Patients with apparently isolated lobular breast carcinoma and poorly cohesive gastric carcinoma with signet ring cells can present a tremendous diagnostic challenge. The tempting explanation for these findings is sporadic, independent, synchronous primaries. Alternative explanations include (1) metastasis from one primary site to the other, classically from the breast to the GI tract for lobular breast cancers or (2) an underlying germline cancer susceptibility allele such as a CDH1 mutation that confers elevated risk for both of these malignancies. Although the most frequent metastatic sites for lobular breast carcinoma are the lymph nodes, liver, lung, and brain,1 several reports document early gastric spread, typically years after breast cancer treatment,2 with or without synchronous colonic3 or rectal4 metastasis. One recent report documented a suspected gastric primary that was found to be metastatic breast cancer after a fortuitous screening mammogram.5 Rare cases even simulate linitis plastica.6
We report a patient in whom synchronous breast and gastric primaries were suspected to result from a germline CDH1 mutation, but which were ultimately explained by metastatic lobular breast carcinoma. To date, no similar accounts have (1) documented a case with a diagnosis-to-metastasis interval of less than 1 year or (2) entertained the incorporation of CDH1 testing as a possible explanation of the clinical findings.
A 75-year-old woman presented with a 1-year history of progressive difficulty swallowing and early satiety, with reported findings consistent with synchronous breast and gastric cancer primary tumors. Nine months previously, at an outside hospital, she underwent mammography that was read as Breast Imaging Reporting and Data System (BI-RADS) 1. Esophagogastroduodenoscopy (EGD) at the time revealed mild gastritis and a hiatal hernia. Routine complete blood count 2 months before presentation demonstrated mild anemia (hematocrit = 31), and repeat EGD 1 month prior noted several umbilicated polyps and nodules within the body and antrum of the stomach. Biopsy was performed at this time, with pathology highly suggestive of poorly cohesive adenocarcinoma with signet ring cells (Fig 1). Positron emission tomography–computed tomography demonstrated heterogenous hypermetabolic activity in the antral wall and a hypermetabolic focus in the medial aspect of the left breast. Subsequent bone scan was unremarkable; however, repeat mammography with ultrasound revealed a 1.7-cm parenchymal lesion at the 9 o'clock position in the left breast, read as BI-RADS 5.
The patient was referred for core breast biopsy (Fig 2), which demonstrated invasive lobular carcinoma, classic type, with the following immunohistochemical profile: cytokeratin (CK)7 positive, BRST2 positive, CK20 negative, E-cadherin negative, estrogen receptor/progesterone receptor (ER/PR) positive, Ki67 positive, human epidermal growth factor receptor 2 (HER2) negative. On transfer to Stanford for her ongoing care, physical examination revealed a well-healed biopsy scar overlying a 3 × 4 cm fixed, nodular mass on the medial aspect of the left breast but was otherwise unremarkable. Past medical and family history were noncontributory, except as detailed above. The patient underwent menarche at age 11, menopause at age 60, used oral contraceptive pills for 20 years, and was gravida 3, para 3. H pylori work-up was negative, and she was referred to a geneticist for CDH1 mutation testing (later found to be negative).
Outside EGD pathology was reviewed and confirmed by a pathologist without specific expertise in GI pathology, and total gastrectomy with Roux-en-Y esophagojejunostomy and D2 lymph node dissection was performed. Pathology from the gastrectomy specimen (Fig 3A), however, demonstrated classic infiltrating lobular carcinoma and foci of poorly differentiated carcinoma with signet ring features, similar to that seen in the preoperative biopsy. A total of 13 of 16 lymph nodes were affected by metastatic carcinoma. Immunohistochemical results from the specimen—CK 7 positive, BRST2 positive(Fig 3B), CK 20 negative, E-cadherin negative, ER/PR positive (Fig 3C), HER2 negative—matched the profile of her lobular breast carcinoma, and the diagnosis was revised to metastatic breast carcinoma. Postoperative recovery was complicated by decreased per os tolerance on day 17 that resolved after a brief inpatient stay with further education around meal planning. The patient is currently on anastrazole treatment.
