A total of 195 unique HCCs were retrospectively identified. After exclusion for loss to clinical follow-up, lack of post-procedural cross-sectional imaging, and the diagnosis of hepatitis B virus-induced cirrhosis, cryptogenic cirrhosis, or hemochromatosis, 138 separate HCCs in 122 unique patients were included. Demographic data are found in Table 1. Mean age at time of treatment was 64.6 ± 8.7 years. Twenty-nine (29/122, 24%) patients were female, and 93 (93/122, 76%) were male. Thirty (30/122, 22%) patients were diagnosed with NALFD/NASH, 52 (52/122, 43%) had HCV, 25 (25/122, 21%) had ALD, and 14 (14/122, 11%) presented in patients with mixed HCV and ALD-induced cirrhosis. When comparing among the cohorts, there was a significant difference in the model for end stage liver disease-sodium (MELD-Na) score (p = 0.036) among the various etiologies of HCC (Table 1). A subgroup analysis comparing patients with NASH/NAFLD to non-NASH/NAFLD patients demonstrated a significantly higher incidence of type 2 diabetes mellitus (p < 0.001) in the NASH/NAFLD cohort. Median follow-up time was 8.5 months in the ALD cohort, 9.6 months in the HCV cohort, 13.5 months in the mixed ALD/HCV cohort, and 8.9 months in the NAFLD/NASH cohort.

For the entire cohort, Kaplan–Meier analysis revealed no significant difference in overall survival (p = 0.928), local PFS (p = 0.339), or hepatic PFS among the cohorts (p = 0.946) (Fig. 2) when evaluated by underlying cause of HCC.

Kaplan Meier survival curves for the entire cohort a Overall survival (p = 0.928) b Hepatic progression-free survival (p = 0.946) c Local progression-free survival (p = 0.339). Abbreviations: NAFLD/NASH non-alcoholic fatty liver disease/non-alcoholic steatohepatitis, ALD alcoholic liver disease, HCV hepatitis C virus, HCV + ALD hepatitis C virus + alcoholic liver disease

After separating the entire cohort into non-NASH/NAFLD and NASH/NAFLD cohorts, there was no significant difference in OS (HR = 1.1, 95%CI: 0.32–3.79, p = 0.886) by log-rank analysis. Additionally, when discrete subgroup analyses comparing each etiology of HCC to the NASH/NAFLD cohort were performed, there was no significant difference in overall survival (Fig. 3, Table 2).

Kaplan Meier survival curves comparing overall survival for a NASH/NAFLD vs non-NASH/NAFLD (p = 0.886) b ALD vs NASH/NAFLD (p = 0.763) c HCV vs NASH/NAFLD and (p = 0.770) d ALD + HCV vs NASH/NAFLD (p = 0.403). Abbreviations: NAFLD/NASH non-alcoholic fatty liver disease/non-alcoholic steatohepatitis, ALD alcoholic liver disease, HCV hepatitis C virus, HCV + ALD hepatitis C virus + alcoholic liver disease.

When examining local PFS, there was no significant difference (HR = 1.2, 95%CI:0.58–2.44, p = 0.639) when the NASH/NAFLD cohort was compared to the non-NASH/NAFLD cohort (Fig. 4, Table 3) by log-rank analysis. Subgroup analyses comparing each etiology of HCC to the NASH/NAFLD cohort also did not demonstrate any significant difference in local PFS (Table 3).

Kaplan Meier survival curves comparing local progression-free survival for a NASH/NAFLD vs non-NASH/NAFLD (p = 0.639) b ALD vs NASH/NAFLD (p = 0.705) c HCV vs NASH/NAFLD (p = 0.407) and d ALD + HCV vs NASH/NAFLD (p = 0.323). Abbreviations: NAFLD/NASH non-alcoholic fatty liver disease/non-alcoholic steatohepatitis, ALD alcoholic liver disease, HCV hepatitis C virus, HCV+ALD hepatitis C virus+alcoholic liver disease

Finally, there was no significant difference in hepatic PFS when the NASH/NAFLD cohort was compared to the entire non-NASH/NAFLD cohort (HR = 1.3, 95%CI: 0.52–3.16, p = 0.595; Fig. 5, Table 4) by log-rank analysis. Similarly, when each cause of HCC was independently compared to the NASH/NAFLD cohort, there was no significant difference in hepatic PFS (Table 4).

Kaplan Meier survival curves comparing hepatic progression-free survival for a NASH/NAFLD vs non-NASH/NAFLD (p = 0.595) b ALD vs NASH/NAFLD (p = 0.579) c HCV vs NASH/NAFLD (p = 0.592) and d ALD + HCV vs NASH/NAFLD (p = 0.995). Abbreviations: NAFLD/NASH non-alcoholic fatty liver disease/non-alcoholic steatohepatitis, ALD alcoholic liver disease, HCV hepatitis C virus, HCV+ALD hepatitis C virus+alcoholic liver disease

There was no significant difference (p = 0.616) in post-radioembolization clinical toxicities as characterized by Common Terminology Criteria for Adverse Events (CTCAE) between each cohort. Two (2/52, 3.8%) patients in the HCV cohort suffered Grade 1 nausea. One (1/25, 4%) patient in the ALD cohort suffered Grade 1 nausea and two additional patients (2/25, 8%) suffered Grade 2 abdominal pain. Two (2/14, 14%) patients in the combined ALD/HCV cohort suffered Grade 1 abdominal pain. Five (5/30, 17%) patients in the NASH/NAFLD cohort suffered Grade 1 abdominal pain. In the HCV cohort, two patients suffered right common femoral access site complications, including one pseudoaneurysm requiring operative management.

There was no difference in best overall radiologic response (progressive disease, stable disease, partial response, complete response) for the entire cohort when separated by underlying cause of HCC (p = 0.687, Table 5). A subgroup analysis comparing the best overall radiologic response rate for each cause of HCC to the NASH/NAFLD cohort did not demonstrate any significant difference (Table 6).