Mutation profiling of appendiceal cancer: Distinguishing tumor grades, comprehensive mutation landscape, and ctDNA as a discovery tool.

Abstract

2564 Background: Appendiceal cancer (AC) encompasses rare tumors with varying clinical behavior. Histologic grade is a key determinant of disease biology and prognosis. This study utilized an in-house circulating tumor DNA (ctDNA) biomarker discovery pipeline to assess genetic determinants of histologic grade in AC, analyzing both peripheral blood and tumor tissue. Methods: Paired peripheral blood and solid tumor samples were collected from 52 patients undergoing surgery for AC (18 low-grade and 34 intermediate/high-grade). Comprehensive genomic profiling (CGP) using the TSO500 assay was performed on ctDNA, tissue-derived DNA and buffy-coat (germline)-derived DNA. Tumor-specific and germline mutations were analyzed using OncoKB, which classifies variants as oncogenic, likely oncogenic, or actionable (Level 1 therapeutic mutations with an approved therapy). The concordance of mutations between solid tumor and plasma CGP assays was assessed, categorizing variants as detected in both tumor and ctDNA, ctDNA only, or tumor only. Results: ctDNA exhibited 82% concordance with tumor tissue for known actionable mutations. Among 26 patients that were identified to have Level 1 therapeutic mutations, 88% (n = 23) had matching mutations in plasma. Frequently detected mutations included KRAS (40%), GNAS (30.8%), SMAD4 (28.8%), and TP53 (28.8%) . Germline analysis revealed additional variants, including RUNX1 (71.2%), NOTCH4 (50%) , and BARD1 (48.1%). Tumor-specific TP53 and SMAD4 mutations correlated with high-grade tumors; while GNAS was more prevalent in low-grade tumors. Germline analysis identified NOTCH3 and SPEN mutations predominantly in high-grade tumors, suggesting that inherited determinants may determine tumor grade (23.5% each). Plasma samples exhibited lower variant allele frequencies, limiting sensitivity for novel biomarker discovery. Concordance analysis revealed some mutations were exclusive to solid tumors, while others were plasma-specific, highlighting the need for a multi-modal genomic assessment. Conclusions: Tumor TP53, GNAS, and SMAD4 mutations serve as molecular classifiers for histologic grade differentiation in AC, while germline NOTCH3, SPEN, RUNX1, NOTCH4, and BARD1 variants may influence histologic grade. ctDNA showed strong concordance for actionable mutations but had reduced efficacy for novel mutation discovery in the plasma samples. These findings underscore the value of integrating tumor and germline profiling for classification and treatment stratification, while refining liquid biopsy methodologies to enhance sensitivity in AC research.