Multicancer Detection Tests: No Evidence for Benefit

Blood-based multicancer detection tests offer patients the promise of simple screening for multiple cancer types, and they are increasingly promoted to the lay public.

But no controlled studies have shown multicancer detection tests to be beneficial for reducing cancer mortality and the evidence for their accuracy and harms remains inconsistent, according to a newly published systematic review of the research.

The analysis identified 20 studies with roughly 109,000 participants, but none were controlled, randomized trials reporting the impact of multicancer detection (MCD) screening on cancer detection, mortality, or quality of life.

Accuracy varied by test and study design, but the available evidence was found to be "insufficient to evaluate harms and accuracy," wrote researcher Leila Kahwati, MD, of RTI International, Research Triangle Park, North Carolina, in Annals of Internal Medicine.

The commercially available multicancer detection tests in the United States have not been approved by the U.S. Food and Drug Administration for multicancer screening.

Kahwati and colleagues noted that three of the MCD tests included in the systematic review are available with prescription as laboratory-developed tests, which do not require FDA approval (OneTest, Galleri, OncoSeek). Another (Cancerguard) is available through an FDA investigational device exemption associated with a real-world evidence registry.

"No tests have published evidence on use for screening," they wrote. "Although a test could be accurate for identifying types of cancer, accuracy evidence alone does not mean MCD tests have a greater benefit than currently recommended screening tests."

The researchers searched Cochrane Library, trial registries, and relevant websites through March 2025 to identify relevant controlled studies of MCD tests in asymptomatic populations reporting cancer detection, mortality, quality of life, and harms (psychosocial, adverse events, decrease in standard-of-care screening); uncontrolled studies for harms of diagnostic evaluation; test accuracy studies.

They found no controlled studies which evaluated the benefits of screening, while 20 studies (n=109,177 participants) reported accuracy for 19 unique MCD tests.

Seven studies reported the accuracy of future cancer detection in asymptomatic people followed for 1 year (prediagnostic performance), but 5 of these studies had a high risk of bias (ROB), and 2 had an unclear ROB. The remaining studies "estimated accuracy from high ROB case-control studies in clinically confirmed cancer cases and healthy, cancer-free, control participants (diagnostic performance)."

Among the findings:

• Sensitivity ranged from 0.095 to 0.998 across the included tests, with specificity ranging from 0.657 to 1.0, and area under the curve (AUC) ranging from 0.52 to 1.0.
• Sensitivity and AUC were higher in diagnostic performance compared with pre-diagnostic performance studies.
• One cohort study reported harms, but the data were limited.

Kahwati et al. noted that the most significant study limitation was the "absence of completed, controlled studies assessing the direct benefits of screening."

"Furthermore, we identified only 1 study with an unclear ROB that reported on harms, providing limited evidence about harms from using MCD tests. Most test accuracy studies were rated as high ROB inherent in the study designs used because of potential for selection and spectrum bias from use of clear-cut cases and healthy control participants"

The researchers further noted that to minimize the risk for false positives across multiple cancer types, specificity in MCDs must be set high, which results in lower sensitivity

"Although specificity is high, given the potential use in a general population with a low prevalence of cancer, this could still result in a high rate of false positives as evidenced by low positive predictive value estimates ranging from 0.01 to 0.43, which will result in potentially unnecessary diagnostic testing," they wrote.

In an editorial published with the study, David Weinberg, MD, of Fox Chase Cancer Center, Philadelphia, wrote that the lack of data identified in the "rigorous, comprehensive" systematic review by Kahwati and colleagues, "should give pause to any patient or clinician ready to embrace MCD tests today."

"Important clinical decision making requires powerful, convincing evidence," Weinberg wrote, while parenthetically adding that "the future of MCD tests is not foretold."

"The already impressive science is likely to improve. For now, a major unmet challenge is how to detect cancer precursor lesions like colon adenomas" Weinberg wrote.

"Unlike standard-of-care cervical or colorectal screening, MCD tests cannot prevent cancer, an outcome always preferable to detecting it earlier. But all screening tests are imperfect. The time and cost of definitive randomized controlled trials to demonstrate clinical benefit are great. Proving that a screening test, new or old, ’bends the cancer curve’ is challenging."

The editorial writer noted that more comprehensive trials of MCD tests are needed, such as the US National Cancer Institute’s recently initiated Vanguard study, which should be adequately powered to address many currently unanswered questions about the impact of MCD testing in different subgroups.

"The prevention potential for MCD tests is exciting, but there are insufficient data demonstrating true clinical benefit and how to weigh the balance of that benefit to potential harms," Weinberg concluded. "Despite market pressures to move faster, we need such data to make the best decisions. Until we have them, it is hard to justify testing outside of a research setting."

This research was funded by the Agency for Healthcare Research and Quality. Kahwati reported no relevant disclosures related to this study.

Kahwati LC, et al "Multicancer detection tests for screening: a systematic review" Ann Intern Med 2025; DOI: 10.7326/ANNALS-25-01877.

Weinberg DS "Multicancer detection tests: not ready for popular use," Ann Intern Med 2025; DOI: 10.7326/ANNALS-25-03530.