GI Surgery Annual: Volume 25

© Indian Association of Surgical Gastroenterology 2019

Peush Sahni and Sujoy Pal (eds.)GI Surgery AnnualGI Surgery Annual25https://doi.org/10.1007/978-981-13-3227-2_1

1. Esophagogastric Junction (EGJ) Carcinoma: An Updated Review

Rajneesh Kumar Singh¹  

(1)

Department of Surgical Gastroenterology, Sanjay Gandhi Post Graduate Institute of Medical Sciences (SGPGIMS), Lucknow, UP, India

Rajneesh Kumar Singh

1.1 Introduction

Epithelial carcinomas constitute the majority of all cases of esophageal cancer. While squamous cell carcinoma (SCC) typically occurs throughout the esophagus (commonest middle third), adenocarcinomas mostly occur in the distal one-third and the esophagogastric junction (EGJ). All adenocarcinomas involving the EGJ are included under the group of EGJ carcinomas; these include esophageal carcinomas, gastric carcinomas and true carcinomas of the cardia. It is rare for lower esophageal SCCs to involve the EGJ; hence all discussion of EGJ carcinomas refers to adenocarcinoma. The incidence of adenocarcinoma has increased, while that of SCC has declined steadily in the Western population, in the last few decades [1, 2]. Hence EGJ carcinoma has become a tumour of increasing importance over the last few decades. The reasons for the increasing focus on these tumours include the rising incidence in the Western world, the controversies in classifications, the generally poor prognosis and major differences in the treatment and outcomes as compared to squamous cell carcinoma of the esophagus.

The other aspect in which esophageal adenocarcinomas differ from SCC is the well-characterized metaplasia-dysplasia-carcinoma sequence for which a large volume of scientific research has accumulated from across the world. This provides an opportunity to study the molecular mechanisms of carcinogenesis and early diagnosis and treatment of some of these tumours [3].

Adding to complexity in case of EGJ adenocarcinoma are the multiple terminologies used by different authors to denote one or all subgroups of EGJ carcinoma, varying from ‘junctional’ carcinoma, ‘cardia’ tumours, gastro-esophageal junction tumours, distal esophageal adenocarcinoma, etc. One often needs have a careful look at the patient cohort represented while interpreting studies including these tumours.

1.2 Classification

As opposed to the usual organ-based neoplasias, EGJ carcinomas are a heterogeneous group of zone-based tumours that arise from or involve the gastro-esophageal junction; these include esophageal, gastric and true cardia carcinomas. The heterogeneity in this group pertains to the epidemiology, etiopathogenesis, molecular pathology, differences in treatment and outcomes of the different subgroups. These within group differences in clinical behaviour were understood quite early, and several attempts were made to subclassify these tumours. Most of these classifications are topographical classifications, and the most commonly referred to is the one provided by Professor Siewert and his group. In the 1990s this classification was adopted by a consensus conference of the International Gastric Cancer Association and the International Society for Diseases of the Esophagus, and experts concluded that this should form the basis of definition, investigation and reporting management of EGJ adenocarcinoma [4].

EGJ carcinomas were defined by Siewert as a group of epithelial carcinomas arising from a zone 5 cm below or 5 cm above the EGJ and mandatorily involving the EGJ [5]. This needs an accurate definition of the location of the EGJ, considering the fact that anatomists, physiologists and endoscopists have all defined the EGJ differently [6]. Adding to this confusion is the shifting of the squamocolumnar junction due to columnar metaplasia of the lower esophagus, at least in some patients. The best accepted definition of EGJ for this purpose is that it lies at the proximal limit of the gastric mucosal folds (rugae). Siewert divided these into three subgroups based on the epicentre of the tumour [5]:

Type 1 tumours: Distal esophageal adenocarcinoma infiltrating the EGJ and mostly associated with intestinal metaplasia, i.e. Barrett’s esophagus (epicentre located more between 1 and 5 cm above the EGJ).

Type 2 tumours: True carcinoma of the cardia arising from the epithelium of the gastro-esophageal junction and often referred to as ‘junctional carcinoma’ (epicentre located between 1 cm above to 2 cm below the EGJ).

Type 3 tumours: Subcardiac gastric carcinoma located below the EGJ and infiltrating the gastro-esophageal junction and distal esophagus (epicentre located between 2 and 5 cm below the EGJ).

