Diagnosis and therapy of cancers of the esophagogastric junction (EGJ) tend to be accompanied by some unique and specific challenges. Some of these result from the historic lack of specific definitions of the anatomic location and extent of the EGJ. Others stem from the findings or limitations of endoscopic diagnosis, where the appearance of a locally advanced, near-obstructing proximal gastric cancer with EGJ involvement may be that of a distal esophageal cancer or that of a true EGJ lesion may be best visualized upon intragastric retroflection giving the impression of only a proximal gastric cancer without more proximal EGJ involvement. Additional challenges may be a function of the specialization pattern within medical disciplines, that is, a clinical therapeutic spectrum of thoracic versus abdominal disorders, without sufficient overlap. Despite significant institutional variations on surgical specialties involved in the care of EGJ cancer, a separation between thoracic (with a focus on “esophageal cancer”) and general/abdominal surgeons (treating “gastric cancer”) remains commonplace. Such difference in perspective or background may even be evident through distinctive terminologic use of “esophagogastric junction” compared to “gastroesophageal junction” (GEJ) cancer. This problem is worsened in settings of specialty-directed postgraduate training pathways, at least in Northern America, through which the trainee may get exposed to either esophageal cancer or to gastric cancer, but rarely to both. This is disturbing, as operative components and multidisciplinary aspects of care for EGJ cancers would clearly benefit from specialty expertise in both areas. In fact, for both esophageal and gastric resections, specialty high-volume settings have been associated with superior outcomes in terms of postoperative mortality or long-term survival after either procedure.^1,2 In addition, significant contributions in the management of EGJ cancer have traditionally originated in programs in which the same providers deliver treatment of esophageal and gastric cancer in parallel.^3–9 Finally, EGJ cancer treatment is challenged by significant shifts in incidence and demographic factors. There is an ongoing trend within the United States of an increased incidence of lower esophageal adenocarcinomas (Siewert Type I) and to a lesser extent of cardia cancers (Siewert Type II), almost entirely reflecting a steep incidence rise in Caucasian and to a lesser extent African-American men.^10–12 At the same time, traditional (distal) gastric adenocarcinoma and esophageal squamous cell cancer are declining. The worse prognostic survival of proximal gastric cancer compared to distal gastric cancer has been recognized earlier and suggests different disease courses based on primary locations even within the stomach itself.¹³ In other countries, EGJ cancer represents a quite different disease spectrum, as for instance in Japan, lower esophageal adenocarcinomas (Siewert Type I) are rather rare, and most EGJ cancers there can be assigned Siewert Type II or III (proximal gastric cancer) status; as a result, the treatment pattern of EGJ cancer in Japan thus follows mainly that of gastric adenocarcinoma.^14–16 Classifying all cancers involving the esophagogastric junction as “esophageal cancer,” as suggested through the seventh edition of the AJCC TNM staging recommendations, therefore remains highly controversial.^17,18

Given these numerous challenges around EGJ cancer, information with specific implications on EGJ cancer is provided in other chapters of this textbook. The intent is not to complicate the discussion around definition, staging, and therapy, but to provide specific highlights within these domains where EGJ cancer deserves special consideration in order to achieve best insight and generate best outcomes.

The anatomic junction between the esophagus and the stomach is defined in a variable way depending on underlying theory and purpose. Even when asked to simply localize the EGJ, anatomists, endoscopists, radiologists, and surgeons may produce different answers based on the variable ability to identify anatomic structures utilized as surrogates for the EGJ location. In the most recent edition of the AJCC Cancer Staging Handbook,¹⁹ it is stated that “the lower thoracic esophagus is bordered inferiorly by the stomach. Because it is the end of the esophagus, it includes the esophagogastric junction.” This description neither reflects the relationship between EGJ and diaphragm or intraabdominal esophagus, nor does it consider anatomic or physiologic functional mechanisms specific to this area (and relevant to malignant tumors) such as lymphatic drainage patterns. It appears to primarily represent an attempt to classify lower third esophageal tumors into a segment within between 30 cm and 40 cm of endoscopic distance from the incisors. The “arbitrary 10 cm segment encompassing the distal 5 cm of the esophagus and proximal 5 cm of the stomach with the EGJ in the middle” therefore indeed runs the risk to create an artificial entity for stage grouping, as this is primarily based on similarities in overall survival outcomes among the various tumor subsites included, without reflecting considerations for, in part, very different tumor progression patterns or therapeutic approaches between subsites. Siewert and Stein²⁰ summarized the results of an international consensus conference regarding the classification of EGJ tumors within 5 cm above and below the anatomic gastric cardia in 1998; therein, three types of EGJ tumors (“Siewert Type I-III”) were outlined based on their anatomic, etiologic, and histopathologic characteristics. Accordingly, Type I tumors are adenocarcinomas of the distal esophagus that usually arise from an area of specialized intestinal metaplasia of the esophagus (i.e., Barrett’s esophagus) and which may infiltrate the EGJ from above; Type II tumors are true carcinomas of the cardia arising from the cardiac epithelium or short segments with intestinal metaplasia at the EGJ (“junctional carcinoma”); Type III tumors represent subcardial gastric carcinomas that infiltrate the EGJ and distal esophagus from below. This classification was corroborated by the delineation of obvious differences between these three types in terms of gender distribution, hiatal hernia prevalence, history of gastroesophageal reflux, and intestinal metaplasia. This three-type classification of EGJ cancer (“Siewert classification”) has since become a widely used system to describe tumors of the EGJ that may require different therapeutic approaches; it should still be considered a useful guiding system for treatment decision making today.

