Quantifying Overdiagnosis for Multicancer Detection Tests: A Novel Method

Area of Science:

• Oncology
• Biostatistics
• Medical Screening

Background:

• Multicancer detection (MCD) tests identify preclinical cancers from blood samples.
• Overdiagnosis, detecting cancers that won't become symptomatic, is a significant concern with screening, potentially leading to harmful treatments.
• Quantifying overdiagnosis, particularly the screen overdiagnosis fraction (SOF), is essential but challenging, especially for rapidly evolving MCD technologies.

Purpose of the Study:

• To introduce a novel method for estimating the average screen overdiagnosis fraction (SOF) for multicancer detection (MCD) screening programs.
• To address the difficulty in estimating SOF due to the unobserved nature of overdiagnosis and the need for short-term data with MCD tests.
• To develop a method applicable to cancers without conventional screening and robust to technological changes.

Main Methods:

• A new method is proposed requiring at least two annual MCD tests in individuals across a range of ages.
• The method assumes an exponential distribution for the sojourn time in the operational preclinical cancer (OPC) state.
• An SOF plot is generated, graphing average SOF against mean sojourn time, using lung cancer screening and synthetic data.

Main Results:

• The proposed SOF plot method demonstrated the ability to distinguish between small and moderate levels of SOF.
• The method's reliance on only one term with the exponential distribution assumption makes its results robust to violations.
• The study provides a new tool for estimating SOF, particularly valuable for MCD tests where short-term data is available.

Conclusions:

• The novel SOF plot method offers a complementary approach to existing models for estimating overdiagnosis in MCD screening.
• This method is particularly useful for assessing the risk of overdiagnosis in new screening technologies like MCD tests.
• Further application of the SOF plot is recommended as more short-term observational data for MCD tests becomes available.