KRAS and GNAS Co-Mutation in Metastatic Low-Grade Appendiceal Mucinous Neoplasm (LAMN) to the Ovaries: A Practical Role for Next-Generation Sequencing

Background

Low-grade appendiceal mucinous neoplasms (LAMNs) are cytologically low-grade tumors of the appendix, which despite their indolent appearance, are a frequent cause of pseudomyxoma peritonei (PMP) [1,2]. LAMNs can represent a diagnostic challenge when they metastasize to the ovaries, where they can be difficult to distinguish from primary ovarian mucinous tumors [3–5]. Here, we present a case involving metastatic LAMN to the ovaries and describe how next-generation sequencing was utilized to support the diagnosis of metastatic disease. We discuss the use of next-generation sequencing as a practical diagnostic tool to resolve the origin of mucinous ovarian tumors when a concurrent LAMN is identified in the appendix.

Case Report

A 49-year-old woman with hypertension and no previous history of malignancy presented with several months of abdominal pain and distension, which had worsened during the previous 3 weeks. A CT scan of the abdomen was read as a single large septated mass (27.0×22.6×12.6 cm) involving both ovaries. Labs revealed elevated serum levels of CA 125, CA 19–9, and CEA. She underwent laparoscopic total abdominal hysterectomy, bilateral salpingo-oophorectomy, complete supracolic omentectomy, and appendectomy. Intraoperatively, there were bilateral ovarian tumors and no mucinous ascites. The appendix was normal in appearance.

The left ovary was sent for intra-operative frozen section analysis. Grossly, the ovary was enlarged (5640 g, 30.0 cm), with a focal surface defect. A diagnosis of “mucinous neoplasm, suspicious for borderline tumor” was rendered. The right ovary was also enlarged (866 g, 16.0 cm) but intact. Both ovaries had rough patches on their surface suspicious for possible surface involvement, and their cut surfaces showed numerous septations without solid areas. The appendix (2.5 cm in length ×0.5 cm in diameter) exuded a minimal amount of luminal mucin upon sectioning but was otherwise grossly un-remarkable. The uterus, cervix, and fallopian tubes were also unremarkable.

Microscopic examination of the appendix (entirely submitted) revealed a 0.5-cm tumor in the body which exhibited a low-grade mucinous epithelial lining and no frankly infiltrative growth (Figure 1A). Focal deposits of acellular mucin were visible within the wall, although the wall was intact. The lamina propria and muscularis mucosae were absent at the periphery of the tumor (Figure 1B). Both ovarian tumors showed similar histologic findings including mucinous morphology with low-grade cytological features, focally complex architecture, abundant pseudomyxoma ovarii, and extensive sub-epithelial clefting (Figure 2).

Immunohistochemical stains were performed on sections of tumor from the right ovary, left ovary, and appendix (Figure 3). The ovarian tumors were strongly and diffusely positive for CK7 and CK20. However, the tumor in the appendix was only focally immunoreactive for both markers. In contrast, all 3 tumors were diffusely and strongly positive for CDX2, while PAX8 was diffusely negative.

In an effort to support the histological impression of metastatic LAMN, DNA isolated from the ovarian tumors, the appendiceal tumor, and matched normal tissue were subjected to an in-house next-generation targeted sequencing panel. This panel is validated to identify more than 2800 genetic variants in 50 genes commonly mutated in cancer [6]. It includes genes that are mutated frequently in primary ovarian mucinous neoplasms and primary colon carcinomas, such as KRAS, TP53, BRAF, CDKN2A, PIK3CA, FBXW7, SMAD4, and NRAS [7,8]. Our panel does not include BRCA1 or BRCA2; however, BRCA mutations including familial mutations are reported to be exceptionally rare in primary ovarian mucinous tumors [7,9]. Samples from all 3 tumors harbored identical point mutations in KRAS (c.35G>T p.G12V) and GNAS (c.602G>A p.R201H), which were absent in the matched normal tissue. No additional clinically significant somatic mutations were identified in the panel.

The patient received a final diagnosis of LAMN metastatic to the ovaries. Because she had both ovarian tumors and the appendix removed at surgery, her disease was considered to have been optimally debulked and she began a period of close observation. She remains free from disease 15 months after surgery.

