Related Experiment Video
Updated: Jun 23, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
Subgroup-Specific Nomogram for Refined Risk Stratification in Hepatocellular Carcinoma Patients with Low LDL-C
Bojun Liu1, Zhixia Gu2,3,4,5, Ling Li6
1Hepatic Disease and Oncology Minimally Invasive Interventional Center, Beijing Youan Hospital, Capital Medical University, Beijing, 100069, People's Republic of China.
Background:
Hepatocellular carcinoma (HCC) exhibits high mortality and clinical heterogeneity. While low low-density lipoprotein cholesterol (LDL-C) is a recognized risk factor for HCC, its continuous dose-response relationship with overall survival (OS) remains unclear. Furthermore, existing prognostic models lack precision for this specific high-risk population, highlighting an urgent need for individualized risk stratification.
Methods:
A retrospective cohort of 486 HCC patients treated at Beijing You'an Hospital (January 1, 2015-January 1, 2021) was analyzed. Restricted cubic splines (RCS) and Cox regression evaluated the LDL-C and OS dose-response relationship. Following quartile stratification (Q1-Q4), the cohort was dichotomized into low (Q1+Q2, n=245) and high (Q3+Q4, n=241) LDL-C groups. Within the high-risk low LDL-C subgroup, least absolute shrinkage and selection operator (LASSO) regression identified core features to construct a specific nomogram, validated via concordance index (C-index), receiver operating characteristic (ROC) curves, calibration, decision curve analysis (DCA), and Kaplan-Meier (KM) curves for risk stratification.
Results:
RCS analysis revealed a non-linear relationship: mortality risk increased as baseline LDL-C decreased below 2.2 mmol/L. Multivariable Cox regression confirmed LDL-C as an independent protective factor for OS (HR=0.714, 95% CI: 0.539-0.947, P=0.019). KM analysis demonstrated significantly inferior OS in the low LDL-C cohort compared to the high LDL-C group (P=0.013). The subgroup-specific nomogram, integrating drinking history, Barcelona Clinic Liver Cancer stage, tumor number, and thrombin time, exhibited good discrimination (C-index=0.686; 3- and 5-year AUCs=0.752 and 0.696, respectively), accurate calibration, and positive clinical benefit. Furthermore, the model achieved secondary risk stratification within this subgroup (P<0.0001).
Conclusion:
Lower baseline LDL-C is an independent predictor of worse OS. The developed subgroup-specific nomogram provides an objective tool for secondary risk stratification, aiding personalized therapeutic management for this high-risk population.
