Noninvasive prediction model for predicting spontaneous tumor necrosis in hepatocellular carcinoma and prognostic study.

Background and objectives: In hepatocellular carcinoma (HCC), patients with spontaneous tumor necrosis have a high recurrence rate and poor prognosis. However, conventional preoperative imaging could not detect the presence of tumor necrosis. Accordingly, we developed and assessed a nomogram to forecast tumor necrosis.

Methods: Clinical data were collected retrospectively from 495 patients with HCC who received a hepatectomy at Zhongnan Hospital of Wuhan University from 1 January 2015 to 31 May 2024. The patients ( n  = 495) were randomly divided in a 7 : 3 ratio into the training cohort (TC, n  = 348) and the validation cohort (VC, n  = 147). The logistic regression analyses were used to identify factors independently predicting tumor necrosis in the patients with TC. The Kaplan-Meier survival analysis was used for comparing and estimating survival rates.

Results: Preoperative clinical tumor-node-metastasis stage, hemoglobin, systemic immune inflammation, alkaline phosphatase, and alpha-fetoprotein levels were identified as hazard factors for predicting tumor necrosis. The area under the receiver operating characteristic curve of the TC, VC, and the full cohort was 0.810, 0.758, and 0.795, respectively. The calibration curves demonstrated a high degree of concordance. The decision curve analysis showed the clinical significance of the nomogram. Both overall survival and recurrence-free survival of patients in the tumor necrosis group were poorer.

Conclusion: Our predictive model could effectively predict the risk of spontaneous tumor necrosis in patients with HCC, and tumor necrosis was related to a worse prognosis.