Counting Advanced Pre-Cancerous Lesions as True Positives When Determining Colorectal Cancer Screening Test Specificity
A l'aide d'une modélisation, cette étude analyse la pertinence des seuils de sensibilité et de spécificité fixés par les "Centers for Medicare & Medicaid Services" pour les biomarqueurs sanguins du cancer colorectal lorsque sont également prises en compte les lésions précancéreuses de stade avancé
The landmark Centers for Medicare & Medicaid Services (CMS) Decision Memo on blood-based biomarkers to screen for colorectal cancer (CRC) sets thresholds of ≥ 74% for sensitivity and ≥90% for specificity for CRC. This approach does not consider detection of advanced precancerous lesions as true positives. We contrasted the impact of counting advanced precancerous lesions as true vs. false positives, and projected CRC outcomes under contrasting tests in a validated model. A test with the threshold performance set by CMS decreased CRC incidence by 30% and CRC mortality by 48% in 45-year-olds. If this test also detected advanced precancerous lesions with 30% sensitivity, CRC incidence decreased by 45% and mortality by 58%, but the test’s CRC specificity of only 88% would not satisfy the CMS threshold. CMS should reconsider its definition of threshold specificity for CRC screening biomarkers. Future coverage determinations on biomarkers to screen for cancer should consider detection of relevant precursor lesions, and projected outcomes.
Journal of the National Cancer Institute , résumé, 2021