Published in:
08-03-2021 | Portal Hypertension | Original Research
The Cook Score: A Novel Assessment for the Prediction of Liver-Associated Clinical Events in a Diverse Population
Published in: Journal of Gastrointestinal Cancer | Issue 2/2022
Login to get accessAbstract
Background and Aims
Transient elastography (TE) provides accurate quantification of liver fibrosis. Its usefulness could be significantly amplified in terms of predicting liver-associated clinical events (LACE). Our aim was to create a model that accurately predicts LACE by combining the information provided by TE with other variables in patients with chronic liver disease (CLD).
Methods
We retrospectively reviewed the electronic medical records of patients who underwent liver elastography, at John H. Stroger Hospital in Cook County, Chicago, IL. The incidences of LACE were documented including decompensation of CLD, new hepatocellular carcinoma, and liver-associated mortality. Significant predicting factors were identified through a forward stepwise Cox regression model. We used the beta-coefficients of these risk factors to construct the Cook Score for prediction of LACE. Receiver-operating characteristic (ROC) curves were plotted for Cook Score to evaluate its efficiency in prediction, in comparison with MELD-Na Score and FIB-4 Score.
Results
A total of 3097 patients underwent liver elastography at our institution. Eighty-eight LACE were identified. Age (hazard ratio (HR) 1.04, p = 0.002), aspartate aminotransferase to alanine aminotransferase ratio (HR 2.61, p < 0.001), platelet count (HR 0.98, p < 0.001), international normalized ration (INR) (HR 17.80, p < 0.001), and liver stiffness measurement (HR1.04, p < 0.001) were identified as significant predictors. The Cook Score was constructed with two optimal cut-off points to stratify patients into low-, intermediate-, and high-risk groups for LACE. The Cook Score proved superior than MELD-Na Score and FIB4 Score in predicting LACE with an area under curve of 0.828.
Conclusion
This novel score based on a large robust sample would provide accurate prediction of prognosis in patients with chronic liver disease and guide individualized surveillance strategy once validated with future studies.