A Case of Synchronous Pancreatic Ductal Adenocarcinoma and Ovarian Mucinous Cystic Neoplasm: Use of Kras Mutation Molecular Phenotyping to Demonstrate Independent Primary Origin

CASE REPORT

A 56-year-old postmenopausal Puerto Rican woman presented with a brief history of obstructive jaundice and rapid weight loss. Her past medical history was significant for well-controlled hypertension and a five-year history of type II diabetes mellitus. She had noted worsening hyperglycemia over several months despite titration of oral hypoglycemic agents and had recently required initiation of insulin therapy. She had a five pack-year history of cigarette smoking and reported minimal alcohol intake. There was no family history of malignancy.

A computer tomography (CT) scan of the abdomen demonstrated a 4.5 × 3.9 cm mass involving the uncinate process of the pancreas causing biliary and pancreatic ductal dilatation. There was approximately 90 degree involvement of the superior mesenteric vein and superior mesenteric artery by the mass, with involvement of the second and third parts of the duodenum. Review of cross-sectional imaging also demonstrated an incidental finding of a 4.3 cm left ovarian cyst of indeterminate significance (see Figure 1). CA 19.9 was elevated at 2638 u/mL; CEA was within normal limits at 3.1 ng/mL.

Figure 1.

Staging CT scan prior to initiation of therapy demonstrating pancreatic and left adnexal masses.

Endoscopic ultrasound confirmed the presence of a mass in the pancreatic head, and fine-needle aspiration of the lesion was performed. Cytology was diagnostic for adenocarcinoma. Staging laparoscopy was performed, and no evidence of metastatic disease was identified. A metal biliary stent was inserted at ERCP with rapid resolution of jaundice. A duodenal stent was subsequently placed for relief of symptoms of duodenal obstruction.

The patient was enrolled on a clinical trial of neoadjuvant systemic therapy. She completed four cycles of gemcitabine and oxaliplatin without significant toxicity. A marked decline from baseline in CA 19.9 was seen. A restaging CT scan of the abdomen and pelvis was performed preoperatively following completion of neoadjuvant chemotherapy. The pancreatic head mass was stable in appearance; however, the left adnexal cyst was found to have increased in size to 6.7 × 4.8 cm (see Figure 2). A pelvic ultrasound confirmed the presence of a cystic mass with internal septations involving the left adnexa (see Figure 3). CA 125 was within normal limits at 8 units/mL.

Figure 2.

CT scan following completion of therapy showing interval stability of pancreatic mass and increase of the left adnexal mass.

Figure 3.

Direct sequencing of KRAS exon 2 was performed in the presence of 10-mer locked nucleic acid (LNA) oligonucleotide to suppress amplification of wild-type KRAS. A G to T transversion is seen in the pancreatic tumor (A), and a G to A transversion is present in the ovarian tumor (B).

Given the significant decline in CA 19.9 and interval stability of the known pancreatic mass, the patient proceeded to a laparotomy by both pancreatic and gynecologic surgical oncologists. The right adnexa appeared normal, and there was no evidence intraoperatively of distant metastases or intraperitoneal spread. A bilateral salpingo-oophorectomy was performed. Frozen section of the resected left adnexal cyst indicated a mucinous tumor; however, it was not possible based on the frozen section to determine the malignant potential of the lesion or whether it was a primary in the ovary vs. a metastasis from the pancreatic carcinoma. A decision was made to proceed with a pancreatoduodenectomy. This was an uncomplicated procedure, and the patient had an uneventful postoperative course.

Pathologic analysis of the resected pancreatic specimen confirmed an invasive pancreatic ductal adenocarcinoma arising from the pancreatic head. The tumor measured 5.5 × 2.6 × 2 cm and extended to involve the common bile duct, duodenal mucosa, and peripancreatic soft tissue. Two of seventeen lymph nodes contained metastatic carcinoma. All final surgical margins were negative. There was no cystic component in the tumor, either cystic invasive glands or a cystic precursor lesion (such as an intraductal papillary mucinous neoplasm or a mucinous cystic neoplasm).

