DEVELOPMENT OF NOMOGRAM PREDICTION MODEL FOR POSTOPERATIVE RECURRENCE OF HCC PATIENTS UNDERGOING LIVER TRANSPLANTATION IN CHINESE POPULATION WITH INTERNAL VALIDATION

Ensi Ma,Jianhua Li,Zhengxin Wang
DOI: https://doi.org/10.1097/01.tp.0000701236.11366.90
2020-01-01
Transplantation
Abstract:Background: Recurrence prediction models for hepatocellular carcinoma liver transplantation in western countries already exist, but the differences in tumor biological characteristics among different populations. So we aim to create a Nomogram prediction model for predicting hepatocellular carcinoma liver transplantation patients in China to predict individual survival after surgery and guide the postoperative treatment. Materials and Methods: Between February 2016 and October 2019, 210 HCC patients who received liver transplantation in the Huashan Hospital Affiliated to Fudan University were collected for this study. Prognostic nomograms predicting post-transplant recurrence were developed from a multivariate cox regression. The prediction power of the nomograms was tested by C-statistic and calibration plots. Results: There were 46(21.9%) patients within the Milan criteria. Before transplantation,24(11.4%) patients received transcatheter arterial chemoembolization (TACE) and 9(4.3%) received tumor targeted therapy. Multivariate predictors of recurrence included Age, Max diameter, thrombus, MVI, ALT on the 7th day after surgery(7d-ALT),AFP on the 7th day after surgery(7d-AFP). Nomogram for recurrence free survival (RFS) have been created. Discrimination and calibration of the nomograms revealed good predictive abilities(C-index:0.796).Compared with Milan criteria(C-index:0.617), UCSF criteria(C-index:0.624), the RFS nomogram improved accuracy in predicting recurrence after transplantation for HCC.Finally, two subgroups were generated based on the total risk points. The 1-, 2-, and 3-year recurrence rate were 16.6%, 23.3%, and 26.9% for the low-risk group, and 52.4%, 80.1%, and 80.1% for the high-risk group, respectively.Conclusion: The nomogram made by using different examination indexes had good ability in predicting prognostic recurrence and guiding the follow-up of patients. It can achieve patient stratification and guide patients for treatment and follow-up. Patients with score above 312.1 points had high risk of early recurrence after LT. For these patients, more strict postoperative targeted treatment and follow-up are needed. In order to be convenient and practical, we have also made web prediction models for convenience and practicality. For example, a 45-year-old male (0 point), tumor maximum diameter of 7 cm (30.3 points), have tumor thrombus (78.8 point), MVI of 1 (36.9 points), ALT of 156u / L (25.1 points), AFP of 83 μ g / L (6.2 points) on the 7th day after operation. For this example, the total score is 177.3 points, so the predicted recurrence rate of 1 and 2 years is 18% and 31%. This calculated value can be used to develop treatment plans and communicate with patients.https://maensi.shinyapps.io/Huashan_surgery/
What problem does this paper attempt to address?