This case underscores the importance of distinguishing primary gastric from metastatic breast disease, as the treatment modalities are markedly different (surgery in the former and systemic therapy in the latter). Two primary factors confound proper diagnosis in this complicated scenario. First, lobular carcinoma has a rare signet ring cell variant that may mimic a primary poorly cohesive signet ring cell carcinoma of the stomach, thus making early pathologic evidence potentially misleading. In exceedingly rare cases, prominent signet ring cell features may be observed in metastases while being absent in the primary.7 In this case, the initial biopsy demonstrated poorly differentiated cells with significant atypia, a morphologic pattern that is more consistent with a gastric primary than metastatic lobular carcinoma. Second, although GI metastases in patients with lobular breast carcinoma can occur (6% to 18% of the time) and the stomach is the most common site, they typically present long after the initial diagnosis, approximately 7 to 10 years later.8 In a study by McLemore et al,9 the mean diagnosis-to-metastasis interval was 7 years in women with lobular breast carcinoma and subsequent GI spread. In the present case, the diagnosis of breast cancer and metastases to the stomach occurred within the same month. This indicates that a recent (within 9 months) BI-RADS 1 mammogram does not preclude the possibility of developing subsequent metastatic disease, especially as there may be radiographically occult breast tumors.
Although pathology and time course may be misleading, GI signet ring cell adenocarcinoma and invasive lobular carcinoma have distinct, highly predictive immunophenotypes. In a 2000 study by Tibor Tot,10 expression profiles for CK20, CK7, and ER were compared in a series of known poorly cohesive GI (signet ring cell) carcinomas (22 primary and 13 metastases) and invasive lobular carcinoma (79 primaries with metastases). Although CK7 stained positively in the majority of the tumors, CK20 and ER analysis effectively stratified tumor origin. CK20 was positive in all GI primaries, whereas it was only present in 5% of metastatic lobular carcinomas. ER positivity, in contrast, was absent in the GI primaries but present in 91% of breast metastases. On the basis of these data, CK20 and ER analysis are an effective panel to distinguish between these tumor origins (CK20 positive/ER negative for GI primary and CK20 negative/ER positive for breast metastasis).
In patients with potentially synchronous neoplasms, it is also reasonable to pursue possible genetic explanations. CDH1 is a tumor suppressor gene that encodes E-cadherin. Germline mutations in this gene have been implicated in hereditary diffuse gastric cancer (HDGC) syndrome, a clinical syndrome associated with diffuse signet ring cell gastric cancer and lobular breast cancer.11,12 E-cadherin inactivation is important in the pathogenesis of HDGC and lobular breast cancer. Dissociation of the cadherin-catenin complex at the cell membrane is present in most cases of infiltrating lobular breast cancer, but not HDGC.13 However, germline mutations in CDH1 that are diagnostic of HDGC are infrequent in patients with lobular breast cancer.14 Patients with HDGC have a more than 80% lifetime risk of gastric cancers, more than 60% of which have signet ring cells,15 and a 60% lifetime risk of developing lobular breast carcinoma by age 80.16 Although the aggressiveness and age of onset of disease in HDGC are notoriously variable, the International Gastric Cancer Linkage Consortium has arrived at the following criteria for CDH1 testing: (1) two or more familial cases of diffuse gastric cancer (at least one diagnosed before age 50), (2) three or more familial cases of diffuse gastric cancer (any age), (3) a personal diagnosis of diffuse gastric cancer before age 35, (4) a personal diagnosis of both diffuse gastric and lobular breast cancer, or (5) one family member with diffuse gastric cancer and another with either lobular breast cancer or signet ring colon cancer.17 On the basis of criterion 4, the present patient was referred for CDH1 testing that ultimately proved negative. Had immunophenotyping results been positive, this may have proved useful for disease characterization, modified screening protocols, and family counseling.
In patients with apparently synchronous lobular breast and poorly cohesive gastric carcinoma with signet ring features, a high index of suspicion for metastatic breast disease must be maintained, irrespective of early pathology and diagnosis-to-metastasis interval. If there is an appropriate family or individual history, CDH1 testing may play a role, but clinicians must remember that metastatic breast disease may mimic HDGC in unusual cases. Regardless, awaiting appropriate immunohistochemical results, especially for CK20 and ER, will likely clarify the diagnosis and prevent unnecessary gastrectomy.
Authors' Disclosures of Potential Conflicts of Interest
The author(s) indicated no potential conflicts of interest.
MJ Borst, JA Ingold: Metastatic patterns of invasive lobular versus invasive ductal carcinoma of the breast Surgery 114: 637– 641,1993 discussion 641-642
T Tot: The role of cytokeratins 20 and 7 and estrogen receptor analysis in separation of metastatic lobular carcinoma of the breast and metastatic signet ring cell carcinoma of the gastrointestinal tract APMIS 108: 467– 472,2000
R Seevaratnam, N Coburn, R Cardoso, etal: A systematic review of the indications for genetic testing and prophylactic gastrectomy among patients with hereditary diffuse gastric cancer Gastric Cancer 15: S153– S163,2012 suppl 1
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.