Hereditary diffuse gastric cancer: the evolution of a cancer syndrome

ABSTRACT

Hereditary Diffuse Gastric Cancer (HDGC) is a high risk cancer syndrome caused predominantly by germline mutations in the CDH1 gene. HDGC is characterised by a lifetime risk of advanced diffuse-type gastric (stomach) cancer of up to 70%, and an additional 40% lifetime risk of lobular breast cancer in women. Since the first description of HDGC in three whānau Māori in 1998, our understanding of this syndrome’s life history and clinical behaviour has steadily evolved, leading to changes to its clinical management. In particular, it is now evident that the signet ring cell carcinomas that develop in the stomachs of CDH1 pathogenic variant carriers have an indolent phase, although the factors that drive progression of these early cancers to advanced disease remain to be identified. This indolent phase provides the opportunity for chemoprevention to be considered as an alternative to prophylactic surgery as a risk reduction strategy. Here, we describe the evolution of our knowledge of HDGC, with particular reference to the syndrome’s penetrance, tumour spectrum and pathology.

Introduction

Approximately 5%–10% of all cancers can be attributed to hereditary cancer syndromes (Nagy et al. 2004; Huang et al. 2018). These syndromes are characterised by a high frequency of specific cancers in families, often over multiple generations, and an early age of cancer onset. They can be caused by germline mutations in one or more of over 100 cancer predisposition genes, with most mutations elevating individual risk 2-4 fold (Mighton and Lerner-Ellis 2022). However, about one quarter of the known cancer predisposition genes increase risk substantially higher, resulting in a clearly discernible inheritance pattern through families.

Such high penetrance syndromes are typified by Lynch syndrome, a relatively common inherited disorder that raises the risk of several cancers, in particular colorectal and endometrial cancers (Lynch et al. 2015). Although the first gene predisposing to Lynch syndrome wasn’t identified until 1993 (Fishel et al. 1993; Peltomäki et al. 1993), a family later shown to be affected by this syndrome was described in the literature in 1913 by the pathologist Aldred Warthin (Warthin 1913). The plight of that family was made clear by a young woman from the family, Pauline Gross, when she said ‘I’m healthy now, but I fully expect to die an early death from cancer … . Most of my relatives are sick and many in my family have already passed on’. Indeed, Pauline died from endometrial cancer at the age of 46yrs (Campos et al. 2024).

The existence of an inherited form of stomach cancer (also known as gastric cancer) has also long been accepted, thanks in large part to Napoleon Bonaparte’s family’s medical history. This history shows an apparently dominant inheritance pattern of early onset stomach cancer, and Napoleon himself was quoted as saying, just prior to his death from this cancer (Lugli et al. 2021a, 2021b), ‘Le pylore, mon père’, referring to his father Charles’ death from the same disease (Figure 1).

Figure 1.

Napoleon Bonaparte’s family tree, showing the elevated incidence of early onset stomach cancer.

Although not widely noted at the time internationally, a 1964 New Zealand Medical Journal publication describing the incidence of stomach cancer in a large whānau Māori (Family A) from the Bay of Plenty in Aotearoa New Zealand provided the strongest evidence for the presence of a ‘stomach cancer gene’. This whānau had lived with this cancer for generations, but it only came to the attention of the medical literature after a young University of Otago medical graduate working at Tauranga Hospital, Dr Ted Jones, noticed that a 21 yr old male and a 24 yr old female, who had both died of diffuse-type gastric (stomach) cancer within one year of each other, were first cousins (Jones 1964). His publication outlined the extended cancer history in this whānau, and noted its concentration in particular family lines (Figure 2). Unfortunately, at that time it was not possible to identify genes for the great majority of human diseases. As a result, there was little that could be done for this family for yet another generation.

Figure 2.

Family A’s family tree, as published in the New Zealand Medical Journal in 1964 (Jones 1964).

In 1994, the affected whānau, supported by gastroenterologists including Dr Robin Scoular at Tauranga Hospital, formed a unique partnership with the Cancer Genetics Laboratory, University of Otago, to identify the causative gene. Using the first human genome-wide genetic map (developed by the French genetic research consortium, Généthon) and a candidate gene approach, the partnership was, in time, able to identify CDH1 as the gene predisposing to DGC in this whānau, leading to the first description of the inherited stomach cancer syndrome Hereditary Diffuse Gastric Cancer [HDGC] (Guilford et al. 1998; Guilford et al. 1999).

