Premium
Model predicting the microvascular invasion and satellite lesions of hepatocellular carcinoma after hepatectomy
Author(s) -
Shen Junyi,
Wen Tianfu,
Chen Weixia,
Lu Changli,
Yan Lvnan,
Yang Jiayin
Publication year - 2018
Publication title -
anz journal of surgery
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.426
H-Index - 70
eISSN - 1445-2197
pISSN - 1445-1433
DOI - 10.1111/ans.14473
Subject(s) - medicine , nomogram , hepatocellular carcinoma , hazard ratio , receiver operating characteristic , proportional hazards model , concordance , hepatectomy , multivariate statistics , oncology , multivariate analysis , stage (stratigraphy) , carcinoma , confidence interval , surgery , urology , statistics , resection , paleontology , mathematics , biology
Background Microvascluar invasion and satellite lesion (MS), important unfavourable pathological factors, significantly contribute to tumour recurrence and impair the prognosis in hepatocellular carcinoma. We aimed to construct a model for the prediction of MS in order to plan treatment better. Methods A total of 1135 consecutive patients with hepatocellular carcinoma who received radical hepatectomy at West China Hospital were randomly assigned to a training set and a validation set. Multivariate analysis was preformed to identify independent risk factors of MS in the training set, and a nomogram was then constructed based on the risk factors. The concordance index (C‐index) and a calibration curve were used to assess the predictive performance of the model. Results The occurrence rate of MS was about 36.5%. Based on the multivariate analysis, the following six variables were incorporated into the nomogram: age (hazard ratio (HR): 0.531), alpha fetoprotein (HR: 1.327), neutrophil‐to‐lymphocyte ratio (>2.8, HR: 1.732), international normalized ratio (>1.07, HR: 1.702), tumour size (HR: 1.116) and tumour number (HR: 1.842). The model showed satisfactory discrimination abilities, with a C‐index of 0.721 for the training set and 0.704 for the validation set. The receiver operating characteristic curve confirmed the predictive power. Meanwhile, the calibration curve presented a goodness of fit between prediction of the model and actual observations. Conclusions The user‐friendly model may be useful for prediction of the occurrence of MS and to plan treatment more rationally preoperatively.