[Value of histological activity in predicting endoscopic relapse among patients of ulcerative colitis in endoscopic remission].

ABSTRACT

Objective: To investigate the value of histological evaluation in predicting endoscopic relapse among patients with ulcerative colitis (UC) who were in endoscopic remission, and to compare the usefulness of various histological scoring systems. Methods: Histological sections from 61 patients with UC who were in endoscopic remission were retrospectively analyzed, at Peking University Third Hospital, Beijing, China from January 2015 to June 2021. They were subdivided into endoscopic persistent remission group (remission group, n=31, Mayo endoscopic score 0) and endoscopic relapse group (relapse group, n=30, Mayo endoscopic score≥1) according to the results of the first endoscopic reexamination after the biopsy. Histological evaluation was performed using the Geboes score (GS) and its simplified version (SGS), the Nancy index (NI) and the Robarts histopathological index (RHI). The median and maximum histological scores for each case in all biopsies were recorded. Univariate comparisons were performed using chi-squares and multivariate analysis using binary logistic regression. The values of four histological evaluation systems for predicting endoscopic relapse among UC patients in endoscopic remission were analyzed using receiver operating characteristic (ROC) curves. Results: Significant differences were observed between the remission and relapse groups. The differences were more pronounced in the maximum histological scores; the mean and highest results of area under the ROC curve scores (AUC) for GS, SGS, NI, and RHI were 0.657, 0.668, 0.682, 0.691, and 0.866, 0.863, 0.864, 0.869, respectively. The differences were statistically significant (P<0.05). The corresponding best cut-offs were GS≥2B.1, SGS≥2B.1, NI≥2, and RHI≥2.5, respectively, which meant mild active inflammation histologically, while there was no statistical difference of AUC among the four histological scoring indices (P>0.05). Univariate and multivariate analyses revealed statistically significant differences in the number of neutrophils in the epithelium and lamina propria (P<0.05). Conclusions: Biopsies from UC patients in endoscopic remission may still have histological active inflammation which appears to correlate with endoscopic relapse. Four commonly used histological scoring systems can be used to assess the risk of endoscopic relapse among UC patients in endoscopic remission. The patients who more likely have endoscopic relapse seem to have a histological score greater than the cut-off value (i.e., mild histological activity). The maximum histological scores can accurately predict the risk of endoscopic relapse, while the presence of epithelial and laminar propria neutrophil infiltrates can independently predict the endoscopic relapse in these patients. Considering the utility and convenience in routine practice, NI is recommended for evaluating histological inflammatory activity.