The first clear evidence for a gastric cancer susceptibility genetic locus was the identification in 1998 of germline inactivating (truncating) mutations in the gene encoding E-cadherin, called CDH1, in 3 large Maori families from New Zealand with diffuse early-onset gastric cancer.¹⁰ The pattern of inheritance was consistent with an autosomal dominant susceptibility gene with high penetrance. Hereditary diffuse gastric cancer (HDGC) caused by germline mutations in CDH1 has subsequently been shown to occur in a number of families with different ethnic backgrounds, including European, African American, Pakistani, Japanese, Korean, and others.^11–17 CDH1 is localized on chromosome 16q22.1¹⁴ and encodes for a protein that comprises 5 extracellular cadherin repeats, a transmembrane region, and a highly conserved cytoplasmic tail. It functions as a calcium-dependent cell-cell adhesion glycoprotein. It is hypothesized that CDH1 acts as a tumor suppressor gene, and loss of function through inactivating mutations or promoter methylation leads to development and progression of cancer by lack of inhibition of cell adhesion.^5,18,19 Its inhibition increases cell proliferation, invasion, and metastases. Besides HDGC, CDH1 mutations are associated with an increased incidence of lobular carcinoma of the breast and, possibly, colorectal carcinoma and prostate cancer.^13,20,21 Missense mutations of CDH1 (unlike truncating ones) are not clearly associated with these hereditary cancer risks.²² The lifetime cancer-specific risks associated with germline CDH1 mutations are thought to be >70% for gastric cancer and up to 40% for lobular breast cancer in women.¹⁷

Each patient underwent total gastrectomy with d-2 lymph node dissection and Roux-en-Y esophagojejunostomy without complication or mortality. Each was done as an open procedure with postoperative epidural pain medication. No patient received a blood transfusion. The average length of postoperative stay was 7 days (range, 6–10 days). Each patient had a normal swallowing evaluation of the esophagojejunostomy on postoperative day 5. Each patient was discharged on a 6 small feed postgastrectomy diet. The pathologic analysis of each surgical specimen showed multifocal invasive diffuse (signet-ring cell) adenocarcinoma of the stomach that was limited to the mucosa (T1) (Fig. 3). In situ carcinoma with signet-ring cell histology was also identified in 2 patients. The non-neoplastic mucosa exhibited mild foveolar hyperplasia and tufting, as previously described. No intestinal metaplasia was detected.¹⁹ Immunohistochemistry for E-cadherin protein demonstrated that signet-ring cell cancers did not have functional E-cadherin protein, while normal intact glands did (Fig. 4). No patient had any evidence of lymph node metastases despite the fact that the average number of lymph nodes examined on each specimen was 24 and the range was 18 to 40. No patient received any other treatment (radiation or chemotherapy).

All patients have remained free of tumor at 1-year follow-up. Each is able to eat normally with more frequent meals and has occasional intermittent episodes of fullness and some discomfort. These episodes appear to be related to amount of food consumed. Each patient has lost weight following the surgery, with a mean weight loss of 40 pounds (range, 20–60 pounds). Each patient maintains normal activities and has returned to work. Each patient states that he or she would undergo the operation again.

A significant issue with HDGC, that limits screening methods to diagnose it early, is its diffuse multifocal infiltrative biology.^9,22,29,30 It is usually not associated with mass lesions or ulcers. In fact, patients with the diffuse signet-ring cell type of gastric cancer commonly present with linitus plastica or diffuse involvement of the entire stomach.^31–33 Recently, other investigators have recommended the use of PET scan or chromoendoscopy directed biopsies in at-risk individuals to screen for early detection of cancer.^34,35 Van Kouwen et al reported a single case in which a documented CDH1 carrier from a HDGC family had 2 abnormal foci identified by FDG-PET in the proximal and pyloric regions of the stomach, and following gastrectomy, signet-ring carcinoma was found in these and other areas.²⁶ Shaw et al have performed chromoendoscopy with both methylene blue and Congo red in 33 CDH1 carriers, and biopsied 56 abnormal appearing lesions. Twenty-three of these showed signet-ring cell carcinoma (10 patients), and subsequent prophylactic gastrectomies from some of these patients demonstrated that only lesions 4 to 10 mm were detected endoscopically, but not those <4mm.³⁵ Our data suggest that these screening procedures have low sensitivity for detecting early gastric cancer in patients with HDGC. In our series of 6 patients, PET scan, chromoendoscopy, and endoscopic ultrasound were negative in every patient despite the fact that each had invasive cancer. Indeed, the stomach and lymph nodes were grossly normal in appearance, and palpation in every patient suggesting that preoperative diagnostic and imaging studies will often fail to detect early cancers.

