Clinical Significance of a Multicancer Screening Trial With Stage-Based End Points

The first randomized screening trial of a multicancer early detection test is ongoing, with the primary end point being the incidence of late-stage cancer. The unprecedented use of a stage-based end point and short 3-year follow-up raise questions about how results should be interpreted and used in developing multicancer screening policy.

To estimate outcomes of the trial and to identify information from the estimates that may aid in interpreting results of short-term trials evaluating multicancer screening tests.

This multicancer decision-analytic model estimated outcomes from registry data from the England National Cancer Registration and Analysis Service for cases diagnosed between January 2013 and December 2018. This model simulated a population-based multicancer screening trial of average-risk participants without a prior cancer diagnosis. The analysis was performed between April 2024 and April 2025.

Three annual multicancer screenings at months 0, 12, and 24. Cancers were assumed detectable 1 or 2 years before clinical diagnosis based on a published analysis, and cancer-specific early-stage sensitivities were set to either 100% or 50% of published sensitivities among clinically diagnosed cases.

Reductions in late-stage (stage III-IV) cancer incidence over 3 years, cancer mortality over 5 years, and contributions of each cancer type to reductions in late-stage incidence and cancer mortality across the range of detectable intervals and early-stage sensitivities.

The model simulated 70 000 participants per arm (screening and control; median age, 66 years), and estimated that the overall late-stage incidence reductions at 3 years ranged from 6% to 23%, with corresponding reductions in 5-year cancer mortality from 6% to 9%. Colorectal cancer contributed the most to the reduction in late-stage incidence (28% to 39%), while lung cancer contributed the most to mortality reduction (40% to 42%).

This independent decision-analytic model study found that the trial could achieve nontrivial cancer downstaging over 3 years, but modest mortality reduction over 5 years. These results were due to a limited number of target cancer types, underscoring the importance of transparent reporting of outcomes by cancer type, consideration of mortality implications, and careful preplanning for subsequent evaluation steps by the cancer research community.