Immune checkpoint inhibitor–based combination therapies are established options for patients with unresectable hepatocellular carcinoma (HCC). Conventional trial endpoints such as objective response rate (ORR), progression-free survival (PFS), and overall survival (OS) may not fully capture the full pattern of tumor response. This study aimed to descriptively evaluate response dynamics — including depth of response, time to response, and duration of response — in patients with unresectable HCC treated with durvalumab plus tremelimumab or atezolizumab plus bevacizumab.
This was a retrospective single-center study that included 148 patients with unresectable HCC. Patients received either durvalumab–tremelimumab (n=53) or atezolizumab–bevacizumab (n=95). Because of baseline imbalances between treatment groups and differences in the observation period, the authors conducted descriptive analyses only and treated findings as hypothesis-generating rather than confirmatory.
The study reported standard oncologic endpoints and additional measures of response dynamics. A reduction of at least 50% in the sum of target lesion diameters was prespecified as DpR50 (depth of response ≥50%). Other endpoints included ORR, disease control rate (DCR), time to response, duration of response, durable response (response lasting ≥6 months), median PFS, and median OS.
Overall response rate was higher in the durvalumab–tremelimumab group at 35.8% compared with 27.4% in the atezolizumab–bevacizumab group. Disease control rates differed in the opposite direction: the DCR was 54.7% for durvalumab–tremelimumab and 71.6% for atezolizumab–bevacizumab. These figures suggest a larger proportion of patients on atezolizumab–bevacizumab experienced stable disease or minor tumor shrinkage, while a higher fraction of durvalumab–tremelimumab patients achieved objective responses.
DpR50 (≥50% reduction in target lesion diameter) was observed in 26.4% of patients treated with durvalumab–tremelimumab versus 11.6% of those treated with atezolizumab–bevacizumab, indicating more pronounced tumor shrinkage among a subset of durvalumab–tremelimumab recipients.
Median time to response was 2.3 months for durvalumab–tremelimumab and 3.3 months for atezolizumab–bevacizumab, demonstrating a shorter interval to initial tumor reduction with the former regimen in responders.
Median duration of response favored durvalumab–tremelimumab at 22.3 months compared with 10.0 months for atezolizumab–bevacizumab. When considering durable responses lasting at least 6 months, these occurred in 24.5% of all patients treated with durvalumab–tremelimumab and 15.8% of all patients treated with atezolizumab–bevacizumab.
Median progression-free survival was reported as 5.2 months for the durvalumab–tremelimumab cohort and 7.0 months for the atezolizumab–bevacizumab cohort. Median overall survival was 17.0 months for durvalumab–tremelimumab and 22.0 months for atezolizumab–bevacizumab. The authors presented these survival outcomes descriptively given the study design limitations.
Within the constraints of a retrospective single-center dataset and unequal group characteristics, the analysis suggested distinct patterns of anti-tumor activity between the regimens. Durvalumab–tremelimumab was associated with deeper and more durable tumor shrinkage among selected responders (higher rates of DpR50 and longer median duration of response). In contrast, atezolizumab–bevacizumab appeared to confer broader disease control across the treated population, primarily through higher rates of stable disease and a higher DCR.
These observations may inform clinical expectations about the nature of response with each combination: one regimen producing less frequent but deeper, longer-lasting responses in some patients, and the other producing more frequent disease stabilization. However, the authors emphasize that results are hypothesis-generating and should not be interpreted as direct evidence of superiority.
The analysis was retrospective and single-center. The authors noted baseline imbalances between treatment groups and differing observation periods, which precluded formal comparative hypothesis testing. Details of patient characteristics, prior therapies, safety outcomes, and statistical adjustment methods were not reported in the abstract; therefore, those specifics cannot be summarized from the source. Findings should therefore be interpreted cautiously and considered exploratory.
In this descriptive cohort of 148 patients with unresectable HCC, durvalumab–tremelimumab produced higher rates of deep and durable tumor shrinkage among responders, while atezolizumab–bevacizumab delivered broader disease control largely via stable disease. Given the retrospective design and group imbalances, these outcomes are hypothesis-generating and require validation in prospective or better-controlled comparative analyses.