Dietary and Nutritional Factors

Drinking from non–tap water has been shown to be an independent risk factor for squamous cell cancer of the esophagus in high-incidence areas of China. Exposure to polycyclic aromatic hydrocarbons commonly present in soot extract of coal-burning stoves, a domestic cooking device used in poorly ventilated kitchens throughout the esophageal cancer belt, has been implicated as a risk factor of squamous cell cancer, particularly in women who do not smoke.

An association with dietary intake of N-nitroso compounds has been implicated in the pathogenesis of squamous cell cancer, particularly in high-incidence areas. N-nitroso compounds are derived from reduced dietary nitrates, often by ubiquitous fungal toxins, and certain food material are particularly susceptible to be contaminated by N-nitroso compounds, such as smoked pickles and a bread-like food called qocho or kocho in Ethiopia.¹⁰ It has been hypothesized that in black South Africans, the rising incidence of squamous cell cancer could be partly related to a recent dietary shift to maize from sorghum.¹¹ Fusarium fungi, which grow sparsely on sorghum, grow freely on maize, producing fumonisins, which reduce nitrates to nitrites and synthesize cancer-producing nitrosamines. N-nitroso compounds are mutagenic in animal models and act by inducing alkyl adducts in deoxyribonucleic acid (DNA).¹² Chewing betel-quid with areca nut and use of gutka (pan masala, a dry powdered mixture of areca nut, catechu, lime, unspecified spices, and flavoring agents), which is very common in Southeast Asia, has been proposed to have a synergistic carcinogenic effect in association with alcohol and tobacco.¹³ In areas of the world endemic for squamous cell cancers, drinking very hot beverages has been shown to have carcinogenic attributes, probably by causing chronic esophageal damage related to repeated thermal injury.

Certain dietary factors such as selenium have been shown to be protective against squamous cell cancers. Low levels of selenium have been associated with squamous cell cancers in high-risk areas of China, and selenium supplementation has been shown to be protective in controlled trials from China.^14,15 Similarly, zinc deficiency, which potentiates the carcinogenic effects of nitrosamines as well as causes overexpression of the cyclo-oxygenase pathway in esophageal carcinogenesis, has been recognized as an independent risk factor of squamous cell cancers.¹⁶

Evidence suggests that low dietary folate intake and impaired folate metabolism due to functional polymorphisms in folate-metabolizing pathways, particularly polymorphisms in the 5,10-methylenetetrahydrofolate reductase (MTHFR) gene, play an important role in the causation of many gastrointestinal cancers, including esophageal cancer. In a recent meta-analysis of seven case-control studies, relative cancer risks for individuals in the highest compared with the lowest category of dietary folate intake were 0.66 for esophageal squamous cell carcinoma and 0.50 for esophageal adenocarcinoma.¹⁷ Also other investigators have shown that the MTHFR 677TT gene variant, which is associated with reduced enzyme activity, was associated with an increased risk of esophageal squamous cell carcinoma.¹⁸

There are clearly unknown interactions of genetic, nutritional, and environmental elements that may be synergistically important in the etiology of squamous cell esophageal cancers. For example, in northern Iran, the traditionally semi-nomadic Turkoman population has one of the highest incidences of squamous cell cancers. Intensive investigation of well-known nutritional and environmental risk factors has failed to identify a specific risk factor that could explain the very high incidence of squamous cell cancers in this population. Prevalence of cigarette smoking and alcohol consumption (discussed following), which are major risk factors for squamous cell cancers in Western populations, was on a negligible scale among men in this population and practically absent among women. Foodstuffs had very few carcinogens such as nitrosamines, aflatoxins, or polycyclic hydrocarbons. It has been postulated that a very restricted diet (low in animal protein, fruit, or fresh vegetables), along with consumption of dry, coarse bread contaminated with sharp seeds and silica fibers, frequent use of hot tea, and the regular smoking or swallowing of opium dross (“sukhtheh”), along with potential genetic predispositions, may act synergistically by an unspecified mechanism to impart a very high risk for developing squamous cell esophageal cancer in this population.

Coffee drinking and increased intake of fruit, fish, and white meat has a protective effect on squamous cell cancers. In contrast, red meat, salted meat, and meat boiled at high temperature may increase the risk of squamous cell cancers.^19,20

Alcohol and Tobacco

Alcohol and tobacco use, particularly in the form of pipe, cigar, or cigarette smoking, is a dominant risk factor of squamous cell cancers in low-incidence areas, such as the United States. On the other hand, in high-incidence areas, tobacco or alcohol does not seem to play a dominant causative role, although in Asian countries, where smoking is becoming increasingly popular, particularly among men, its relative importance as a risk factor is increasing. Epidemiologic studies suggest that the amount of alcohol consumed is more important than type of alcohol consumed, and the most prevalent alcoholic beverage consumed in a particular region tends to be the one with the highest risk of squamous cell esophageal cancers in that population.²¹ Multiple epidemiological studies carried out in regions with different incidences of squamous cell cancers suggest that specific polymorphisms in genes encoding for alcohol metabolizing enzymes may determine individual susceptibility to the carcinogenic role of alcohol.^22,23

