Validation and optimization of AFP-based biomarker panels for early HCC detection in Latin America and Europe
Boris J. B. Beudeker, Siyu Fu, Domingo Balderramo, Ângelo Zambam de Mattos, Enrique Carrera, Javier Díaz, Jhon Prieto, Jesús M. Bañales, Arndt Vogel, Marco Arrese, Jeffrey Oliveira, Zwier M. A. Groothuismink, Gertine van Oord, Bettina E. Hansen, Robert A. de Man, José D. Debes, André Boonstra
Abstract
BACKGROUND: HCC is a major cause of cancer death worldwide. Serum biomarkers such as alpha-fetoprotein (AFP), protein induced by vitamin K absence-II, and the Gender, Age, AFP-L3, AFP, Des-gamma-carboxy prothrombin (GALAD) score have been recommended for HCC surveillance. However, inconsistent recommendations in international guidelines limit their clinical utility. METHODS: In this multicenter study, over 2000 patient samples were collected in 6 Latin American and 2 European countries. The performance of the GALAD score was validated in cirrhotic cases, and optimized versions were tested for early-stage HCC and prediagnostic HCC detection. RESULTS: The GALAD score could distinguish between HCC and cirrhosis in Latin American patients with an AUC of 0.76, sensitivity of 70%, and specificity of 83% at the conventional cutoff value of -0.63. In a European cohort, GALAD had an AUC of 0.69, sensitivity of 66%, and specificity of 72%. Optimizing the score in the 2 large multicenter cohorts revealed that AFP-L3 contributed minimally to early-stage HCC detection. Thus, we developed a modified GALAD score without AFP-L3, the ASAP (age, sex, AFP, and protein induced by vitamin K absence-II), which showed promise for early-stage HCC detection upon validation. The ASAP score also identified patients with cirrhosis at high risk for advanced-stage HCC up to 15 months before diagnosis (p < 0.0001) and differentiated HCC from hemangiomas, with a specificity of 100% at 71% sensitivity. CONCLUSION: Our comprehensive analysis of large sample cohorts validates the GALAD score's utility in Latin American, Spanish, and Dutch patients for early-stage HCC detection. The optimized GALAD without AFP-L3, the ASAP score, is a good alternative and shows greater promise for HCC prediction.