Another reason for recommending other screening procedures rather than total gastrectomy is the less than complete penetrance of gastric cancer in patients with CDH1 mutations, with most studies suggesting that 3 of 4 carriers of the CDH1 mutation will develop gastric cancer.^2,23,24 Indeed, in a previous report, 1 patient with the CDH1 mutation underwent total gastrectomy and was found not to have cancer.¹ However, in the current study, each patient with the mutation had multifocal invasive diffuse gastric cancer with pure signet-ring cell histology, and in our experience at Stanford with 2 additional HDGC families, 6 other CDH1 carriers undergoing prophylactic gastrectomy have had identified signet-ring cell carcinoma. Nevertheless, for those choosing not to pursue a prophylactic gastrectomy at this time, we do recommend careful screening. Our current approach for individuals in known HDGC families is to recommend genetic counseling and testing for CDH1 carriage by the early 20s, and biannual chromoendoscopy and biopsies by the age of 25 or at least 10 years earlier than the youngest family member affected by cancer. Furthermore, for women with CDH1 mutations, we recommend a similar approach to screening as for other hereditary breast cancer syndromes, including annual mammography and breast MRI starting at the age of 25 years.²

In this series of 6 consecutive patients, multifocal invasive diffuse gastric cancer with pure signet-ring cell histology was present in every instance and none had lymph node metastases. Therefore, this was not a prophylactic gastrectomy but a therapeutic one. The expected long-term 5-year survival of patients with presumably sporadic T1, N0, M0 adenocarcinoma of the stomach is 78%.³⁶ In each of the patients diagnosed by the CDH1 mutation, the gastric cancer was stage 1a. This means that the cancer has invaded under the epithelium into the mucosa, but not into the submucosa or other tissues including lymph nodes. This pathology result is favorable, but approximately 20% of these patients may be at risk for recurrence. We are following these patients for gastric cancer recurrence with annual PET/CT, physical examination, routine laboratory, and tumor markers. However, although we have chosen to follow with PET/CT, it is not clearly indicated, but rather our preference. These patients are also undergoing upper gastrointestinal endoscopy every 2 years. We think that the prognosis of these patients may be better than predicted for sporadic gastric cancer, and we have not chosen to add adjuvant chemotherapy or radiation despite the proven benefit for stage II to IV gastric cancer patients.³⁷ Another potential drawback for “prophylactic” surgery for individuals with CDH1 mutations is the morbidity and mortality of total gastrectomy.^4,38,39 Certainly, the current study argues in favor of surgery as the immediate short-term morbidity and mortality was excellent in that no patient had any complications. The short hospital stay, no blood transfusion, large number of lymph nodes, and lack of morbidity make the current approach the method of choice. The expected death-rate with total gastrectomy is as high as 2% to 4%.⁴⁰ In a prophylactic gene-directed operation, death is clearly an unacceptable outcome.⁴¹ The long-term complications of total gastrectomy include diarrhea, dumping, and weight loss. The expected incidence is nearly 100%.² Each patient in this study developed weight loss, and each reports intermittent eating difficulties. This will require additional monitoring before weare sure of the long-term adverse effects of total gastrectomy in this population. At present, no patient has developed recurrent gastric cancer or other tumors, and each appears to be functioning reasonably well as all have returned to a productive life.

We chose to perform an open total gastrectomy with d-2 lymph node dissection as part of this protocol. We did it with no morbidity and mortality, but only 6 patients were treated. Since no patients had lymph node metastases and the morbidity of d-2 lymph node dissection has been shown in previous studies⁴² to be greater than routine gastrectomy, we do not recommend a d-2 lymph node dissection for these patients. We recommend removing the nodes along the left gastric artery and the greater omentum. Further, recent evidence from surgery for morbid obesity extrapolated to this condition supports the attempted use of laparoscopic total gastrectomy in patients with CDH1 mutations. Clearly, the entire stomach must be removed in every case as each patient had multiple diffuse sites of gastric cancer. However, if the stomach can be completely removed laparoscopically with similar low morbidity and mortality, it may improve the recovery and significantly reduce the pain and time to normal activity of the procedure.