Preexisting Diseases of the Esophagus

Several underlying esophageal diseases are known to increase risk of subsequent development of squamous cell cancers of the esophagus. There is a fairly well-established association of esophageal cancer with achalasia. In a recent epidemiologic study from Sweden involving 2896 patients with achalasia, excess risks for squamous cell carcinoma and adenocarcinoma of the esophagus were observed, predominantly in men. There was no association with esophagomyotomy.²⁴ Esophageal strictures caused by ingestion of lye, a caustic corrosive agent, are associated with a very high risk of squamous cell cancer, which develops three to five decades after the initial event. In achalasia and lye strictures, relative stasis, stagnation of food, and chronic inflammation have been implicated, but no definite carcinogenic mechanisms have been established. Tylosis, particularly the inherited form, which manifests in an autosomal dominant fashion with hyperkeratosis of palms and soles, has been associated with squamous cell cancer. For an affected family member, the estimated lifetime risk of esophageal cancer varied from 40% to 92% by the age of 70 (see Chapter 22). Although rare, the carcinogenetic events in tylosis has received recent attention in that a tylosis esophageal cancer (TOC) gene locus has been mapped to chromosome 17q25 by linkage analyses. Interestingly, loss of heterozygosity at this same locus, which contains the promoter sequence of the cytoglobin gene, has been detected in a significant number of patients with sporadic squamous cell cancers.²⁵ The association of chronic mucocutaneous candidiasis with squamous cell carcinoma of the oral cavity and esophagus has been described in patients with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED).²⁶

A rare syndrome of iron deficiency anemia, dysphagia, and postcricoid esophageal web known as Plummer-Vinson syndrome in the United States and Patterson-Kelly syndrome in the United Kingdom has been reported to have an association with squamous cell cancers (see Chapter 41). The anecdotally reported association of squamous cell cancers with partial gastrectomy has not been substantiated.

Patients with history of squamous cell cancer of the upper aero-digestive tract have an increased risk of synchronous or metachronous squamous cell carcinoma of the esophagus, most likely due to shared risk factors. In prospective controlled studies, 3% to 14% of patients with squamous cell cancers of head and neck area developed synchronous or metachronous squamous cell cancers of the esophagus.^27,28 Radiation therapy following mastectomy moderately increases the risk of squamous cell esophageal cancer in the upper and middle thirds of the esophagus, starting approximately 5 years after exposure, with the risk persisting after 10 years; no similar increase in the risk of esophageal adenocarcinoma has been reported. This finding appears to be a function of the portals used for postmastectomy radiation therapy, which typically do not expose the lowest third of the esophagus, where adenocarcinoma commonly arises.²⁹

Given that human papillomavirus (HPV) has been implicated in the etiopathogenesis of squamous cell cancers of the oropharynx, its association with squamous cell cancers of the esophagus has been studied. Patients with squamous cell cancers of the esophagus from the United States, Europe, and Japan are infrequently HPV positive. On the other hand, in certain high-incidence areas, such as China and South Africa, HPV DNA can be demonstrated in a significant proportion of patients with squamous cell cancer of the esophagus. The interactive effect of genetic, environmental, and dietary risk factors in determining geographic susceptibility to the oncogenetic potential of chronic HPV infection needs further study.³⁰ There is no clear association of esophageal cancer with chronic infection with other viruses such as herpes simplex virus or Epstein-Barr virus.

Dietary and Nutritional Factors

Increased intake of cereal fiber has been shown to have a dose-dependent inverse association with gastric adenocarcinoma and, to a lesser extent, with distal esophageal adenocarcinoma, but not with squamous cell cancer of the esophagus. Diets high in fiber, beta-carotene, folate, and vitamins C, E, and B₆ may be protective, whereas diets high in cholesterol, animal protein, and vitamin B₁₂ may be associated an with increased risk of esophageal adenocarcinoma.³¹ Antioxidant intake appears to be protective against esophageal adenocarcinoma but not gastric adenocarcinoma.³² Carbonated soft drinks may have an inverse relationship, but drinking tea and coffee is not associated with esophageal adenocarcinoma.³³

Alcohol and Tobacco

Numerous epidemiologic studies, including prospective studies, suggest that association between alcohol and tobacco use is less consistent with esophageal adenocarcinoma than with squamous cell cancers. In general, smoking is considered a moderate risk factor for esophageal adenocarcinoma, whereas alcohol use has no association with esophageal adenocarcinoma. Relative to persons who had never smoked, current cigarette smoking was found in a prospective study to be associated with increased risk of squamous cell cancers (hazard ratio [HR], 9.27), esophageal adenocarcinoma (HR, 3.70), adenocarcinoma of the gastric cardia (HR, 2.86), and noncardia gastric adenocarcinoma (HR, 2.04).³⁴

Obesity

Pooled results from observational studies support a positive association between an increased body mass index (BMI) greater than 25 kg/m² and the risk for esophageal adenocarcinoma and possibly for adenocarcinoma of the cardia. In one study, increasing abdominal waist size was associated with an increased risk of esophageal adenocarcinoma, independent of BMI.³⁵ Most recent published information support the hypothesis that obesity, in particular, abdominal obesity, contributes to GERD, which may in turn increase the risk of Barrett’s esophagus (see Chapters 43 and 44).^36–38 In contrast to adenocarcinoma, in a large population-based prospective study of Chinese men, a low BMI was associated with an increased risk of squamous cell carcinoma of the esophagus.³⁹

Gastroesophageal Reflux Disease

Barrett’s esophagus, which is a complication of gastroesophageal reflux disease (GERD), is one of the strongest risk factors for esophageal adenocarcinoma (see Chapter 44). However, there is controversy whether GERD without Barrett’s esophagus is by itself a risk factor for esophageal adenocarcinoma. Many patients with esophageal adenocarcinoma do not report preexisting GERD symptoms and only a minority of patients with GERD develop Barrett’s esophagus. However, it is also thought that chronic mucosal inflammation in patients with recurrent GERD may predispose to esophageal adenocarcinoma. A large population-based Swedish study showed that in persons with recurrent symptoms of reflux, as compared with people without such symptoms, the odds ratio for esophageal adenocarcinoma was 7.7 and was 2 for adenocarcinoma of the cardia. The more frequent, more severe, and longer-lasting the symptoms of reflux, the greater the risk.⁴⁰ A recent Australian study suggested that obesity and GERD may play a synergistic role in the pathogenesis of esophageal adenocarcinoma.⁴¹ Overall, epidemiologic evidence suggests that chronic, severe GERD may predispose susceptible individuals to esophageal adenocarcinoma.

