Non granular and granular mixed Laterally Spreading Tumor: where submucosal invasion lies beneath

Aims Current guidelines suggest that colorectal laterally spreading tumors (LSTs) with stigmata of superficial submucosal invasion (SMI) should be treated with en bloc endoscopic submucosal dissection (ESD). Since prediction of SMI risk is still challenging, in this study we evaluated which endoscopic features associated with presence of SMI in LSTs.

Methods In this monocentric retrospective study, we included all patients with resected LSTs between January 2018 and November 2023. Lesions were classified according to morphology into granular (G), granular mixed (GM) and non granular (NG) phenotype. Paris Classification, Kudo pit Pattern, location and lesion diameter were evaluated as predictors of SMI. Univariate and multivariate logistic regression analysis was performed to assess predictors of SMI.

Results Overall, 459 lesions were analyzed: n=172 NG-LST (37,5%), n=172 GM-LST (37,5%) and n=115 G-LST (25%). Superficial SMI was detected in 35 NG-LSTs (20,3%) , 21 GM-LSTs (12,2%) and 4 G-LSTs (3,5%). In the NG-LST group the prevalence of SMI was non-significantly higher in distal compared to proximal colon (24,5% vs 17% , OR=1.59 [0.70-3.62] p=0.27). According to Paris classification, the risk of SMI was significantly higher in depressed NG-LSTs (Paris IIc- IIa+IIc) compared to non-depressed lesions (31% vs. 12%, OR=3.62 [1.59-8.26]; p=0.002).

In GM-LSTs cohort, a not significantly higher SMI prevalence was reported comparing distal with proximal lesions (12,5% vs 5,9%, OR=2,26 [0.61-8.30] p=0.22), whereas rectal lesions showed significant higher risk than proximal lesions (22,9% vs 5,9%, OR=5.87 [1.94-17.73]; p=0.002). Kudo pit pattern Vi lesions in comparison to Kudo pit pattern IIIs/IIIL/IV lesions showed a significantly higher SMI prevalence both in NG-LSTs (32% vs. 2%, OR=36.57 [4.80-278.60]; p<0.001) and in GM-LSTs (73% vs 5%, OR=40.36 [11.96-136.18]; p<0.001). Besides, in GM-LST group a lesion diameter≥4 cm was associated with higher SMI prevalence only in rectal lesions (40% vs 5%, OR 11.33 [1.33-96.81]; p=0.03).

At the multivariate analysis, independent predictors of SMI presence were Kudo pit pattern Vi (OR 35.02 [4.31-284.85] p<0,001) in NG-LSTs, and Kudo pit pattern Vi (OR 74.15 [14.08-390.43], p<0.001) and rectal lesions (OR 13.45 [2.00- 90.25], p=0,007) in GM-LSTs.

Conclusions Our study confirmed that NG-LST lesions deliver a higher overall risk of SMI than GM-LSTs and G-LSTs, independently on the location. GM-LSTs carried an intermediate risk of SMI and, considering only rectal lesions, size was an important risk factor for SMI. Furthermore, the Kudo pit pattern was an independent risk factor for SMI in both groups. As a consequence, our study stresses that a proper optical diagnosis is paramount in order to stratify SMI risk and choose the appropriate treatment for these lesions.

Publikationsverlauf

Artikel online veröffentlicht:
15. April 2024