Hereditary Diffuse Gastric Cancer: Surveillance Endoscopy is Not Enough to Save Lives

Abstract

Kaurah P, MacMillan A, Boyd, N, et al. (Hereditary Cancer Program and the Center for Translational and Applied Genomics, British Columbia Cancer Agency, Vancouver, British Columbia, Canada). Founder and recurrent CDH1 mutations in families with hereditary diffuse gastric cancer. JAMA 2007;297:2360–2372. Hereditary diffuse gastric cancer (HDGC) is a cancer susceptibility syndrome defined by the early onset of diffuse gastric cancer and lobular breast cancer. The diffuse gastric cancers are characterized by signet ring cell histology, and the tumors are often multifocal and clinically silent until patients present with advanced and incurable disease. The syndrome is caused by a germline mutation in the gene CDH1, which codes for the cell-to-cell adhesion protein, E-cadherin. The stated goal of the study by Kaurah et al was to determine whether or not recurring germline CDH1 mutations occurred as a consequence of independent mutational events or common ancestry. The study also sheds light on the prevalence of germline CDH1 mutations in patients meeting clinical criteria for the syndrome, and provides important data regarding the incidence of gastric and breast cancer in CDH1 mutation carriers, and important data regarding the benefits of prophylactic gastrectomy in asymptomatic gene carriers. In the study, 38 families diagnosed based on clinical criteria with HDGC were analyzed for CDH1 mutations using state-of-the-art methods to increase the sensitivity of mutation detection, and to resolve whether or not missense and splice site alterations were pathogenic or merely benign polymorphisms. In addition, haplotype analysis was undertaken to determine whether the recurring mutations detected in different families were due to common ancestry or chance. The investigators found germline CDH1 mutations in 15 of the 38 families (40%). The mutation detection rate would have exceeded 50% if testing had been limited to families satisfying a more stringent definition of HDGC, in which ≥2 family members had to have diffuse gastric cancers with 1 cancer occurring before 50 years of age. As for the recurring mutations detected in different families, sometimes this was the result of common ancestry, and sometimes it was not. One mutation (2398delC) was found in 4 families from Newfoundland, suggesting a founder effect, one that might account in part for a known increased incidence of gastric cancer in this Canadian province. Penetrance data from this small subset of families revealed the risk for symptomatic gastric cancer, occurring by age 75 years, was 40% in men with the mutant gene (95% confidence interval [CI], 12%–91%), and 63% in women (95% CI, 19%–99%). The earliest occurrence of gastric cancer in the entire series of families was in an 18-year-old. The risk for breast cancer among women was 52% (95% CI, 29%–94%). Predisposition gene testing was carried out in 129 asymptomatic individuals from among all of the families, and 51 of the 129 (40%) were found to be mutation carriers. Among the 18 asymptomatic gene carriers who underwent prophylactic gastrectomy with available pathology results, 12 of 18 had occult cancer, but only a single individual had a diffuse gastric cancer detected by endoscopy with random biopsies before gastrectomy. Gastric adenocarcinoma is the second leading cause of cancer death worldwide, accounting for >10% of the cancer deaths across the globe (World J Gastroenterol 2006;12:354–362). The prevalence varies, with the highest incidence rates being found in Asia, Latin America, and Eastern Europe. Major risk factors for sporadic disease include Helicobacter pylori infection, diets rich in smoked, salted, or pickled foods, and use of tobacco or alcohol (Ann Surg 2005;241:27–39; Curr Oncol Rep 2006;8:183–191). Family history also is an important risk factor, with a 1.5- to 3.0-fold increase in the risk of the disease among first-degree relatives of gastric cancer patients (Int J Cancer 2001;93:148–152; Gastric Cancer 2006;9:9–13). Gastric adenocarcinoma occurs as 1 of 2 major histologic types—intestinal or diffuse. The intestinal type often occurs in the distal stomach and is more common in older men. The diffuse type of cancer is more likely to be proximal, and occurs in young patients and in women (Ann Surg 2005;241:27–39). Both cancer types occur in the