Multi-Cancer Early Detection Tests: State of the Art and Implications for Radiologists

Abstract

Multi-cancer early detection (MCED) tests are already being marketed as noninvasive, convenient opportunities to test for multiple cancer types with a single blood sample. The technology varies—involving detection of circulating tumor DNA, fragments of DNA, RNA, or proteins unique to each targeted cancer. The priorities and tradeoffs of reaching diagnostic resolution in the setting of possible false positives and negatives remain under active study. Given the well-established role of imaging in lesion detection and characterization for most cancers, radiologists have an essential role to play in selecting diagnostic pathways, determining the validity of test results, resolving false-positive MCED test results, and evaluating tradeoffs for clinical policy. Appropriate access to and use of imaging tests will also factor into clinical guidelines. Thus, all clinicians potentially involved with MCED tests for cancer screening will need to weigh the benefits and harms of MCED testing, including consideration of how the tests will be used alongside or in place of other screening options, how diagnostic confirmation tests should be selected, and what the implications are for policy and reimbursement decisions. Further, patients will need regular support to make informed decisions about screening using MCED tests in the context of their personal cancer risks, health-related values, and access to care.

© RSNA, 2025

Summary

Radiologists have a crucial role in advancing science, clinical care, and policy related to multi-cancer early detection tests.

Essentials

• ■ Many cancers indicated by multi-cancer tests will require imaging confirmation.

• ■ Appropriate confirmation pathways have not been identified for all suspected cancers, particularly those without established screening programs.

• ■ Radiologists have a major responsibility to develop clinical guidelines for multi-cancer tests by synthesizing evidence, delineating the potential diagnostic pathways, and identifying the clinical tradeoffs of these pathways compared with more established screening pathways.

• ■ Patients weighing options for cancer screening will need evidence-based support to navigate available tests in the context of their personal cancer risks, health-related values, and access to care.

Introduction

New and emerging technologies that assay a single medium—blood, urine, saliva, breath, or stool—to screen for any of multiple cancer types simultaneously are known collectively as pan-cancer, multi-cancer detection, or multi-cancer early detection (MCED) tests. These tests are intended for screening asymptomatic individuals in the general, average-risk population. If shown to be successful in reducing the incidence of late-stage cancers and/or reducing cancer mortality without engendering an unacceptable level of harm, use of MCED tests could expand the number and types of cancers for which screening is available and effective. Separately, some individuals may consider MCED tests more acceptable as a first-line screening test than currently recommended screening modalities, potentially leading to increased screening rates and reduced mortality for underscreened cancer types.

Many MCED tests are in development, and at least two are already available and marketed to the U.S. public, approved as laboratory developed tests via one of the available approval pathways of the U.S. Food and Drug Administration (1–4). However, questions about their ultimate impact on cancer-specific survival and overall survival have not yet been answered using randomized trials. Moreover, outside of industry sponsorship, few researchers have investigated appropriate strategies for confirming positive MCED tests or for educating patients, physicians, and other stakeholders about their possible benefits and harms. In this review, we describe the current state of the art in MCED, the implications of MCED test use on the demand for diagnostic imaging, and considerations for developing patient-centered approaches to MCED testing.

MCED Tests Vary in Technology and Phase of Development

To identify original studies on MCED test development and evaluation, we searched PubMed (Medline) using the following search string: “((Circulating Tumor DNA[MeSH Terms]) OR (Cell-Free Nucleic Acids[MeSH Terms]) OR (Antigens, Neoplasm/blood[MeSH Terms]) OR (DNA Methylation[MeSH Terms]) OR (Biomarkers, Tumor/blood[MeSH Terms]) OR (multi-cancer early detection) OR (multicancer early detection) OR (liquid biopsy) OR (MCED)) AND ((early detection of cancer[MeSH Terms]) OR (early diagnosis[MeSH Terms]) OR “cancer screening”).” Results were updated until December 1, 2023. Original studies that evaluated the test performance of cell-free assays for at least two cancer types were reviewed. Systematic reviews were screened for additional references related to MCED test performance characteristics and clinical studies. We also searched ClinicalTrials.gov to identify trials, searched Google using individual test names or “multicancer early detection” to identify relevant meeting proceedings, and reviewed the websites of MCED test manufacturers for the most current performance data. In this review, rather than providing an exhaustive description of technologies in a rapidly changing field, we focus on MCED tests with clinical validation data and ongoing trials.

