development-bank-group-approves-10-million-
(accessed Feb 13, 2020). equity-razorite-healthcare-fund-africa-35206 (accessed Dec 1, 2020).

e169 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

76 The World Bank. Guarantees program. https://www.worldbank.org/ 98 Bettmann MA, Holmberg O, del Rosario Perez M, et al.
en/programs/guarantees-program (accessed Jan 12, 2021). International Collaboration on Clinical Imaging Guidelines: many
77 World Bank Group. World Bank Group guarantee products. April, hands make light work. J Am Coll Radiol 2015; 12: 43–44.
2016. https://ppp.worldbank.org/public-private-partnership/sites/ 99 American College of Radiology. ACR Appropriateness Criteria.
ppp.worldbank.org/files/documents/PPPCCSA_WBGGuarantees_ https://www.acr.org/Clinical-Resources/ACR-Appropriateness-
Final%20_%20English%20_Printed%20Oct%202016.pdf (accessed Criteria (accessed Dec 1, 2020).
Jan 15, 2021). 100 European Society of Radiology. ESR iGuide: clinical decision
78 Atun R, Knaul FM, Akachi Y, Frenk J. Innovative financing for support using European imaging referral guidelines. https://www.
health: what is truly innovative? Lancet 2012; 380: 2044–49. myesr.org/esriguide (accessed Dec 1, 2020).
79 Atun R, Silva S, Knaul FM. Innovative financing instruments for 101 Remedios D, Brink J, Holmberg O, et al. Clinical imaging
global health 2002–15: a systematic analysis. Lancet Glob Health guidelines part 1: a proposal for uniform methodology.
2017; 5: e720–26. J Am Coll Radiol 2015; 12: 45–50.
80 Social Finance. Investing in social outcomes: development impact 102 The Royal College of Radiologists. Introducing iRefer8 – updated
bonds. http://www.socialfinance.org.uk/investing-in-social- diagnostic imaging guidelines for referring clinicians. May 24, 2017.
outcomes-development-impact-bonds/ (accessed June 2, 2020). https://www.rcr.ac.uk/posts/introducing-irefer8-updated-
81 Finance S. Social Impact Bonds. http://www.socialfinance.org.uk/ diagnostic-imaging-guidelines-referring-clinicians (accessed
wp-content/uploads/2016/07/SIBs-Early-Years_Social- Dec 1, 2020).
Finance_2016_Final3.pdf (accessed June 2, 2020). 103 Society of Nuclear Medicine & Medical Imaging. Appropriate Use
82 International Monetary Fund. 2020. World economic outlook: Criteria (AUC). http://www.snmmi.org/ClinicalPractice/content.
a long and difficult ascent. October, 2020. https://www.imf.org/en/ aspx?ItemNumber=15666 (accessed Dec 1, 2020).
Publications/WEO/Issues/2020/09/30/world-economic-outlook- 104 European Association of Nuclear Medicine. Guidelines. https://
october-2020 (accessed Nov 27, 2020). www.eanm.org/publications/guidelines/ (accessed Dec 1, 2020).
83 UN. Sources and effects of ionizing radiation. 2008. https://www. 105 Dondi M, Paez D, Torres L, et al. Implementation of quality systems
unscear.org/docs/publications/2008/UNSCEAR_2008_Report_ in nuclear medicine: why it matters. an outcome analysis (quality
Vol.I.pdf (accessed Oct 6, 2020). management audits in nuclear medicine part III). Semin Nucl Med
84 National Council on Radiation Protection and Measurements. 2018; 48: 299–306.
Report No. 184. Medical radiation exposure of patients in the United 106 International Atomic Energy Agency. Quality management audits
States. Nov 15, 2019. https://ncrponline.org/shop/reports/report- in nuclear medicine practices, second edition. 2015. https://www.
no-184-medical-radiation-exposure-of-patients-in-the-united- iaea.org/publications/10714/quality-management-audits-in-nuclear-
states-2019/ (accessed Oct 6, 2020). medicine-practices (accessed Oct 6, 2020).