In this review, we will describe how our understanding of HDGC has evolved as hundreds of new families affected by this syndrome have been identified worldwide, and the syndrome’s underlying pathology elucidated. This evolution is evident across all aspects of HDGC, including our understanding of its high frequency in the Māori population, its tumour spectrum, penetrance, and preferred clinical management strategies, leading to initial perceptions being replaced by a more nuanced and complex picture of the disease.

Initial perceptions of HDGC

Prior to the identification of the underlying genetic cause in 1997 (Guilford et al. 1998), Family A’s medical history indicated the presence of a highly penetrant, dominantly inherited cancer syndrome characterised by early onset, diffuse-type gastric cancer (DGC), with little evidence of other associated cancers (Figure 3). This apparently narrow tumour spectrum could be explained by two possibilities: (i) the predisposing mutation affected a cancer pathway that was of particular importance to stomach cells, or (ii) the mutation also increased the risk of other cancers, but these cancers were slower to develop. In this second scenario, carriers would die from DGC before other cancers had time to develop. A necessary consequence of the risk of multiple cancers would be the complexity of clinical management in carriers of the predisposing variant(s). A second striking observation prior to 1997 was the presence of two other large whānau Māori (Families B and C) with the same DGC presentation as Family A, suggesting the likelihood of a founder effect in the Māori population.

Figure 3.

The incidence of gastric (stomach), colorectal and breast cancer in Family A in 1998. Square symbols, males; Circular symbols, females. First published in Guilford et al (1998).

Tumour spectrum

The first pathogenic germline CDH1 variant, c.1008G > T, was identified in Family A by direct DNA sequencing after genetic linkage analysis had indicated the high statistical probability that the CDH1 locus was involved (Guilford et al. 1998). CDH1 encodes E-cadherin, a cell–cell adhesion protein expressed in all epithelial tissues (van Roy and Berx 2008). Related to its cell adhesion function, E-cadherin is also an important determinant of cell polarity, cell migration, tension sensing and tissue architecture (Mendonsa et al. 2018).

E-cadherin’s broad expression pattern reinforced concerns that carriers of pathogenic CDH1 variants were at elevated risk of multiple carcinomas, not just stomach cancer, and that interventions that prevented an early death from stomach cancer would expose these additional predispositions. That is, carriers might survive one cancer just to later die of another. Fortunately, 30 years later, this worst case scenario has not eventuated. We now know that, other than DGC, HDGC is only significantly associated with one other cancer, lobular breast cancer (LBC). There is currently no evidence for other common epithelial cancers such as pancreatic, colorectal or prostate cancer occurring at rates significantly higher than expected in the general population (Decourtye-Espiard and Guilford 2023; Garcia-Pelaez et al. 2023). However, it cannot be excluded that CDH1 germline mutations can accelerate the progression of sporadic (i.e. non-inherited) cancers by contributing to a more invasive phenotype. As more data is gathered on HDGC families, it will be important to determine if there has been a shift in the age of incidence, or histological subtypes, of sporadic cancers in these families.

Notably, LBC was not obvious in the first three Māori HDGC families prior to any clinical interventions. This initial absence may be explained by protective reproductive lifestyle factors in these whānau, such as age of first birth (Kotsopoulos et al. 2010). However, LBC is becoming more evident in these families today (unpublished observations), perhaps because female carriers of pathogenic CDH1 variants are now living longer lives, thanks to the successful clinical management of the gastric cancer risk in these whānau.

The lifetime penetrance of LBC in female pathogenic CDH1 variant carriers worldwide is now estimated at 37%–42% (Hansford et al. 2015; Xicola et al. 2019; Ryan et al. 2024). This high penetrance has led to the inclusion of CDH1 on commercial breast cancer genetic testing panels, resulting in the identification of families with pathogenic CDH1 variants that are predominantly affected by LBC rather than DGC. Further, the high LBC penetrance has led reference organisations including Online Mendelian Inheritance in Man (OMIM) and ERN GENTURIS to rename HDGC ‘Diffuse Gastric and Lobular Breast Cancer Syndrome’.