By recent accounts, at least 15 MCED tests are in various phases of development in the United States and abroad (5,6). There is substantial variability in the number of cancers targeted, the specific cancer types targeted, and the technologies involved. Major types of analytes include cell-free or circulating tumor DNA, RNA, and extracellular vesicle–derived nucleic acids and proteins; there are also multianalyte approaches. Most MCED tests with completed clinical studies are based on cell-free DNA assays (Fig 1). Cell-free DNA is DNA that is circulating in plasma and that, while found in small amounts in healthy patients, is elevated and altered in individuals with cancer due to cellular apoptosis and tissue necrosis (7,8). Circulating tumor DNA is circulating DNA that derives from tumor cells specifically, and specific profiles by tumor type (eg, methylation patterns and mutations from whole-genome sequencing) enable identification of the cancer signal origin (CSO). RNA sequencing has also been employed, with utility derived from the fact that the RNA of platelets changes depending upon the presence of different types of cancer (9). Proteins and nucleic acids from exosome sources have also been evaluated for MCED test development (10).

Figure 1:

Examples of multi-cancer early detection tests and reported performance. Color coding for organ systems or diseases evaluated: gray = cancer site included but not individually reported, red = sensitivity of less than 50%, orange = sensitivity of 50%–69%, yellow = sensitivity of 70%–89%, green = sensitivity of 90% or greater. a = Organ system results reported in original phase II case-control study; a subsequent prospective cohort study of asymptomatic women with fixed specificity using CancerSEEK and protein markers reported a sensitivity of 27.1% (95% CI: 18.5, 37.1) and a specificity of 98.9% (95% CI: 98.7, 99.1) (13). b = Number of significant figures is presented as originally reported. c = Specificity was fixed at this value. d = A subsequent prospective cohort study in October 2023 reported a positive predictive value of 43.1% (95% CI: 31.2, 55.9). e = DNA fragmentation profiles. f = Sensitivity reported by stage of cancer; stage with largest sample size is represented in figure unless sample size is less than 10. g = Independent validation separately performed in a phase III study in China with reported sensitivity of 38% (95% CI: 31, 44) at a fixed specificity of 95%; validation in patients with early-stage cancer and healthy individuals had a reported sensitivity of 86.1% at 94.7% specificity. h = Sensitivity reported without specificities for each organ system. MCDBT = multicancer detection blood test, N/A = not available, SRFD = semi-reference-free deconvolution.

It is notable that, regardless of the underlying technology, all MCED tests that have reported performance by cancer stage have shown lower sensitivity for earlier stages (Fig 1). At least three MCED tests have been evaluated in phase III studies (assessment using cases vs noncases identified retrospectively), and two tests—Galleri from GRAIL and CancerSEEK from Exact Sciences—have been the subject of phase IV studies (assessment using cases vs noncases identified prospectively). Our review is largely centered on these data (11–13).

CancerSEEK (the forerunner to Cancerguard) was the first MCED test to receive Breakthrough Device designation from the U.S. Food and Drug Administration, in 2018 (13,14). CancerSEEK assesses for numerous DNA mutations through multiplex polymerase chain reaction and levels of particular proteins through a supplemental analysis for improved sensitivity and identification of probable CSO (15). The Breakthrough Device designation is granted when a device in development could provide “more effective treatment or diagnosis of life-threatening or irreversibly debilitating diseases or conditions” (16). Breakthrough Device designation also has implications for reimbursement that are under consideration by Congress. Medicare and Medicaid typically do not cover new health technology, and it can take approximately 2 years to build evidence for clinical utility and qualify for public reimbursement coverage. With Breakthrough Device designation, some devices or diagnostics could qualify for more rapid Medicare and Medicaid coverage under a proposed Transitional Coverage for Emerging Technologies pathway until standard public coverage begins (17).

