Efficacy of the ATR inhibitor ceralasertib in patients with ARID1A-deficient gynecologic and other solid tumor malignancies.

Abstract

ARID1A is frequently mutated in cancer. Motivated by the preclinical synthetic lethality between ARID1A loss and ATR inhibition, we performed an investigator-initiated phase 2 study of the ATR inhibitor ceralasertib in ARID1A-deficient solid tumors (NCT03682289). This was a phase 2, Simon two-stage study with a planned sample size of 29 evaluable patients. Eligible patients had locally advanced or metastatic solid tumors with measurable disease by RECIST 1.1 and radiographic progression at study entry. Patients were required to have ARID1A loss as determined by immunohistochemistry analysis of tumor tissue. Patients received ceralasertib 160 mg twice daily on days 1-14 every 28 days. RNA-seq and cyclic immunofluorescence were performed on tumor tissue to identify potential biomarkers of treatment response. The confirmed objective response rate (ORR) was 14% among the 29 efficacy-evaluable patients. All 4 responses, including 3 complete responses, occurred in endometrioid endometrial carcinoma or ovarian clear cell carcinoma, with an ORR of 33% and a median duration of response of 33.7 months in this subset. Including 1 patient with uterine carcinosarcoma who had stable disease, ORR was 31% among the 13 patients with gynecologic malignancies. Exploratory RNA-seq analysis of pretreatment archival tumor tissue identified upregulated G2/M checkpoint and DNA double-strand break-sensing pathway in patients who responded. Immune profiling revealed tumor immune microenvironment changes associated with response to ceralasertib. Ceralasertib monotherapy demonstrated promising anti-tumor activity in ARID1A-deficient gynecologic malignancies. Tumor molecular and immune correlates may inform the further development of ATR inhibitors in this patient population.