The Siewert classification was based on data from their large experience. The salient features separating the three types of tumours are as in Tables 1.1 and 1.2. Siewert type 1 adenocarcinoma is quite similar to esophageal adenocarcinoma, including a male preponderance, a strong history of reflux disease and mainly intestinal-type (Lauren) histology. The majority of these tumours are associated with chronic gastro-esophageal reflux disease, and most arise from Barrett’s metaplasia. Siewert type 3 adenocarcinomas, however, are similar to distal (non-cardia) gastric cancers with only slight male majority, an almost equal proportion of intestinal and diffuse histological types and an insignificant association with reflux.

Table 1.1

Demographic and morphologic tumour differences according to Siewert tumour type

Adapted from [260]

Table 1.2

Pattern of nodal spread according to the Siewert tumour type

Adapted from [261]

Lymphographic studies have shown that the lymphatic drainage from the lower esophagus goes both ways, upwards towards the mediastinum and downwards towards the celiac axis, while the lymphatic drainage from the gastric cardia and subcardiac region mostly drains towards the abdomen (celiac axis and the para-aortic lymph nodes) [4, 7]. The epicentre of the EGJ tumour determines the distribution of the nodal metastasis. The overall frequency of lymph node metastasis is about 90% for type 3 carcinoma, 70% for type 2 carcinoma and 65% for type 1 carcinoma. Type 1 tumours metastasize to nodes both in the mediastinal and upper abdomen, whereas the type 2 tumours mostly drain towards the abdominal nodes, especially the paracardial, lesser curvature and left gastric nodes, and only occasionally to the mediastinal nodes (Table 1.2). The recurrence pattern also varies according to the site, with peritoneal and nodal recurrence being more common with type 3 as compared to type 1 and 2 carcinomas [8].

Although the Siewert classification is useful in defining the prognosis, treatment and outcome of EGJ tumours, there are several practical difficulties encountered, in part due to the limitations of the investigations and often locally advanced nature of these tumours. In a Dutch study, Grotenhuis et al. had found that the overall accuracy in correctly predicting tumour location (Siewert type) was not very high (70% for endoscopy/EUS and 72% for CT) [9]. In this study in 22% of patients, large tumours obscured the landmarks of the gastric folds on preoperative investigations and could not be compared with the pathologic assessment [9]. In another study from Italy, only 72.5% of patients could be accurately assigned a Siewert subtype using EUS and endoscopy [10]. In a Dutch randomized controlled trial on esophageal and EGJ adenocarcinomas, the authors found major differences between the classification of the tumour on endoscopy and on pathology of the resection specimen, in several patients [11].

AJCC in the 7th edition had named all tumours in 10 cm zone straddling the EGJ as EGJ carcinoma, and these were staged as esophageal carcinoma [12]. These included tumours whose epicentre was in the lower thoracic esophagus or EGJ or within the proximal 5 cm of the stomach cardia, which also involved the EGJ or esophagus.

However the eighth edition of AJCC has changed this to include only Siewert type 1 and 2 in the esophageal carcinoma staging schema [13]. Siewert type 3 tumours (2–5 cm below the EGJ) are to be staged as gastric carcinoma according to this recent classification. This change is viewed as an interim topographical classification of EGJ tumours till more genetic and molecular profile data enables these tumours to be classified according to more discerning criteria rather than the inaccurate topographical criteria presently in use.

1.3 Epidemiology and Risk Factors

Esophageal carcinoma is the eighth most common cancer worldwide and the sixth leading cause of cancer-related mortality, according to the GLOBOCAN database [14]. Most of the available epidemiologic data considers esophageal adenocarcinoma as a whole, and most databases do not categorize EGJ carcinoma separately. The incidence of esophageal adenocarcinoma has surpassed that of esophageal SCC in a number of Western countries, while SCC continues to dominate in Asian and African countries [15]. Esophageal adenocarcinoma is typically a disease of the obese Caucasian male often with chronic GERD. The rising incidence of esophageal adenocarcinoma has mirrored the increasing incidence of obesity and the high incidence of GERD in the Western countries [16, 17].

