Impact of subsequent screening episodes on the positive predictive value for advanced neoplasia and on the distribution of anatomic subsites of colorectal cancer: a population-based study on behalf of the French colorectal cancer screening program

Background: the anatomic distribution of advanced colorectal neoplasia is increasingly important for choosing screening strategies and treatment options. We sought to evaluate the impact of repeated screening on the positive predictive value (PPV) for advanced colorectal neoplasia (advanced adenoma, AA, and colorectal cancer, CRC) and their distribution according to anatomic subsite distribution in average-risk adults. Method: the study included 98,031 men and women aged 50-74 who had a positive g-FOBT in 2010 and 2011 and underwent total colonoscopy. The PPV for detection of AA and CRC and the relative risks were determined with log-binomial models, and the distribution of anatomic subsites was estimated according to screening history. Result: the median age was 61 years (62 years for participants with AA and 64 for those with CRC). The PPV for detection of advanced neoplasia was 24.5%, substantially higher in men than women (30.7% vs 17.7%), and it increased with age. It also fell at all screening episodes after the first. Subsequent screening episodes were associated with an increased RR for proximal AA (RR 1.13, 95% CI 1.16 1.20). Advancing age (RR 1.28, 95% CI 1.19 1.39 for every 10-year increase in age), female gender (RR 1.31, 95% CI 1.19 1.44), and subsequent screening (RR 1.15, 95% CI 1.04 1.27) were significantly and independently associated with detection of proximal adenocarcinoma. The latter was also detected at an advanced stage more often (RR, 1.24, 95% CI: 1.09 1.42). Early stages of invasive adenocarcinoma (stages I and II) was more likely to be detected in a subsequent than an initial screening (RR 1.07, 95% CI 1.01 1.13). Conclusion: this study found that subsequent screening episodes using g-FOBT were associated with an increase in the detection rate of proximal AA and CRC, especially among women. The more frequent detection of proximal invasive adenocarcinoma at an advanced stage in subsequent screenings suggests that some of these tumors may well not be real incident lesions, but are likely to include lesions that were missed on the previous screens. Although modest, the increase in the rate of detection of invasive adenocarcinoma at early (and more curable) stages from the first to subsequent screenings, together with this potential for missed diagnoses on initial screening and the increased detection rate for proximal or rectal AA in subsequent screening episodes, underlines the need to reinforce the population's awareness of the importance of regular consistent screening, after negative results.