85 Kashyap R, Dondi M, Paez D, Mariani G. Hybrid imaging 107 International Atomic Energy Agency. comprehensive clinical audits
worldwide-challenges and opportunities for the developing world: of diagnostic radiology practices: a tool for quality improvement.
a report of a technical meeting organized by IAEA. Semin Nucl Med 2010. https://www.iaea.org/publications/8187/comprehensive-
2013; 43: 208–23. clinical-audits-of-diagnostic-radiology-practices-a-tool-for-quality-
86 Paez D, Becic T, Bhonsle U, et al. Current status of nuclear medicine improvement (accessed Oct 6, 2020).
practice in the Middle East. Semin Nucl Med 2016; 46: 265–72. 108 European Society of Radiology (ESR). The ESR Audit Tool
87 Paez D, Orellana P, Gutierrez C, et al. Current Status of nuclear (Esperanto): genesis, contents and pilot. Insights Imaging 2018;
medicine practice in Latin America and the Caribbean. J Nucl Med 9: 899–903.
2015; 56: 1629–34. 109 European Commission. The rules governing medicinal products in
88 Holmberg O, Czarwinski R, Mettler F. The importance and unique the European Union, volume 4. Feb 3, 2010. https://ec.europa.eu/
aspects of radiation protection in medicine. Eur J Radiol 2010; health/sites/health/files/files/eudralex/vol-4/2009_06_annex13.pdf
76: 6–10. (accessed Oct 6, 2020).
89 Shiralkar S, Rennie A, Snow M, Galland RB, Lewis MH, 110 International Atomic Energy Agency Tecdoc Series. Good practice
Gower-Thomas K. Doctors’ knowledge of radiation exposure: for introducing radiopharmaceuticals for clinical use. 2018. https://
questionnaire study. BMJ 2003; 327: 371–72. www.iaea.org/publications/10829/good-practice-for-introducing-
90 IAEA. Clinical training of medical physicists specializing in radiopharmaceuticals-for-clinical-use (accessed Oct 6, 2020).
diagnostic radiology. Training course series no. 47. Vienna, 2011. 111 International Atomic Energy Agency Tecdoc Series. Quality Control
https://www-pub.iaea.org/MTCD/Publications/PDF/TCS-47_web. in the Production of Radiopharmaceuticals. 2018. https://www.iaea.
pdf (accessed Jan 15, 2020). org/publications/13422/quality-control-in-the-production-of-
91 International Atomic Energy Agency. International Safety radiopharmaceuticals (accessed Oct 6, 2020).
Standards. https://www.iaea.org/resources/rpop/resources/ 112 ICRP. Radiological protection in therapy with radiopharmaceuticals.
international-safety-standards (accessed Dec 1, 2020). ICRP Publication 140. Ann ICRP 2019; 48: 1–106.
92 ICRP. Radiological Protection in Medicine. ICRP Publication 105. 113 International Atomic Energy Agency. Nuclear Medicine Physics:
Ann ICRP 2007; 37: 1–61. a handbook for teachers and students. December, 2014. https://
93 International Atomic Energy Agency. Radiation protection and www-pub.iaea.org/MTCD/Publications/PDF/Pub1617web-1294055.
safety in medical uses of ionizing radiation. 2018. https://www.iaea. pdf (accessed Dec 1, 2020).
org/publications/11102/radiation-protection-and-safety-in-medical- 114 European Commission, Directive 2013/35/EU on the minimum
uses-of-ionizing-radiation (accessed Oct 6, 2020). health and safety requirements regarding the exposure of workers
94 International Atomic Energy Agency. Governmental, legal and to the risks arising from electromagnetic fields (2013). https://eur-
regulatory framework for safety. 2016. https://www.iaea.org/ lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32013L0035
publications/10883/governmental-legal-and-regulatory-framework- (accessed Jan 15, 2020).
for-safety (accessed Oct 6, 2020). 115 Kanal E, Barkovich AJ, Bell C, et al. ACR guidance document on
95 International Atomic Energy Agency. Radiation protection and safety MR safe practices: 2013. J Magn Reson Imaging 2013; 37: 501–30.
of radiation sources: international basic safety standards. General 116 Cross NM, Hoff MN, Kanal KM. Avoiding MRI-related accidents:
safety standards requirements part 3. 2014. https://www.iaea.org/ a practical approach to implementing MR safety. J Am Coll Radiol
publications/8930/radiation-protection-and-safety-of-radiation- 2018; 15: 1738–44.
sources-international-basic-safety-standards (accessed Oct 6, 2020). 117 Panch T, Szolovits P, Atun R. Artificial intelligence, machine
96 International Atomic Energy Agency. Establishing the infrastructure learning and health systems. J Glob Health 2018; 8: 020303.
for radiation safety. 2018. https://www.iaea.org/publications/11085/ 118 Pisano ED, Garnett LR. Big data and radiology research.
establishing-the-infrastructure-for-radiation-safety (accessed J Am Coll Radiol 2019; 16 (9 pt B): 1347–50.