setting of a family history, and a small percentage of both, although probably <5% overall, occur as part of a dominantly inherited, highly penetrant cancer susceptibility syndrome. Some of the familial intestinal type gastric adenocarcinomas occur as part of well-described family cancer syndromes, such as Lynch syndrome, familial adenomatous polyposis (FAP), Peutz–Jeghers syndrome, or Li–Fraumeni syndrome, but familial intestinal gastric cancer apart from these known cancer susceptibility syndromes has been described as well, although the genetic cause remains unknown (Int J Surg Pathol 2006;14:21–33; J Clin Pathol May 18, 2007;Epub ahead of print). With respect to diffuse gastric cancer, in 1998 three large Maori families with multiple early onset diffuse gastric cancers were described, and the cancer susceptibility in these families was found to be secondary to inactivating mutations in CDH1, the gene encoding E-cadherin (Nature 1998;392:402–405). E-cadherin is a transmembrane glycoprotein that plays a major role in epithelial architecture, cell adhesion, and cell invasion, and inactivating somatic mutations in E-cadherin are known to be common in sporadic diffuse gastric cancer (Clin Gastroenterol Hepatol 2006;4:416–425). Patients with HDGC inherit a nonfunctioning copy of CDH1 from a parent (autosomal-dominant transmission), but occasionally the germline mutation may be spontaneous. When the second allele of the gene is inactivated by a somatic event in a gastric epithelial cell, neoplastic transformation can occur (Clin Gastroenterol Hepatol 2006;4:262–275). In a process analogous to what occurs in the colon of patients with FAP, who have an inherited inactivating germline mutation in the APC gene, those with a mutant CDH1 gene develop numerous early gastric neoplasias, with ≥1 of these lesions progressing to invasive, and ultimately advanced, cancer in a high percentage of patients. The histology of these early lesions is characterized by tiny clusters of transformed signet ring cells that often underlie normal mucosa. Most patients, however, do not become symptomatic until they develop advanced, metastatic linnitis plastica, which is incurable (Clin Gastroenterol Hepatol 2006;4:262–275). HDGC is rare, but may account for up to 1% of all gastric cancers (Int J Surg Pathol 2006;14:21–33). The report by Kaurah et al on HDGC is important for several reasons (JAMA 2007;297:2360–2372). It nearly doubles the number of families with HDGC previously reported in the medical literature, and therefore serves to confirm and refine several important aspects regarding the genetic and clinical features of HDGC (JAMA 2007;297:2410–2411). The majority of cases of HDGC, regardless of ethnicity, are caused by germline CDH1 mutations. In Western populations, when genetic testing is limited to families with ≥2 cases of diffuse gastric cancer, 1 of which occurred before the age of 50 years, CDH1 mutations can be detected in >50% (Clin Gastroenterol Hepatol 2006;4:262–275). The authors suggest that genetic testing be undertaken when this clinical situation occurs in low gastric cancer incidence populations, but caution that in populations in which gastric cancer is very common, such as Asia, this criterion may be too permissive (JAMA 2007;297:2360–2372). Other cases of HDGC may be secondary to undetectable germline CDH1 mutations or epimutations, or are due to mutations in other genes. These recommendations for genetic testing are sound, and CDH1 gene testing is commercially available. Genetic testing is of enormous value in CDH1 families because it will allow discrimination among nonaffected, asymptomatic relatives as to who requires extraordinary cancer prevention measures (see below) and who does not. As described, the rate of diffuse gastric cancer in CDH1 gene carriers is very high. In the study by Kaurah et al, the lifetime risk in men was 40% and 63% in women (JAMA 2007;297:2360–2372). These numbers are slightly lower than previously reported (Clin Gastroenterol Hepatol 2006;4:262–275), but the penetrance is high nonetheless. The reason for the different incidence rates in men and women remains a mystery. As in previous reports, the gastric cancers can occur at very young ages, as young as 18 years in the present study, although most occurred after 30 years of age. The rate of breast cancer in women with a CDH1 mutation also was very high, exceeding 50% in this study, with