Models to predict disease-free survival for hepatocellular carcinoma patients with surgical resections

J Surg Oncol. 2020 Dec;122(7):1444-1452. doi: 10.1002/jso.26169. Epub 2020 Sep 1.

Abstract

Background and objectives: Precise prognostic prediction for an individual hepatocellular carcinoma (HCC) patient before and after liver resection is important. We aimed to establish simple prognostic models to predict disease-free survival (DFS) for these patients.

Methods: Six hundred and ninety-eight HCC patients with liver resections were reviewed. Preoperative (model 1) and postoperative (model 2) nomogram-based scoring systems were constructed by multivariate analyses, and DFS was estimated.

Results: Among 698 patients, 490 (70.2%) patients had tumor recurrence at a median follow-up of 84.4 months. Risk factors of tumor recurrence in model 1 included viral hepatitis, platelet count, albumin, indocyanine green retention rate, multiplicity of tumor, and radiologic total tumor volume (TTV). Prognostic variables identified in model 2 were viral hepatitis, platelet count, multiplicity of tumor, cirrhosis, microvascular invasion, and pathologic TTV. By nomogram in model 1, the patients were classified into three groups with 5-year DFS of 61.0%, 35.7%, and 21.1%, respectively (P < .0001). In model 2, the patients were divided into five groups with 5-year DFS of 58.0%, 43.7%, 24.0%, 15.4%, and 0.0%, respectively (P < .0001).

Conclusion: Based on nomogram models, DFS for the patients who had liver resection for HCC can be predicted before liver resection and re-assessed after liver resection.

Keywords: hepatocellular carcinoma; nomogram; prediction; tumor recurrence.

MeSH terms

  • Adult
  • Aged
  • Carcinoma, Hepatocellular / mortality
  • Carcinoma, Hepatocellular / pathology
  • Carcinoma, Hepatocellular / surgery*
  • Disease-Free Survival
  • Female
  • Hepatectomy*
  • Humans
  • Liver Neoplasms / mortality
  • Liver Neoplasms / pathology
  • Liver Neoplasms / surgery*
  • Male
  • Middle Aged
  • Neoplasm Recurrence, Local
  • Nomograms
  • Prognosis
  • Tumor Burden