Even if overall survival prediction of these EGJ cancer subtypes would suggest similarities to esophageal cancers of the upper or middle thirds of the esophagus, several diagnostic and therapeutic considerations require specific attention. Local, regional, and possibly hematogenous distant progression patterns appear to create special challenges. Local extension frequently occurs in submucosal plains and thus is not easily determined on direct endoscopy; proximal esophageal extension of a Type III lesion or significant distal gastric extension of a Type I lesion would mandate adjustments in therapy plans specific to EGJ tumors that are rarely encountered in esophageal tumors of other sites. Transserosal extension by definition is a risk factor for peritoneal progression and recurrence; this is a commonly encountered problem of locally advanced “gastric” tumors, but hardly ever occurs in true “esophageal” lesions without serosal wall components.^21,22 Relevant lymphatic drainage patterns and thus predominant locations of lymph nodes at risk for regional metastasis for any of the three Siewert types are in stark contrast to those of the proximal or middle esophagus.^23,24 This carries important implications for the appropriate choice of regional lymphatic dissection and its impact on staging assessment as well as possible influence on regional disease control and resulting cancer-specific survival. Finally, options for multimodality therapy of EGJ cancer may differ based on “esophageal” or “gastric” protocols, especially regarding the proper choice of preoperative induction therapies. Adenocarcinomas are the predominant histologic EGJ subtype, and therapy response differences in comparison to squamous cell cancers of the esophagus are also documented.^22,25

For the reasons mentioned, surgeons involved in the care of EGJ cancer are expected to understand specifics of diagnosis and therapy that sets this disease entity apart from other esophageal or gastric cancers. In addition, differences between the three EGJ subtypes themselves may carry important implications for treatment choices and thus need to be considered also.

Esophagogastric junction cancers are usually identified and localized though upper gastrointestinal endoscopy, which also most often at the same time allows for biopsy and achieves histologic confirmation. It is recommended to describe all tumors involving the EGJ based on their suspected epicenter in relation to the cardia (i.e., Types I, II, or III) and their proximal and distal mucosal disease extent, with endoscopic distance from the incisors being a useful reference point. In addition, other mucosal abnormalities should be evaluated, since the extent of Barrett’s metaplasia may impact on subsequent treatment decisions such as resection extent. Endoscopic ultrasound (EUS) is recommended for clinical staging, specifically tumor depth for the T category, submucosal tumor extension within esophagus or stomach, and the identification of ultrasonographically abnormal mediastinal or retrogastric lymph nodes (N category).^26,27 Other abnormalities that can be well identified via EUS include splenic artery or hilar lymph nodes and abnormalities within the left lateral liver parenchyma. For cross-sectional imaging, contrasted computed tomography (CT) of chest, abdomen, and pelvis is performed most often; although it adds relatively little regarding T staging, relationships between primary tumor and large vascular structures are well defined. In addition, regional and distant lymph nodes can be identified, and abnormalities concerning for distant metastases (pulmonary and hepatic lesions, ascites, etc.) are evaluated.²⁶ Positron emission tomography (PET) scanning may be helpful to define the extraregional disease extent but is not universally performed; it has more limited accuracy for diffuse-type histology and smaller primaries. PET-CT has shown some increased diagnostic accuracy over PET scanning alone.²⁸ Metabolic tumor evaluation of preoperative chemotherapy or radiation effects via PET scanning requires knowledge of pretreatment PET findings and treatment administered, as negative PET results may indicate metabolic response but not absence of tumor. Most authors recommend repeat staging via CT evaluation after preoperative therapy prior to possible resection unless the treatment plan is affected by the degree of metabolic response. Repeat EUS is not routinely performed as treatment artifacts or responses are not easy to clearly distinguish.^29,30 For all reasons, it appears prudent to perform all quality staging tests prior to conducting multidisciplinary treatment planning and before any therapy component is initiated.

