CDH1-mutated clinically advanced urothelial bladder cancer (UBC): A genomic landscape and real-world clinical outcome study (RWCOS).

Abstract

4587 Background: CDH1 mutated bladder cancers are characterized by plasmacytoid histology and are associated with an aggressive clinical course. Methods: Cohort 1: 6,676 clinically advanced UBC patients underwent hybrid capture based comprehensive genomic profiling (CGP) to evaluate all classes of genomic alterations (GA) as well as microsatellite instability (MSI), tumor mutational burden (TMB), and genomic loss of heterozygosity (gLOH, high ≥16%). Tumor cell PD-L1 expression was determined by immunohistochemistry (Dako 22C3). GAs were compared between CDH1 mutated and wild-type (WT) patients using Chi-square. Cohort 2: 586 UBC patients underwent a RWCOS using the de-identified Flatiron Health-Foundation Medicine urothelial clinicogenomic database. The de-identified data originated from approximately 280 US cancer clinics (~800 sites of care) between January 2011 and April 2022. Differences in real-world progression-free survival (rwPFS) and overall survival (rwOS) were evaluated by Cox proportional hazard models. Results: Cohort 1: 217 UBC patients featured a CDH1 short variant mutation with 65.2% featuring plasmacytoid histology (PLC). The PLC cohort was slightly younger, with a higher proportion of male patients, and MSI-High status. Cell-Cycle regulatory GAs were significantly higher in the non-PLC cohort, specifically CDKN2A (25.5% vs 9.7%, p=0.01) and CDKN2B (23.6% vs 5.8%, p<0.01). CDH1-mutated UBCs featured a higher MSI high frequency (2.7% vs 0.8%; p=.002), mean TMB (14.8 vs 9.9mut/Mb p<.0001), RB1 GA (52.5% vs 20.3%; p<.0001), PTEN GA (9.2% vs 4.3%; p=.006) and PIK3CA GA (29.5% vs 21.8%; p=.02), but lower gLOH (6.8% vs 15.9%; p=.009), CDKN2A loss (12.4% vs 38.3%; p<.0001), MTAP loss (10.1% vs 25.1%; p<.0001), and FGFR3 GA (9.7% vs 18.1%; p=.002). TP53 GAs and PD-L1 expression levels were similar between groups. From cohort 2: 22 (3.7%) patients featured CDH1 mutations. When compared with the CDH1 WT patients, the age, gender, ethnicity and ECOG status were similar. CDH1 mutation was associated with less favorable outcomes for 270 UBC patients treated with immune checkpoint inhibitors (ICPI) including rwPFS (2.8 vs 3.5 months; p=.096) and rwOS (3.3 vs 9.5 months; p=.03). Similar comparisons for 316 UBC patients treated with chemotherapy showed no significant adverse impact of CDH1 mutation status on either rwPFS (7.9 vs 6.2 months; p=.11) and rwOS (13.4 vs 13.4 months; p=0.83). Conclusions: In addition to its classic association with PLC histology, CDH1-mutated UBC features an unique genomic landscape including higher MSI and TMB, activating GAs in the MTOR pathway, but lower frequency of FGFR3 GAs. RWCOS further supports that CDH1 mutation predicts resistance to ICPI-based treatment, but does not systemic chemotherapy. These findings support CGP to guide therapeutic approaches based on the personalized genomic signature of UBC.