Abstract / Summary
The baseline CRAFITY score, derived from AFP and C-reactive protein, is a validated prognostic biomarker for hepatocellular carcinoma (HCC) patients receiving immunotherapy. However, whether its prognostic value can be improved by on-treatment reassessment remains unclear. We evaluated the two-cycle on-treatment CRAFITY score in 108 advanced-HCC patients who received immune-based combination therapy as first- or second-line treatment. The score was calculated at baseline and after two cycles. Kaplan–Meier, Cox regression, time-dependent ROC, landmark, and sensitivity analyses were performed. The two-cycle score significantly stratified overall survival (median OS from landmark: 34.6, 12.9, and 6.7 months for scores 0, 1, and 2; p < 0.0001) and remained independently associated with OS after adjustment for vascular invasion, treatment line, and concomitant transarterial chemoembolization (TACE) (HR 2.36, 95% CI 1.70–3.29, p < 0.001). The baseline score showed only borderline associations (OS p = 0.07; PFS p = 0.09). The two-cycle score had an apparent AUC of 0.709 (95% CI 0.586–0.832) at 6 months and 0.745 (95% CI 0.658–0.832) at 12 months. The corresponding values for the baseline score were 0.645 and 0.638. Adding the two-cycle score to a clinical reference model increased the 12-month AUC from 0.699 to 0.811 (ΔAUC = 0.112). Bootstrap internal validation yielded an optimism-corrected C-index of 0.698 (95% CI 0.650–0.776). In conclusion, the two-cycle CRAFITY score provides complementary posttreatment prognostic information for early risk stratification, but prospective external validation is required before clinical application.