Antidrug Antibodies Against Atezolizumab Associated With Poor Outcomes in HCC

Highly elevated levels of antidrug antibodies against atezolizumab were associated with poorer clinical outcomes for patients who received atezolizumab plus bevacizumab for advanced hepatocellular carcinoma.

Highly elevated levels of antidrug antibodies (ADA) against atezolizumab (Tecentriq) were associated with poorer clinical outcomes for patients who received atezolizumab plus bevacizumab (Avastin) for advanced hepatocellular carcinoma (HCC), according to findings published in JAMA Oncology.

Data showed that patients with ADA levels of at least 1000 ng/mL at cycle 2, day 1 (C2D1) of treatment exhibited reduced systemic exposure to atezolizumab and impaired proliferation and activation of peripheral CD8-positive T cells. Investigators hypothesized that the presence of elevated ADA levels early in treatment could diminish the immunotherapeutic efficacy of atezolizumab.

Investigators set 1000 ng/mL as the ADA cutoff for grouping patients as ADA high (≥1000 ng/mL) or ADA low (<1000 ng/mL) to focus on posttreatment ADA response and to minimize false-positive results. In the discovery cohort of 50 patients with advanced HCC treated with atezolizumab/bevacizumab, 9 patients exhibited high ADA levels at C2D1. In the validation cohort of 82 evaluable patients, 14 were deemed ADA high at C2D1.

The overall response rate (ORR) for patients grouped as ADA low in the discovery cohort was 34%, compared with 11% for patients who were ADA high. Notably, 5 of 9 patients who were ADA high experienced progressive disease, compared with 7 of 41 patients who were ADA low. In the validation cohort, the ORR was 29% and 7% for patients who were ADA low and ADA high, respectively. Eight of 14 patients who were ADA high had progressive disease, compared with 11 of 68 patients who were ADA low.

In the discovery cohort, progression-free survival (PFS; HR, 2.84; 95% CI, 1.31-6.13; P = .005) and overall survival (OS; HR, 3.30; 95% CI, 1.43-7.64; P = .003) were both superior for the ADA low group. The same was true in the validation cohort for PFS )HR, 2.52; 95% CI, 1.27-5.01; P = .006) and OS (HR, 5.81; 95% CI, 2.70-12.50; P = .001)

Investigators conducted this study in 2 phases. In the discovery cohort, patients with HCC who received atezolizumab/bevacizumab prospectively enrolled at the CHA Bundang Medical Center in Seongnam, South Korea, from June 2020 to July 2021. For the validation cohort, patient enrollment was extended to 3 additional tertiary cancer centers in South Korea.

Among all 132 patients included in this analysis from the discovery and validation cohorts, the median age was 61 years (range, 55-69) and 84.1% of the population was male. Additionally, 53.8% of patients had an ECOG performance status of 0, and the remaining 46.2% had a ECOG performance status of 1.

Furthermore, 18.2% of patients had Barcelona Clinical liver cancer stage B disease and 81.8% had stage C disease. The most common etiology was hepatitis B (67.4%).

At C2D1, the overall median ADA level was 45.95 ng/mL (interquartile range, 0-257.9; P <.001). Most patients did not develop atezolizumab ADAs or exhibited very low ADA levels at this point. However, investigators identified a fraction of patients who presented a very rapid and robust ADA response, particularly in patients with progressive disease compared with those who had complete response, partial response, or stable disease (discovery cohort: median, 265.05 ng/mL vs. 0 ng/mL; P = .008; validation cohort: median, 62.6 vs. 0 ng/mL; P = .01).

Additional data showed that high ADA levels remained independently associated with shorter PFS and OS, according to a multivariable analysis of both cohorts. High ADA levels were associated with shorter PFS and OS even after adjusting for age, sex, ECOG performance status, Child-Pugh score, α-fetoprotein, macroscopic vascular invasion, extrahepatic spread, and neutrophil-to-lymphocyte ratio.

Overall, ADA levels of 1000 ng/mL or greater had a favorable predictive value for OS in the time-dependent receiver operating characteristic curve (area under curve, 0.78; 95% CI, 0.63-0.82).

Investigators found an inverse correlation between concentration of atezolizumab at C2D1 was

ADA levels at C2D1; atezolizumab concentration was significantly lower in patients with ADA

values of 1000 ng/mL or greater. They also found that the fraction of Ki-67– and CD8-positive proliferating T cells was higher in the ADA low group. This fraction was not significantly different in the ADA high group, but the ADA high group displayed reduced interferon-γ and tumor necrosis factor-α from CD8-positive T cells.

“This study enrolled a limited number of East Asian [Korean] patients in an endemic region of hepatitis B virus. Therefore, it is necessary to confirm this in a larger number of patients, including other ethnic groups,” lead study author Chan Kim, MD, PhD, Medical Oncology, Department of Internal Medicine, CHA Bundang Medical Center, CHA University School of Medicine, and colleagues, wrote. “Moreover, this study focused on ADA levels at 3 weeks, an early time point, but did not evaluate the prevalence of neutralizing antibody, which is a relatively later event. Furthermore, a cutoff point for ADA positivity should be optimized and validated in future studies.”

Another trial (NCT05173298) has been initiated to address the limitations of this study.

Reference

Kim C, Yang H, Kim I, et al. Association of high levels of antidrug antibodies against atezolizumab with clinical outcomes and t-cell responses in patients with hepatocellular carcinoma. JAMA Oncol. Published online October 20, 2022. doi:10.1001/jamaoncol.2022.4733