DGC penetrance

The first estimate of lifetime penetrance of DGC in the three affected whānau Māori was 70% (Guilford et al. 1998). In 2001, a combined analysis of these three families with eight further North American/European HDGC families estimated an 83% DGC risk in females and 67% risk in males (Pharoah et al. 2001). A larger 2015 study examined the penetrance in 75 families from around the world meeting specific clinical criteria based on the number of gastric cancer cases in first and second degree relatives (Hansford et al. 2015). That study saw a reduction in the female lifetime risk estimate to 56%, but the male risk steady at 70%. However, in 2019, two US-based studies that drew heavily on families identified through commercial breast cancer gene panel testing, with little reference to HDGC clinical criteria, found the lifetime DGC penetrance in their cohorts to be (female-male) 33%–42% (Roberts et al. 2019) and 25%–37% (Xicola et al. 2019). Finally, in 2024, the overall penetrance in a cohort of 213 CDH1 families who were identified by both clinical criteria and breast cancer panel testing, and were undergoing clinical management for HDGC risk at the NIH (Bethesda, USA), was reported to be 7%–10% (Ryan et al. 2024). Together, these studies demonstrate the effect of ascertainment biases, with earlier estimates driven by the identification of families meeting specific clinical criteria related to DGC incidence, and the more recent US studies equally biased by the impact of breast cancer gene panel testing. Care will be required to prevent the assumption that any particular study is exclusively correct, and others are under– or over-estimates. Instead, these different studies simply highlight the variability of DGC risk in families with pathogenic germline CDH1 variants, further complicating HDGC’s clinical management. The underlying reasons for this variability are not yet known, but are unrelated to the location of the mutation with the CDH1 gene (Garcia-Pelaez et al. 2023). Most probably, the observed variability reflects the existence of risk modifier genes, as observed for other cancer syndromes (Campos et al. 2024). Possible HDGC modifier genes include genes involved in immune surveillance and inflammation, APOBEC variants, and different members of the cadherin family. Segregation of unknown HDGC modifier genes in and out of families highlights the need for caution when trying to estimate an individual CDH1 variant carrier’s DGC risk. This lack of predictability would be particularly pronounced if few genes modify CDH1-driven DGC risk, but less remarkable if it transpires that large numbers of independent variants influence the baseline risk associated with a CDH1 mutation. In addition to genetic modifiers, there also may be unknown environmental triggers that differ between, or within, families. For example, H. pylori infection has been shown to increase the risk associated with pathogenic variants in homologous recombination genes, although there is still no compelling evidence that this bacteria increases HDGC penetrance (Usui et al. 2023).

High incidence in Māori

Direct DNA sequencing demonstrated that the two other HDGC whānau Māori that were known prior to 1998 had different pathogenic CDH1 variants (c.2386dupC and 2095C > T) to that found in Family A (c.1008G > T), excluding the possibility of a single founder in the Māori population (Guilford et al. 1998). We are now aware of 14 different pathogenic germline CDH1 variants in whānau Māori with HDGC. Many of these variants can be traced back to the mid nineteenth century, affecting large branches of several iwi. Despite the absence of a single founder mutation, germline CDH1 mutations appear to be enriched in the Māori population, accounting for 14/22 (64%) of the different New Zealand CDH1 mutations we are aware of, despite Māori comprising only 17% of the total population (www.stats.govt.nz). Based on our knowledge of the affected New Zealand families, we estimate that there are 650–700 CDH1 pathogenic variant carriers in the country, with 500 being Māori.