Barrett’s Esophagus (see also Chapter 44)

The reported prevalence of Barrett’s esophagus in the general population varies widely from 0.9% to 4.5%.^42,43 Such fivefold variations in prevalence are mostly related to the type of population studied and also the definition of Barrett’s esophagus that is used. For example, experts disagree whether histopathologic demonstration of intestinal metaplasia is a requirement for diagnosing Barrett’s esophagus.^44,45 Recent studies have suggested an increasing incidence of Barrett’s esophagus.^46,47

The reported annual risk esophageal adenocarcinoma in patients with Barrett’s esophagus varies from 0.2% to 2% (see Chapter 44). In more recent studies with long-term follow-up of patients with histologically confirmed long-segment Barrett’s esophagus, the annual risk of developing high-grade dysplasia (HGD) or adenocarcinoma is approximately 1%.^48,49 Although the annual incidence of cancer in those with low-grade dysplasia (LGD) was 0.6% to 1.3%, the incidence rate in patients with HGD is almost 10 times higher.^50,51 In a meta-analysis of 236 patients with Barrett’s esophagus and HGD, esophageal adenocarcinoma was reported in 69 patients over 1241 patient-years of follow-up, with a weighted incidence rate of 6.58 per 100 patient-years.⁵² In another recently published meta-analysis, the overall pooled risk for developing adenocarcinoma in all patients with Barrett’s esophagus was 6.1 per 1000 person-years and was just 4.1 per 1000 person-years when early incident cancers and HGD at baseline were excluded.⁵³ The risk of esophageal adenocarcinoma in patients with short segment Barrett’s esophagus or with specialized intestinal metaplasia of the esophagogastric junction is not known, but may be less compared with long-segment Barrett’s esophagus. However, it is important to note that these two entities are several-fold more common than long-segment Barrett’s esophagus and could account for a large number of patients with gastroesophageal junctional adenocarcinoma.

In addition to segment length and dysplasia, hiatal hernia, central obesity, and possibly smoking may be other factors favoring progression from nondysplastic Barrett’s esophagus to adenocarcinoma.^54,55

Helicobacter pylori Infection

Epidemiologic studies show that as the prevalence of Helicobacter pylori infection has decreased in Western societies, the prevalence of GERD, Barrett’s esophagus, and distal esophageal and gastroesophageal junctional adenocarcinoma has rapidly increased. Although there are no data to suggest that H. pylori plays any role in esophageal mucosal resistance to acid-induced injury, in esophageal acid clearance, or in competence of the lower esophageal sphincter, it has been suggested that eradication of H. pylori in patients with corpus-predominant gastritis may exacerbate acid reflux. It has been postulated that the alkaline ammonia produced by H. pylori colonizing the cardiac mucosa could protect the neighboring distal esophageal squamous mucosa from damage by acidic reflux. There appears to be a negative association between H. pylori infection and GERD symptoms, although an association between successful H. pylori therapy and development of new or recurrent reflux symptoms has not been established.^56,57 In a recent meta-analysis, patients with Barrett’s esophagus and adenocarcinoma, but not those with squamous cell esophageal cancer, were less likely to be infected with H. pylori, particularly cagA-positive strains.⁵⁸ The clinical implication of such inverse correlations is unclear and will require careful evaluation, given that H. pylori infection is considered an important etiologic factor for gastric cancer.

Diverse Preexisting Conditions

Patients with acid hypersecretory states or conditions associated with severe GERD (e.g., scleroderma) may be at increased risk for esophageal adenocarcinoma. In a population-based study from Sweden, cholecystectomy was associated with a moderately increased risk of subsequent esophageal adenocarcinoma, but not squamous cell cancer. Increased duodeno-gastroesophageal reflux and a toxic effect on bile acids and bile salts on the esophageal mucosa was thought to be the mechanism. A case control study failed to identify an association between diabetes and esophageal adenocarcinoma.⁵⁹ GERD is common in premature infants and in infants who are small for gestational age, and epidemiologic data suggest that preterm birth and low birth weight may be risk factors for development of esophageal adenocarcinoma. Down syndrome is not associated with esophageal cancer.⁶⁰

Drugs

There is conflicting evidence regarding whether long-term use of drugs that relax the lower esophageal sphincter, such as anticholinergics, β-adrenergic agonists, theophylline or aminophylline, and benzodiazepines increase risk of esophageal adenocarcinoma.⁶¹ A few epidemiologic studies and a recent meta-analysis reported a protective effect of aspirin use (and to a lesser extent nonsteroidal anti-inflammatory drug [NSAID] use) with both types of esophageal cancer, an effect that appears to be dose dependent.⁶²

Autonomous Growth (Growth Factors)

