RISK FACTORS OF POST-COLONOSCOPY COLORECTAL CANCER IN A LARGE INTEGRATED HEALTHCARE SYSTEM

Background: Post-colonoscopy colorectal cancer (PCCRC), defined as cancer diagnosed after a colonoscopy in which no cancer is found, offers unique insight into our current screening and surveillance practices and how these may need to be adjusted to account for specific patient characteristics and suboptimal colonoscopy quality. Prior work from our group and others has demonstrated that 54-96% of PCCRC is likely due to missed lesions or inadequate exams. However, the specific risk factors for PCCRC remain poorly characterized.

Methods: We assembled a longitudinal cohort of patients who underwent an index colonoscopy between 2007-2018 in a large healthcare system (Mass General Brigham) by obtaining detailed endoscopic, pathologic, and other clinical data from electronic health records. Among individuals who had a CRC-free index colonoscopy, we identified those who were subsequently diagnosed with CRC between 6 and 48 months (PCCRC<4y) and at least 48 months (PCCRC≥4y) later. Multivariable logistic regression was used to compare patient characteristics (including demographics and polyp features) and quality indicators of index colonoscopy (including cecal intubation, bowel prep quality, adenoma detection rate (ADR), and serrated polyp detection rate (SDR)) between PCCRC<4y and PCCRC≥4y cases (Figure 1).

Results: Among 214,189 patients with a CRC-free index colonoscopy, we documented 218 PCCRC<4y and 255 PCCRC≥4y cases (females: 46% vs. 47%; mean age: 69.2 vs. 71.5 years; median interval between index colonoscopy and diagnosis: 2.3 vs. 6.7 years). We did not find a statistically significant difference in cecal intubation and bowel prep quality between PCCRC<4y and PCCRC≥4y cases (Table 1). While previous studies based in Europe, Asia, and California have identified an inverse relationship between ADR and risk of interval colorectal cancer, our current data demonstrated that PCCRC<4y did not have a lower ADR or SDR than PCCRC≥4y cases. On the other hand, compared to PCCRC≥4y, we found that patients with PCCRC<4y were more likely to have a large (≥1 cm) polyp (multivariable OR=3.02, 95% CI=1.38-6.61, p=0.006) or a proximal polyp (multivariable OR=1.95, 95% CI=1.06-3.60, p=0.03) in the index colonoscopy.

Conclusions: Our findings demonstrate that patients with a large or proximal polyp in the index colonoscopy were more likely to develop PCCRC<4y compared to PCCRC≥4y. In contrast, the commonly used colonoscopy quality indicators (e.g., cecal intubation, bowel prep quality, ADR, and SDR) were not associated with PCCRC<4y. Efforts are currently underway to collect and analyze additional procedural factors including endoscopic resection technique and completeness as well as other quality-related indices (e.g., advanced ADR and proximal ADR/SDR) in relation to PCCRC.