Li Fraumeni Syndrome predisposes to gastro-esophageal junction tumours

Li-Fraumeni Syndrome (LFS) is defined by germline pathogenic variants in the tumour suppressor gene TP53 and is associated with increased rates of many different malignancies, some with up to 100% lifetime risk without appropriate intervention [1]. While the strongest associations have been demonstrated with sarcoma, breast cancer, brain tumours and adrenocortical carcinomas, the increased risk of colorectal and, more recently, gastric adenocarcinomas have also been acknowledged [2]. The relative risk of esophageal cancer is less well characterised with conflicting reports [1, 3]. An analysis of the International Agency for Research on Cancer database identified 0.5% of individuals with esophageal cancer (15/3043) and 3.3% (101/3043) with gastric cancer [3, 4]. Gastro-esophageal junction (GEJ) tumours arise at the histological transition between esophagus and stomach, and the commonly used Siewert classification divides GEJ adenocarcinomas into three categories for which oncological management differs: Type 1 arising from distal esophagus (1 to 5 cm proximal to GEJ), Type 2 located at the true junction (between 1 cm proximal to 2 cm distal of the GEJ) and Type 3 located 2 to 5 cm distal to the GEJ [5]. In contrast, the American Joint Committee on Cancer stages all GEJ tumors with epicentre ≤ 2 cm into the proximal stomach as esophageal cancers and those > 2 cm as gastric cancers [5]. Consequently, GEJ tumors have been variably categorised in the literature as esophageal, gastric or a separate entity.

In this context, expert consensus guidelines differ in their recommendations for upper endoscopy, or esophagogastroduodenoscopy, screening in LFS: North American guidelines support upper endoscopy every 2–5 years from the age of 25 alongside colonoscopy, while European guidelines do not routinely recommend upper endoscopy screening [2, 6]. The uptake of these recommendation is also unclear.

There is a biological basis to suspect pathogenic variants in TP53 as a driver of upper gastrointestinal cancers. The Cancer Genome Atlas Program classified 559 sporadic esophageal and gastric cancers into distinct molecular phenotypes. When analysing adenocarcinoma, four subclasses were identified, but only one, a chromosomal instability (CIN) subclass, accounted for the vast majority of lower esophagus and GEJ tumours [7]. CIN molecular subgroup accounted for 49% of adenocarcinomas from esophagus and stomach and was characterised by ERBB2 amplification, VEGFA amplification and, importantly, pathogenic variants in TP53.

The familial cancer centre at our institution offers a dedicated gastrointestinal risk management clinic for patients with high-risk genetic predisposition to gastrointestinal malignancy, and we routinely offer upper endoscopy at the same time as colonoscopy to patients with LFS. We sought to assess the rates of upper gastrointestinal cancers and their characteristics in our cohort, with comparison to results from the TCGA cohort.

Separately, data of each subject analysed in the TCGA study to characterise the molecular characteristics of sporadic esophageal and gastric cancers was obtained from the supplementary table in the initial publication. Corresponding details of pathogenic variants in TP53 were then obtained from the openly available dataset on cBioportal (accessed 27 June 2023) [7, 8]. From 559 subjects, 90 esophageal squamous cell cancers, 13 gastric adenocarcinomas of unclear location and 2 undifferentiated esophageal tumours were excluded. The remaining 454 cases of adenocarcinoma were then subdivided based on subtype (esophageal/probable esophageal GEJ adenocarcinoma, indeterminate GEJ adenocarcinoma, gastric/probable gastric GEJ adenocarcinoma) and proportion of cases with somatic pathogenic variants in TP53 calculated.

Sixty-five with LFS (57% female, mean at diagnosis 34.5, median follow-up 51 months) seen via the high-risk cancer genetics clinic over the study period, 35 (53.8%) of whom had at least one upper endoscopy. Four patients (6.2%) were diagnosed with cancer at the GEJ (Table 1), with no cancers elsewhere in the upper digestive tract and one separate case of colorectal cancer on colonoscopy. Two cases, a White man and North-East Asian woman in their 30s, were asymptomatic and undergoing screening, at their second and first procedures respectively. Both cancers were at early stage allowing for resection with curative intent. The lesions were subtle and best appreciated on close inspection of the GEJ in forward-view (Fig. 1). In the former case, index upper endoscopy revealed moderate reflux esophagitis (Los Angeles Classification grade B) with no dysplasia on biopsy, but a decision was made for early repeat surveillance at 18 months revealing subsequent carcinoma in situ. No cases of cancer were associated with prior family history of digestive tract malignancy, or abdominal radiotherapy. Furthermore, five patients undergoing screening (14.3% of those who had upper endoscopy) were diagnosed with a premalignant lesion (Barrett’s esophagus, non-Helicobacter associated gastric intestinal metaplasia and a proximal fundic gland polyp with low grade dysplasia).

On review of the data for these patients from the TCGA paper via cBioportal, we found that amongst esophageal or probable esophageal adenocarcinomas, 76.4% (55/72) had somatic pathogenic variants in TP53, with decreasing prevalence in the indeterminate GEJ (22/36, 61.1%), gastric/probable gastric GEJ (29/63, 46.0%) and gastric non-cardia (113/283, 39.9%) adenocarcinoma groups (Fig. 2).

It is still not globally recognised that upper endoscopy should be a routine component of surveillance in LFS. Moreover, in most guidelines recommending upper endoscopy, the focus has been on gastric adenocarcinoma, perhaps because some reports have presumably grouped GEJ cancers with gastric cancers [9], even though some cases are better characterised as esophageal or a distinct entity [5, 7]. The distinction is important as inspection of the gastric mucosa is performed within the stomach and on retroflexion during upper endoscopy, while the GEJ is best assessed in forward-view.

Our findings suggest that the GEJ and lower esophagus may be a particularly vulnerable region in LFS, in keeping with the molecular phenotyping seen in sporadic GEJ tumours. While it has been suggested that Asian carriers may be at higher risk than non-Asian carriers, and that risk of both gastric and colon cancers may run in families [9, 10], the risk in our cohort was not isolated to these specific groups. Both cases seen on surveillance were early stage, facilitating likely curative resection. This data suggests a harmonisation of germline and somatic observations that are highly suggestive of a biological effect of p53 involving the GEJ. There are limitations to the data given this is a retrospective analysis of a cohort of LFS, but the high prevalence of GEJ cancer in this group is compelling and warrants a prospective evaluation in patients who carry germline pathogenic variants of TP53.

Additionally, premalignant lesions in the upper gastrointestinal tract appeared relatively frequently in our cohort. Given the presence of gastric intestinal metaplasia in four of the five individuals with premalignant lesions, we would also endorse routine biopsy of the gastric antrum and body and eradication of Helicobacter pylori if present, similar to recommendations for Lynch Syndrome [11].