D. R. Cherry1, N. Krishnamurthy2, and M. Buckstein1; 1Department of Radiation Oncology, Icahn School of Medicine at Mount Sinai, New York, NY, 2Icahn School of Medicine at Mount Sinai, New York, NY
Purpose/Objective(s):Alpha fetoprotein (AFP) plays an important role in diagnosis, therapeutic course, and prognosis for the 60% of hepatocellular carcinoma (HCC) patients whose tumors secrete this biomarker. Here we compare metrics of AFP kinetics in patients with advanced HCC who have undergone treatment with concurrent radiation and immunotherapy as a predictor of outcomes. Materials/
Methods: We identified patients treated with definitive radiation for HCC with portal vein thrombus within 6 weeks of concurrent immunotherapy from 2017 to 2023. We excluded patients with normal range pretreatment AFP (< 9.0 ng/mL) and patients without at least one pretreatment and one post treatment AFP. We recorded AFP at 3, 6 and 12 months after treatment and absolute AFP nadir during the 12 months after radiation. We calculated the absolute and percent reductions in AFP compared to the patients’ baseline. AFP velocity was defined as reduction in AFP versus baseline divided by the time in years. We used Cox proportional hazards modeling to analyze the relationships of AFP levels and other clinicodemographic covariates on progression-free (PFS) and overall survival (OS). Covariates with p-values < 0.20 in univariable analysis were included in a multivariable Cox regression. All analysis was conducted in R version 4.3.3. Results: The 50 included patients had a median age of 64.5. Forty-two patients (84%) were men. Twenty-nine patients (58%) were Child-Turcott-Pugh (CTP) class A and 21 (42%) were CTP class B. In initial univariable analysis of percent reduction, absolute reduction, and velocity of AFP at 12-month nadir, only percent reduction in AFP at nadir was correlated with PFS and OS (HR = 0.9991, p = 0.013; HR = 0.9986, p = 0.002). Percent reduction in AFP was likewise correlated with PFS and OS at 3, 6 and 12 months post treatment (HR = 0.9991, p = 0.032; HR = 0.9990, p = 0.006; HR = 0.9850, p = 0.0157).In subsequent univariable analysis, normalization of AFP was correlated with PFS and OS (HR = 0.3908, p=0.0375; HR = 0.9991, p = 0.013; OS HR = 0.230, p = 0.009 and HR = 0.9986, p = 0.002). In a multivariable analysis of percent reduction in AFP from baseline to nadir and CTP class, only CTP was significant, though percent reduction approached significance (HR = 2.03, p = 0.041 and HR = 0.999, p = 0.0570). With respect to OS, only CTP and non-zero ECOG were significant (HR 2.842, p = 0.014 and HR = 2.979, p = 0.029). In a separate multivariable analysis, CTP but not post-treatment normalization in AFP was associated with PFS (HR = 2.101, p = 0.027; HR = 0.525, p = 0.070). However, AFP normalization, CTP, and non-zero ECOG were all correlated with OS (HR = 0.304, p = 0.041; HR 2.742, p = 0.014; HR = 2.979, p = 0.029). Conclusion: This study shows that AFP kinetics do potentially correlate with outcomes. In particular, not achieving normalization of AFP is a strong predictor for recurrence for PFS and OS. Larger patient cohorts are needed to validate the utility of AFP kinetics in HCC in a variety of clinical settings.
