A practical nomogram from the SEER database to predict the prognosis of hepatocellular carcinoma in patients with
Kai Zhang1, Changcheng Tao1, Fan Wu1
1Department of Hepatobiliary Surgery, National Cancer Center/National Clinical Research Center for Cancer/Cancer Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China.
Background:
The presence of lymph node (LN) metastases is associated with poor survival outcomes in hepatocellular carcinoma (HCC) patients. Because of the low probability of LN metastasis, research into the prognoses of these patients is difficult. The present study developed a nomogram model to predict the prognosis of HCC patients with LN metastasis.
Methods:
This retrospective, noninterventional study enrolled patients from the Surveillance, Epidemiology, and End Results (SEER) database from 2010 to 2015. The following inclusion criteria were used: (I) site recode ICD-O-3 (International Classification of Diseases for Oncology, Third Edition) of 8170-8175 and malignant histological behavior; (II) seventh edition American Joint Committee on Cancer (AJCC) stage N1; (III) older than 18 years; and (IV) available information. Potential prognostic factors were collected from the SEER database; the primary outcomes of interest were overall survival (OS) and disease status. Cox and Lasso regression were used to investigate independent prognostic factors for survival. A prognostic nomogram using these independent risk factors was constructed. The concordance index (C-index) and calibration curves were used to evaluate the model's predictive performance. The clinical benefit was assessed via decision curve analysis (DCA).
Results:
Patients were randomized into a training group (944 patients) and a validation group (402 patients) in a 70:30 ratio. Grade, T stage, liver surgery, chemotherapy, radiation recode, alpha-fetoprotein level, fibrosis score, tumor size group, and M stage were selected as independent prognostic factors, and a nomogram was developed using these variables. The C-indices of the training and validation groups were 0.70 and 0.73, respectively. Calibration curves for the probability of survival showed good agreement. DCA indicated that the nomogram had positive net benefits.
Conclusions:
The constructed nomogram may assist clinicians in predicting the prognosis of HCC patients with LN metastasis and may provide a rationale for treatment options.
