Development and Validation of a Clinical Risk Score to Predict Immune-mediated Liver Injury Caused by Sintilimab: Assessed for Causality Using Updated RUCAM.

Abstract

Immune-mediated liver injury is a fatal side effect of sintilimab. This study aimed to shed light on the associated risk factors and characteristics of this adverse event. The clinical records of 772 patients treated with sintilimab were retrospectively reviewed to investigate risk factors associated with sintilimab immune-related hepatotoxicity, as well as its incidence and outcome. The Roussel Uclaf Causality Assessment Method was used to identify cases of sintilimab-induced hepatotoxicity. Furthermore, logistic regressions were performed to compare the clinical and bloodwork characteristics of patients with and without immune-mediated liver injury caused by checkpoint inhibitors. Of the 585 patients included in the study, 71 (12.1%) developed liver injury during sintilimab use. The median RUCAM score with interquartile range was 7 (6, 8). Hypoproteinemia, dyslipidemia, and the presence of thyroid peroxidase antibodies were risk factors for sintilimab-related hepatotoxicity. A nomogram model was constructed for sintilimab-induced immune-mediated liver injury based on these risk factors, which had a C-index value of 0.713 and a good calibration curve. When applied to patients with grade ≥3 and ≥4 sintilimab-induced immune-mediated liver injury, it achieved C-index values of 0.752 and 0.811, respectively. The nomogram model also showed a good prediction potential in patients ≥65 years and males. Six of the patients with sintilimab-related hepatotoxicity showed improved liver function upon treatment with steroids. This study demonstrated that hypoproteinemia, dyslipidemia, and the presence of thyroid peroxidase antibodies were clinically feasible prognostic biomarkers to predict liver injury in patients treated with sintilimab.