Increased expression of epidermal growth factor (EGF) and transforming growth factor-α (TGF-α) has been demonstrated in esophageal adenocarcinoma and has been associated with invasive disease and possibly poorer outcome.^68,69 Furthermore, increased concentrations of TGF-β in azygos vein blood and increased expression of connective tissue growth factor (CTGF) and endoglin, a member of the TGF-β receptor family, have been correlated with angiogenesis, angiolymphatic invasion, and metastasis in patients with esophageal cancer.⁷⁰ On the other hand, increased expression of SMAD has been associated with better survival in patients with squamous cell esophageal cancer.⁷¹ C-erbB2, a proto-oncogene coded by the HER2/neu gene and sharing a significant homology with EGF receptor, is amplified in patients with Barrett’s esophagus and adenocarcinoma, is clinically associated with a poorer outcome, and may play an early role in disease invasiveness.^72,73 HER2/neu gene overexpression may promote carcinogenesis through multiple pathways, such as inhibition of apoptosis and enhanced cell proliferation, increased mitogen activated protein kinase (MAPK) activity and increased matrix metalloproteinase (MMP) activity, and by potent induction of vascular endothelial growth factor (VEGF).

The restriction point (R-point) is a critical gatekeeping checkpoint that controls G1- to S-phase transformation in the cell cycle (see Chapter 3). Cyclin D and E are proteins that are important regulators of the R-point and serve as the final step in many proliferation cascades. CCND1, the gene that encodes cyclin D1, is associated with an increased risk for esophageal adenocarcinoma.⁷⁴ Also cyclin B has been shown to be overexpressed in esophageal squamous cell cancer, with possibly with a negative prognosis.⁷⁵

Attenuation of Antiproliferative Pathways

Attenuation of antiproliferative cellular signals is a hallmark of cancer cells. Most cellular antiproliferative pathways converge on the retinoblastoma protein (Rb) pathway, which plays an important antiproliferative role by having a regulatory role on transition of cells from G0 to S phase (see Chapter 3). Loss of heterozygosity (LOH) of Rb protein and its functional repression by hypermethylation of p16, a tumor suppressor gene, has been described in patients with esophageal adenocarcinoma.^76,77 Expression of p21, another cell cycle regulator that is activated after DNA damage by ionizing radiation, has been correlated with responsiveness of esophageal cancers to chemoradiotherapy.⁷⁸ Although mutations of the APC gene is uncommon in patients with esophageal cancer, allelic deletion of 5q where APC resides, as well as functional repression of APC by hypermethylation of the promoter region of this gene, is quite common in patients with esophageal adenocarcinoma.⁷⁹ In fact, plasma levels of hypermethylated APC DNA may be a biomarker of esophageal cancer.⁸⁰

Disordered Apoptosis

Programmed cell death, or apoptosis, is an important defense mechanism against cancer and is regulated by TP53 and bcl-2 family of genes and their proteins (see Chapter 3). Alterations in TP53 in patients with esophageal cancer are seen in up to 50% to 99% of patients and are usually associated with a more aggressive form of tumor.⁸¹ Similarly, the bcl-2 family of proteins (Bcl-2, Bcl-xl, Bax) maintains an intricate balance of pro- and antiapoptotic factors in the cellular milieu and may play an important yet unclear role in the progression of non-dysplastic Barrett’s esophagus to high-grade dysplasia and adenocarcinoma.⁸² Nuclear factor kappa B (NF-κB), an antiapopotic factor, is expressed in 60% of patients with esophageal adenocarcinoma and has prognostic significance.⁸³ The protein kinase Akt (also known as protein kinase B), a recently described serine-threonine kinase, is an important mediator of growth and antiapoptotic signals in esophageal adenocarcinoma. Increased Akt activation in the basal epithelium has been associated with metaplastic progression from squamous epithelium to nondysplastic Barrett’s esophagus, and eventual progression to high-grade dysplasia and adenocarcinoma.⁸⁴ The hormone leptin and short-term acid exposure activate Akt in esophageal adenocarcinoma cells in vitro.⁸⁵

Unlimited Replication

Telomeres are composed of several thousand repeats of short six-base-pair sequence elements and are located at the ends of chromosomes. At each cell replication the telomeres are shortened so that telomere length serves as counters of cycles of replication. After a finite number of cell replications, the telomeres are short enough to trigger the cell to exit from G1 to G0 phase, thus arresting further replication. Cancer cells, which are characterized by ability for limitless replication, achieve this ability partly by increased expression of telomerase, a ribonucleoprotein reverse transcriptase that counteracts shortening of telomeres. Increased expression of telomerase has been associated with progression of metaplastic epithelium to dysplasia to adenocarcinoma in patients with Barrett’s esophagus.⁸⁶

Factors Promoting Neoangiogenesis, Invasion, and Metastasis

The ability of cancer cells to invade and disseminate is determined by their ability to disrupt intercellular adhesion as well as by altering the delicate balance between extra cellular matrix degrading proteases, such as urokinase-type plasminogen activator (uPA) of the serine protease system, MMP, and antiproteinases, such as tissue inhibitory metalloproteinase (TIMP) (see Chapter 3). Not unexpectedly, esophageal cancer, one of the most aggressive cancers with respect to invasiveness and propensity to metastasis, is associated with molecular abnormalities related to cell-cell adhesion molecules (CAMs), such as E-cadherins, integrins, and CD44 transmembrane glycoproteins, and overexpression or altered localization of uPA, the cysteine protease, cathepsin B (CTSB) protein, MMP and TIMP. Evidence suggests that many of these may be useful prognostic markers.^87,88

