Enhancing the prognostic accuracy of the RETREAT score with AFP-L3 and DCP tumor markers.

Abstract

The RETREAT (Risk Estimation of Tumor Recurrence After Transplant) Score is a validated tool to predict post-transplant HCC recurrence risk. Alpha-fetoprotein (AFP) bound to Lens culinaris agglutinin (AFP-L3) and des-gamma-carboxyprothrombin (DCP) measured at transplant predict worse post-LT survival and may improve the RETREAT score. Our cohort comprised 284 patients transplanted for HCC who were within or downstaged to Milan, with 23 (8.1%) experiencing HCC recurrence. The modified RETREAT (mRETREAT) score assigns AFP-L3 ≥15% 2 points and DCP ≥7.5 ng/mL 3 points. Patients with a modified RETREAT score ≥4 showed a 3-year recurrence-free survival of 73.2% versus 97.8% recurrence-free survival if <4. In comparison, the original RETREAT score had a 3-year recurrence-free survival of 80.0% if ≥2 versus 98.0% if <2. mRETREAT demonstrated a superior AUC of 0.86, compared to the original RETREAT's 0.82, and enhanced calibration and accuracy with a lower Brier score (0.04). The integration of AFP-L3 and DCP into the RETREAT score appears to enhance the prediction of post-LT HCC recurrence. Given these promising results, further study in a larger multicenter cohort is warranted for empiric derivation and validation of a modified RETREAT score, including AFP-L3 and DCP.