The prospective DETECT-A (Detecting Cancers Earlier through Elective Mutation-based Blood Collection and Testing) study (13) of CancerSEEK enrolled 10 006 U.S. women aged 65–75 years who were highly adherent with standard-of-care cancer screening. Participants received an initial blood test, and participants with a positive test received a confirmatory blood test. Participants with a DNA or biomarker abnormality on both tests were considered to have a positive MCED result, and their cases were referred for review by a multidisciplinary committee. Those participants whose results could not be attributed to a “potential non-cancer-related cause” were invited to undergo contrast-enhanced full-body PET/CT diagnostic imaging to confirm disease status and identify the CSO. Of the 9911 participants with a baseline test, 134 (1.35%) had a positive MCED result on both the baseline and confirmatory blood tests. Of the 127 participants who underwent diagnostic imaging—including 116 who underwent full-body PET/CT—following multidisciplinary review, 26 were found to have cancer that was first detected with the blood test.

The Galleri test identifies circulating tumor DNA and methylation patterns. Galleri was granted Breakthrough Device designation in 2019 (2) and has been marketed by GRAIL since 2021 as a laboratory developed test under a Clinical Laboratory Improvement Amendments certification from the Centers for Medicare and Medicaid Services. Published evaluations include the Circulating Cell-free Genome Atlas study (18), a prospective case-control study, and PATHFINDER (11), a prospective cohort study. The Circulating Cell-free Genome Atlas study included an independent clinical validation of the panel developed and validated in prior phases of the study. The clinical validation analysis included assays in 4077 adults, of whom 2823 had confirmed cancer and 1254 were cancer-free. Test specificity was 99.5% (1248 of 1254). Across all cancers, test sensitivity was 51.5% (1453 of 2823) overall but varied by disease stage: 16.8% (143 of 849) for stage I, 40.4% (284 of 703) for stage II, 77.0% (436 of 566) for stage III, and 90.1% (557 of 618) for stage IV. Sensitivity also varied considerably by cancer type; sensitivity was 80% or higher for only 10 of more than 25 confirmed cancer types (excluding unknown primaries). In addition to providing a binary indicator for the presence of cancer, the Galleri test also provides a prediction of the CSO, localizing the result to a particular tissue type. The overall reported accuracy of the predicted CSO was 88.7% (1236 of 1393) in the clinical validation of the Circulating Cell-free Genome Atlas study (18).

PATHFINDER was a prospective study of the Galleri test in a cohort of 6662 U.S. adults aged 50 years or older, with cancer status confirmed at 1 year for all participants (11). Health care providers were not given specific instructions regarding the follow-up of “signal detected” results in these asymptomatic patients, there were no protocol-required diagnostic procedures, and a participant’s treating physician was responsible for all diagnostic decisions. Overall, 1.4% (92 of 6621) of evaluable participants had a positive test, and cancer was confirmed in 38% (35 of 92) of those cases (11). Test specificity was 99.1% (6235 of 6290). More than 92% of participants with a positive MCED result had at least one imaging test, and this proportion was similar regardless of ascertained cancer status. PATHFINDER 2 (19), another prospective interventional study, enrolled nearly 36 000 asymptomatic adults aged 50 years or older from numerous sites across North America. Participants are being followed for 3 years to evaluate safety and diagnostic performance of the Galleri test.

In addition, in partnership with the National Health Service of England and Wales, the Galleri test was evaluated prospectively in the SYMPLIFY trial (20). Unlike the PATHFINDER studies, which focus on disease screening in asymptomatic individuals, SYMPLIFY involves participants who were referred with nonspecific cancer symptoms or symptoms potentially associated with gynecologic, lung, or gastrointestinal cancers. Among 5461 participants with both an evaluable MCED test result and a diagnostic outcome, the test indicated cancer in 244 of the 368 individuals who were diagnosed with cancer (sensitivity, 66.3%), and the positive predictive value was 75.5% (244 of 323). Among the 5093 participants without cancer, 79 had a positive test (specificity, 98.4%). Accuracy of the predicted CSO among those with a positive test was 85.2% (208 of 244). In a separate study, the National Health Service is randomizing 140 000 asymptomatic adults to receive or not receive the Galleri test (21). The study will evaluate differences in late-stage cancer diagnosis and cancer-specific mortality 3.5 years after randomization.

