Clinicopathologic and allele-specific analysis of germline <i>ATM</i> alterations in a pan-cancer cohort.

10615 Background: ATM is a tumor suppressor gene involved in DNA repair and telomere maintenance. ATM biallelic pathogenic or likely pathogenic germline variants (gPV) are associated with ataxia-telangiectasia syndrome. Monoallelic ATM gPV are associated with increased cancer risk; however, their contribution to carcinogenesis has not been elucidated. We sought to characterize the genomic landscape of gPV ATM -associated cancers in a pan-cancer cohort. Methods: ATM alterations (germline and somatic) were identified in patients (Pts) with a solid tumor diagnosis sequenced with MSK-IMPACT, an FDA-approved, tumor-normal paired targeted NGS. Genetic-imputed ancestry clinicopathologic characteristics and FACETS estimated allele-specific copy number profiles were evaluated. Samples with purity &lt;0.2 or with undetermined copy number profiles were excluded. HRD score was evaluated using FACETS profiles. Both somatic and germline sequencing data were analyzed within protocol NCT01775072. Results: Among 40,136 Pts with cancer who underwent germline testing, 1.1% (n=442) harbored an ATM gPV, inclusive of 2 Pts with biallelic ATM gPVs. The most frequent mutations were: R2547_S2549del (n=17), K2756* (n=14) and V1268* (n=11). Among these, R2547_S2549del and K2756* were identified exclusively in Pts with European ancestry, while E343Ifs*2 (n= 9/10) and c.1065+1G&gt;T (n= 6/7) were most prevalent in Pts with Ashkenazi Jewish (AJ) ancestry. Concomitant gPVs in other genes were observed in 10% (n=47), with the most frequent ones being APC (I1307K), MUTYH , BRCA2, CHEK2 , and BRCA1 . 63,270 out of 86,039 tumor samples had usable FACETS profiles. ATM somatic allele-specific information was available for 67% (297/442) of Pts. 62% (197/317) of tumor samples from Pts with monoallelic ATM gPVs exhibited somatic biallelic (Bi) ATM inactivation, while 38% (120/317) retained monoallelic (Mono) ATM status. The underlying mechanism of Bi-ATM inactivation was loss of heterozygosity in 75% (149/197) and additional somatic ATM mutations in 25% (48/197). Samples from Pts with tumors known to be associated with ATM gPVs (Breast, Pancreatic and Prostate cancers) had significant enrichment in Bi-ATM inactivation compared to tumors without strong association with ATM gPVs (88% Vs 58%; p&lt;0.01). Although samples with Bi- ATM had higher overall HRD-scores compared to Mono- ATM (median = 34 Vs 20; p&lt;0.01), no significant enrichment in HRD-High phenotype was seen in Bi- ATM (23 Vs 18; p=0.49). Conclusions: Evaluation of a pan-cancer Pt population with ATM gPVs demonstrated a high prevalence of biallelic somatic inactivation. Most Pts with ATM gPVs who had malignancies implicated in ATM -associated cancer risk, had biallelic somatic inactivation in their tumors suggestive of their contribution to tumorigenesis. While biallelic ATM inactivation is associated with higher genomic instability, no enrichment in HRD phenotype was observed.