Rather than being caused by a major founder effect, other possibilities may explain the high incidence of HDGC in Māori. Firstly, it may represent an ascertainment bias caused by the high visibility of the very large affected Māori kindred, an effect accentuated by the strong awareness amongst many whānau and hapu of their kindred relationships. Alternatively, it may indicate the presence of an evolutionary advantage for CDH1 variants, as observed for other disease-causing genes (Esoh and Wonkam 2021). For example, the protein encoded by CDH1, E-cadherin, is the cell surface receptor for the bacteria Listeria monocytogenes (Mengaud et al. 1996). Listeria causes listeriosis, a form of food poisoning associated with a high rate of miscarriage in pregnant women (Kraus et al. 2024). Hypothetically, the loss of one CDH1 allele may reduce the severity of listeria infection, leading to positive selection through evolution (da Silva Tatley et al. 2003). Alternatively, E-cadherin loss disrupts voltage-gated ion channel function (Godwin et al. 2019). Hypothetically, this effect could reduce the impact of neurotoxins such as the ciguatera toxin produced by the dinoflaggellate Gambierdiscus toxicus. Episodes of high ciguatera concentration in reef fish species are relatively frequent in coral reef ecosystems in the Pacific, affecting the reliability of fish food sources. It has even been suggested that these episodes were important drivers of the waves of migration across Polynesia between AD1000 and AD1450 (Rongo et al. 2009).

Identification of carriers

A priority for HDGC researchers has been to identify previously unknown HDGC families in order to clinically manage their risks. Most effort to find such families has focussed on defining clinical criteria that would trigger CDH1 genetic testing. For example, the original International Gastric Cancer Linkage Consortium (IGCLC) criteria published in 1999 proposed genetic testing for families meeting either of the following criteria (Caldas et al. 1999): (i) Two or more documented cases of DGC in first or second degree relatives, with at least one diagnosed before the age of 50, or (ii) three or more cases of documented DGC in first/second degree relatives, independent of age of onset. Subsequent updates to the IGCLC guidelines have progressively widened the testing criteria, balancing the likelihood of identifying new families with the financial cost, genetic counselling workforce implications, and the emotional cost of unnecessary genetic testing (Fitzgerald et al. 2010; van der Post et al. 2015; Blair et al. 2020). LBC incidence in 1st and 2nd degree relatives has increasingly been included in proposed criteria (Garcia-Pelaez et al. 2023). Notably, the 2020 IGCLC testing criteria included a specific criterion promoting genetic testing in people of Māori ethnicity with confirmed DGC.

By linking genetic testing to predominantly DGC-based clinical criteria, the search for new families identified, by definition, HDGC families with high DGC penetrance. However, the more recent inclusion of CDH1 on familial cancer genetic testing panels, the increasing uptake of direct-to-consumer genetic testing, lower cost, and the increasing public familiarity with genetic risk, have seen the net widened further. As noted earlier, this has led the identification of HDGC families with greater variability in disease penetrance.

Mutation spectrum

Over 680 pathogenic/likely pathogenic germline CDH1 variants have now been reported in ClinVar (ClinVar 2024). These variants are all truncating or splice site mutations, and spread across the gene with no significant hotspots, consistent with CDH1’s classification as a tumour suppressor gene. Greater than 2000 CDH1 missense variants have also been identified, however, the vast majority of these are benign changes. The few that are associated with the HDGC phenotype appear to increase cancer risk via impact on exon/intron splice sites rather the amino acid substitution itself. One possible exception is the c.2273A > G (p.Glu758Gly) variant identified in a family meeting the 2020 HDGC clinical criteria for genetic testing (P. Guilford, unpublished). This variant directly disrupts the interaction between E-cadherin, P120 catenin and the mitotic spindle protein NuMA (Monster et al. 2022). Although the majority of the reported mutations in HDGC are substitutions and small insertions/deletions, large deletions encompassing all or part of the CDH1 gene have also been described (Oliveira et al. 2009; Sugimoto et al. 2014; Yamada et al. 2014; Feroce et al. 2017).

Pathogenic germline CDH1 variants have been reported widely around the world, although the incidence is highly variable from country to country (Corso et al. 2023). Notably, the reported frequency of pathogenic CDH1 variants in Asia is relatively low, an observation that is in contrast to the high incidence of sporadic gastric cancer in that region (Kim et al. 2013; Liu et al. 2022). This low rate suggests the prevalence of protective variants in as yet unknown risk modifier genes in the Asian population, although the high frequency of sporadic cases is likely to have hindered the detection of families with genuine inherited risk.

