Genomic abnormalities in invasive endocervical adenocarcinoma correlate with pattern of invasion: biologic and clinical implications

Abstract

The pattern-based classification system for HPV-related endocervical adenocarcinoma, which classifies tumors based on the destructiveness of stromal invasion, is predictive of the risk of nodal metastases and adverse outcome. Previous studies have demonstrated clinically important molecular alterations in endocervical adenocarcinoma, including KRAS and PIK3CA mutations; however, correlation between the molecular landscape and pathological variables including pattern of invasion has not been thoroughly explored. In this study, 20 endocervical adenocarcinomas were classified using the pattern-based classification system and were subjected to targeted sequencing using the Ion AmpliSeq Cancer Hotspot Panel v2 (ThermoFisher Scientific, Waltham, MA, USA) that surveys hotspot regions of 50 oncogenes and tumor suppressor genes. Single-nucleotide polymorphisms were correlated with clinical and pathologic variables including pattern of invasion. Five (25%), six (30%), and nine (45%) cases were classified as patterns A, B, and C respectively. Lymph node metastases, advanced stage at presentation and mortality from disease were exclusively seen in destructively invasive tumors (patterns B or C). Prevalent mutations in the cohort involved PIK3CA (30%), KRAS (30%), MET (15%), and RB1 (10%). Most (94%) relevant genomic alterations were present in destructively invasive tumors with PIK3CA, KRAS, and RB1 mutations seen exclusively in pattern B or C subgroups. KRAS mutations correlated with advanced stage at presentation (FIGO stage II or higher). Our findings indicate that the pattern of stromal invasion correlates with genomic abnormalities detected by next-generation sequencing, suggesting that tumors without destructive growth (pattern A) are biologically distinct from those with destructive invasion (patterns B and C), and that pattern B endocervical adenocarcinoma is more closely related to its pattern C counterpart. The pattern-based classification may be used as a triage tool when considering molecular testing for prognostic or therapeutic purposes.