Establishing novel mutation subtypes in peritoneal carcinomatosis of appendiceal origin.

Abstract

200 Background: Appendiceal cancer (AC) is a rare disease process with complex treatment strategies. The objective of this study was to identify mutation-based genetic subtypes that may differ from the current histological classification, compare genetic make-up of primaries and metastases, and to find novel targetable alterations. Methods: The analysis involved the curation and normalization of mutation panels from adenocarcinoma and mucinous adenocarcinoma (n = 196) stored in the AACR GENIE Database v6.0. Genes mutated in less than two patients and tumors profiled with incomplete mutation panels were excluded from the study. The optimal number of AC subtypes was established using the Non-negative Matrix Factorization algorithm. Statistical comparisons of mutation frequencies were performed by using the Pearson’s χ2 test. Results: AC patients were stratified into five mutation subtypes. AC0 had no mutations in the 41 genes in the study. The most frequently mutated genes varied between the subtypes. AC1: KRAS (91.9%) and GNAS (77.4%); AC2: KRAS (52.5%), APC (32.5%), and GNAS (30%); AC3: KMT2D (38.7%), TP53 (38.7%), KRAS (35.5%), EP300 (22.6%); and AC4: TP53 (97.2%), KRAS (77.8%), and SMAD4 (36.1%). Additionally, AC3 was less likely to be mucinous (22.6% v 50.0-74.2%, p < 0.001) and had a higher mutation frequency (3.6 v 0-3.1, p < 0.001). There were no differences between primaries and metastases (p = 0.35). Conclusions: Characterization of these subtypes suggest a need for molecular rather than anatomic staging for AC. Histone regulation by KMT2D and EP300 may be considered for targeted therapy for patients with AC3. A prospective comparison of subtype prognosis and response to surgery and adjuvant treatment is needed.