Effects of subspecialty signout and group consensus on the diagnosis of microscopic colitis

Abstract

Microscopic colitis (MC) includes lymphocytic colitis (LC) and collagenous colitis (CC). Microscopic changes are required to establish these diagnoses. While criteria exist, interobserver variability has been reported previously. This has not been evaluated in the context of subspecialty signout (SSSO) or a consensus conference. We identified 133 colon biopsies diagnosed as LC, CC, MC, or normal but with mild changes insufficient for MC. All predated the introduction of SSSO at our institution. They were independently reviewed by three gastrointestinal (GI) pathologists. Cases lacking independent consensus were reviewed by the same pathologists in consensus conference to establish a final diagnosis. Individual diagnoses were compared with the consensus diagnoses, and consensus diagnoses were compared with original diagnoses made by GI and non-GI pathologists. Consensus diagnoses were normal (n = 34), LC (n = 57), and CC (n = 42). "Normal" was the diagnosis most commonly agreed upon independently (27/34 cases, P = 0.0073 versus LC, P = 0.0172 versus CC). The reviewing pathologists independently agreed with 80%, 80%, and 94% of consensus diagnoses (κ = 0.70, 0.69, and 0.91). The group consensus agreed with the diagnoses in 49 of 58 (84%) cases originally signed out by non-GI pathologists (κ = 0.77) and in 44 of 57 (77%) cases originally signed out by GI pathologists (κ = 0.63). Good interobserver agreement exists for MC, though whether GI subspecialty training improves agreement remains unclear. Group consensus may aid in diagnosis of difficult/borderline MC cases.

Keywords: Collagenous colitis; Interobserver variability; Lymphocytic colitis; Microscopic colitis; Subspecialty signout.