Sensitivity, effect and overdiagnosis in screening for cancers with detectable pre-invasive phase

Area of Science:

• Oncology
• Public Health
• Biostatistics

Background:

• Cancer screening studies often rely on indirect performance indicators.
• Detecting pre-invasive lesions in screening can represent both benefits (sensitivity, effect) and harms (overdiagnosis).

Purpose of the Study:

• To develop and define formulae for sensitivity, effect, and overdiagnosis in cancer pre-invasive lesion screening.
• To establish distinct indicators for laboratory, clinical, and population-level screening purposes.

Main Methods:

• Formulated sensitivity as the ability to identify progressive lesions at test, episode, and program levels.
• Defined effect as the reduction in cancer incidence (efficacy and effectiveness).
• Quantified overdiagnosis as the detection rate of pre-invasive lesions minus prevented invasive cancers, with corrections for attendance and risk.

Main Results:

• Developed distinct formulae for sensitivity, effect, and overdiagnosis in cancer screening.
• Demonstrated the application using cervix cancer screening data (Papnet®).
• Showed that estimated sensitivity and effect align with screening's goal of preventing invasive disease.

Conclusions:

• The developed formulae provide a robust framework for evaluating cancer screening performance.
• The study defines six specific estimators for laboratory, clinical medicine, and public health contexts.
• Accurate measurement of sensitivity, effect, and overdiagnosis is crucial for optimizing cancer screening programs.