Alteration of the Cyclooxygenase Pathway

Cyclooxygenases (COX) are key enzymes in the prostaglandin metabolism. COX-2, an inducible enzyme, has been shown to be an important mediator of tumorigenesis and angiogenesis. Selective inhibition with COX-2 inhibitors induces apoptosis and reduces angiogenesis. Activation of the COX-2 pathway can inhibit apoptosis by lowering the intracellular level of arachidonic acid, a fatty acid precursor of prostaglandins, which is known to promote apoptosis. A second mechanism that can be targeted by COX-2 inhibitors is prostaglandin-E₂ (PGE₂) dependent, and appears to be mediated by increase in intracellular Bcl-2 as well as intracellular cyclic adenosine monophosphate (cAMP), both of which suppress apoptosis. The angiogenetic potential of the COX-2 pathway occurs through increased expression of VEGF, increased endothelial survival by bcl-2 and Akt signaling, induction of MMP, EGF receptor–mediated angiogenesis, and suppression of interleukin 12 (IL-12) expression. Studies have shown that patients with esophageal cancer who express a high level of COX-2 in their tumors are more likely to have metastatic disease at presentation, to have more local tumor recurrences after definitive treatment, and to have a worse survival than patients not overexpressing COX-2.^89,90

Microsatellite Instability

Several investigators have reported microsatellite instability (MSI) in less than 20% of esophageal and gastroesophageal junctional adenocarcinomas. Early unconfirmed evidence, however, suggests that MSI may be an early event in the progression of intestinal metaplasia to esophageal adenocarcinoma.⁹¹

Chromosomal Abnormalities

Defects in mitotic checkpoint genes lead to aneuploidy, or an abnormal number of chromosomes. Although a vast majority of esophageal cancers have aneuploidy, the exact gene defects are not characterized. DNA aneuploidy detected by flow cytometry has been considered to be a marker of progression of intestinal metaplasia to esophageal adenocarcinoma.⁹² A host of specific chromosomal abnormalities have been described in patients with esophageal cancer, and recent studies based on a comparative genomic hybridization technique using different types of esophageal cancer and gastric cardiac adenocarcinoma may lead to identification of genes important for disease-specific pathogenesis.⁹³

Molecular Events in Neoplastic Progression of Barrett’s Esophagus

Research has focused on the complex molecular events that lead to development of esophageal adenocarcinoma in the setting of Barrett’s esophagus.⁹⁴ It is postulated that the initial turn-on event, which is related to chronic exposure to noxious stimuli, such as acid reflux and bile salt exposure, is the activation of the inflammatory cascade mediated by cytokines such as tumor necrosis factor-α (TNF-α) and IL-1b. This cytokine activation eventually leads to ectopic expression of CDX1 and CDX2 homeobox proteins that play a major role in the development of intestinal epithelium in the embryonic stage and in this setting cause intestinal metaplasia to develop.^95–97 With the development of metaplastic intestinal epithelium, multiple poorly characterized genetic (e.g, aneuploidy, loss of heterozygosity, mutation) and epigenetic events (e.g., hypermethylation) set off the metaplasia to dysplasia to adenocarcinoma sequence, which eventually leads to genetic instability with clonal expansion of cells with genetic errors. The exact sequence of this multistep process is not known, but several early events in this transition sequence have been identified.⁹⁸ Abnormalities of p16 and possibly overexpression of cyclin D1 are relatively early events that lead to loss of control of cell cycle regulatory genes. Telomerase may also play in early role. Prevalence of TP53 damage such as loss of heterozygosity and mutation increases with advancing grade of dysplasia. Similarly, progression of dysplasia is characterized by decreased expression of E-cadherin and membranous β-catenin expression. Growth signals such TNF-α and COX-2 expression also increase with progression of grades of dysplasia and warrant further evaluation as potential biomarkers.

Future Molecular Directions

Exciting developments in the diverse field of genome analysis technology, stem cell signaling pathways, microarray analysis, and proteomics will hopefully lead to clearer understanding of the pathobiology of carcinogenesis and the diverse but interrelated molecular events responsible for development and progression of dysplasia involved into the eventual development of esophageal cancers. Such developments will also likely lead to maturation of our evolving knowledge on a host of potential molecular biomarkers, prognostic indicators, and therapeutic targets for chemopreventive and chemotherapeutic strategies for incorporation into standard clinical practice.

Squamous Cell Cancer

Esophageal squamous cell cancer typically develops by progression from premalignant (dysplastic) precursor lesions (Fig. 46-2). Many pathologists describe dysplasia using a two-tiered system: low-grade dysplasia, which includes mild and moderate dysplasia; and high-grade dysplasia, which includes severe dysplasia and cancer in situ.⁹⁹ In the World Health Organization (WHO) classification the term “intraepithelial neoplasia” is preferred over “dysplasia,” although dysplasia is by far more popular.¹⁰⁰ Dysplasia includes both cytologic and architectural abnormalities of varying severity and extent. Low-grade dysplasia often involves the basal part of the epithelium and high-grade dysplasia usually involves the entire epithelial layer and, in a small proportion of patients, dysplasia extends into the ducts of the esophageal glands, simulating stromal invasion or spread in a horizontal pagetoid pattern.¹⁰¹ Dysplasia is often multifocal, supporting a field defect hypothesis in the pathogenesis of esophageal squamous cancer. Cytologic changes are typically seen as coarse chromatin, increased nuclear to cytoplasmic ratio, nuclear hyperchromasia, nuclear pleomorphism, and mitotic figures. Architectural changes of dysplasia include disorganization, loss of polarity, overlapping nuclei, and lack of surface maturation. The presence of these features is often used to distinguish dysplasia from non-neoplastic (reactive) processes typically seen with esophageal mucosal inflammation.