Tests not based on cell-free DNA also exist. Unlike the Galleri and CancerSEEK tests, the GAGome test, developed by Elypta, uses plasma and urine glycosaminoglycan (GAG) profiles, or GAGomes, as biomarkers reflective of tumor metabolism. This approach could potentially be better than other technologies at detecting cancers that shed little cell-free DNA, such as brain and genitourinary cancers. In a case-control study using GAG profiles to test for 14 cancer types in more than 700 plasma samples and more than 500 urine samples (22), the urine GAG score achieved a sensitivity of 35% (33 of 95) at 98% specificity (153 of 156) in validation, and the plasma GAG score achieved a sensitivity of 41% (73 of 178) at 98% specificity (137 of 140). Prospective studies of the GAGome test include current enrollment of 2000 U.S. active-duty firefighters (23) (a group with elevated cancer risk compared with the general adult population [24]) and active enrollment of 9170 asymptomatic adults (25).

Other MCED tests in development rely primarily on analysis of cell-free DNA using various techniques (detection of methylation vs fragmentation) or use a combination of cell-free DNA and one or more types of biomarkers, including circulating proteins, RNA sequences, serum metabolites, and extracellular vesicles. Studies of a number of these tests are enrolling prospective observational cohorts in the United States (eg, a test from Adela) or abroad (eg, PanSEER, from Singlera Genomics, in China) (26,27). As of late 2023, two other tests received Breakthrough Device designation but have not yet been launched commercially (28,29). Many more so-called liquid biopsies are being developed and studied for single cancer screening and for detection of residual disease or disease recurrence following cancer treatment (30–33).

In summary, there are more than 15 tests in various phases of development, and no randomized controlled trials have been completed on cancer outcomes in a screening population. The sensitivity and specificity of these tests differ based on the specific technology, collection method, and cancer type and stage. The phases of MCED test evaluation and translation to clinical use may be conceptualized using the framework for early detection tests described below.

A Framework for MCED Test Evaluation

While the possibility of screening for numerous cancer types before symptom onset with a single blood test has engendered hope and excitement among clinicians, scientists, and disease advocates, the available evidence is incomplete for supporting widespread use of any MCED test. The current phase of development of MCED tests can be characterized using an established framework for cancer early detection tests (Fig 2) (34,35). Existing frameworks differ primarily in terms of the granularity of the study features composing the sequential phases. For simplicity, we focus on a framework with three phases: analytical validity, clinical validity, and clinical utility (34,36).

Figure 2:

Diagram of two established multiphase frameworks for developing (left) and evaluating (right) cancer early detection tests (34,36).

The analytical validity of a test is its accuracy and reliability, where accuracy reflects the probability that the measured value will be within a predefined range of the true activity or concentration, and reliability is the probability of replicating the same result. Clinical validity reflects measures of test performance, including the probability that the test will be positive in people with the disease (sensitivity), the probability that the test will be negative in people without the disease (specificity), and the probabilities that people with positive test results have the disease (positive predictive value) and that people with negative results do not have the disease (negative predictive value). Predictive value depends on sensitivity and specificity, which are characteristics of the test, but also on disease prevalence, which is a characteristic of the group being tested. It is now increasingly recognized that sensitivity and specificity can also depend on disease prevalence and on spectrum effects in the population under study (37). Spectrum effects reflect the disease severity and clinical presentation among patients in the diagnostic performance evaluation. Clinical utility is the ability of a test to prevent or reduce adverse health outcomes as a function of actions taken in response to test results. As Grosse and Khoury (38) note, “a screening or diagnostic test alone does not have inherent utility; because it is the adoption of therapeutic or preventive interventions that influences health outcomes, the clinical utility of a test depends on effective access to appropriate interventions.”3 We would add that clinical utility also depends on the effective delivery or administration of those interventions. In its broadest definition, clinical utility reflects an acceptable balance of benefits and harms across all health attributes considered important to individuals and families.

