OR20-04 Genetic Profiling Using A Gene Panel Identifies Pathogenic Variants And Copy Number Variations In Pituitary Adenomas

Abstract

Abstract Disclosure: T. Silva: None. D. Mohan: None. J. Mebarak: None. D. Meredith: None. W. Bi: Speaker; Self; Stryker. R. Carroll: None. U.B. Kaiser: None. A. Marcondes Lerario: None. A. Abreu: None. Introduction: Pituitary adenomas (PA) are the second most common central nervous system neoplasm and arise from different cell lineages. PA are benign tumors, although a subset of them may present with aggressive behavior, resulting in significant morbidity. Several genetic alterations have been implicated in PA tumorigenesis, such as USP8 mutations in Cushing disease (CD) and GNAS mutations in somatotroph adenomas (SA). However, the genetic profiles of PA are not completely understood. The aim of this study was to identify genetic alterations associated with PA using a gene panel. Methods: Targeted next generation exome sequencing (OncoPanel) was performed to identify single nucleotide variants (SNV) and copy number variations (CNV) associated with PA. OncoPanel version 1 included 275 genes, version 2 had 300 genes, and the third, current version 447 genes. USP8 was included only in the most recent version. Patients undergoing surgery at our institution between 2013-2020, who agreed to participate in the OncoPanel initiative and who had samples with good quality DNA were enrolled. Classification of PA into subtypes was performed by combining clinical and histopathological data. A strict algorithm was used to assess SNV pathogenicity. Chi-square, Mann Whitney and Kruskal-Wallis tests were used for statistical analyses. Results: Our cohort included 171 patients with following PA subtypes: TPIT lineage - 18 CD, 14 silent ACTH; PIT-1 lineage - 29 SA, 21 lactotroph and 4 silent poorly differentiated PIT-1 adenomas; SF-1 lineage - 77 gonadotroph and 8 null cell adenomas (negative immunostaining for hormones and TF). GNAS mutations were found in 28% of SA and USP8 mutations in 50% of CD screened for USP8. Pathogenic variants in MECOM were associated with silent ACTH, in FGFR4 and NF2 with CD, and in NOTCH2 with lactotroph adenomas, whereas pathogenic variants in TP53, EGFR, MEN1, BRCA1, FH, ATR, BLM, and XPC were identified in multiple PA subtypes. The number of CNV varied significantly between different subtypes of adenomas and were higher in lactotroph (median, 46), poorly differentiated PIT-1 (41), SA (26) and silent ACTH (17) compared to CD (3), gonadotroph (1) and null cell (0.5) (p< 0.001). MIB-1 proliferation index >3% was associated with higher CNV burden (p=0.022). Interestingly, copy number loss of chromosome 16 and 1q combined were exclusively identified in SA without GNAS mutations. Conclusion: OncoPanel analyses identified pathogenic variants in PA, some well recognized variants such as USP8 and GNAS, and some rare events such as NF2 in CD. In addition, our findings highlight subtype-specific patterns of genomic instability. The association between high CNV burden and the MIB-1 proliferation index may help to predict adenoma behavior. Genomic profile using a gene panel has the potential to better classify PA and contribute to clinical management. Presentation: Saturday, June 17, 2023