The histology of the ovarian cyst was consistent with a well-differentiated mucinous cystic neoplasm. Given the potential that this may represent a cystic metastasis from the pancreatic primary, parallel immunohistochemical staining was performed on both tumor specimens. Both tumors stained positively for CK7 and focally for MUC 1. Staining was negative for CK 20, MUC 2, CDX2, ER, and PR. SMAD4 staining was retained in both tumors; p53 was more strongly expressed in the pancreatic tumor, and focal labeling was also seen in the ovarian tumor. This immunohistochemical profile in the ovarian tumor was thought to be consistent with either a metastasis from the pancreatic carcinoma or an ovarian primary, because insufficient differences were seen between the staining profiles of the pancreatic and ovarian tumors. Molecular analysis for KRAS mutation was then performed using polymerase chain reaction on both tumors. The pancreatic tumor was found to harbor a mutation in exon 2, mutation G12V. The ovarian tumor also had an activating KRAS mutation in exon 2; however, this was a different mutation, G12D. This finding of distinctly different activating point mutations in exon 2, codon 12 of KRAS in the two tumors strongly supported that the ovarian tumor was a primary mucinous cystadenoma, rather than a metastasis from the pancreatic adenocarcinoma. The pancreatic tumor was finally staged as T3N1M0, AJCC stage IIB.

The patient made an excellent postoperative recovery and is currently undergoing adjuvant chemotherapy with gemcitabine; postoperative CT imaging showed no evidence of residual or recurrent disease at either primary site.

DISCUSSION

The differentiation between a primary mucinous neoplasm of ovarian origin and a metastasis to the ovary from a primary adenocarcinoma of the gastrointestinal tract frequently poses a diagnostic challenge to the clinician and pathologist. Metastatic adenocarcinomas involving the ovary may become cystic, even if the primary tumor does not have cystic elements. Metastatic adenocarcinomas may differentiate in the ovary to epithelium that has the histologic appearance of a borderline ovarian tumor or even a benign cystadenoma, referred to as the maturation phenomenon.¹ In fact, in a patient with a known primary pancreatic carcinoma, mucinous tumors of the ovary are judged to be metastases until proven otherwise. Histopathological features favoring a diagnosis of ovarian metastasis rather than a primary ovarian tumor include the presence of bilateral ovarian tumors, gross and microscopic surface involvement, infiltrative pattern of stromal invasion, or a nodular invasive pattern with intervening normal stromal ovarian tissue between the nodules of mucinous tumor.² Immunohistochemical staining may also aid the differentiation between primary and secondary neoplasms; however, in our case the staining patterns were not substantially different between the two tumors. Furthermore, the immunophenotype of primary mucinous neoplasms of the ovary overlaps considerably with that of gastrointestinal and (in particular) pancreatic adenocarcinomas.

Although ovarian metastases from pancreatic adenocarcinoma are relatively uncommon in clinical practice, several single-institution reviews have reported on small numbers of patients with the ovary as a site of metastasis from a pancreatic primary adenocarcinoma, either as a single site of disease or in the setting of disseminated metastases.³ The pancreas was identified as the primary site in 10% of cases of ovarian metastases from extragenital sites in two published case series,⁴ and autopsy series have reported the finding of ovarian metastases in 4–6% of patients with pancreatic adenocarcinoma.⁵ Falchool et al described the largest case series to date of ovarian metastases from pancreatic cancer. Eighteen cases were identified in a single institution review, of which 11 underwent resection of ovarian metastases as primary treatment. There was a trend toward increased survival in the group of patients who underwent surgical resection compared with those who did not; however, this small series of patients included those with multiple sites of metastases as well as oligometastastic disease, so limited conclusions can be drawn from these data. Mucinous histology was identified in 5 out of 18 pancreatic tumors and in 8 of 16 ovarian metastases in this series. Given the potential for large, bulky ovarian metastases to cause significant symptoms and impact adversely on quality of life, consideration of palliative resection of isolated ovarian metastases is reasonable in selected patients. In this case, given the uncertainty regarding the nature of the ovarian lesion, the decision to proceed with resection of the pancreatic primary offered the patient the best chance of achieving a meaningful long-term survival.

K-RAS is a member of the RAS family of proteins, which play a key role in intracellular signal transduction pathways mediating cell proliferation, angiogenesis, and survival. The protein is produced in an inactive form in the cytoplasm and requires translocation to the cell membrane to exert its signaling activity. It may be activated by a variety of extracellular signals, causing it to bind to GTP-ase-activating protein (GAP) to stimulate the RAF/mitogen-activated protein kinase (MAPK) pathway.⁶ The RAF/MAPK protein complex in turn causes phosphorylation of MEK1 and MEK2, resulting in stimulation of downstream effector proteins involved in the regulation of cell growth and apoptosis.