Under this framework, current evidence provides reasonable support for the analytical validity, and preliminary support for the clinical validity, of some MCED tests for cancer screening. The tests furthest along in their development have demonstrated analytical validity for the entities they measure and are being evaluated for clinical validity in the intended-use population. Appropriately, most developers have prioritized test specificity over sensitivity to constrain the likelihood of false-positive results. Consequently, test specificity and positive predictive value are generally high. Test sensitivity and negative predictive value, however, are generally lower, although precise performance varies considerably by cancer type and stage. Perhaps most importantly, we have no evidence yet of the clinical utility of MCED tests in asymptomatic individuals. Until long-term results from large randomized studies of clinical utility are published, we cannot assume that MCED tests will reduce cancer morbidity and mortality (39). In the absence of evidence of clinical utility, the promotion and dissemination of MCED tests is premature and potentially harmful. Randomized trials, simulation modeling studies based on them, and assessments by the U.S. Preventive Services Task Force and other recommendation-forming organizations (40,41) should be the basis for determining whether patients at average or high risk for particular cancers may benefit from MCED testing and whether such testing should be offered alongside or in place of existing screening, or possibly given in alternating sequence (Fig 3). In the same vein, establishing diagnostic pathways for positive MCED tests is crucial both for facilitating evaluation of clinical utility and for managing patients who are increasingly being tested ahead of clinical recommendations.

Figure 3:

Schematic for resolving a positive multi-cancer early detection (MCED) test and for evaluating its clinical validity. If an initial follow-up imaging or other confirmation test is negative, interval testing may be needed to rule out cancer. After a negative confirmation test, the time interval to a subsequent confirmation test may be lengthened.

Diagnostic Confirmation Pathways

A positive MCED test cannot be resolved until either a cancer is diagnosed or there is sufficient investigation to determine that no cancer is actually present (ie, that the test result was a false positive). A minimum of one other test with high sensitivity and specificity is needed to assess the predicted CSO and determine disease status. Confirmation tests are thus key to effective and efficient diagnosis. In addition, the type, number, and sequence of diagnostic tests—whether imaging, laboratory-based examination, endoscopy, or biopsy—may vary widely across patients if providers are not given clear guidance about appropriate ways to resolve the MCED test result. Further, definitive resolution of the test result as a false positive or identification of its exact source may be challenging, creating a pernicious uncertainty in some respects akin to genetic tests yielding variants of uncertain significance (42).

Given the current uncertainties about preferred pathways for diagnostic workup, defining a reasonable process to determine definitive disease status is urgently needed. Without consensus or expert guidance, judicious use of advanced imaging and avoidance of unnecessary invasive diagnostic procedures may be challenging for the health care system and for patients. Even if one reasonably accurate test is negative for a neoplasm (eg, CT for pancreatic cancer), additional testing (eg, PET) may be pursued to enhance sensitivity for rapid resolution. The risk-benefit considerations of each potential diagnostic pathway should be weighed according to the clinical consequences and burden faced by patients, providers, and the health care system. The type of imaging test, the possible need for repeat or additional imaging studies, and the time frame for conclusions about disease status are all key considerations for establishing the appropriate confirmation testing pathway and, ultimately, the clinical utility of an MCED test (Fig 4). Imaging regimens will partially determine the impact of MCED tests on late-stage cancer diagnoses and population-level health gains as well as the harms, including the costs and inconvenience of unnecessary tests and possible exposure to invasive diagnostic procedures associated with false-positive test results.

Figure 4:

Applied framework for evaluating multi-cancer early detection (MCED) tests and determining how to incorporate them into routine screening. Evidence may or may not indicate sufficient performance characteristics of a test to support its use in average-risk populations. For high-risk populations, combined or alternating MCED and standard-of-care testing may be considered, even if it is not recommended in the general, average-risk population.

For example, in the PATHFINDER study (11), the diagnostic pathways for positive Galleri results were not prescribed by the protocol and were thus selected by individual clinicians. Investigators analyzed the use of whole-body imaging and the rate of diagnostic resolution. Among 90 participants with a Galleri-predicted CSO, 83 (92%) underwent at least one imaging test. Fifty-five patients (61%) underwent PET/CT and 35 patients (39%) underwent full-body CT, with some patients receiving multiple tests. These groups included 35 patients who underwent PET/CT and 20 patients who underwent CT for what were ultimately determined to be false-positive results. And in 17 of 57 patients (30%) with false-positive results, the diagnostic workup included invasive procedures. While the first or second CSO prediction was still 97% accurate among diagnoses (33 of 34 diagnoses), and patients who did have cancer had a median time to resolution of 57 days, patients in whom cancer was not found had a median time to resolution of 162 days. The positive predictive value was 38% (35 of 92), but this increased to 43% (24 of 56) with use of a “refined” version of the test (11).