Although HDGC is predominantly attributed to pathogenic/likely pathogenic variants in the CDH1 gene, pathogenic variants in one other gene, CTNNA1, have now also been found (Majewski et al. 2013; Lobo et al. 2021). Although few CTNNA1-related HDGC families have been published to date, the phenotype appears less severe than CDH1-related HDGC, with a probable lower lifetime penetrance of DGC (Coudert et al. 2022). CTNNA1, encodes α-catenin, a 100KDa protein that forms part of a complex at the adherens junction that includes E-cadherin, β-catenin and p120 catenin. This complex effectively links the F-actin cytoskeletons of adjacent epithelial cells (Decourtye-Espiard and Guilford 2023). Together, mutations in CDH1 and CTNNA1 suggest that HDGC can be functionally defined as a defect in the epithelial adherens junctions.

There are currently no other known bone fide HDGC genes, although a South Korean family with a strong history of DGC was recently identified with a R129W RhOA mutation (Oh et al. 2024). Although yet to be confirmed in other families, RHOA is promising candidate HDGC gene due to its frequent mutation in sporadic DGC (Wang et al. 2014; Schaefer et al. 2023) and a function that is closely associated with adherens junction and cytoskeleton function (Nam et al. 2019). However, for specific mutations in this gene to predispose to HDGC, it would likely be necessary to show that those mutations are phenotypically silent during embryonic development and therefore not embryonically lethal.

Genome-wide sequencing will be central to the identification of new HDGC genes and/or previously unrecognised pathogenic non-coding variants in the regulatory regions of CDH1 and CTNNA1. Demonstrating the causality of novel genetic variants identified in small affected families will, however, be challenging until functional assays can be developed that directly test the ability of a variant to initiate DGC or LBC.

Some gastric cancer families carry variants in well-established cancer susceptibility genes associated with DNA damage repair, including PALB2, MSH2, ATR, ATM, RAD51C, and BRCA1/2 (Hansford et al. 2015; Sahasrabudhe et al. 2017; Fewings et al. 2018; Momozawa et al. 2022). Although mutations in these genes lead to a general increase in cancer risk without enrichment for DGC and/or LBC, detecting their presence in gastric cancer families without identifiable CDH1/CTNNA1 pathogenic variants remains valuable for the clinical management of such families.

Tumour pathology and the natural history of HDGC

HDGC is characterised by multifocal stage pT1a signet ring cell carcinomas (SRCC) that are identified in the gastric mucosa of >95% of CDH1 pathogenic variant carriers (Rocha et al. 2018), and which can develop from a young age (Falgout and Gensler 2023). In some carriers, several hundred independent foci have been identified, with most <1 mm in diameter (Charlton et al. 2004). However, the number of SRCC foci observed in gastrectomy specimens varies greatly between individuals (Blair et al. 2020). These SRCC foci are initiated by the somatic loss of the 2^nd CDH1 allele by mutation, or more frequently, DNA promoter hypermethylation (Humar et al. 2009). The importance of epigenetic inactivation of the 2nd allele is illustrated by the enrichment of foci in the transition zone between the body and antrum of the stomach in some carriers with large foci numbers (Charlton et al. 2004). Notably, stage pT1a gastric SRCC foci are also common in the gastrectomy specimens of CDH1 carriers who have a personal or family history of LBC but not DGC (Gamble et al. 2022) suggesting that the absence of advanced DGC in these families is due to the failure of early stage gastric cancer to progress to late stage disease, perhaps due to enhanced immune surveillance.

Sporadic stage pT1 gastric SRCCs are generally treated aggressively (Sreeram et al. 2024). However, it is now clear that most HDGC pT1a SRCC are relatively indolent. Many may even be transient, as suggested by the observation that the number of stage pT1a foci in individual mutation carriers is not correlated with increasing age (Figure 4). However, a subset of foci progress to a more aggressive state that is characterised by a pleomorphic, poorly differentiated appearance, KI67 positivity and abnormal accumulation of TP53 (Gullo et al. 2021; van der Post et al. 2023). These cells are highly proliferative and invade diffusely through the stomach muscle layers, leading in time to Linitis plastica and metastatic disease that frequently involves peritoneal dissemination. It remains unclear what drives this switch to an aggressive phenotype. However, it has been associated with a partial epithelial–mesenchymal transition (Humar et al. 2007). These changes may be initiated by paracrine signalling from neighbouring stromal cells, the acquisition of oncogenic driver mutations or changes in immune surveillance. SRCC lesion behaviour may also be influenced by the cell of origin, with lesions originating from immortal tissue stem cells more likely to persist than those beginning with a progenitor cell.