The KRAS oncogene is located on chromosome 12p and is frequently mutated in pancreatic adenocarcinoma, with an activating point mutation seen in up to 90% of cases.⁷ This results in a mutated form of KRAS, which is insensitive to GAP, rendering it constitutively active and causing unregulated cell growth via downstream intracellular signaling pathways. Most of the activating point mutations seen in pancreatic carcinomas are located at codon 12, with a minority occurring at codon 13 or 61, thereby making testing for KRAS mutation status relatively straightforward.⁸ Activating KRAS mutations are seen early in the development of pancreatic adenocarcinoma, prior to the progression to invasive carcinoma, and can be detected in up to 87% of noninvasive precursor lesions of pancreatic ductal adenocarcinoma (pancreatic intraepithelial neoplasia [PanIN]). PanIN is graded as 1–3 and demonstrates progression of histologic and molecular changes during the evolution from normal pancreatic ductal epithelium through PanIN 1–3 to invasive carcinoma.⁹ Mouse models of PanIN have shown that KRAS mutation is one of the earliest genetic events occurring in the PanIN to invasive adenocarcinoma sequence. Mutational activation of KRAS also appears to play a key role in the maintenance of invasive pancreatic adenocarcinoma via its effect on cell proliferation, differentiation, and apoptosis through the RAS/MAPK pathway.¹⁰

In primary ovarian tumors, activating KRAS mutation is associated with mucinous and low-grade histology and with low FIGO stage.¹¹ The largest analysis of KRAS mutation status in primary ovarian tumors was reported in 2009, with analysis performed on 403 samples from 398 patients; 381 tumors were malignant (including borderline malignancy) and 22 were benign. Mutation frequency was 50% in mucinous borderline tumors and 60% in mucinous carcinomas. Serous histology was associated with a lower rate of KRAS mutation: 35% in borderline tumors and just 9% in invasive serous carcinomas, in line with previously published reports. The G12D mutation, as identified in the case described above, was the most commonly identified mutation, accounting for 40% of all KRAS mutations identified.

In PDAC, KRAS mutations are most commonly found on codon 12, followed by codon 13. Rare point mutations have also been described on codons 17, 34, and 61. By far the most commonly identified mutation is the G12D mutation, accounting for 52% of the 2,440 KRAS mutations in PDAC recorded in the Cosmic Database of Somatic Mutations in Cancer.¹² This is followed by the G12V mutation, which accounts for 33%, and the G12R mutation, which accounts for 12%. These three mutations together make up over 75% of all KRAS mutations identified in pancreatic cancer specimens to date.

In the current case, the primary pancreatic carcinoma was found to harbor an activating mutation of KRAS in exon 2, codon 12 (G12V). This exact mutation would also be expected to be present in tumors at any site of metastases from this pancreatic carcinoma. The Rapid Autopsy Study performed at Johns Hopkins reported on patients with pancreatic adenocarcinoma who underwent rapid autopsy, 88% of whom had metastatic disease. Activating mutations of KRAS were found in 95% of the 59 tumors analyzed.¹³ The rate of KRAS mutation did not differ significantly between tumors with widespread metastases compared with those that remained localized, in contrast to the DPC4 gene, which was inactivated more frequently in metastatic disease. Importantly, the same exact mutation was found in all sites of metastatic disease as in the primary pancreatic carcinoma.

However, the G12D mutation has been identified as the most common activating mutation of KRAS in both pancreatic adenocarcinoma and in primary ovarian tumors. It is therefore possible, given the narrow spectrum of KRAS mutations identified in both malignancies, that two different malignancies may have the same mutation by chance. Identification of the G12D KRAS mutation in the ovarian specimen therefore in this case would not be sufficient to support a diagnosis of metastatic pancreatic carcinoma.

The patient described in this report had synchronous primary neoplasms of the pancreas and ovary. Several hereditary conditions predispose to the development of both ovarian and pancreatic cancer, including the Hereditary Non Polyposis Colorectal Cancer (HNPCC), Familial Adenomatous Polyposis (FAP), Hereditary Breast and Ovarian Cancer (BRCA 1 or 2 mutation), Peutz Jaeger Syndrome, and L-Fraumeni syndromes.¹⁴ This patient's personal and family history was not suggestive of an inherited genetic predisposition; however, consideration may be given to genetic counseling given her relatively young age at diagnosis of pancreatic cancer and the presence of a second primary ovarian neoplasm.

In the case illustrated above, identification of two distinct KRAS mutations in a primary pancreatic ductal adenocarcinoma and a synchronous ovarian mucinous cystic neoplasm was successfully used to differentiate between a metastatic deposit and a primary ovarian tumor. The use of molecular pathology to differentiate between these possibilities has not previously been reported; for this patient it provided clinically meaningful diagnostic information with important implications for her clinical management and prognosis. Recent advances in molecular techniques have enabled the complete sequencing of the pancreatic cancer genome. In parallel with the identification of common somatic mutations associated with various cancers, the application of molecular pathology techniques to diagnostic challenges such as in this case soon may have an even greater role to play.