Correlation between the endoscopic and histologic score in assessing the activity of ulcerative colitis

Abstract

Background: In routine practice, scoring of activity and severity of ulcerative colitis is based on combined clinical and endoscopic assessment. Histology also allows sensitive scoring of ulcerative colitis activity. The correlation between endoscopy and histology has not been investigated thoroughly. It is still unknown how well they correlate and whether scoring both endoscopy and histology better reflects the true disease activity than each of the methods separately.

Methods: Two hundred and sixty-three biopsy sets from 131 known patients with ulcerative colitis were reviewed by an experienced gastrointestinal pathologist and scored using the Geboes and Riley histologic scoring systems. Endoscopic scoring had been performed previously by inflammatory bowel disease specialists using the Mayo endoscopic subscore. Bidirectional comparison of the Mayo endoscopic subscore with the full and converted histologic scores was then performed.

Results: We found a statistically significant overall correlation between the Mayo endoscopic subscore and the histologic scores (highest correlation: Kendall's τ = 0.482, P < 0.0001). Although a very high concordance was found for inactive and severely active disease, a high diversity was detected between these extremes. For example, endoscopic mildly active disease (Mayo 1) was distributed over all different histologic grades (37%, grade 0; 21%, grade 1; 28%, grade 2; and 14%, grade 3).

Conclusions: Both extremes of the histologic and endoscopic activity scores neatly correlate, but important misclassifications exist for mild disease. Microscopy may detect more severe disease than endoscopically suspected, possibly altering the clinical follow-up scheme. We also infer from our results that histologic scoring should be used in addition to endoscopy when scoring disease activity for clinical trials.