The Landscape of ctDNA in Appendiceal Adenocarcinoma

Abstract

Purpose: Appendiceal adenocarcinoma is a rare malignancy with distinct histopathologic subtypes and a natural history with metastasis primarily limited to the peritoneum. Little is known about the molecular pathogenesis of appendiceal adenocarcinoma relative to common tumors.

Experimental design: We analyzed molecular data for patients within the Guardant Health database with appendix cancer (n = 718). We then identified patients with appendiceal adenocarcinoma at our institution (from October 2004-September 2022) for whom ctDNA mutation profiling (liquid biopsy) was performed (n = 168) and extracted clinicopathologic and outcomes data. Of these 168 patients, 57 also had tissue-based tumor mutational profiling, allowing for evaluation of concordance between liquid and tissue assays.

Results: The mutational landscape of ctDNA in appendiceal adenocarcinoma is distinct from tissue-based sequencing, with TP53 being the most frequently mutated (46%). Relative to other tumors, appendiceal adenocarcinoma seems less likely to shed ctDNA, with only 38% of patients with metastatic appendiceal adenocarcinoma having detectable ctDNA (OR = 0.26; P < 0.0001 relative to colorectal cancer). When detectable, the median variant allele frequency was significantly lower in appendiceal adenocarcinoma (0.4% vs. 1.3% for colorectal cancer; P ≤ 0.001). High-grade, signet ring, or colonic-type histology, metastatic spread beyond the peritoneum, and TP53 mutation were associated with detectable ctDNA. With respect to clinical translation, patients with detectable ctDNA had worse overall survival (HR = 2.32; P = 0.048). In the Guardant Health cohort, actionable mutations were found in 93 patients (13.0%).

Conclusions: Although metastatic appendiceal adenocarcinoma tumors are less likely to shed tumor DNA into the blood relative to colorectal cancer, ctDNA profiling in appendiceal adenocarcinoma has clinical utility.

Figure 1.. The landscape of circulating tumor DNA in appendiceal adenocarcinoma

A) Oncoprint of the top 11 genes noted to be mutated in the peripheral blood of patients with appendiceal adenocarcinoma in the Guardant Health cohort of 718 patients, along with microsatellite instability status and TMB. These were the only mutations noted in greater than 1% of samples. B) Pie chart of the types of circulating mutations noted in appendiceal adenocarcinoma patients. C) Histogram of the number of alterations noted within this tumor agnostic ctDNA assay run on circulating plasma of patients with AA. D) Histogram of the tumor mutational burden (TMB) score noted within the peripheral blood of appendiceal adenocarcinoma patients. E) Histogram of the clinical actionable mutations identified in circulating plasma. F) Co-mutational and mutual exclusivity plot of mutations noted in the peripheral blood of AA patients. Statistically significant co-occurrences are denoted with an * while there were no statistically significant mutually exclusive associations.

Figure 2:. Clinical and mutational predictors of detectable circulating tumoral DNA in appendiceal adenocarcinoma

A) Percent of total samples with positive ctDNA comparing AA and CRC. Rates of ctDNA positivity were significantly higher in CRC (OR=0.26, p<0.0001). B) Circulating mutational profiles when for AA (n=153) along with tumor mutational status in AA (n=569) from the MD Anderson cohort. C) Percent of total samples with positive ctDNA in AA by tumor grade comparing well, moderate, and poorly differentiated tumors. A trend exists for increasing ctDNA positivity by grade and there was a significantly higher rate comparing well and poorly differentiated tumors (p=0.01). D) Percent of total samples with positive ctDNA in AA by tumor histology comparing mucinous, goblet, non-mucinous, and signet ring histologies. Rates of ctDNA positivity was lower in mucinous tumors as compared to goblet (p=0.02) and signet ring tumors (p=0.03), all other associations were non-significant (p>0.05). E) Percent of total samples with positive ctDNA in AA by site of metastatic disease shows distant metastases had a higher rate of ctDNA positivity as compared to local disease (p=0.005). F) Circulating tumoral variant allele frequency comparing AA (n=58) to CRC (n=241) demonstrating a significantly higher VAF in CRC (p<0.001).

Figure 3.. Concordance between tissue and circulating mutational status within the MDACC cohort.

A) Mutational concordance between circulating and tumoral mutations comparing AA and CRC demonstrates significantly lower concordance in AA (p<0.0001). B) Mutational concordance between circulating and tumoral mutations comparing AA by grade demonstrates significantly lower concordance in well or moderately differentiated cancers as compared to poor (p=0.0001 for both associations). C) Mutational concordance between circulating and tumoral mutations comparing in AA by mutation status demonstrates higher concordance for TP53 mutations as compared to GNAS or KRAS mutational status. All comparisons of sensitivities described above calculated using McNemar’s Chi-squared.

Figure 4:. Associations between circulating mutational status and survival.

A) Overall survival of AA patients by ctDNA status demonstrates significantly impaired overall survival in the group of patients with detectable ctDNA (HR=2.324, p=0.0482). B) Overall survival of poorly differentiated AA patients by ctDNA status demonstrates significantly impaired overall survival in the group of patients with detectable ctDNA (HR=2.7, p=0.0351). C) Overall survival of AA patients by circulating KRAS status demonstrates significantly impaired overall survival in the group of patients with detectable ctDNA (HR=6.38, p<0.0001). All Hazard ratios calculated using Kaplan-Meier statistics.