Methods: We conducted a systematic search in PubMed, Embase, the Cochrane Library, and Scopus for published articles until July 2017. We estimated pooled prevalence or odds ratios (ORs) with 95 % confidence intervals (CIs), using random-effects models. We classified endoscopic subtypes into granular LST, which comprises the homogeneous and nodular mixed subtypes, and non-granular LST, which comprises the flat elevated and pseudodepressed subtypes.
Results: We identified 2949 studies, of which 48 were included. Overall, 8.5 % (95 %CI 6.5 % - 10.5 %) of LSTs contained SMI. The risk of SMI differed among the LST subtypes: 31.6 % in non-granular pseudodepressed LSTs (95 %CI 19.8 % - 43.4 %), 10.5 % in granular nodular mixed LSTs (95 %CI 5.9 % - 15.1 %), 4.9 % in non-granular flat elevated LSTs (95 %CI 2.1 % - 7.8 %), and 0.5 % in granular homogenous LSTs (95 %CI 0.1 % - 1.0 %). SMI was more common in distally rather than in proximally located LSTs (OR 2.50, 95 %CI 1.24 - 5.02). The proportion of SMI increased with lesion size (10 - 19 mm, 4.6 %; 20 - 29 mm, 9.2 %; ≥ 30 mm, 16.5 %). The pooled prevalence of patients with one or more LSTs in the general colonoscopy population was 0.8 % (95 %CI 0.6 % - 1.1 %).
Conclusion: The majority of LSTs are non-invasive at the time of colonoscopic detection and can be treated with (piecemeal) endoscopic mucosal resection. Pretreatment diagnosis of endoscopic subtype, specifying areas of concern (nodule or depression), determines those LSTs at highest risk of containing SMI, where en bloc resection is the preferred therapy.