A simplified algorithm to evaluate the risk of submucosal invasive cancer in large (≥20 mm) nonpedunculated colonic polyps

Abstract

Background and aims: Recognition of submucosal invasive cancer (SMIC) in large (20mm) non-pedunculated colonic polyps (LNPCPs) informs selection of the optimal resection strategy. LNPCP location, morphology and size influence the risk of SMIC, however currently no meaningful application of this information has simplified the process to make it accessible and broadly applicable. We developed a decision-making algorithm to simplify the identification of LNPCP subtypes with increased risk of potential SMIC. Methods: Patients referred for LNPCP resection from September 2008-November 2022 were enrolled. LNPCPs with SMIC were identified from endoscopic resection specimens, lesion biopsies, or surgical outcomes. Decision tree analysis of lesion characteristics identified in multivariable analysis was used to create a hierarchical classification of SMIC prevalence. Results: 2451 LNPCPs were analysed. 1289 (52.6%) were flat, 1043 (42.6%) nodular and 118 (4.8%) depressed. SMIC was confirmed in 273 (11.1%) of LNPCPs and associated with depressed and nodular versus flat morphology (OR 35.7 CI 22.6-56.5 and 3.5 CI 2.6-4.9, p<0.001 respectively); left versus right colon location (OR 3.2, CI 2.5-4.1, p<0.001); non-granular (NG) versus granular (G) (OR 2.4 CI 1.9-3.1, p<0.001) and size (OR 1.12 per 10mm increase CI 1.05-1.19, p<0.001). Decision tree analysis targeting SMIC identified 8 terminal nodes: SMIC prevalence was 62% in depressed LNPCPs, 19% in nodular left colon LNPCPs and 20% in nodular right colon NG LNPCPs. Conclusions: This decision-making algorithm simplifies identification of LNPCPs with an increased risk of potential SMIC. When combined with surface optical evaluation, it facilitates accurate lesion characterisation and resection choices.