Oct 6, 2020). 119 Liu B, Cheng J, Guo DJ, et al. Prediction of prostate cancer
97 Muhogora W, Rehani MM. Review of the current status of radiation aggressiveness with a combination of radiomics and machine
protection in diagnostic radiology in Africa. J Med Imag 2017; learning-based analysis of dynamic contrast-enhanced MRI.
4: 031202. Clin Radiol 2019; 74: 896.e1–8.

www.thelancet.com/oncology Vol 22 April 2021 e170

 The Lancet Oncology Commission

120 Lou R, Lalevic D, Chambers C, Zafar HM, Cook TS. Automated 143 US Food and Drug Administration. Radio frequency identification
detection of radiology reports that require follow-up imaging using (RFID). https://www.fda.gov/radiation-emitting-products/
natural language processing feature engineering and machine electromagnetic-compatibility-emc/radio-frequency-identification-
learning classification. J Digit Imaging 2020; 3: 131–36. rfid (accessed Aug 8, 2020).
121 Munir K, Elahi H, Ayub A, Frezza F, Rizzi A. Cancer diagnosis 144 Arunachalam SP, Asan O, Nestler DM, et al. Patient-care team
using deep learning: a bibliographic review. Cancers 2019; 11: e1235. contact patterns impact treatment length of stay in the emergency
122 Nasrullah N, Sang J, Alam MS, Mateen M, Cai B, Hu H. Automated department. Annu Int Conf IEEE Eng Med Biol Soc 2019;
lung nodule detection and classification using deep learning 2019: 345–48.
combined with multiple strategies. Sensors 2019; 19: e3722. 145 Álvarez López Y, Franssen J, Álvarez Narciandi G, Pagnozzi J,
123 Shen L, Margolies LR, Rothstein JH, Fluder E, McBride R, Sieh W. González-Pinto Arrillaga I, Las-Heras Andrés F. RFID technology
Deep learning to improve breast cancer detection on screening for management and tracking: e-Health applications. Sensors 2018;
mammography. Sci Rep 2019; 9: 12495. 18: 2663.
124 WHO. M-health: new horizons for mobile health through mobile 146 Shan H, Padole A, Homayounieh F, et al. Competitive performance
technologies: second global survey on e-health. 2011. http://www.who. of a modularized deep neural network compared to commercial
int/goe/publications/goe_mhealth_web.pdf (accessed Dec 1, 2020). algorithms for low-dose CT image reconstruction. Nat Mach Intell
125 Littman-Quinn R, Mibenge C, Antwi C, Chandra A, Kovarik CL. 2019; 1: 269–76.
Implementation of m-health applications in Botswana: 147 Mollura DJ, Soroosh G, Culp MP, et al. 2016 RAD-AID Conference
telemedicine and education on mobile devices in a low resource on International Radiology for Developing Countries: gaps, growth,
setting. J Telemed Telecare 2013; 19: 120–25. and United Nations sustainable development goals.
126 WHO. Global diffusion of eHealth: making universal health J Am Coll Radiol 2017; 14: 841–47.
coverage achieveable. https://www.who.int/goe/publications/ 148 Lacson R, Cochon L, Ip I, et al. Classifying safety events related to
global_diffusion/en/ (accessed Oct 6, 2020). diagnostic imaging from a safety reporting system using a human
127 Frehywot S, Vovides Y, Talib Z, et al. E-learning in medical factors framework. J Am Coll Radiol 2019; 16: 282–88.
education in resource constrained low- and middle-income 149 Lacson R, Harris K, Brawarsky P, et al. Evaluation of an automated
countries. Hum Resour Health 2013; 11: 4. information extraction tool for imaging data elements to populate a
128 Wahl B, Cossy-Gantner A, Germann S, Schwalbe NR. Artificial breast cancer screening registry. J Digit Imaging 2015; 28: 567–75.
intelligence (AI) and global health: how can AI contribute to health 150 Masud R, Al-Rei M, Lokker C. Computer-aided detection for breast
in resource-poor settings? BMJ Glob Health 2018; 3: e000798. cancer screening in clinical settings: scoping review.