Figure 4.

The number of foci of stage pT1a signet ring cell carcinoma in the gastrectomy specimens of members of Family A (mutation CDH1 c.1008G > T). Age (years) refers to the age of the carrier at the time of surgery.

Once the 2^nd CDH1 copy has been inactivated, the loss of E-cadherin reduces cohesion between neighbouring cells and disrupts the orientation of the mitotic spindle of dividing cells (Monster et al. 2022). This lost orientation results in some cells dividing out of the epithelial plane, either into the lumen where they are predicted to form in situ carcinomas, or through the basement membrane into the Lamina propria, leading to the initiation of a pT1a SRCC lesion (Decourtye-Espiard and Guilford 2023). This life history is important as it explains the initial indolence of the gastric SRCC foci – they are simply normal, non-transformed cells in the wrong place.

Clinical management

The clinical management of HDGC has become more nuanced since 1998, although the fundamentals remain the same. In New Zealand families with a known CDH1 pathogenic variant, genetic testing is recommended from the age of consent (16 years), although it is occasionally carried out 2–3 years younger, on a case by case basis. This minimum age is standard practice in genetic counselling for conditions that are considered to be adult onset, and for which little can be done prior to that age to modify individual risk. However, Māori HDGC whānau have frequently expressed their dissatisfaction with this minimum age, pointing out that Māori were never consulted on this policy and it is in conflict with the principle of tino rangitiratanga. Work is currently underway at the University of Otago to better understand the views of both Māori and New Zealand genetic service providers towards this policy, and to explore opportunities to revise the policy in a way that better incorporates Te Ao Māori.

Consecutive IGCLC clinical guidelines have recommended annual endoscopic surveillance with both targeted and random biopsies for mutation carriers from 16 years, followed by a total gastrectomy if SRCC are identified in a biopsy specimen. For carriers in their 20s or older, a total prophylactic gastrectomy is recommended. However, if the carrier does not wish to have a gastrectomy, annual surveillance endoscopies are recommended (Blair et al. 2020). The earlier versions of the guidelines were strongly influenced by the high (56%–70%) penetrance of DGC in known HDGC families, the modest ability of endoscopy to detect early stage DGC, and assumptions about the expected aggressive behaviour of SRCC lesions. However, the 2020 IGCLC guidelines and the 2024 1st edition NCCN guidelines have softened the early emphasis on prophylactic surgery, with increasing weight being put on surveillance. This shift has been driven by an appreciation of the disease’s early indolence, improvements in endoscopic methods (in particular the uptake of narrowband imaging), and the lower observed penetrance of DGC in many families. If carriers are interested in prophylactic surgery, they are encouraged not to rush the decision, with delays of a few years considered safe. Moreover, the histological characteristics of any SRCC foci identified by targeted biopsy are now being considered before a decision is made to proceed to surgery (van der Post et al. 2023). Overall, this shift in emphasis reflects a change in the role of HDGC surveillance from simply finding evidence of disease in order to justify proceeding to a total gastrectomy, to now, when the goal is to extend the time before surgery becomes necessary.

The management of the LBC risk in female carriers is also enabled through a combination of prophylactic surgery and surveillance, although the emphasis has always fallen on the side of surveillance. It is recommended that female carriers have annual MRI scans and a clinical breast examination from age 30yrs, and additional annual mammograms from about 40yrs (Blair et al. 2020). If MRI is unavailable, ultrasound is recommended. Other radiological imaging approaches such as ⁶⁸Ga-FAPI-PET/CT have not yet been explored in HDGC, although the high expression of fibroblast activating protein (FAP) in both LBC and DGC suggests a possible future role (Sahin et al. 2024; Sun et al. 2024).