Although epidemiologic data based on subtypes of EGJ carcinoma is difficult to come by, the Siewert type 1 EGJ carcinoma probably has an epidemiology similar to esophageal adenocarcinoma, and the incidence rise has paralleled that of esophageal adenocarcinoma [18]. There is less reliable data about the incidence trends of Siewert type 2 and 3 EGJ carcinoma. A study from Sweden showed (among men) a much larger increase in the incidence of esophageal adenocarcinoma as compared to gastric cardia adenocarcinoma (10% versus 2.3%) [19]. However the different subtypes of EGJ carcinoma probably have different incidence patterns. A study based on the Surveillance, Epidemiology, and End Results (SEER) data from 1970 to 2010 has shown that the incidence of esophageal adenocarcinoma has shown a sharp rise, while that of EGJ carcinoma has increased only a modestly, and there has been a sharp reduction in the incidence of (non-cardia) gastric carcinoma [20]. The parts of the world like China or Iran, with a high incidence of esophageal carcinoma in general, do not report a jump in the incidence of esophageal adenocarcinoma unlike that reported by the Western countries [21, 22]. A multi-ethnic study from the USA showed that among people of Asian origin, the incidence of esophageal adenocarcinomas was quite low [23]. This stands in contrast to the much higher incidence of gastric cardia adenocarcinoma and esophageal SCC in the same subgroup.

When publications of EGJ adenocarcinoma from the East and the West are compared, quite a few differences can be observed. The proportions of the three Siewert EGJ carcinoma subtypes are very different, being almost equal in European series (one-third each), while EGJ type 2 and 3 tumours are much more common in series from Korea and Japan, and the proportion of Siewert type 1 is less than 5% in the Eastern series [24]. These comparisons indicate that, in actual practice the presentation and hence the management of EGJ adenocarcinoma are quite different between these parts of the world. Indian reality is probably closer to the Eastern data rather than the Western data, as type 1 tumours are an unusual sight in our country.

1.3.1 GERD and Barrett’s Metaplasia

A major risk factors of esophageal adenocarcinoma is Barrett’s metaplasia of the esophageal epithelium. Barrett’s metaplasia has been defined as intestinal type columnar metaplasia of the (lower) esophagus, seen at endoscopy and proven by biopsy, associated with chronic GERD [25–27]. Barrett’s metaplasia is considered a precancerous lesion, and it is generally accepted that the intestinal metaplasia component is responsible for this risk, even though there is some disagreement on the malignant potential of non-intestinal columnar metaplasia [28]. Barrett’s progression to carcinoma proceeds through a well-studied sequence of Barrett’s metaplasia—low-grade dysplasia—high-grade dysplasia—adenocarcinoma. In a large prospective study, the annual incidence of these changes in Barrett’s mucosa was as follows: low-grade dysplasia, 4.3%; high-grade dysplasia, 1.3%; and adenocarcinoma, 0.5% [29]. However a large number of patients still present with advanced stage of carcinoma at the time of diagnosis. One reason postulated is that up to 40% of patients do not report symptoms of GERD prior to diagnosis [30].

In EGJ carcinoma, however, the association with Barrett’s metaplasia varies with the subgroup of tumours (Table 1.1). Siewert type 1 tumours have a strong association with Barrett’s metaplasia as in the case of esophageal adenocarcinoma. Chronic GERD and Barrett’s metaplasia have been found in 70–97% of patients with type 1 tumours [31]. On the other hand, type 2 tumours have a very low prevalence of Barrett’s that is slightly more than type 3 tumours (Table 1.1). Type 3 tumours are thought to be similar in pathology to gastric carcinoma and do not have an etiologic background of Barrett’s metaplasia.

1.3.2 Obesity

Several studies have documented a high incidence of GERD in obese patients. The risk is to the magnitude of 16% for every 1 kg/m² increase in BMI as calculated in one study [32]. These patients have a high risk of esophageal adenocarcinoma, probably through a mechanism of chronic GERD and metaplasia, while cardiac adenocarcinoma is only weakly associated with reflux disease [33]. Hiatus hernia also has a similar close association with GERD and esophageal adenocarcinoma [34].

1.3.3 Helicobacter pylori

H. pylori infection (especially CagA strain) is considered an important risk factor for adenocarcinoma of the distal stomach. This is believed to proceed through an orderly sequence of events such as chronic active gastritis, atrophic gastritis, intestinal metaplasia and gastric cancer [35]. Siewert type 3 adenocarcinoma seems to have a similar association with H. pylori. On the other hand, an infection with H. pylori seems to have a protective effect for esophageal adenocarcinoma and type 1 EGJ carcinoma, probably through its inverse effect on GERD [36]. With regard to the role of H. pylori in causation of cardia carcinoma (type 2 EGJ), however, the data is quite conflicting and inconclusive [18]. Eurogast Study Group meta-analysis found that there was lack of a consistent association between junctional cancers and H. pylori across the world [37]. While most studies from the West showed a negative association, quite a few studies from the East showed a positive association between EGJ adenocarcinoma and H. pylori infection [37].

