Sensitivity of pre-operative imaging and radiologist inter-rater reliability in detecting lesions in Crohn's disease.

Abstract

Computed tomography (CT), computed tomography enterography (CTE) and magnetic resonance enterography (MRE) are commonly used pre-operatively in surgical planning in Crohn's Disease (CD). The findings on CT, CTE and MRE may not, however, correlate with operative findings. This study aims to establish the sensitivity of these imaging modalities and analyse radiologist inter-rater reliability by comparing imaging findings of strictures, fistulas and abscesses with intra-operative findings. A retrospective review of CD patients who had either CT, CTE and/or MRE imaging and CD related surgical intervention at a public health service from 2010 to 2019 inclusive. The number and locations of strictures, fistulas and abscesses on pre-operative original radiology reports (OR) were recorded. Subsequently, all scans were re-read by two specialist abdominal radiologists and consensus recorded (SR). Lesions recorded from both OR and SR were compared to those found intra-operatively. Eighty-three patients were included. For strictures, sensitivity was 67%, 74% and 79% for OR and 88%, 71% and 87% for SR for CT, CTE and MRE respectively. The frequency of fistulas and abscesses were small hence a conclusion could not be drawn. The level of agreement between radiologists ranged from 44% to 82% for strictures and 64 to 100% for fistulas and abscesses across all three imaging modalities. CT and MRE have similarly high sensitivities for the identification of strictures pre-operatively when read by specialist radiologists. Inter-rater reliability calculations found similar agreement levels between specialist radiologists and between OR and SR for strictures, fistulas and abscesses across CT, CTE and MRE.