A randomised, double blind, phase II clinical trial of maintenance cabozantinib following chemotherapy for metastatic urothelial carcinoma (mUC): Final analysis of the ATLANTIS cabozantinib comparison.

LBA4505 Background: Platinum-based chemotherapy (PBC) is an active first line therapy in mUC, but duration of response is usually short. We hypothesised that switch maintenance therapy with the tyrosine kinase inhibitor cabozantinib, in patients who have derived clinical benefit from PBC, would improve outcomes for mUC. Methods: ATLANTIS is an adaptive, multi-comparison, phase II trial platform. It tests multiple maintenance therapies for mUC patients who complete 4 to 8 cycles of PBC without disease progression. Whilst some patients were selected for randomisation into alternative comparisons in the platform based on putative predictive biomarkers, those who were not selected were randomised 1:1 to commence either cabozantinib 40mg once-daily or matching placebo within 10 weeks of completing PBC until progression. The primary endpoint was progression free survival (PFS) and 114 events in 140 patients (90% power, 20% 1-sided level alpha) were required to target a hazard ratio of 0.67. Due to a change in standard of care (maintenance avelumab), this was modified to 50 events in 57 patients (72.3% power, 20% 1-sided alpha) prior to the study closing early. PFS was compared between trial arms, by intention to treat, within a Cox model incorporating baseline minimisation factors. Secondary endpoints included overall survival (OS), response rate, maximum percentage decrease in measurable disease, safety and tolerability. Results: 30 patients were randomised to cabozantinib and 31 to placebo from 25 sites (Feb 2017-March 2021). Patient characteristics (median age 69.0 years; 75.4% male; 75.4% bladder primary; 67.2% ECOG PS 0; 70.5% prior cisplatin; 36.1% visceral metastases) were balanced by treatment arm. 25 (83.3%) and 26 (83.9%) PFS events had occurred in the cabozantinib and placebo arms. Median PFS was 13.7 weeks (80% confidence interval (CI) 12.1-23.3) with cabozantinib and 15.8 weeks (80% CI 11.3-23.6) with placebo (adjusted hazard ratio 0.89 favouring cabozantinib, 80% CI 0.61-1.3, 1-sided p = 0.35). There was no difference in overall survival (hazard ratio 0.80 (0.52-1.23) p=0.25). In the safety population (n = 61) treatment related adverse events were mostly low grade. The most frequent (all grades), fatigue (56.7% vs 32.2%, p = 0.02), hypertension (43.3% vs 12.9%, p = 0.01), nausea (30% vs 19.4%, p = 0.36) and diarrhea (40.0% vs 6.5%, p < 0.01), were more common with cabozantinib. Overall, cabozantinib was tolerable with median duration of treatment of 13 28-day cycles of cabozantinib or 10 of placebo. 13 (43.%) vs 3 (9.7%) patients required dose reduction in the cabozantinib and placebo arms respectively. Conclusions: Though tolerable, cabozantinib did not show a significant benefit compared to placebo when used in the switch maintenance setting following platinum based chemotherapy. Clinical trial information: ISRCTN25859465.