1.3.4 Tobacco Smoking

Tobacco smoking is a well-established and moderately strong risk factor for esophageal adenocarcinoma in both men and women, with ever smoking conferring an approximately doubled risk of adenocarcinoma compared with never smoking (OR, 1.96) [38]. Further, the Northern Ireland Barrett’s register reported an approximate twofold increased progression risk from Barrett’s esophagus to adenocarcinoma associated with tobacco smoking. A similar association was reported with cardia cancer (type 2) as well [39].

1.3.5 Alcohol Consumption

A large study confirmed no association between alcohol intake and increased risk of esophageal adenocarcinoma [40].

1.3.6 Dietary Factors

The most comprehensive global report of diet, nutrition and esophageal cancer, published by Continuous Update Project of the World Cancer Research Fund International/American Institute for Cancer Research recently, found no good evidence for linking any conventional dietary factors with esophageal adenocarcinoma, except that vegetable intake had limited suggestive evidence for a reduced risk of adenocarcinoma [41].

1.3.7 Genetic and Molecular Studies

Over the last decade, attempts to classify the EGJ tumours based on genetic/molecular characteristics have covered a lot of ground and have provided a lot of exciting data. This is a rather complex area in which the picture has started becoming clearer only recently. The Cancer Genome Atlas (TCGA) project group had done extensive work on molecular profiling of gastric cancer in 2014 [42]. This study had classified gastric carcinoma into four subtypes on the basis of—(1) Epstein-Barr virus (EBV) infection, (2) microsatellite instability (MSI), (3) chromosomal instability (CIN) and (4) genomic stability (GS) [42]. Further work by the same authors has clearly shown that EGJ adenocarcinoma is distinct from esophageal SCC and needs to be viewed separately for therapeutic targets [43]. EGJ carcinomas were found to be quite similar to CIN type gastric carcinoma as opposed to other types of gastric carcinoma. These investigators found that among adenocarcinomas, there was an increasing prevalence of CIN as the location of the tumour moved proximally up to the esophagus and none of the esophageal adenocarcinomas was positive for MSI or EBV, unlike gastric carcinoma (Table 1.3). Some EGJ carcinoma were, however, MSI-positive and EBV-positive. With more and more genome-wide studies becoming available, it is becoming clear that topographical subgrouping (Siewert classification) of EGJ adenocarcinoma is a rather inaccurate way of classifying these tumours. In the not too distant future, the molecular profile of EGJ tumours will, possibly, determine the subgrouping, prognosis and treatment strategies adopted for these tumours.

Table 1.3

Molecular mutation profile of CIN type gastro-esophageal adenocarcinomas by anatomic location

Adapted from TCGA [43]

1.4 Clinical Presentation

The majority of patients at presentation already have advanced disease. The commonest symptoms are dysphagia and odynophagia (i.e. painful swallow). It has been estimated that dysphagia occurs only after 75% of the lumen is obstructed by the tumour, though a small tumour may sometimes cause a tight stenosis through intense fibrosis. Hoarseness or Horner’s syndrome occur with the invasion of the recurrent laryngeal nerve or cervical ganglia, respectively, and such patients are almost always inoperable. Cervical or supraclavicular lymphadenopathy is associated with distant spread and indicates inoperability in EGJ adenocarcinoma. Occult or overt GI bleeding can occur especially with ulcerated tumours. Other important symptoms that indicate advanced disease are chest pain, back pain, excessive weight loss (more than 10%) and long duration of dysphagia (more than 6 months). In addition severe comorbid illnesses, particularly cardiopulmonary, carry a risk of poor overall outcome. Submucosal infiltrating carcinoma at the EGJ may mimic achalasia, and as such is termed pseudoachalasia.