129 Akhlaq A, McKinstry B, Muhammad KB, Sheikh A. Barriers and JMIR Med Inform 2019; 7: e12660.
facilitators to health information exchange in low- and middle- 151 Sippo DA, Warden GI, Andriole KP, et al. Automated extraction of
income country settings: a systematic review. Health Policy Plan BI-RADS final assessment categories from radiology reports with
2016; 31: 1310–25. natural language processing. J Digit Imaging 2013; 26: 989–94.
130 Labrique AB, Wadhwani C., Williams KA, et al. Best practices in 152 Stolz LA, Muruganandan KM, Bisanzo MC, et al. Point-of-care
scaling digital health in low and middle income countries. ultrasound education for non-physician clinicians in a resource-
Global Health 2018; 14: 103. limited emergency department. Trop Med Int Health 2015;
131 Brenn K, Ergun AS, Firouzi K, et al. Advances in capacitative 20: 1067–72.
micromachines ultrasonic transducers. Micromachines 2019; 10: 152. 153 Yang S, Gao X, Liu L, et al. Performance and reading time of
132 National Institutes of Health. Tablet-based mobile health ultrasound automated breast US with or without computer-aided detection.
for point-of-care breast cancer diagnosis in Nigeria. Aug 10, 2020. Radiology 2019; 292: 540–49.
https://clinicaltrials.gov/ct2/show/NCT04501419 (accessed 154 Gu Y, Lu X, Yang L, et al. Automatic lung nodule detection using a
Dec 1, 2020). 3D deep convolutional neural network combined with a multi-scale
133 Azubuike SO, Muirhead C, Hayes L, McNally R. Rising global prediction strategy in chest CTs. Comput Biol Med 2018; 103: 220–31.
burden of breast cancer: the case of sub-Saharan Africa (with 155 Masood A, Yang P, Sheng B, et al. Cloud-based automated clinical
emphasis on Nigeria) and implications for regional development: decision support system for detection and diagnosis of lung cancer
a review. World J Surg Oncol 2018; 16: 63. in chest CT. IEEE J Transl Eng Health Med 2019; 8: 4300113.
134 Cramer A, Hecla J, Wu D, et al. Stationary computed tomography 156 Wu W, Gao L, Duan H, Huang G, Ye X, Nie S. Segmentation of
for space and other resource-constrained environments. Sci Rep pulmonary nodules in CT images based on 3D-UNET combined
2018; 8: 14195. with three-dimensional conditional random field optimization.
135 Cooley CZ, Stockmann JP, Armstrong BD, et al. Two-dimensional Med Phys 2020; 47: 4054–63.
imaging in a lightweight portable MRI scanner without gradient 157 Chilamkurthy S, Ghosh R, Tanamala S, et al. Deep learning
coils. Magn Reson Med 2015; 73: 872–83. algorithms for detection of critical findings in head CT scans:
136 Cooley CZ, Haskell MW, Cauley SF, et al. Design of sparse halbach a retrospective study. Lancet 2018; 392: 2388–96.
magnet arrays for portable MRI using a genetic algorithm. 158 Kuo W, Häne C, Mukherjee P, Malik J, Yuh EL. Expert-level detection
IEEE Trans Magn 2018; 54: 1–12. of acute intracranial hemorrhage on head computed tomography
137 Reynes-Llompart G, Gámez-Cenzano C, Romero-Zayas I, et al. using deep learning. Proc Natl Acad Sci USA 2019; 116: 22737–45.
Performance characteristics of the whole-body discovery IQ PET/CT 159 Adambounou K, Adjenou V, Salam AP, et al. A low-cost tele-imaging
system. J Nucl Med 2017; 58: 1155–61. platform for developing countries. Front Public Health 2014; 2: 135.
