ClustALL: A robust clustering strategy for stratification of patients with acutely decompensated cirrhosis

Abstract

AbstractPatient heterogeneity represents a significant challenge for both individual patient management and clinical trial design, especially in the context of complex diseases. Most existing clinical classifications are based on scores built to predict patients’ outcomes. These classical methods may thus miss features that contribute to heterogeneity without necessarily translating into prognostic implications.To address patient heterogeneity at hospital admission, we developed ClustALL, a computational pipeline designed to handle common clinical data challenges such as mixed data types, missing values, and collinearity. ClustALL also facilitates the unsupervised identification of multiple and robust stratifications. We applied ClustALL to a prospective European multicentre cohort of patients with acutely decompensated cirrhosis (AD) (n=766), a highly heterogeneous disease. ClustALL identified five robust stratifications for patients with AD, using only data at hospital admission. All stratifications included markers of impaired liver function and number of organ dysfunction or failure, and most included precipitating events. When focusing on one of these stratifications, patients were categorized into three clusters characterized by typical clinical features but also having a prognostic value. Re-assessment of patient stratification during follow-up delineated patients’ outcomes, with further improvement of the prognostic value of the stratification. We validated these findings in an independent prospective multicentre cohort of patients from Latin America (n=580).In conclusion, this study developed ClustALL, a novel and robust stratification method capable of addressing challenges tied to intricate clinical data and applicable to complex diseases. By applying ClustALL to patients with AD, we identified three patient clusters, offering insights that could guide future clinical trial design.