Discussion

Our case highlights the difficulty in distinguishing metastatic LAMN from concurrent primary mucinous ovarian tumors. Determining the tissue of origin in low-grade mucinous tumors involving the ovaries and appendix can be challenging. However, establishing the tissue of origin is important because metastatic LAMN may run an aggressive clinical course, while bilateral low-grade mucinous ovarian tumors and an incidental LAMN represent surgically curable disease.

These 2 scenarios often can be resolved on histologic grounds, but this is not always the case. LAMNs may be grossly unremarkable, with microscopic evidence of invasion limited to a subtle, pushing type of invasion with loss of the lamina propria and muscularis mucosae (Figure 1B) [10]. The presence of acellular mucin within the wall of the appendix also can be a helpful feature [11]. Ovarian tumors can display subtle histologic features which favor a diagnosis of metastatic LAMN.

These features include abundant pseudomyxoma ovarii without an exuberant inflammatory response (Figure 2A), prominent sub-epithelial clefting (Figure 2A), and a complex glandular architecture with tall mucinous cells (Figure 2B, 2C) [5]. Notably, overt mucinous ascites is not the only manifestation of pseudomyxoma peritonei (PMP), because ovarian involvement by LAMN also is a form of PMP [11]. While these features were present in our case, the prevailing opinion in our department was that the ovarian tumors represented bilateral primary mucinous borderline tumors and the minute appendiceal LAMN was an incidental finding. This idea was supported by the absence of overt mucinous ascites and discordant immunohistochemical staining between the tumors in the ovaries and the appendix (Figure 3).

In cases where 2 or more concurrent tumors are identified at different anatomic sites, DNA sequencing has been utilized to determine whether the masses represent metastatic disease or synchronous primary tumors [12–14]. We believe this approach may have particular utility for low-grade tumors such as LAMNs which tend to be genetically stable. Our institution favors next-generation sequencing (NGS) for this purpose. However, since genetic variants associated with LAMN are limited primarily to several hot-spot regions in KRAS and GNAS, Sanger sequencing also would be a viable alternative. NGS is currently available at most academic medical centers and accessible to community hospitals through reference laboratories. To date, the expense of NGS has limited its use primarily to the molecular characterization of treatment-refractory tumors. In these cases, it often identifies actionable therapeutic targets and may help match patients with available clinical trials. However, as the cost of sequencing continues to fall, NGS will be viewed increasingly as a practical tool to resolve common diagnostic dilemmas in pathology rather than as an assay of last resort. Its utility will include identifying the tissue origin of tumors.

In the present case, we submitted samples of tumor from both ovaries, the appendix, and matched normal tissue for analysis by our in-house cancer gene mutation panel. The NGS was completed at no cost to the patient and we did not seek reimbursement from the patient’s insurance. The results identified clinically significant somatic point mutations in KRAS and GNAS in the appendix, which also were identified in both ovarian tumors. Both of these mutations are associated with oncogenesis. KRAS is a proto-oncogene encoding a GTPase that acts as a molecular switch to control cellular pathways of proliferation and differentiation [15]. Mutation of KRAS at G12 greatly reduces GTP hydrolysis, leading to constitutive KRAS activity [15]. GNAS encodes the stimulatory alpha subunit of G-protein-coupled receptors [16]. Mutation of GNAS at R201 has been reported to result in loss of GTPase activity, leading to constitutive downstream pathway activation [17].

LAMNs have related but demonstrably different genetic signatures compared to primary mucinous ovarian tumors. While most LAMNs are defined by co-mutation of KRAS and GNAS, primary mucinous ovarian tumors exhibit more varied genetics with mutations in multiple known tumor drivers [7,18,19]. In a series of 82 mucinous ovarian tumors, fewer than 3% contained KRAS and GNAS co-mutations [7]. Thus, we propose that when identical KRAS and GNAS mutations are identified concurrently in a LAMN and ovarian mucinous tumor, the diagnosis of metastatic LAMN can be made with a high degree of certainty.

Conclusions

Careful attention to morphologic features can be coupled with the practical application of next-generation sequencing to resolve challenging cases involving concurrent appendiceal and ovarian mucinous tumors of unclear etiology. In our case, next-generation sequencing provided a rapid and practical solution to distinguish between metastatic disease and concurrent primary tumors.