Re-evaluating the diagnostic value of α-fetoprotein for hepatocellular carcinoma in the direct-acting antiviral era.

[BACKGROUND] Direct-acting antivirals (DAAs) achieve sustained virological response (SVR) in most hepatitis C virus (HCV) patients, reducing hepatic inflammation and baseline α-fetoprotein (AFP). Concurrent shifts toward nonviral hepatocellular carcinoma (HCC) may further improve AFP specificity. This study re-evaluated the diagnostic performance of AFP and des-γ-carboxy prothrombin (DCP).

[METHODS] The authors retrospectively analyzed 388 consecutive patients undergoing curative hepatectomy for HCC. Tumor marker levels measured 4 months postoperatively in recurrence-free patients at 1 year (n = 257) served as non-HCC reference values. Cohorts were stratified by era (pre-DAA vs. post-DAA). Diagnostic accuracy was assessed via area under the receiver operating characteristic curve (AUROC). Subgroup analyses used multivariable modeling to test fibrosis and viral status as predictors of baseline (postoperative) AFP.

[RESULTS] Baseline (postoperative) AFP was lower post-DAA (3 [2-4] ng/mL) than pre-DAA (4 [3-7] ng/mL; p < .001). AFP AUROC improved from 0.702 to 0.793 (p = .014), whereas DCP performance was unchanged. The proportion of HCV-related HCC declined (30% to 20%; p = .012) and nonviral HCC increased (58% to 66%; p = .093). HCV non-SVR and advanced fibrosis independently predicted higher baseline (postoperative) AFP. An updated AFP cutoff of 5 ng/mL is supported; the conventional DCP threshold of 40 mAU/mL remains appropriate.

[CONCLUSIONS] In the current era, lower baseline AFP levels are associated with improved diagnostic performance of AFP for HCC. Adoption of a 5 ng/mL AFP cutoff and 40 mAU/mL DCP cutoff appears appropriate for current clinical practice.