Landscape of somatic alterations and real world outcomes of ovarian cancer associated with ATM mutations

Background
ATM germline mutation carriers have a two-fold increased risk of ovarian cancer, but little is known about tumor biology and outcomes of ATM-associated ovarian cancer.

Methods
Using the Tempus Lens Platform, we identified ovarian cancer patients tested with Tempus xT (DNA, 648 genes), tumor-normal match testing, and receipt of 1L platinum-based therapy. Carriers had a pathogenic/likely pathogenic germline or somatic mutation in ovarian cancer-associated genes. Somatic alterations were compared by gene, among carriers and non-carriers using adjusted p-values (q-values). Real-world progression-free survival (rwPFS) was calculated from 1L treatment start to 2L, progression, death or censoring, with maximum follow-up of 3 years. Hazard ratios (HR) were estimated using Cox proportional hazard models and compared using a Wald test.

Results
1810 patients were eligible: 314 carriers (132 BRCA1, 78 BRCA2, 37 ATM, 21 BRIP1, 46 others) and 1496 non-carriers. Median diagnosis age was 64; most were diagnosed at advanced stages (69%) and with high-grade serous histology (72%). The most common 1L regimens were chemotherapy alone (50.3% carriers, 60.6% non-carriers), chemotherapy with PARP inhibitors (29.3% carriers, 15.5% non-carriers) or chemotherapy with bevacizumab (12.7% carriers, 17.3% non-carriers). Compared to non-carriers, ATM carriers had lower frequency of somatic TP53 alterations (37.8%, vs. 79.0%, q<0.001) and a trend towards higher somatic KRAS alterations (24.3%, vs. 8.8%, q=0.049). Compared to ATM carriers, BRCA1 carriers had longer rwPFS (median not reached vs.18.1 months (mo), HR 0.46, 95% confidence interval (CI) 0.22-0.94, p=0.034) with a similar but non-significant trend for BRCA2 carriers (median 28.8 mo, HR 0.56, CI 0.26-1.20, p=0.13). rwPFS did not differ between non-carriers and ATM carriers (median 18.3 vs 18.1 mo, HR 0.91, CI 0.50-1.65, p=0.75).

Conclusions
Despite fewer TP53 mutations, ATM associated ovarian cancers have shorter PFS compared to BRCA1/2 associated disease. This suggests ATM-associated ovarian is biologically distinct and has implications for trial design and prevention and treatment strategies.