Even when diagnostic pathways have been prescribed by investigators based on clinical judgment, the testing cascade in trials has been complex. For evaluation of CancerSEEK results in the DETECT-A study (13), which enrolled 10 006 women with no prior history of cancer, the protocol set pathways for diagnostic resolution that entailed confirmatory evaluation with the baseline test as well as multidisciplinary committee discussion to determine whether any signal was likely due to non-cancer-related confounders or required diagnostic imaging evaluation. CancerSEEK tests for the presence of mutations in specific regions of 16 genes. For 4.9% of participants (490 of 9911 women with a baseline test), a positive test result prompted an additional blood test for confirmation, and after this additional test and multidisciplinary discussion, 127 participants underwent full-body imaging (mostly PET/CT). Among these women, 64 had results that were concerning for cancer, and of these, 26 were confirmed to have cancer; 63 had no sign of malignancy and were determined to be negative for disease. Associated performance characteristics were 27.1% sensitivity (26 of 96), 98.9% specificity (9707 of 9815), 19.4% positive predictive value (26 of 134), and 99.3% negative predictive value (9707 of 9777). Based on these results, to diagnose one cancer, there need to be 661 blood tests with confirmatory PET/CT or 381 blood tests with any confirmatory scan. For every five diagnosed cancers, two additional cases of cancer would be undetected with the test and potentially found using another available screening modality. Further, in DETECT-A, repeat imaging was performed after 6 months when recommended by the multidisciplinary committee. The time to confirmation for negative disease status (ie, confirmation that an unresolved positive test was a false-positive test) was set to 12 months after enrollment, and a subsequent longer-term study with at least 4 years of follow-up confirmed the low rate (<1%) of cancers developing in the putative false-positive group (12). These controlled conditions suggest that substantial time and effort is required to evaluate initial screening results, which may impact the value of broad MCED test use.

Under real-world conditions, the imperfect performance of the Galleri test in identifying the CSO has led to accuracy being reported for the top two predictions, rather than the top one prediction, such that additional diagnostic testing may be needed to evaluate both possibilities (11). The need to investigate two predicted organs of origin, if reported, expands the number of diagnostic pathways and increases the potential overuse of full-body imaging when the indicated organs are anatomically distant. Notably, several single-cancer liquid biopsies or anatomy-based MCED tests (eg, for thoracic cancers) are also under development, with the intention of focusing the clinical evaluation on particular selected cancers and avoiding the tangle of testing and decisions that arises when investigating multiple possible cancer sites (30–33).

Implications of MCED Tests for Current Imaging-based Screening

Another important facet of MCED test dissemination is considering the merits of these tests in the context of other screening tests that may already be in widespread use or under investigation for widespread use. The clinical utility of using an MCED test when no other screening test is established will continue to be assessed in prospective cohort studies and randomized clinical trials. Imaging alternatives to MCED tests could offer comparable or better performance and warrant comparative effectiveness studies whether screening is used in the general population or targeted to particular patients with increased risk. When an established screening test is recommended by the U.S. Preventive Services Task Force and covered by health plans, comparative effectiveness studies should evaluate whether and how to combine MCED tests with the established screening test and schedule. These studies may find that a certain screening approach performs best, such as using existing tests only, MCED tests only, synchronous combination, alternating combination, or MCED tests as an occasional substitute (Fig 3). An important facet of such evaluations will be risk-benefit profiles of MCED testing in series or combined with imaging depending upon patient risk strata. Patients with high risk may have different numbers of incident cases detected with screening and may experience different treatment patterns or mortality rates compared with patients with average risk.

