Validation of Quantitative Magnetic Resonance Cholangiopancreatography Metrics in Prediction of Transplant-free Survival in Primary Sclerosing Cholangitis

Background and aims Qualitative magnetic resonance cholangiopancreatography (MRCP) scoring models in primary sclerosing cholangitis (PSC) are hampered by interobserver variability and evidence for quantitative MRCP has so far been limited by cohort sizes, follow-up time and lack of validation. This study aimed to validate the prognostic value of quantitative MRCP metrics in PSC in a large multicentre cohort. Method Retrospective, cross-sectional, clinical and quantified MRCP data by MRCP+ were collected from a non-transplant and transplant centre and randomised (1:1 ratio) into a derivation and validation set. Transplant-free survival, a composite of liver transplantation and PSC-related mortality (excluding colorectal carcinoma), was the primary endpoint. Least absolute shrinkage and selection operator analysis with manual guidance was used to compose a risk classifier. Prognostic performance and risk stratification were expressed by C-statistic and hazard ratios (HRs), and were validated in the validation set. Results A total of 224 patients were included with a median 6.8 years (Q1,Q3: 4.5, 9.8) of follow-up from MRCP onwards. Analysis identified number of strictures, proportion of 3–5 mm diameter ducts, years from diagnosis to MRCP and centre type as prognostic. The derived risk classifier showed a C-statistic of 0.72 (95% confidence interval [CI]: 0.60–0.81) and stratified effectively, with high-risk patients having threefold higher HR than low-risk patients (HR, 3.2; 95% CI: 1.6–6.4; P = 0.001) in the validation set. Conclusion This study confirms the prognostic value of quantitative MRCP (number of strictures and proportion of 3–5 mm diameter ducts) on long-term transplant-free survival in PSC and warrants further study on incorporating quantitative MRCP metrics into existing prognostic risk models.