Approximately 50% to 60% of squamous cell esophageal cancers occur in the middle third of the esophagus, 33% involve the distal esophagus, and 10% occur in the proximal esophagus. The gross appearance of squamous cell esophageal cancers ranges from a flat, relatively inconspicuous lesion to elevated, plaque-like, polypoid, ulcerated, exophytic, and infiltrative lesions. Superficial squamous cell esophageal cancers invade the submucosa but not the muscularis propria, whereas advanced cancers usually invade the muscularis propria and beyond, either with an infiltrative pattern of small isolated tumor nests or in an expansile pattern of a solid advancing sheet of tumor cells, often with a prominent peripheral lymphocytic infiltration. Approximately one third of squamous cell esophageal cancers are well differentiated, whereas two thirds are either moderately or poorly differentiated, without any features of keratinization.

Squamous cell cancers are aggressive and local lymph nodal metastasis occurs early, partly related to the presence of lymphatic channels in the esophageal lamina propria. Because of the prognostic importance of lymph nodal metastasis and its effect in deciding management options, early squamous cell cancer has been categorized into six subcategories based on depth of tumor infiltration. Intramucosal carcinomas are divided into three groups (m1, m2, and m3), and carcinomas invading the submucosa also are divided into three groups (sm1, sm2, and sm3) (Fig. 46-3). Data from Japan suggest that although m1 and m2 early squamous cell cancers do not usually have nodal metastasis, m3 disease has up to an 8% incidence of nodal metastasis, and this incidence progressively increases as the tumor infiltrates the submucosa, with sm1, sm2, and sm3 tumors having incidences of nodal involvement of 17%, 28%, and 49%, respectively. Similarly the incidence of vascular invasion progressively increases, with m1 tumors having none, whereas almost 90% of sm3 tumors show vascular involvement.¹⁰²

Figure 46-3. A schematic diagram of depth of invasion of a superficial esophageal neoplastic lesion, as evaluated in a resected mucosal specimen. Depths m1, m2, and m3 denote invasion limited to the epithelial layer, lamina propria, and muscularis mucosae, respectively. Similarly, sm1, sm2, and sm3 denote superficial, intermediate, and deep submucosal invasion, respectively.

Invasion of local structures, such as mediastinal pleura, the trachea, the bronchi, and the aorta as well as distant metastases to liver, lung, bone, and other sites may be present in more than a third of these patients at presentation and signifies a poor prognosis. Relatively rapid invasion of the tumor into neighboring structures in the mediastinum has been attributed to the absence of a true serosal layer in the esophageal wall.¹⁰³

Adenocarcinoma

Compared with squamous cell esophageal cancer, a larger body of information exists related to the pathobiology of dysplasia in Barrett’s esophagus and its progression to esophageal adenocarcinoma. Grossly, esophageal dysplasia in Barrett’s esophagus is usually flat and inconspicuous. Dysplasia may be described as low grade or high grade and generally follows the same patterns of cytological or architectural changes described earlier with squamous cell dysplasia. There is considerable interobserver variation in the interpretation and grading of dysplasia and adenocarcinoma among pathologists.¹⁰⁴ Furthermore, progression of dysplasia from low grade to high grade is a morphologic continuum and thus may not follow identifiable features of orderly progression for definitive categorization.¹⁰⁵ “Adenoma-like” dysplastic changes show minimal crypt distortion and mild cytological abnormalities. This is in contrast to the much less common “nonadenoma-like” dysplasia that typically shows crowded glands containing cuboidal cells with a high nuclear to cytoplasmic ratio, round nuclei with irregular contours, vesicular chromatin, and prominent nucleoli.¹⁰⁶ Goblet cells are usually depleted in dysplastic epithelium, and even the surrounding nondysplastic mucosa may show a decrease in goblet cells.¹⁰⁷

Many pathologists often use the term “indefinite for dysplasia” when true dysplasia cannot be accurately differentiated from reactive changes in the presence of esophageal inflammation. Although lack of surface maturation is normally considered a cardinal feature of dysplasia, this has been recently debated.¹⁰⁸ Adjunct immunohistochemical markers (e.g., proliferating cell nuclear antigen (PCNA) and Ki67, cyclin D1, and TP53) have been used to improve the accuracy of diagnosis of dysplasia in this setting.^109,110 Recent studies have reported the utility of immunostaining for alpha-methylacyl-CoA racemase (AMACR) for differentiating non-dysplastic epithelium from low-grade and high-grade dysplasia and also from adenocarcinoma.^111,112 In one study, AMACR staining was negative in all cases of Barrett’s esophagus considered negative for dysplasia, whereas 38% of cases of low-grade dysplasia, 81% of cases of high-grade dysplasia, and 72% of cases of adenocarcinoma were positive.¹¹¹ Another recent study also reported a high negative predictive value for AMACR immunostaining.¹¹²

Almost all esophageal adenocarcinomas arise in the setting of Barrett’s esophagus and typically occur in the distal third of the esophagus, including the esophagogastric junction (see Chapter 44). Adenocarcinomas unrelated to Barrett’s esophagus are extremely rare and usually arise from foci of gastric heterotopia in the cervical esophagus (inlet patch; see Chapter 41). Esophageal adenocarcinoma may have a flat, inconspicuous appearance or can be polypoid, ulcerated, or infiltrative. The majority are well or moderately differentiated, usually comprising cystic or tubular glands in solid nests and irregular clusters and often in a cribriform pattern with considerable stratification. In poorly differentiated carcinomas, the tumor cells infiltrate the esophageal wall with sheets of poorly formed glands with a prominent desmoplastic stroma. Signet ring cells and bizarre pleomorphic tumor cells may be present. One of the common problems encountered by pathologists is distinguishing a tumor of the gastric cardiac from an esophageal adenocarcinoma. In these instances, correlation of the biopsy site with endoscopic anatomic landmarks is crucial.

