Development and Validation of a Pre-Transplant Risk Score (LT-MVI Score) to Predict Microvascular Invasion in...

Background/Objectives: MVI is a relevant prognostic factor among patients with hepatocellular carcinoma (HCC) receiving liver transplantation (LT). The preoperative assessment of the risk for MVI is relevant to pre-LT patient management and selection. The objective of this study was to create and validate a model to predict microvascular invasion (MVI) based on preoperative variables in the LT setting. Methods: A total of 2170 patients from 11 collaborative centers in Europe, Asia, and the US, who received transplants between 1 January 2000 and 31 December 2017, were enrolled in the study. The entire cohort was split into a training and a validation set (70/30% of the initial cohort, respectively) using random selection. Results: MVI was reported in 586 (27.0%) explanted specimens. Using the training set data, multivariable logistic regression identified three preoperative parameters associated with MVI: α-fetoprotein (lnAFP; odds ratio [OR] = 1.19; 95% confidence interval [CI] = 1.13-1.27), imaging tumor burden score (lnTBS; OR = 1.66; 95%CI = 1.39-1.99), and a fast-track approach before LT due to the availability of a live donation (OR = 1.99; 95%CI = 1.56-2.53). In the validation set, the LT-MVI c-index was 0.74, versus 0.69 for the MVI score proposed by Endo et al. (Brier Skill Score +75%). The new score had a relevant net reclassification index (overall value = 0.61). Stratifying the validation set into three risk categories (0-50th, 51st-75th, and >75th score percentiles), a very good stratification was observed in terms of disease-free (5-year: 89.3, 75.5, and 50.7%, respectively) and overall survival (5-year: 79.5, 72.6, and 53.7%, respectively). Conclusions: The preoperative assessment of MVI using the proposed score demonstrated very good accuracy in predicting MVI after LT.