Importance: Administration of atezolizumab could be immunogenic and induce undesirable antidrug antibody (ADA) responses. This may interfere with atezolizumab-mediated actions, affecting drug clearance and serum concentration or inducing antibody neutralization.
Objective: To determine the clinical and immunological associations of highly elevated ADA levels with clinical outcomes after atezolizumab/bevacizumab (Atezo/Bev) treatment in patients with advanced hepatocellular carcinoma (HCC).
Design, setting, and participants: This cohort study prospectively enrolled 174 patients with advanced HCC treated with first-line Atezo/Bev (discovery cohort: 61 patients from 1 center; validation cohort: 113 patients from 4 centers).
Exposures: Serum ADA levels at pretreatment and 3 weeks (cycle 2 day 1 [C2D1]) were analyzed using competitive enzyme-linked immunosorbent assays. In addition, samples were subjected to serological and flow cytometric analyses.
Main outcomes and measures: Overall, ADA positivity was associated with treatment outcomes and T-cell functions.
Results: After excluding patients with inadequate samples, follow-up loss, or consent withdrawal, 132 patients (discovery cohort: 50 patients; 41 [82.0%] men; median age [IQR], 61 [55-70] years; validation cohort: 82 patients; 70 [85.4%] men; median age [IQR], 61 [53-68] years) were analyzed, and robust ADA (≥1000 ng/mL) responses at C2D1 were identified in 23 (17.4%) of the patients. Patients with progressive disease exhibited higher ADA levels (median [IQR], 65.2 [0-520.4] ng/mL) at C2D1 than in responders (median [IQR], 0 [0-117.5] ng/mL). In both discovery and validation cohorts, patients with high ADA levels at C2D1 were associated with a reduced response rate (discovery cohort: 34% vs 11%; validation cohort: 29% vs. 7%) and worse progression-free survival (discovery cohort: hazard ratio [HR], 2.84; 95% CI, 1.31-6.13; P = .005; validation cohort: HR, 2.52; 95% CI, 1.27-5.01; P = .006) and overall survival (discovery cohort: HR, 3.30; 95% CI, 1.43-7.64; P = .003; validation cohort: HR, 5.81, 95% CI, 2.70-12.50; P = .001) with Atezo/Bev compared with those with low ADA levels. In multivariable Cox regression, the clinical implication of high ADA levels persisted even after adjusting for various confounding factors and was most significant at 1000 ng/mL or greater. Compared with patients with low ADA levels, patients with high ADA levels exhibited reduced serum atezolizumab concentrations, impaired CD8-positive T-cell proliferation, and had decreased interferon-γ and tumor necrosis factor-α from CD8-positive T cells compared with patients with low ADA levels.
Conclusions and relevance: This cohort study found that highly elevated ADA levels at C2D1 may be associated with poor clinical outcomes in patients with advanced HCC treated with Atezo/Bev. High ADA levels may reduce atezolizumab exposure and attenuate the anticancer efficacy of the drug.