Another important question when developing clinical policy around MCED use, with or without currently recommended screening, is when and whether to consider the extent of current patient and health system burdens, particularly as they relate to costly follow-up imaging, barriers to access, and disparities in early detection. Of particular interest has been the possibility that an MCED product could help to address health inequities in cancer screening and cancer outcomes by increasing access to early detection programs (2,43). Proposed mechanisms include reducing barriers, such as time needed to undergo first-line screening tests, and offering an alternative to endoscopy-based screening, for which procedural risks, anesthesia, and necessity of an escort home generate patient hesitancy. However, a full understanding of the benefit and burden of MCED test use for patients must also include consideration of geographic and financial access to diagnostic confirmation tests. In rural areas, for example, a patient might need to travel quite far or face a substantial wait time for advanced imaging to confirm a positive MCED test result. And cost sharing for advanced imaging to confirm positive tests could result in positive tests going unconfirmed or could burden patients with substantial out-of-pocket expenses simply to rule out cancer after a positive test.

MCED tests that are complementary to established screening modalities could be evaluated for use in special populations, such as those with increased cancer risks or those that prefer to avoid repeat imaging (Table). For example, if a patient with chronic obstructive pulmonary disease and elevated risk of smoking-related cancers feels that repeat CT scans are burdensome and indicates a strong preference for screening with a blood test, such a test might replace CT scans in alternating years if there is an informed choice about the associated tradeoffs (ie, knowledge of the potential cascade of tests after an MCED test and a strong preference to lessen the physical burden of screening). Evaluations of clinical utility will likely require further trials with cancer-specific mortality as the primary end point for one or more testing strategies. Such studies would likely be complemented by mathematical or computer simulation–based analyses to evaluate a wide variety of screening modalities, frequencies, combinations, and sequences (44,45).

Possible Screening Pathways Using Existing Tests and/or MCED Tests

Patient-centered Screening Considerations

Once the clinical consequences of practical MCED testing strategies are better characterized, the inevitably imperfect performance of MCED tests will require educational initiatives to inform the public about the benefits and drawbacks of different options. Patients will need accurate and accessible information about the risks of false-positive and false-negative results and the comparative performance of different screening tests to make informed decisions. Furthermore, patients would ideally participate in shared decision-making about the use of these tests (46). From an ethical perspective, informed decision-making involving patients is particularly necessary in the early phases of development of MCED tests, when there is greater uncertainty about diagnostic performance and long-term utility. At the same time, primary care providers will need education and guidance, ideally from the radiologists and other specialists who conduct the diagnostic confirmation tests, about the types and timing of such tests, so they can help patients make informed decisions.

A patient-centered approach to selecting appropriate cancer screening tests should include clear communication about risk-benefit tradeoffs as well as uncertainty about immediate and long-term outcomes (47). Discussions necessary for informed decision-making should address screening options and the subsequent diagnostic pathways, and incorporate patients’ health-related values and priorities. For example, some patients may have strong aversions to unnecessary imaging or invasive procedures and may feel that the potential benefit of an MCED test does not outweigh the potential harms. Other patients may be concerned about cancer morbidity and mortality to an extent that they are willing to accept the possibility of a false-positive screening result, and the follow-up required, if it affords them even a very small—and as yet unproven—reduction in cancer mortality.

Critical Evidence Gaps and Further Steps for Evaluation

Critical gaps remain in our understanding of the clinical validity and utility of the current generation of MCED tests, but data from randomized trials in the intended-use population and follow-up long enough to identify impacts on cancer mortality will generally be unavailable for evidence-based screening recommendations in the near future. Meanwhile, Galleri and OneTest (20/20 GeneSystems) are already available for purchase by consumers, although not covered by insurance.

Analysis of the consequences of MCED-based screening should include consideration of the costs of the MCED test and associated diagnostic tests, procedures, or treatments and other services related to MCED results, alongside health outcomes. For example, if many lethal cancers are found and treated early as a result of MCED testing, and early detection and treatment leads to reduced morbidity and palliative care for advanced disease, then costs may be materially reduced. In contrast, if MCED testing yields many spurious findings, and many indolent cancers are found and treated, then costs may be materially increased. This latter outcome, commonly known as overdiagnosis, would lead to increased cost without additional improvement in morbidity or mortality.