138 Runge VM, Richter JK, Heverhagen JT. Motion in magnetic 160 Olpin J, Chuang L, Berek J, Gaffney D. Imaging and cancer of the
resonance: new paradigms for improved clinical diagnosis. cervix in low- and middle-income countries. Gynecol Oncol Rep 2018;
Invest Radiol 2019; 54: 383–95. 25: 115–21.
139 Yang Q, Yan P, Zhang Y, et al. Low-dose CT image denoising using 161 Akkus Z, Cai J, Boonrod A, et al. A survey of deep-learning
a generative adversarial network with Wasserstein distance and applications in ultrasound: artificial intelligence-powered
perceptual loss. IEEE Trans Med Imaging 2018; 37: 1348–57. ultrasound for improving clinical workflow. J Am Coll Radiol 2019;
140 Lin DJ, Johnson PM, Knoll F, Lui YW. Artificial intelligence for 16 (9 Pt B): 1318–28.
MR image reconstruction: an overview for clinicians. 162 Nasief HG, Rosado-Mendez IM, Zagzebski JA, Hall TJ.
J Magn Reson Imaging 2020; published online Feb 12. https://doi.org/ A quantitative ultrasound-based multi-parameter classifier for
10.1002/jmri.27078. breast masses. Ultrasound Med Biol 2019; 45: 1603–16.
141 Häggström I, Schmidtlein CR, Campanella G, Fuchs TJ. DeepPET: 163 Liu X, Faes L, Kale AU, et al. A comparison of deep learning
a deep encoder-decoder network for directly solving the PET image performance against health-care professionals in detecting diseases
reconstruction inverse problem. Med Image Anal 2019; 54: 253–62. from medical imaging: a systematic review and meta-analysis.
142 Liang G, Fouladvand S, Zhang J, et al. GANai:standardizing CT Lancet Digit Health 2019; 1: e273–97.
images using generative ad-versarial network with alternative 164 Kong Z, Jiang C, Zhu R, et al. 18F-FDG-PET-based radiomics
improvement. 2019 IEEE International Conference on Healthcare features to distinguish primary central nervous system lymphoma
Informatics; Xi’an, China; June 10–13, 2019 (abstr 1–11). from glioblastoma. Neuroimage Clin 2019; 23: 101912.

e171 www.thelancet.com/oncology Vol 22 April 2021

 The Lancet Oncology Commission

165 Kang D, Park JE, Kim YH, et al. Diffusion radiomics as a diagnostic 180 Loucaides EM, Fitchett EJA, Sullivan R, Atun R. Global public and
model for atypical manifestation of primary central nervous system philanthropic investment in childhood cancer research: systematic
lymphoma: development and multicenter external validation. analysis of research funding, 2008-16. Lancet Oncol 2019;
Neuro-oncol 2018; 20: 1251–61. 20: e672–84.
166 Ba-Ssalamah A, Muin D, Schernthaner R, et al. Texture-based 181 Maruthappu M, Head MG, Zhou CD, et al. Investments in cancer
classification of different gastric tumors at contrast-enhanced CT. research awarded to UK institutions and the global burden of
Eur J Radiol 2013; 82: e537–43. cancer 2000-2013: a systematic analysis. BMJ Open 2017; 7: e013936.
167 Ronneberger O, Fischer P, Brox T. U-net: convolutional networks 182 Abdel-Wahab M, Zubizarreta E, Polo A, Meghzifene A. Improving
for biomedical image segmentation. International Conference on quality and access to radiation therapy–an IAEA perspective.
Medical Image Computing and Computer-assisted Intervention Semin Radiat Oncol 2017; 27: 109–17.
2015. Published online Nov 18. https://doi.org/10.1007/ 183 Konert T, Vogel WV, Paez, D, et al. Introducing FDG PET/CT-
978-3-319-24574-4_28. guided chemoradiotherapy for stage III NSCLC in low- and middle-
168 Cook GJR, Goh V. What can artificial intelligence teach us about the income countries: preliminary results from the IAEA PERTAIN
molecular mechanisms underlying disease? trial. Eur J Nucl Med Mol Imag 2019; 46: 2235–43.
Eur J Nucl Med Mol Imaging 2019; 46: 2715–21. 184 International Atomic Energy Agency. Radiation therapy planning of
169 Du Y, Qi Y, Jin Z, Tian J. Noninvasive imaging in cancer non-small cell lung cancer based on PET/CT. Studies number
immunotherapy: The way to precision medicine. Cancer Lett 2019; E13042 and E33038. https://www.iaea.org/projects/crp/e13042
466: 13–22. (accessed Dec 1, 2020).
