Abstract
BACKGROUND: The prognosis is extremely troubling in advanced hepatocellular carcinoma (HCC). Prognostic scores have been developed. Yet, the positive predictive values might appear inadequate. This retrospective study aimed to develop a quick and efficient risk score to assess prognosis and clinical response. METHODS: A total of 391 hCC patients were enrolled and were divided into training and validation groups between 2015 and 2024. Patients were separated into high-risk and low-risk groups using X-tile software. Using the COX proportional risk model analysis method, we then created a risk score and examined them using Kaplan-Meier, time-dependent receiver operating characteristics (ROC) curve, and nomogram analysis. RESULTS: In predicting overall survival (OS), free fatty acid/high-density lipoprotein cholesterol (FFHL), tumor size, and BCLC stage were independent prognostic variables. A new risk score was developed just above and used as a prognostic factor (p < 0.001 in the training and validation groups) and had a high time-dependent ROC for progress-free survival (PFS) (area under the curve [AUC] 0.688-0.789 in the training group; AUC 0.592-0.741 in the validation group) and OS (AUC 0.812-0.918 in the training group; AUC 0.692-0.981 in the validation group). In comparison to the best overall response (BOR), the score offered a more accurate evaluation of durable clinical benefit (DCB) (p < 0.001 in the training and validation group; p = 0.061 vs 0.001 in the training and validation group). CONCLUSION: A new score based on lipid markers is a useful tool for evaluating prognosis and distinguishing patients with DCB.