Pretransplant Alpha-Fetoprotein Slope and Milan Criteria Are Strong Predictors of Hepatocellular Carcinoma Recurrence after Transplantation

Abstract

Listen

Translate article

Introduction: Hepatocellular Carcinoma (HCC) is a major cause of orthotropic liver transplants (OTL) worldwide. However, tumor recurrence remains a major concern. Our group has shown that a rising natural alpha-fetoprotein slope (NAS) correlates with tumor characteristics. We want to assess if a rising PAS predicts tumor recurrence in a logistic model. Methods: Patients who underwent first OTL for HCC and survived 90 days post-OTL (n=144) at our institution from 1992 to 2010 were reviewed. Patients with less than two alpha-fetoprotein (AFP) values prior to any treatment, whether therapy or OTL, were excluded (n=52). Of the remaining 92 patients, 12 recurred. A positive NAS was defined as >0.1 (n=28); the remainder (n=64) presented a stable or negative NAS. Demographics, pre-transplant trans-arterial chemoembolization (TACE), adherence to Milan's criteria, number of lesions (< 5, ≥5), total tumor size (< 5 cm, ≥5 cm), tumor grade (< 2/4 or more), and presence of micro-vascular invasion were collected. Statistical analysis was done using ANOVA for demographics, logistic regression for recurrence and Chi-square or Fisher's exact tests for univariate analysis. Results: Demographics were similar among the recurrence and non-recurrence groups. Patients who recurred post-OTL received more TACE therapy (50.0%vs17.5%, OR 4.71,95%CI 1.06-20.22,p-value=0.020), had a higher number of lesions (41.7%vs13.8%,OR 4.48,95%CI 1.23-19.65,p-value=0.032), a greater total tumor size (75.0%vs25.0%,OR 9.00,95%CI 1.94-55.12,p-value=0.001) and a greater incidence of microvascular invasion (58.3%vs23.8%,OR 4.49,95%CI 1.07-19.8,p-value=0.013). More patients exceeded Milan's criteria (75.0%vs31.3%,OR 6.60,95%CI 1.45-4.05,p-value=0.008) and had a rising NAS (58.3%vs26.3%,OR 3.20,95%CI 1.11-9.22,p-value=0.024) amongst the recurrence group. However, tumor grade was comparable between recurring and non-recurring patients (33.3%vs17.5%,p-value=0.241). Furthermore, NAS is a strong predictor of microvascular invasion found in pathology (46.15%vs24.24%,OR 1.96,95%CI 1.04-3.68,p-value=0.040). Given that Milan's criteria is based on tumor size and the number of lesions, we used it as a representative of these values in the logistic model. After correcting for age and sex, both a rising NAS (OR 3.84,95%CI 0.99-14.82,p-value=0.051) and non-adherence to Milan's criteria (OR 7.64,95%CI 1.68-35.70,p-value=0.008) were strong predictors of recurrence post-OTL. Conclusion: The natural AFP slope is a predictor of microvascular invasion, a finding exclusive to pathology and in itself a predictor of HCC recurrence. The NAS and Milan's criteria were able to predict the risk of recurrence after OTL, both separately and in a logistic model. Although sample size is small, these results encourage a frequent monitoring of AFP variations prior to liver transplantation.