1.5 Investigations

1.5.1 Upper GI Endoscopy

The cornerstone of diagnosis, screening and surveillance is endoscopy by skilled observers. Barrett’s mucosa is seen as an extension of the salmon-pink velvety gastric mucosa proximal to the squamocolumnar junction. Biopsies are mandatory to enable a pathologic diagnosis of intestinal metaplasia and any associated dysplasia. Any visible lesion in the mucosa should be biopsied, and in addition four-quadrant biopsies should be taken at every 2 cm along the Barrett’s mucosa. To increase the accuracy of endoscopy, additions have been made like chromoendoscopy, high-magnification endoscopy, narrow-band imaging, autofluorescence, light-scattering spectroscopy, optical coherence tomography and confocal endomicroscopy. These techniques have been evaluated in individual studies and incorporated in various endoscopy systems. Having shown benefit in individual studies, these technologies have not been adopted across the board for reasons of high cost, absence of high-quality evidence of benefit and poor penetration across the world.

Endoscopy allows accurate characterization of the tumour’s configuration, length and localization. At least six biopsies from non-necrotic areas of the tumour increase the yield to nearly 100%. Endoscopic views while crossing the EGJ and then the retroflexed views after entering the stomach are a good way of preoperatively subgrouping the tumours as per the Siewert classification. However about half the tumours in the Indian subcontinent are not passable with an endoscope due to the severity of the obstruction.

1.5.2 Endoscopic Ultrasound (EUS)

Most of the literature on EUS and esophageal carcinoma pertains to esophageal SCC and adenocarcinoma [44]. There is paucity of studies on the role of EUS in EGJ carcinoma. EUS is used for staging esophageal cancer and in the evaluation and management of patients with high-grade dysplasia (HGD) in Barrett’s metaplasia. It enables the endosonographer to evaluate the wall-layer pattern of the esophagus and to detect the presence of regional and celiac lymph nodes. EUS-guided FNA permits directed tissue sampling of adjacent nodes. The endoscope-based systems are divided into radial and linear array scanning systems. The radial echoendoscope uses a mechanically rotated transducer to generate a real-time 360-degree cross-sectional image perpendicular to the long axis of the instrument. The linear array echoendoscope has an electronically operated transducer that produces a 270-degree real-time image parallel to the long axis of the endoscope and is used to carry out a guided FNAC of adjacent nodes or lesions.

Tumours that do not admit an EUS scope due to stenosis are locally advanced in the majority of cases [45]. In these situations, dilatation of the tumour for the purpose of EUS staging is fraught with the risk of tumour perforation and should be weighed against the low benefit of EUS staging in therapeutic decision-making. These patients are best referred for neoadjuvant therapy in view of the locally advanced nature. EGJ tumours present a unique challenge to the EUS operator because of its location and frequent extension across the EGJ. The trouble is probably due to the imprecise results in the evaluation of gastric invasion because of the difficulty in positioning the probe on the entire circumference of the cardia region and in the gastric fundus. The invasion of the stomach is therefore frequently studied by the retroflexed endoscopy view alone. Hence the EUS assessment of T-stage of EGJ tumours is often inaccurate. A recent study on the role of EUS in EGJ carcinoma found a 48% concordance between EUS uT-stage and pathologic pT-stage (under-staged 23%, over-staged 29%) [46].

EUS is recommended to be performed in all patients with only loco-regional disease, and it may be helpful in the following clinical scenarios:

Selected patients with high-grade dysplasia and early (T1a) tumours for non-surgical treatment—For accurate T and N staging [47, 48].

Locally advanced esophageal carcinoma—Staging of T4 tumours to determine resectability [49].

Locally advanced esophageal carcinoma—Staging for remote nodal disease and selecting out patients who may not undergo a R0 resection, e.g. upper mediastinal nodes in EGJ carcinoma [50, 51].

Locally advanced esophageal carcinoma—To select patients for neoadjuvant therapy. Stage 2 and 3 patients are usually selected to undergo neoadjuvant treatment prior to surgery [52].

1.5.3 CT Scan

CT scan of the neck, thorax, abdomen and pelvis with intravenous and oral contrast is the standard of care investigation for staging of esophageal carcinoma. The fissure of the ligamentum venosum is seen on the CT separating the caudate lobe from the lateral segment of the left lobe of liver; it points directly at the EGJ [53].

The key findings on CT scan include:

Wall thickening greater than 5 mm (circumferential or part of the wall).

Dilated esophagus proximal to an obstructing lesion.

Tumours infiltrating outside the wall may appear as soft tissue and fat stranding around the esophagus.