the earliest occurrence in a 38-year-old woman. These extraordinary rates of cancer in CDH1 gene carriers demand effective cancer prevention strategies, and this remains the greatest problem in the management of HDGC families (Clin Gastroenterol Hepatol 2006;4:262–275). In the current study, 23 of 51 asymptomatic CDH1 gene carriers elected to have surgery before publication of the study. Pathology reports were available for 18, and remarkably in 12, occult gastric cancer was found. Only a single patient in the current study was found to have an occult cancer during an upper endoscopy with biopsy performed before the stomach was removed (JAMA 2007;297:2360–2372). Although chromoendoscopy-directed biopsies have been recommended as a surveillance method in HDGC, the literature to date does not substantiate the efficacy of any noninvasive, or minimally invasive, test to detect early and curable diffuse gastric cancer (Clin Gastroenterol Hepatol 2006;4:262–275). In a recent report of 6 CDH1 gene carriers from a single family, comprehensive preoperative surveillance by upper endoscopy with random biopsies, high-magnification endoscopy with biopsy, endoscopic ultrasonography, computed tomography, and positron emission scanning, failed to detect any sign of gastric cancer, yet all 6 had early stage (T1) multifocal invasive diffuse signet ring cell type gastric cancers found on histologic examination of their stomachs after resection (Ann Surg 2007;245:873–879)! The situation may be similarly grim with respect to breast cancer prevention. Lobular breast cancer, which also has a diffuse growth pattern, can be very hard to detect with mammography. Whether surveillance methods such as ultrasound or magnetic resonance imaging (MRI) are good enough, or whether or not women with CDH1 mutations should be counseled to undergo prophylactic bilateral mastectomies, remains to be seen (Clin Gastroenterol Hepatol 2006;4:262–275). The data from the study by Kaurah et al, added to previous reports, leads one to the inescapable conclusion that the only reasonable recommendation for gastric cancer prevention in an HDGC gene carrier is prophylactic gastrectomy. Improved endoscopic imaging may obviate the need for this recommendation at some point, but not yet. The time to intervene is uncertain, and it also remains uncertain if the recommendation should differ between men and women. Although cases of gastric cancer have occurred in teenagers, the data overall suggest that the risk of gastric cancer death exceeds the risk of mortality from the operation once individuals reach their mid to late 20s (Clin Gastroenterol Hepatol 2006;4:262–275). As for breast cancer, surveillance should commence at approximately age 30, and should include yearly ultrasound or possibly MRI. Prophylactic bilateral mastectomy should be strongly considered, and may soon prove to be a more prudent strategy than any of the available surveillance methods. The extreme nature of the cancer prevention strategies required to save lives in HDGC families further reinforces the critical role of genetic testing. Unfortunately, only 50% or so of HDGC families test positive for a CDH1 mutation, and therefore, predisposition gene testing is not possible in many families with a noninformative gene test in a cancer-affected individual. Management of these mutation-negative families is an enormous problem. It is not easy to imagine sending an asymptomatic, endoscopy-negative patient for prophylactic gastrectomy, without first being able to confirm the presence of a CDH1 mutation. The final lesson of the study by Kaurah et al is that, although HDGC may be rare, in isolated populations it may be more common than expected (JAMA 2007;297:2360–2372; JAMA 2007;297:2410–2411). The authors’ elegant investigation demonstrated that some recurrent mutations in different families are the result of chance and some are the result of common ancestry. For example, the authors detected a founder mutation in the Newfoundland population that may account for the increased incidence of gastric cancer in that province compared with the rest of Canada. The founder mutation would lead to the occurrence of many more cases of HDGC in Newfoundland than expected. This same situation may occur in other discrete populations around the globe, rendering HDGC not simply a curiosity for physicians working in these areas, but a public health problem of importance.