The incidence of lymph node metastasis in esophageal adenocarcinoma is related to depth of tumor infiltration and appears to be equal or less than in squamous cell esophageal cancer. Whereas nodal disease is rare with adenocarcinomas limited to the esophageal mucosa, the rate of nodal metastases for tumors invading into the submucosa is reported to be 27% to 41% for all patients, and 67% to 78% for those with tumors infiltrating the deep submucosa.^113–115 Involvement of celiac and perihepatic lymph nodes is more common with esophageal adenocarcinoma than squamous cell carcinoma because of the more common occurrence of these former tumors at or near the gastroesophageal junction.

Laboratory Tests

Laboratory abnormalities in patients with esophageal cancer are nonspecific and most often show anemia and hypoalbuminenia. Anemia can have features of iron deficiency or anemia of chronic disease. Hypercalcemia due to osteolytic metastasis or, less commonly, as a response to circulating parathyroid hormone–related protein (in squamous cancer) may be present in up to a third of patients. Hypercalcemia is more common in patients with squamous cell cancers than adenocarcinomas and may indicate advanced disease and a poor prognosis. There are no specific serologic markers for esophageal cancer.

Imaging Tests

Contrast Esophagography

The role of contrast esophagography in diagnosis of esophageal cancer has diminished over the years with the increasing use of endoscopy for primary diagnosis. Esophagography should be performed only when this test is likely to affect decision making. Double-contrast barium radiographs often show early cancers as small polypoid lesions, plaque-like lesions, or focal irregularity of the wall. Advanced cancers commonly appear as areas of irregular luminal narrowing, ulceration, and stricture, with abrupt shoulders. Esophagogastric junctional cancers typically extend into the gastric cardia, and for this reason the fundus and cardia of the stomach should always be included in an optimal contrast radiographic study of the esophagus. To evaluate a suspected esophagorespiratory fistula, contrast radiographs are very useful for delineation of the anatomy prior to endoscopic stenting. Such contrast radiographic studies should be performed using barium instead of a hyperosmolar contrast agent such as meglumine ditrizoate because of the risk of pulmonary edema with the hyperosmolar contrast agent (Fig. 46-4).

Figure 46-4. This barium esophagogram in a patient complaining of cough and dysphagia demonstrates complete esophageal obstruction and a tracheoesophageal fistula. Note the presence of ingested barium in the airways.

Endoscopy and Biopsy

During endoscopy, the location of the esophageal tumor, its relation to anatomic landmarks such as the upper esophageal sphincter and the esophagogastric junction, and the degree of luminal obstruction are typically assessed. Also endoscopic visualization combined with endoscopic biopsy provides an accurate assessment of malignant (and premalignant) lesions of the esophagus. A Paris classification has been proposed for endoscopic assessment of superficial neoplasia of the gastrointestinal tract, including the esophagus, but its clinical acceptance has been limited particularly in North America.¹¹⁷ Advanced esophageal cancers appear endoscopically as fungating, friable, often ulcerated mass lesions occupying some or all of the luminal circumference, usually with indistinct margins (Fig. 46-5). Less frequently, both types of cancers (squamous and adenocarcinoma) can present with a submucosal infiltrative pattern and may not have a prominent luminal component. If the esophagogastric junction is involved by such a tumor, the clinical presentation is often suggestive of peseudoachalasia (see Chapter 42). Typically a higher number of biopsies translates to a higher accuracy rate. Tumors with a submucosal infiltrative pattern may require an aggressive biopsy technique to obtain deeper tissue, using a bite-on-bite technique or endoscopic ultrasonography (EUS)-guided fine-needle aspiration of submucosal tissue.

Figure 46-5. Advanced squamous cell carcinoma of the esophagus that is nearly completely occluding the lumen.

Brush cytology is a useful and quick technique to increase the yield in establishing a diagnosis of esophageal cancer, particularly with the availability of advanced cytologic techniques such as fluorescence in situ hybridization (FISH).¹¹⁸

Endoscopic Detection of Dysplasia and Early Cancer

Numerous endoscopic techniques and newer imaging technologies have evolved over the past decade for endoscopic detection of early or superficial esophageal cancer.

Electronic Chromoendoscopy

Electronic chromoendoscopy by spectral manipulation of white light, such as narrow band imaging using narrow bandwidth filters or the use of postprocessing spectral estimation technique, has been used during esophageal endoscopy to highlight the surface texture such as pit pattern, as well as the mucosal capillary vascular pattern (Fig. 46-6).^123,124 Although initial reports of higher accuracy for early detection of dysplasia and early esophageal cancer were encouraging,^125–128 a later report questioning the effect of electronic chromoendoscopy in this setting has dampened the initial enthusiasm.¹²⁹

Figure 46-6. Images of nondysplastic Barrett’s esophagus using a high-resolution endoscope without (A) and with (B) narrow band imaging. In the lower panels (C, D) an area of early intramucosal cancer in the background of high-grade dysplasia associated with Barrett’s esophagus is shown. Note the irregular and distorted pit and vascular pattern in the area of high-grade dysplasia or intramucosal cancer compared with the nondysplastic Barrett’s esophagus (A, B), which has a regular pit and vascular pattern.

