Survival analysis of patients with hepatocellular carcinoma based on the ratio of platelet count to spleen diameter

Background: In China, 80% of Hepatocellular Carcinoma (HCC) is associated with cirrhosis. Portal hypertension, the most common outcome of cirrhosis progression, has a high incidence. Platelet count/spleen diameter ratio (PSL) with a cut-off value of 909 can predict the presence of esophagogastric varices and thus portal hypertension, which is also an independent risk factor for early recurrence and late recurrence of hepatocellular carcinoma after resection. Therefore, the effect of PSL on the overall survival (OS) of patients with HCC is necessary. The aim of this study was to apply a new method to establish and validate a model for predicting the prognosis of patients based on PSL with HCC.

Methods: A total of 1,104 patients with clinical diagnosed with HCC following non-surgical therapy randomly divided the patients into a primary cohort and a validation cohort in a ratio of 7:3, in which 772 HCC patients were in the primary cohort and a total of 332 HCC patients were in the validation cohort. Through Lasso-Cox analysis, the independent predictors of OS of training cohort were included in nomogram1, and the independent predictors of Cox regression analysis were included in nomogram2. Nomogram1 and nomogram2 used consistency index (C-index), AUC and time-dependent ROC curves in the training cohort, respectively, and the calibration curves were plotted. All suggest that nomogram1 is better than nomogram2. We get similar results in the validation cohort.

Results: The C-index of nomogram1was 0.792 (95%CI: 0.772-0.812), which was superior to nomogram2 (0.788) and traditional modes (0.631-0.712). The AUC of nomogram1 was 0.866 (95%CI: 0.840-0.889). In the validation cohort, the nomogram1 still gave good discrimination (C-index: 0.769, 95%CI: 0.740-0.798; AUC: 0.867, 95%CI: 0.826-0.902). Calibration plots for 3-year OS probabilities showed the good agreement between nomogram1 predictions and actual observations. In addition, we found that the decision curve analysis of nomogram1 and nomogram2 was also meaningful.

Conclusion: Novel nomogram containing PSL, based on LASSO Cox regression, had higher predictive efficacy for 3-year overall survival in patients with HCC.