Background: Multicancer early detection tests currently under development could be used to screen for multiple cancer at the same time, including many cancers for which no screening test currently exists. However, screening may lead to overdiagnosis of indolent tumours that would not have become clinically evident during that person’s lifetime.

Objectives: We assessed the potential for screening overdiagnosis in the context of population-based screening with a generic multicancer early detection test.

Methods: We used a microsimulation model of cancer health trajectories and survival to assess the potential population-level impacts of routine screening in Canada at ages 50-75 years with a multicancer early detection test. The model reproduces cancer incidence and survival outcomes across 25 different cancer sites based on cancer epidemiological data. We assumed cancers would have a preclinical screen-detectable period of 2-5 years, and test sensitivity that increases with cancer stage (ranging from 20-80% sensitivity), and over 99% test specificity. A pool of latent indolent cancers is modeled, representing cancers that would not normally be diagnosed within that person’s lifetime, but can be overdiagnosed through screen-detection.

Results: Assuming a range of screen-detectable periods and sensitivities, up to 2.1-6.0% of all screen-detected cancers with a multicancer screening test would be expected to be overdiagnoses. The most important factor determining the potential for overdiagnosis was the preclinical screen-detectable period, with longer screen-detectable periods leading to a higher fraction of overdiagnosis. The proportion of overdiagnosis was expected to be higher in men (7.3%) than in women (4.5%), and strongly increased with age, going from only 1% at age 50 to over 10% of all screen-detected cancers by age 75. Overdiagnosis proportions varied by cancer site, and were predicted to be lowest for testicular cancer (2.4%) and highest for esophageal cancer (7.7%). Introduction of the screening program led to a short-term increase in overall cancer incidence from screen-detection but returned to pre-screening incidence levels by 5 years after implementation of the screening program. The positive predictive value of a screening test was predicted to range from 15.9-77.6% across sensitivity analyses.

Conclusions: Population-level multicancer screening with a multicancer early detection test would likely detect many preclinical cancers without leading to substantial screen-related overdiagnosis. Potential for overdiagnosis should be considered when assessing the balance of benefits and harms of multicancer screening in different populations, as harms from overdiagnosis are predicted to increase with age.