Locally advanced tumours may cause displacement of the tracheobronchial tree. Unfortunately loss of fat plane between the airway and the esophageal tumour cannot be used as an indication of invasion, as no fat plane is normally evident even in patients without a tumour. It is known that the posterior tracheobronchial wall/membrane is unsupported by incomplete cartilage rings and hence normally indent during expiration. CT scans should therefore be acquired in full inspiration to avoid getting a false impression of a compression due to a mass lesion.

Aortic invasion may be shown in the following findings on the CT scan:

The Picus angle is the angle of contact (loss of fat plane) between the esophageal mass and aorta. Angle of contact more than 90° is highly suggestive of invasion of aorta, an angle less than 45° is associated with no invasion, and angle in between 45° and 90° is indeterminate. Accuracy of these findings is about 80% [54].

Tumour invasion of the triangular space between the spine, esophagus and aorta may also be indicative of aortic invasion.

Node metastasis—While nodes can be seen on CT scan, only a mediastinal node with a short-axis diameter exceeding 1 cm is considered abnormal, except for the nodes in the subcarinal region. However lymph nodes may harbour metastases without being enlarged, and hence the location of all visualized nodes should be noted. In addition, it is important to remember that nodes may be enlarged because of inflammatory or infectious etiologies. In a meta-analysis of staging investigations for carcinoma esophagus, the sensitivity and specificity of CT scan for nodal metastases were found to be rather low (0.50 (95% CI 0.41–0.60) and 0.83 (95% CI 0.77–0.89), respectively) [55].

Distant metastasis can be present in advanced tumours. In a study Quint et al. found the pattern of distant metastasis as follows: abdominal nodes (45%), liver (35%), lung (20%), cervical and/or supraclavicular nodes (18%), bone (9%), adrenal glands (5%), brain and peritoneum (2% each), and stomach, pancreas, pleura, skin or body wall, pericardium or spleen (1% each) [56].

1.5.4 PET-CT Scan

Combined PET and CT scan has a higher sensitivity and specificity for tumour staging than 18F-FDG PET alone [57]. In these integrated scans, the CT scan has two main purposes. It provides an attenuation map to correct for the greater attenuation of photons coming from the deeper structures (as opposed to the photons coming from the more superficial structures). This correction is not only important to improve the quality of the image and but also allows for an accurate quantitative measurement of metabolic activity. This is denoted as the standardized uptake value (SUV). The SUV is the ratio of metabolic activity in the region of interest to the decay corrected activity of injected 18F-FDG. The other purpose of the CT scan is to provide anatomic reference data that improves the interpretation of the metabolic findings on PET imaging by fusing anatomical with metabolic findings.

PET-CT for EGJ carcinoma faces a unique problem of varying avidity for 18F-FDG depending on various histologic features. While Siewert type 1 and type 2 tumours show intestinal differentiation in the majority of patients, type 3 tumours have pathology more like gastric cancer diffuse differentiation in the majority (Table 1.1).

Poor uptake of FDG (i.e. FDG non-avid tumours) is usually associated with diffuse Lauren type tumours, small tumour size, mucinous content and good differentiation. Up to one-third of gastric tumours can be PET non-avid [58]. These facts should be considered before interpreting PET literature for carcinoma esophagus as a whole.

A study on esophageal adenocarcinoma and EGJ carcinoma from India showed that PET-CT findings led to change in management in 16% of patients [59]. The utility of PET-CT can be summed up as follows:

Prognostic value—Several studies have shown that there is a good correlation between higher maximum SUV (SUVmax) and poor overall and disease-free survival [60, 61]. Though the pre-treatment SUV values may have prognostic implication, there is a wide range of cut off values of SUVmax that are reported as significant across studies. In published literature there is no clear agreement on the optimal cut off value of the SUVmax.

Staging—18F-FDG PET is less accurate than EUS for determining the T-stage and is not much better than EUS or CT scan for nodal staging [62]. Uptake in the primary lesion may obscure the involved loco-regional nodes. However 18F-FDG PET-CT is the best investigation for diagnosis of unsuspected distant metastasis and extra-regional involved nodes. In a meta-analysis van Vliet et al. showed that the sensitivity and specificity for detecting distant metastases by 18F-FDG PET were 71% and 93%, respectively, and by CT scan it was 52% and 91%, respectively [55].