170 Sun R, Limkin EJ, Vakalopoulou M, et al. A radiomics approach to 185 Britton N, Miller MA, Safadi S, Siegel A, Levine AR, McCurdy MT.
assess tumour-infiltrating CD8 cells and response to anti-PD-1 or Tele-ultrasound in resource-limited settings: a systematic review.
anti-PD-L1 immunotherapy: an imaging biomarker, retrospective Front Public Health 2019; 7: 244.
multicohort study. Lancet Oncol 2018; 19: 1180–91. 186 Atun R, Cavalli F. The global fight against cancer: challenges and
171 Lambin P, Leijenaar RTH, Deist TM, et al. Radiomics: the bridge opportunities. Lancet 2018; 391: 412–13.
between medical imaging and personalized medicine. 187 United Nations. United Nations Sustainable Development Goals.
Nat Rev Clin Oncol 2017; 14: 749–62. Goal 3: ensure healthy lives and promote well-being for all at all
172 Jha AK, Li Z, Orav EJ, Epstein AM. Care in U.S. hospitals--the ages. https://www.un.org/sustainabledevelopment/health/
Hospital Quality Alliance program. N Engl J Med 2005; 353: 265–74. (accessed Dec 1, 2020).
173 FitzGerald TJ, Bishop-Jodoin M, Followill DS, et al. Imaging and 188 WHO. Tackling NCDs: ‘Best buys’ and other recommended
data acquisition in clinical trials for radiation therapy. interventions for the prevention and control of noncommunicable
Int J Radiat Oncol Biol Phys 2016; 94: 404–11. diseases. 2016. https://apps.who.int/iris/bitstream/
174 Hofman MS, Violet J, Hicks RJ, et al. [177Lu]-PSMA-617 radionuclide handle/10665/259232/WHO-NMH-NVI-17·9-eng.pdf?sequence=1
treatment in patients with metastatic castration-resistant prostate (accessed Dec 1, 2020).
cancer (LuPSMA trial): a single-centre, single-arm, phase 2 study. 189 WHO. WHO report on cancer: setting priorities, investing wisely
Lancet Oncol 2018; 19: 825–33. and providing care for all https://www.who.int/publications/i/item/
175 Murphy P, Koh DM. Imaging in clinical trials. Cancer Imaging 2010; who-report-on-cancer-setting-priorities-investing-wisely-and-
10: S74–82. providing-care-for-all (accessed Dec 1, 2020).
176 Strosberg J, El-Haddad G, Wolin E, et al. Phase 3 trial of 190 Abdel-Wahab M, Lahoupe B, Polo A, et al. Assessment of cancer
177
Lu-Dotatate for midgut neuroendocrine tumors. N Engl J Med control capacity and readiness: the role of the International Atomic
2017; 376: 125–35. Energy Agency. Lancet Oncol 2017; 18: e587–94.
177 Barrington SF, Kirkwood AA, Franceschetto A, et al. PET-CT for 191 St. Jude Children’s Research Hospital. St. Jude Global: where we
staging and early response: results from the response-adapted are. https://www.stjude.org/global/regions.html#3ea93d515bc377e2
therapy in advanced Hodgkin lymphoma study. Blood 2016; 6bebf197d148b877b6142ca37b47cf28379b8649302933b8=0 (accessed
127: 1531–38. Jan 18, 2020).
178 Petkova E, Antman EM, Troxel AB. Pooling data from individual 192 Hopkins J, Burns E, Edena T. International twinning partnerships:
trials in the COVID-19 era. JAMA 2020; 324: 543–45. an effective method of improving diagnosis, treatment and care for
179 Roach PJ, Francis R, Emmett L, et al. The impact of 68Ga-PSMA children with cancer in low-middle income countries.
PET/CT on management intent in prostate cancer: results of an J Cancer Policy 2013; 1: e8–19
Australian prospective multicenter study. J Nucl Med 2018;
© 2021 Elsevier Ltd. All rights reserved.
59: 82–88.

www.thelancet.com/oncology Vol 22 April 2021 e172