Combination Biomarker Is Predictive of Response to Immunotherapy in Metastatic Urothelial Carcinoma

Sangeeta Goswami, MD, PhD

The combination of ARID1A mutations and the immune signaling protein CXCL13 proved to be predictive of clinical response and prolonged survival after treatment with checkpoint inhibitors in patients with metastatic urothelial carcinoma (mUCC), according to results from a study recently published in Science Translational Medicine.¹

“Most biomarker studies have been limited to a single biomarker, such as tumor mutational burden or PD-L1 expression,” Sangeeta Goswami, MD, PhD, assistant professor of Genitourinary Medical Oncology, stated in a press release.² “Our study indicates that combinatorial biomarkers that reflect both the tumor mutational status and immune response will improve predictive capability of the biomarker and may enable better patient selection for treatment with immune checkpoint therapy.”

Through the use of multiplatform analyses of tumor tissue samples, investigators from The University of Texas MD Anderson Cancer Center revealed that in patients who responded well to treatment with checkpoint inhibitors, both the presence of ARID1A mutations in tumor cells and CXCL13 expression in surrounding immune cells had been enhanced.

Patients with mUCC typically have a poor prognosis disease, with a 5-year OS rate of approximately 5%.2 Although the approval of checkpoint inhibitors has resulted in improved outcomes for this patient population, only 15% to 20% will experience benefit with this approach. Predictive biomarkers are needed to determine for which patients this approach will be most beneficial.

For this analysis, The University of Texas MD Anderson Cancer Center provided samples for the discovery cohort from 2 previous clinical trials: CheckMate-275 (NCT02387996), which evaluated the efficacy of nivolumab (Opdivo) in unresectable or mUCC, and a phase 1/2 trial (NCT01928394), which had evaluated nivolumab alone or combined with ipilimumab (Yervoy) in patients with advanced or metastatic solid tumors. For the confirmatory cohort, investigators used data sets from CheckMate-275 and IMvigor210 (NCT02108652), which examined atezolizumab (Tecentriq) in patients with locally advanced or mUCC.

Within the discovery cohort, investigators confirmed the expression of immune cytokine CXCL13 in baseline tissues and an ARID1A mutation in tumor cells (n = 31). This led investigators to examine the clinically relevant correlation between CXCL13 expression and ARID13 mutation in baseline tissues, which revealed an improved response in both cohorts, respectively. Additionally, they examined how these combined biomarkers performed in predicting patient responses to immune checkpoint therapy and, once again, found an improved OS compared with both biomarkers by themselves.

Analyses from CheckMate-275 and IMvigor210 demonstrated a significant link between the presence of the ARID1A mutation and overall survival (OS; P = .03 and P = .03, respectively). In the CheckMate-275 cohort (n = 139), a median OS of 11.4 months (95% CI, 5.48–not attained [NA]) was reported in patients with an ARID1A mutation versus 6.0 (95% CI, 4.6-11.3) in those without the mutation. In the IMvigor210 cohort (n = 275), a median OS of 15.4 months (95% CI, 9.2-NA) was reported in those with the mutation versus just 8.2 months (95% CI, 6.7-10.9) in those without the mutation. Based on these data, the presence of the ARID1A mutation in the baseline bladder tumor tissue was concluded to be predictive of favorable responses to immune checkpoint inhibitors in patients with mUCC.

CXCL13 was also examined as a potential immunological biomarker in the confirmatory cohorts from both of the phase 2 trials. Here, investigators reported higher CXCL13 expression in patients who responded to immune checkpoint inhibitor treatment compared with nonresponders in CheckMate-275 (n = 118; P <.01) and IMvigor210 (n = 237; P <.0001).

Moreover, when the cohort of patients in the CheckMate-275 trial were further divided into 1 of 3 groups—high, medium, and low—based on CXCL13 expression, higher CXCL13 expression was also found to correlate with improved OS (P = .007). In the CXCL13-high group, the median OS was 13.5 months (95% CI, 7.22-NA), 6.6 months in the CXCL13-median subgroup (95% CI, 4.6-11.4) and 5.7 months in the CXCL13-low subgroup (95% CI, 3.15-9.49).

Notably, a significant link between CXCL13 expression and OS was also reported in IMvigor210 (P <.0001). Here, the median OS was 17.1 months (95% CI, 10.9-NA), 6.7 months (95% CI, 5.5-8.8), and 8.0 months (95% CI, 5.8-10.8) in the high, medium, and low cohorts, respectively.

“Together, we identified CXCL13 expression in pretreatment tumor tissue as an immunologic biomarker of favorable responses to immune checkpoint therapy in patients with mUCC, the study authors added.

“We hope that our study will highlight the importance of developing combinatorial biomarkers that consider both tumor cells and immune cells, Padmanee Sharma, MD, PhD, professor of Genitourinary Medical Oncology and Immunology at MD Anderson Cancer Center, added in the release. “This approach may identify better biomarkers that can reliably predict response to immune checkpoint therapy across various tumor types.”

A prospective clinical trial is currently being planned to examine outcomes in patients who are positive for the combination biomarker following PD-1 inhibition.

References

1. Goswami S, Chen Y, Anandhan S, et al. ARID1A mutation plus CXCL13 expression act as combinatorial biomarkers to predict responses to immune checkpoint therapy in mUCC. Sci Transl Med. 2020;12(548):eabc4220. doi:10.1126/scitranslmed.abc4220
2. Combination biomarker predicts response to immune checkpoint therapy in patients with advanced bladder cancer. News release. University of Texas MD Anderson Cancer Center. June 17, 2020. Accessed August 12, 2020. https://bit.ly/3iC5ait.