Development of a prognostic MRCP-score (DiStrict) for individuals with large-duct primary sclerosing cholangitis

• Published in: JHEP reports 2022-12, Vol.4 (12), p.100595-100595, Article 100595
• Main Authors: Grigoriadis, Aristeidis, Imeen Ringe, Kristina, Bengtsson, Johan, Baubeta, Erik, Forsman, Cecilia, Korsavidou-Hult, Nafsika, Rorsman, Fredrik, Nilsson, Emma, Kartalis, Nikolaos, Bergquist, Annika
• Format: Article
• Language: English
• Subjects: Bile ducts; Cholangiopancreatography; Cholangitis sclerosing; Magnetic resonance; Medicin och hälsovetenskap; Prognosis
• ISSN: 2589-5559; 2589-5559
• DOI: 10.1016/j.jhepr.2022.100595

Magnetic resonance cholangiopancreatography (MRCP) is used for the diagnosis and follow-up of individuals with primary sclerosing cholangitis (PSC). The aim of our study was to develop an MRCP-score based on cholangiographic findings previously associated with outcomes and assess its reproducibility and prognostic value in PSC. The score (DiStrict score) was developed based on the extent and severity of cholangiographic changes to intrahepatic and extrahepatic bile ducts (range 0–8) on 3D-MRCP. In this retrospective, multicentre study, three pairs of radiologists with different levels of expertise from three tertiary centres applied the score independently. MRCP examinations of 220 consecutive individuals with PSC from a prospectively collected PSC-cohort, with median follow-up of 7.4 years, were reviewed. Inter-reader and intrareader agreements were assessed via intraclass correlation coefficient (ICC). After consensus, the prognostic value of the score was assessed using Cox-regression and outcome-free survival rates were assessed via Kaplan-Meier estimates. Harrell's C-statistic was calculated. Forty patients developed outcomes (liver transplantation or liver-related death). Inter-reader agreement between experienced radiologists was good (ICC 0.82; 95% CI 0.74–0.87, and ICC 0.81; 95% CI 0.70–0.87, respectively) and better than the agreement for the pair of experienced/less-experienced radiologists (ICC 0.48; 95% CI 0.05–0.72). Agreement between radiologists from the three centres was good (ICC 0.76; 95% CI 0.57–0.89). Intrareader agreement was good to excellent (ICC 0.85–0.93). Harrell's C was 0.78. Patients with a DiStrict score of 5–8 had 8.2-fold higher risk (hazard ratio 8.2; 95% CI 2.97–22.65) of developing outcomes, and significantly worse survival (p