In terms of international policy, national health programs in the United Kingdom, Australia, Canada, and many other countries may limit possible diagnostic paradigms incorporating MCED tests based on cost-effectiveness analyses that capture relevant benefits and harms (48–50). MCED tests and the diagnostic testing pathways required to evaluate positive results may be subject to formal technology assessment before they are covered by public health insurance systems. In the United States, variability in insurance coverage and available tests is likely to persist until one or more MCED tests are recommended or endorsed by the U.S. Preventive Services Task Force for population-based cancer screening, and even then, confirmatory test coverage may vary widely.

While large prospective cohort studies and randomized clinical trials are underway to investigate clinical validity and utility, a prudent approach to diagnostic confirmation of positive MCED tests at this time might be evidence-informed consensus recommendations based on the current diagnostic imaging literature. Although the available literature often does not reflect clinical validity in the intended-use asymptomatic population, published performance metrics still provide useful upper bounds for the plausible impact of tests on cancer-specific outcomes (6). Agreement on the key missing pieces of evidence and the framework under which to rigorously assess these measures, gather feedback, and iterate recommendations for screening will be essential for establishing best practices and making efficient use of available health care resources.

Finally, little is known about what strategies will be needed to promote the widespread uptake of MCED tests. Implementation science, or the study of methods for improving the uptake of evidence-based interventions (51), involves a theory-informed, multistep process of evaluating the current evidence, examining barriers and facilitators to implementation, and finally designing and evaluating implementation strategies (eg, bundles of interventions that target barriers to the use of MCED tests). While efforts to build the evidence for MCED tests are underway, there have been few studies examining determinants of MCED test uptake.

A small survey found that 93% (25 of 27) of primary care physicians would be supportive of patients participating in clinical MCED trials (52). However, in a scoping review of perceptions of liquid biopsy and MCED testing (53), no systematic evaluations of the perspectives of primary care physicians on using MCED tests in clinical practice were found, and nearly all patient perspectives focused on patient views of liquid biopsy for a single cancer (eg, a colorectal cancer screening test) compared with other screening options, with patients citing noninvasiveness and perceived safety, convenience, and effectiveness as favorable features of testing. An industry-funded publication (54) suggested that patients were receptive to MCED testing, particularly as a complement to currently recommended cancer screening tests. Patients preferred MCED testing to no screening at all, with attitudes, priorities, and perceptions surrounding cancer screening mediating willingness to opt into an MCED test. Test accuracy and ability to detect multiple forms of cancer were high patient priorities; over 90% (1562 of 1700) of the patients surveyed were White, and the majority had a college education (54).

Patients on one community cancer advisory board highlighted several unanswered questions relating to MCED test acceptability (eg, accuracy, side effects), unintended consequences (eg, loss of insurance, false assurance if negative), access, affordability, and accountability, as well as several strategies that will need to be in place to support uptake (eg, navigation programs to address financial, follow-up testing, and mental health needs that emerge from positive tests) (55). Overall, we found few if any examples of equity-informed, multi-level (ie, patient, provider, system), theory-informed stakeholder perspectives in the literature. Rigorous mixed methods studies that examine whether attitudes, barriers, facilitators, and decision-making related to MCED testing differ among patient and clinician groups with varying access to resources will be integral to ensuring equitable implementation (55). Ongoing effectiveness and cost-effectiveness studies will also offer an opportunity to concurrently evaluate implementation outcomes (56), such as acceptability, feasibility, appropriateness, and fidelity to various MCED screening programs.

Conclusion

Radiologists have a crucial role in advancing science, clinical care, and policy related to multi-cancer early detection (MCED) tests. In real-world assessment, imagers can provide unique patient-specific insights when estimating the likelihood of cancer at imaging after positive MCED test results. Accumulating evidence about sources and frequencies of false-positive and false-negative results will provide imaging context for investigating the sources of these errors and further evidence on the clinical validity of these tests. At the health system level, many clinicians, including radiologists, will likely be asked to contribute to the development of guidelines and clinical pathways and to provide input for insurance coverage and reimbursement decisions. Further, patients and physicians weighing options for screening will need ongoing decision support regarding diagnostic innovations to determine appropriate screening options based on patients’ personal cancer risks, health-related values, and access to care.