Diagnostic laparoscopy is recommended for midstage EGJ cancers prior to resection.^31–33 Based on the preoperative therapy choice, it can be considered prior to preoperative chemoradiation or chemotherapy, too, if findings of peritoneal metastases or positive peritoneal washing cytology will alter treatment choices. Benefit of staging laparoscopy is well established for “gastric” cancers including Type II and III EGJ cancers;³⁴ for Type I lesions, peritoneal metastases are rare and a benefit to routine laparoscopy in the setting of otherwise normal imaging can be questioned. However, a minimally invasive approach to esophagectomy in this setting invites a laparoscopic overview to rule out metastatic spread.

Discussions still persist over which lymph node regions represent “regional” (i.e., resectable) versus “distant” nodes (without anticipated resection benefit). Nodal spread and frequency of nodal station involvement have been described in detail in several clinical series.^23,24 For all EGJ subtypes, paracardial, proximal perigastric, and retrogastric (left gastric artery, celiac artery) lymph nodes are generally at greatest risk for metastasis. Type I cancers carry a greater likelihood for proximal, mediastinal nodal progression, while Type II and III lesions rarely involve mediastinal sites unless intraabdominal lymph node involvement is extensive. Exceptions to these generalized guidelines certainly exist; they may influence the therapeutic choice significantly, such as in case of a Type II or III cancer with mediastinal node involvement for which an esophagectomy with mediastinal and celiac lymphadenectomy would appear most sensible. Cervical or distant retroperitoneal (paraaortic, iliac, mesenteric) lymph nodes documented to harbor metastases in cases of an EGJ cancer have to be considered distant disease for which no level 1 evidence data currently support curative intent resective therapy.

The decision to apply the same staging criteria for esophageal and EGJ cancers has been supported with citing similarities in OS outcomes: “Although Siewert and colleagues subtype EGJ cancers (type I, II, III), not only do their data support a single-stage grouping scheme across this area, but also they demonstrate that prognosis depends on cancer classification (T, N, M, G) and not Siewert type.”¹⁹ Unfortunately, detailed population-based data on OS of EGJ subtypes do not exist in published form. However, preliminary analyses of this question show differences in postresection survival between EGJ subtypes and also between EGJ cancers and proximal esophageal adenocarcinomas when TNM criteria are controlled for; these differences are especially obvious for cancer-specific survival (Nelson and Schwarz, unpublished data). Given these findings, a single uniform staging classification based on presumed survival similarities between EGJ cancer subsites and other esophageal cancers would appear less sensible and more difficult to uphold. Should the anatomic location that certainly influences operative therapy decisions have an impact on the staging mechanism as it does for instance in cancer of the biliary system, or is this best avoided like it is in colorectal cancers? As will be briefly addressed below, therapeutic similarities between Type I EGJ cancers and other proximal esophageal cancers, and those between Type III lesions and gastric cancer are greatest. Resection approaches to Type II lesions vary among institutions or even individual surgeons between esophagectomy- and gastrectomy-driven techniques, so that anatomic definition criteria would not be able to completely resolve this staging controversy at least for this specific subtype. It is hoped for that at least regarding the definition of T, N, and M categories, homogenized criteria will be developed for all esophageal and gastric adenocarcinomas in the eighth AJCC staging edition.

The “best therapeutic strategy” for EGJ cancer remains challenging and somewhat elusive for several reasons. In general, all midstage EGJ cancers should be considered high-risk malignancies, as recurrence rates even after complete, state-of-the-art therapy remain high. It is by now fairly well established that multimodality therapy leads to superior results when compared to surgical resection alone.^25,35–39 Nevertheless, timing, duration, and actual regimens used remain highly variable, and trials comparing different multimodality strategies for EGJ cancer specifically are sparse. Many trials primarily focus on either esophageal or gastric cancer patients, and therefore conclusions specific to EGJ primaries are difficult to draw (Table 92-1). Certainly not all patients with midstage EGJ cancer qualify for trimodality therapy, with the operative component frequently being considered the limiting factor due to postoperative morbidity predictions; in addition, not all patients who would qualify for trimodality therapy are being offered this option in lieu of definitive chemoradiation. Trimodality therapy, however, appears to be generally superior to lesser treatment efforts.^40,41 Furthermore, operative approaches to the resection of EGJ cancer have varied significantly, and gastroesophageal resection and regional lymphadenectomy extent remain a topic of some debate. For early-stage esophageal and gastric cancers, endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) has become available.^42,43 EMR and ESD applicability and feasibility for T1N0 EGJ cancer still have to be more specifically examined.