Response assessment during neoadjuvant therapy—Tumour response to neoadjuvant therapy can be quite variable, and only in about half of the patients, it may show a major response. Early PET-CT during neoadjuvant therapy allows early recognition of non-responders and institution of salvage therapy for them. While phase 2 studies have shown feasibility and good outcome of such an approach, randomized studies are awaited to adopt this widely [63, 64].

Response assessment after neoadjuvant therapy—Schollaert et al. in a systematic review of 26 studies of post-treatment response assessment suggested that post-treatment 18F-FDG PET has good predictive value for long-term outcomes [65]. However, these studies are difficult to interpret because PET-CT was done at varying time-periods after neoadjuvant therapy (22 days to 6 weeks) and with widely different criteria for response measurement.

Follow-up—PET can detect recurrent/metastatic disease in 8–17% of patients, sometimes even before disease can be diagnosed on standard imaging [66].

Radiation planning—Good radiotherapy planning needs accurate delineation of gross tumour volume. Clearly distinguishing a small primary tumour from normal esophagus can be difficult with CT alone. Compared to CT for radiation planning, the addition of PET results in major changes in the gross tumour volume (GTV) and also influences the radiotherapy dose delivered to the neighbouring normal organs [67]. This is a relatively new field of work in which new data is emerging by the day.

1.5.5 Staging Laparoscopy and Peritoneal Lavage Cytology

CT scan and PET-CT scan have low sensitivity for small peritoneal metastatic nodules. Older series have reported sensitivity and specificity of CT scan in gastro-esophageal adenocarcinoma for the diagnosis of liver metastases, 74% and 99%, respectively, and for the diagnosis of peritoneal carcinomatosis, 34% and 94%, respectively [68]. More recently, with advancement in technology, the sensitivity of CT scan has improved in this regard. In one series with 15% of peritoneal carcinomatosis, the reported sensitivity for CT scan was 75% for diagnosing peritoneal seeding with an impressively low rate of unsuspected peritoneal metastasis (less than 4%) [69]. In another large series, staging laparoscopy changed the treatment plan in 20% of patients [70].

Peritoneal lavage cytology seeks to diagnose free-floating tumour cells shed in the peritoneal cavity. In one large study, staging laparoscopy revealed overt peritoneal metastases in 22.6% of patients with gastric adenocarcinoma and 11.8% in those with esophageal adenocarcinoma. In the same study, positive peritoneal cytology in the absence of obvious peritoneal metastases was identified in another 3.1% of patients with gastric adenocarcinoma and 4.4% of patients with esophageal adenocarcinoma [71]. In another study with potentially resectable esophagogastric adenocarcinoma, 7.2% of patients had positive peritoneal lavage cytology aside from those with obvious peritoneal seedings [72]. The prognosis of positive peritoneal lavage cytology is as grim as grossly obvious peritoneal metastasis. Even though this is a subject of ongoing study, most such patients are offered only palliative chemotherapy.

Others have tried to identify patients at high risk of peritoneal spread who should be offered staging laparoscopy and peritoneal lavage cytology. In a study from MD Anderson, multivariate analysis showed that peritoneal carcinomatosis was associated with poorly differentiated histology, linitis plastica and suspicious CT findings, such as trace peritoneal fluid or nodules [73]. In a French series of esophagogastric adenocarcinoma, the factors associated with positive staging laparoscopy were signet ring cell carcinoma, poorly differentiated carcinoma, T3/T4 tumours and linitis plastica-type tumour [74]. The National Cancer Care Network (NCCN) recommends laparoscopic staging with peritoneal washings (lavage cytology) for patients with Siewert type 2 and type 3 advanced tumours, clinical stage T3 or more or clinical node-positive tumours [75].

1.6 Staging

The eighth edition of the AJCC TNM classification of esophagogastric adenocarcinoma has made some important changes over the seventh edition (Tables 1.4, 1.5, 1.6 and 1.7) [13]. In an effort to overcome the limitations of the seventh edition, which was based entirely on patients treated by esophagectomy alone (without preoperative or postoperative chemotherapy and/or chemoradiotherapy), the dataset used to develop the eighth edition TNM stage groupings included patients who had received preoperative induction therapy (neoadjuvant) and/or postoperative adjuvant therapy. The availability of these data led to the ability to explicitly define cTNM and ypTNM cohorts and stages (Tables 1.5, 1.6 and 1.7).

Table 1.4

AJCC 8th edition TNM categories