Consistent with previous studies, the most prevalent genomic alterations in disseminated AMNs involved genes within the MAPK signaling pathway and GNAS. MAPK/GNAS mutations were present in 92% of cases and found in all three AJCC/WHO histologic grades of disseminated AMNs. In comparison, genomic alterations in TP53, SMAD4, ATM, CDKN2A, and the mTOR genes were identified in 43% of disseminated AMNs and preferentially seen in moderately-differentiated and poorly-differentiated mucinous adenocarcinomas. Furthermore, mutations in the latter group of genes were clinically associated with decreased patient OS and an adverse prognostic factor for OS that was independent of AJCC/WHO grade, PCI, and CC score.
To date, the number of reports that have evaluated the genomic landscape of disseminated AMNs and their prognostic significance has been limited. One of the largest series to molecularly profile appendiceal neoplasms was by Ang et al.23 Analogous to the results reported herein, the authors found MAPK/GNAS mutations to be the most frequent genomic alterations in disseminated AMNs, and inactivation of TP53 and SMAD4 were preferentially detected in mucinous adenocarcinomas and signet ring cell carcinomas. While the number of patients with prognostic follow-up was small, Ang et al showed TP53 alterations were associated with decreased patient OS. Similarly, Raghav et al published TP53 mutations were enriched in moderately-differentiated and poorly-differentiated mucinous adenocarcinomas, and TP53 mutational status correlated with poor patient OS after CRS-HIPEC.33 However, mutations in TP53 were not an independent prognostic factor of OS when adjusted for AJCC/WHO grade. The authors therefore hypothesized that disseminated AMNs represents a network-based disease, arising not from a single dominant gene, but rather through dysregulation of multiple genomic pathways that converge to a specific oncologic state. This supposition is supported by our findings that collective genomic alterations involving not only TP53, but also SMAD4, ATM, CDKN2A, and the mTOR genes are an independent, predictive factor for poor patient OS.
Despite the genetic heterogeneity seen among disseminated AMNs, a potential neoplastic progression model can be inferred from the results of targeted NGS. Based on the tenet that gastrointestinal tract cancers develop through a linear sequence of accumulating genomic alterations34, mutations in the MAPK genes and/or GNAS represent the earliest genomic event in disseminated AMNs and, as mentioned previously, detected in both LAMNs and mucinous adenocarcinomas. Late genomic events include those involving TP53, SMAD4, ATM, CDKN2A, and the mTOR genes, and predominantly identified in mucinous adenocarcinomas. Thus, our data would suggest that mucinous adenocarcinomas of the appendix arise from LAMNs. This potential model is not unfounded when considering other neoplasms that harbor similar genomic alterations, such as pancreatic ductal adenocarcinomas (PDACs) arising from intraductal papillary mucinous neoplasms (IPMNs).35 In addition to LAMNs, IPMNs with low-grade dysplasia harbor frequent mutations in the MAPK genes and GNAS, and acquire genomic alterations in TP53, SMAD4, CDKN2A, and the mTOR genes, as they transition to high-grade dysplasia and PDAC, analogous to appendiceal mucinous adenocarcinomas. Interestingly, many of the acquired genomic alterations in TP53 and the mTOR genes are documented to occur prior to the histologic features of high-grade dysplasia/PDAC.32 Comparatively, a subset of LAMNs also harbored TP53, SMAD4, ATM and mTOR gene alterations. For both LAMNs and IPMNs with low-grade dysplasia, the identification of these “advanced” mutations in histologically low-grade neoplasms suggest an increased risk of neoplastic transformation and potentially necessitate intense surveillance and/or more aggressive management. Hence, our findings further underscore the clinical utility of incorporating targeted NGS to the evaluation of disseminated AMNs.
We acknowledge that there are several limitations to this study. It is retrospective by design and not all patients received the same form of HIPEC or systemic chemotherapy regimen, if indicated. In addition, the number of genes evaluated by targeted NGS was relatively small. Disseminated AMNs are typically paucicellular neoplasms in a background of abundant mucin, which represents a significant challenge for current NGS workflows that require at least 30% tumor cellularity. Consequently, it is not uncommon for targeted NGS of disseminated AMNs using broad gene panels to be negative for genomic alterations. To obviate this issue, we assessed a minimal panel of gastrointestinal tract cancer-specific genes with a high depth of sequencing coverage and, as a result, 98% of disseminated AMNs within this study had detectable genomic alterations. Raghav et al used multiple broad gene panels and identified genomic alterations in only 86% of cases.33 Another issue with this study is we evaluated tissue specimens taken at the time of the patient’s first attempted debulking procedure, and, therefore, it does not adequately assess the genomic evolution of disseminated AMNs upon multiple rounds of successive therapy and after a long follow-up period. Finally, we note that a subset of poorly-differentiated mucinous adenocarcinomas were negative for MAPK/GNAS mutation, which would argue against the proposed LAMN-to-mucinous adenocarcinoma sequence. However, one of the challenges with pathologic evaluation of AJCC/WHO grade 3 tumors is differentiating between a poorly-differentiated mucinous adenocarcinoma of the appendix and a goblet cell adenocarcinoma.27 Based on pathologic assessment alone, a poorly-differentiated appendiceal mucinous adenocarcinoma, such as a mucinous signet ring cell carcinoma, can be histologically indistinguishable from a goblet cell adenocarcinoma. Considering mutations in MAPK/GNAS are often absent in goblet cell adenocarcinomas, it is plausible that a small subset of goblet cell adenocarcinoma mimicking a poorl-differentiated mucinous adenocarcinoma of the appendix may have inadvertently been included within this study to explain this discrepancy.36
In summary, targeted NGS of disseminated AMNs identified a high prevalence of mutations in the MAPK genes and GNAS, and less frequent genomic alterations in TP53, SMAD4, ATM, CDKN2A, and the mTOR genes. However, alterations TP53, SMAD4, ATM, CDKN2A, and mTOR gene alterations were associated with a mucinous adenocarcinoma, the presence of lymphovascular and perineural invasion, and poor patient OS. Moreover, these “advanced” mutations were an adverse prognostic factor for OS that was independent of AJCC/WHO grade, PCI, and CC score. While focused studies are required to further delineate the genetic underpinnings of disseminated AMNs and their neoplastic progression with respect to time after initial treatment, our molecular findings do suggest a linear progression model from LAMN to mucinous adenocarcinoma with a transition that is characterized by the dysregulation of multiple genomic pathways.