Abstracts of the 2023 Annual Meeting of the ALEH
More infoNew hepatocellular carcinoma (HCC) surveillance approaches including PIVKA-II, AFP, and the GALAD panel of serum biomarkers are linked to HCC, but inconsistent use in guidelines limits their value. This study aimed to determine the validity of known and novel serum biomarkers to detect liver cancer in a Latin American cohort.
Materials and MethodsIn a multi-center study, 2045 patient samples were retrospectively or prospectively collected from 7 countries in Latin America and Europe and analyzed for cancer and liver disease etiology. The performance of multivariable models based on AFP and PIVKA was tested for early-stage HCC detection, low AFP HCC, 12 months pre-diagnostic HCC (n=92, range 9-15 months), and compared to cirrhosis and other liver tumors.
ResultsThe GALAD model showed excellent ability to differentiate HCC from liver cirrhosis in our prospective Latin American cohort, with an AUC of 87.9. Sub-analysis of early HCC still demonstrated excellent performance in Latin American cohort. A novel multivariable model was developed to detect early-stage HCC with low AFP levels, by combining sex, age, AFP, and PIVKA-II (also called GAAD), which resulted in AUC of 87.3. Both GALAD and GAAD effectively differentiated low AFP HCC from cirrhosis in both European and Latin American patients, with AUCs of 82.8 and 81.6, respectively. Importantly, GAAD differentiated non-cirrhotic HCC (n=243) from other malignant and benign liver tumors with an AUC of 91.9, and it was 100% sensitive and specific in hemangioma cases (n=64).
ConclusionsWe validated for the first time the GALAD model in a large cohort of Latin American HCC patients. We demonstrated comparable performance of GALAD model with the GAAD model developed on data from our European Latin American cohorts. Our findings provide additional information for consideration of these markers in international guidelines for HCC surveillance.
