Validation of a novel BCM model for recurrence risk prediction after mucosectomy of colorectal lateral spreading tumors in a European cohort

Background and aim

Piecemeal endoscopic mucosal resection (pEMR) is the best approach to resect large lateral spreading tumors (LST, > 20 mm width). However, it is associated with early recurrence (ER) and late recurrence (LR). This study aims to assess the risk factors associated with ER and LR and to validate different predictive scores (SMSA, SERT, and BCM) in identifying the risk of ER and LR after LST resected by pEMR in a European cohort.

Methods

Retrospective observational cohort study, based on a prospectively collected database, of large LST submitted to pEMR.

Results

A total of 108 patients were included in the study and the incidence rates of ER and LR were 22 % and 8 %, respectively. The lesion's size, SERT, and BCM scores were independent predictor factors of ER (p-value < 0.05), while the lesion's site and BCM score were independent predictor factors of LR (p-value < 0.05). For the prediction of ER, the SERT score (cut-off > 1) presented the highest AUROC (0.758 vs 0.697 from BCM and 0.647 from SMSA). Regarding LR, the BCM model (cut-off > 2) presented the highest AUROC (0.817 vs 0.708 from SERT and 0.691 from SMSA).

Conclusions

We present the first external validation of the three scores mentioned in an European cohort. SERT and BCM scores had an acceptable performance in predicting ER and LR. However, the BCM model was the only score that proved to be an independent predictor of both ER and LR, proving to be valuable for both applications.