The potential of chemoprevention

The choice between either a total prophylactic gastrectomy or annual endoscopic surveillance is a uneasy dilemma for many CDH1 pathogenic variant carriers. Gastrectomies eliminate risk, but cause lifelong morbidities in all patients, including cramping and diarrhoea after eating (‘dumping syndrome’), nutritional intolerances, bile reflux, fat malabsorption, gallstones and excessive weight loss. In contrast, endoscopic surveillance is relatively straightforward, but, unfortunately, DGC is able to progress underneath an intact mucosal surface, increasing the risk that an advancing cancer will be missed by endoscopy. Moreover, HDGC surveillance generally only delays surgery, with most carriers of pathogenic CDH1 variants undergoing a gastrectomy within 3–4 years of surveillance starting due to either a concerning biopsy result or growing anxiety about the ever-present DGC mortality risk (Lee et al. 2023). Notably, the decision to have a prophylactic total gastrectomy is getting more difficult for the younger members of large affected whānau (Kahurangi Tipene, pers. comm.). Often, this younger generation has not seen the death that their parents saw, and therefore they feel detached from the risks they face. Moreover, this detachment is likely to negatively impact their recovery post-surgery.

This dilemma in choosing between a prophylactic gastrectomy or ‘watchful waiting’ would be resolved if an effective chemoprevention treatment could be found. Chemoprevention would reduce the need for surgery, while at the same time mitigating the psychological burden associated with surveillance. Ironically, the advantages of chemoprevention are starker for those families with apparently lower DGC penetrance due to the increased likelihood of over-treatment (i.e. unnecessary prophylactic surgery). HDGC chemoprevention will also be an important option for carriers of CDH1 variants of unknown significance (VUS). CDH1 VUS are three times more common than high risk mutations (with over 2,000 reported in ClinVar). Prophylactic surgery is not recommended for CDH1 VUS carriers, but they are advised to undergone endoscopic surveillance annually for at least 2 years, then less frequently depending on biopsy findings (Blair et al. 2020). Although these recommendations are pragmatic, CDH1 VUS carriers are often anxious about their unquantified DGC risk, particularly when they have lost a close relative to the disease. A final group to benefit from HDGC chemoprevention would be individuals with ‘Chromosome 16p21.1 Deletion Syndrome’. This deletion comprises an entire CDH1 allele and neighbouring genes associated with cognitive development. Prophylactic surgery is ill-advised for this group due to their severe developmental delay and anticipated troubled recovery from surgery.

Several studies using isogenic cell lines and gastric organoids have identified vulnerabilities in CDH1-null cells that affect fundamental cellular processes including cytoskeletal organisation, sphingolipid metabolism, membrane trafficking, and the initiation of extracellular signals. These vulnerabilities create opportunities to use targeted drugs to selectively kill CDH1-null cells. For example, CDH1-null cells show increased sensitivity to inhibitors of histone deacetylases, SRC/FAK signalling, DDR2/AKT3 signalling, sphingolipid metabolism and the cholesterol synthetic pathway (Telford et al. 2015; Bougen-Zhukov et al. 2019; Godwin et al. 2019; Bougen-Zhukov et al. 2021; Brew et al. 2022; Decourtye-Espiard et al. 2021). Each of these sensitivities presents a potential chemoprevention approach for HDGC. Furthermore, the accumulation of inhibitory Treg cells in the SRCC tumour microenvironment suggests the potential of immunotherapies to eliminate early stage HDGC (Green et al. 2023).

Concluding remarks

30 years of international experience with HDGC has seen significant evolution in our understanding of the disease. In particular, we now know there is considerable variability in DGC risk between families that is driven by unidentified risk modifier genes and/or environmental risk factors. Its optimal clinical management has also shifted as the natural history of the disease has been elucidated and surveillance methods improved. Overall, the accumulated experience of affected families, clinicians, genetic counsellors, and researchers has drastically reduced the risks associated with this syndrome. Future research into chemoprevention, and ultimately gene therapies, will ensure this syndrome will one day be seen as an inconvenience, not a threat.

Disclosure statement

No potential conflict of interest was reported by the author(s).