Primary Sclerosing Cholangitis Risk Estimate Tool (PREsTo) Predicts Outcomes of the Disease: A Derivation and Validation Study Using Machine Learning

• DOI: 10.1002/hep.30085
• Publication Date: 2018-05-09T23:50:22Z
• Authors: John E. Eaton; Mette Vesterhus; Bryan M. McCauley; Elizabeth J. Atki...; Erik M. Schlicht; Brian D. Juran; Andrea A. Gossard; Nicholas F. LaRusso; Gregory J. Gores; Tom H. Karlsen; Konstantinos N. L...
• Keywords: Primary Sclerosing Cholangitis; Decompensation; Clinical endpoint; Surrogate endpoint; Liver disease

Improved methods are needed to risk stratify and predict outcomes in patients with primary sclerosing cholangitis (PSC). Therefore, we sought derive validate a prediction model compare its performance existing surrogate markers. The was derived using 509 subjects from multicenter North American cohort validated an international (n = 278). Gradient boosting, machine‐based learning technique, used create the model. endpoint hepatic decompensation (ascites, variceal hemorrhage, or encephalopathy). Subjects advanced PSC cholangiocarcinoma (CCA) at baseline were excluded. estimate tool (PREsTo) consists of nine variables: bilirubin, albumin, serum alkaline phosphatase (SAP) times upper limit normal (ULN), platelets, aspartate aminotransferase (AST), hemoglobin, sodium, patient age, number years since diagnosed. Validation independent confirms that PREsTo accurately predicts (C‐statistic, 0.90; 95% confidence interval [CI], 0.84‐0.95) performed well compared Model for End‐Stage Liver Disease (MELD) score 0.72; CI, 0.57‐0.84), Mayo 0.85; 0.77‐0.92), SAP <1.5 × ULN 0.65; 0.55‐0.73). continued be accurate among individuals bilirubin <2.0 mg/dL 0.82‐0.96) when reapplied later course disease 0.82; 0.64‐0.95). Conclusion: (HD) exceeds other widely available, noninvasive prognostic scoring systems.