Prognostic model for identifying candidates for hepatectomy among patients with hepatocellular carcinoma and hepatic vein invasion.
2020
BACKGROUND: Hepatic vein tumour thrombus (HVTT) is a major determinant of survival outcomes for patients with hepatocellular carcinoma (HCC). An Eastern Hepatobiliary Surgery Hospital (EHBH)-HVTT model was established to predict the prognosis of patients with HCC and HVTT after liver resection, in order to identify optimal candidates for liver resection. METHODS: Patients with HCC and HVTT from 15 hospitals in China were included. The EHBH-HVTT model with contour plot was developed using a non-linear model in the training cohort, and subsequently validated in internal and external cohorts. RESULTS: Of 850 patients who met the inclusion criteria, there were 292 patients who had liver resection and 198 who did not in the training cohort, and 124 and 236 in the internal and external validation cohorts respectively. Contour plots for the EHBH-HVTT model were established to predict overall survival (OS) rates of patients visually, based on tumour diameter, number of tumours and portal vein tumour thrombus. This differentiated patients into low- and high-risk groups with distinct long-term prognoses in the liver resection cohort (median OS 34.7 versus 12.0 months; P < 0.001), internal validation cohort (32.8 versus 10.4 months; P = 0.002) and external validation cohort (15.2 versus 6.5 months; P = 0.006). On subgroup analysis, the model showed the same efficacy in differentiating patients with HVTT in peripheral and major hepatic veins, the inferior vena cava, or in patients with coexisting portal vein tumour thrombus. CONCLUSION: The EHBH-HVTT model was accurate in predicting prognosis in patients with HCC and HVTT after liver resection. It identified optimal candidates for liver resection among patients with HCC and HVTT, including tumour thrombus in the inferior vena cava, or coexisting portal vein tumour thrombus.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
32
References
5
Citations
NaN
KQI