CIO 2022-8 Predictors of Overall Survival in Patients with Hepatocellular Carcinoma Treated with Transarterial Chemoembolization

Purpose: To evaluate predictors of overall survival in patients with hepatocellular carcinoma (HCC) treated with lipiodol transarterial chemoembolization (cTACE).

Materials and Methods: A total of 228 patients (75% male, 25% female) with treatment-naïve HCC were reviewed. All patients were treated exclusively with TACE. Baseline laboratory studies included liver function tests and complete blood count with differential to calculate neutrophil-to-lymphocyte ratio (NLR). Median NLR was 2.5. The causes of cirrhosis included hepatitis B (5%), hepatitis C (50%), alcohol (30%), and nonalcoholic steatohepatitis (20%). The baseline median sum of the two largest tumors was 3.9 cm. Most patients (67%) were Child-Pugh (CP) class A, and 33% were CP class B to C. Barcelona Clinic Liver Cancer (BCLC) stages were 0 or A for 59% of the patients, B for 35%, and C for 6%. Albumin-bilirubin (ALBI) scores were 1, 2, and 3 for 27%, 69%, and 4% of the patients, respectively. Median total bilirubin, international normalized ratio, and albumin were 1.1 mg/dL, 1.2, and 3.6 g/L, respectively. Overall survival (OS) and progression-free survival (PFS) were assessed using chi-squared and Kaplan-Meier analysis. Cox proportional hazards (CPH) model was performed to gauge the effects of NLR, CP, BCLC, ALBI, age, gender, tumor number, and cause of cirrhosis on OS and PFS.

Results: Median (95% confidence interval) OS was 25 (21.6–31.4) months. Median PFS was 8.0 (5.1–13.4) months. Median OS was longer for CP A (31.3 [24.9–42.0] months) compared with CP B or C patients (15.6 [10.3–25] months) (X2 = 23.1; P < 0.0001). Median OS was longer for BCLC 0 or A patients (29.4 [24.7–41.7] months) than BCLC B (20.4 [17.2–32.4] months) and BCLC C (12.9 [10.1–NR] months) (X2 = 15.2, P < 0.0001) patients. Median OS was not different for ALBI 1 patients (28.9 [23.5–NR] months) and ALBI 2 patients (23 [19–32.2] months) (P = 0.08). Median PFS was longer for CP A (12.2 [7.8–NR] months) patients than CP B patients (5.0 [3.3–11.5] months) (X2 = 9.5; P = 0.002). PFS was longer for BCLC 0 or A patients (11.8 [8.0–NR] months) than BCLC B (5.1 [3.8–NR] months) and BCLC C (3.6 [3.2–NR] months) (X2 = 11.6; P = 0.003) patients. ALBI was not predictive of PFS (P > 0.05). In the CPH model, NLR (hazards ratio [HR], 7.3; P = 0.007), CP B score (HR, 6.3; P = 0.01), and BCLC C score (HR, 13.8; P = 0.003) predicted worse outcomes for OS. For PFS, baseline NLR (HR, 6.8; P = 0.009), CP B score (HR, 5.9; P = 0.02) and BCLC C score (HR, 9.3; P = 0.03) predicted worse outcomes. The other factors were not predictive.

Conclusions: cTACE resulted in a median OS of 25 months for the whole cohort and 31 months for CP A patients. NLR, CP B, and BCLC C scores predict shorter OS and PFS.