High-Resolution Endoscopic Imaging

High-resolution endoscopes with pixel densities of approximately 850,000 (compared with approximately 100,000 to 300,000 in conventional endoscopes) and magnification endoscopes, which can optically zoom the image from 1.5 to 150 times, have been used in prospective studies to increase the yield of detection of early esophageal cancers and foci of high-grade dysplasia, often performed in conjunction with chromoendoscopy.¹³⁰ Fluorescent-aided confocal endomicroscopy, which uses principles of confocal imaging, has been made commercially available, as an endoscope and as an endoscopic probe. Initial experience in early esophageal neoplasia has been encouraging, particularly for subsurface imaging (Fig. 46-7).¹³¹ One of the limitations of confocal endomicroscopy is the requirement of a fluorescent dye administered either topically or parenterally for fluorescent contrast. Endocytoscopy is based on the technology that incorporates optical contact microscopy on a probe-based platform for real-time microscopic examination of the esophageal epithelial cells at a magnification ranging from 450 to 1100 times.^132,133

Figure 46-7. Confocal endomicroscopic images (Pentax EC-3870CIK, Pentax of America, Montvale, N.J.) after intravenous administration of 10% fluorescein sodium. A, Columnar epithelium with (dark arrow) and without goblet cells. B, Goblet cells (black arrows) and narrow lumen of a Barrett’s gland (white arrow). C, High-grade dysplasia with loss of basal border (black arrows) of dysplastic cells. D, Disorganized architecture of a gland consistent with invasive cancer (black arrow). The white arrow shows a nondysplastic Barrett’s gland with a single layer of columnar epithelium.

EUS and optical coherence tomography (OCT) are endoluminal cross-sectional imaging techniques. The former uses acoustic waves and the latter coherent light waves to quantitatively measure the degree of back-scattering from the tissue to image tissue. High-resolution EUS using catheter miniprobes has limited accuracy for detecting invasive adenocarcinoma in patients with Barrett’s esophagus and high-grade dysplasia or intramucosal carcinoma.^134,135 The utility of EUS seems mostly to be for staging esophageal cancer rather than early detection. Several investigators have reported the utility of OCT in detecting areas of high-grade dysplasia and early esophageal cancer; further technical developments, such as availability of ultrahigh-resolution and Doppler-enabled OCT, as well as optical coherence microscopy, which marries features of OCT and confocal microscopy for 3-D imaging, may further enhance the usefulness of endoscopic OCT.^136–138

Staging System

Staging of esophageal cancer is not only important for planning appropriate stage-specific therapy but also provides crucial information regarding prognosis. For example, Surveillance, Epidemiology, and End Results (SEER) data from 2005 showed that one-year survival in patients with local, regional, and distant (metastatic) disease were 68%, 54%, and 28%, respectively. The five-year survival for esophageal cancers confined to the mucosa (T1m, N0, M0) may approach 90%.¹⁵⁸

The TNM staging system for esophageal cancer is outlined in Table 46-2. The prefixes “c,” and “p” can be used to designate whether the information for a given TNM stage grouping is based solely on clinical information or includes surgical pathology findings, respectively. Some investigators also use a “u” prefix to specify the use of EUS in determining the cancer stage.

Table 46-2 American Joint Committee on Cancer Staging System for Cancers of the Esophagus

Endoscopic Therapy

Esophageal resection has been the standard of care for early esophageal cancer. However, evidence has accumulated that early esophageal cancers with m1 or m2 depth of invasion have little or no risk of nodal disease. In such patients, endoscopic resection may be an attractive alternative to surgery. One endoscopic resection technique is similar to saline-assisted polypectomy using a simple polypectomy snare. A second technique (lift-and-cut) uses a double-channel endoscope and a biopsy forceps to lift up the lesion prior to polypectomy. A third and common technique is the cap-assisted technique, which uses a transparent cap attached to the endoscope to suction the target lesion (typically after submucosal injection of saline below the lesion to form a pseudopolyp), which is then resected using a pre-looped snare. A fourth ligate-and-cut (band-and-cut) technique uses a banding device to suction the target lesion, followed by application of a rubber band ligation prior to polypectomy. This particular technique has become popular with the introduction of a commercially available endoscopic mucosal resection (EMR) kit, which is a modified variceal band ligation device that allows multiple resections with a single endoscopic intubation of the esophagus (Figs. 46-11 and 46-12). The latter two techniques have been compared in a controlled trial and were similar in terms of size of the resected tissue specimens and complications.¹⁷⁹

Figure 46-11. A schematic diagram showing different techniques of endoscopic mucosal resection. A, The simplest technique, similar to saline-assisted polypectomy using a polypectomy snare. B, Lift-and-cut technique, through a double-channel endoscope, using a biopsy forceps to lift the lesion prior to snare polypectomy. C, Cap-assisted technique, in which a transparent cap is attached to the tip of the endoscope with a snare that is pre-looped inside the cap. With gentle suction the lesion is sucked into the cap to create a pseudopolyp. The snare is closed, suction released, the closed snare is pushed out of the endoscope, and the pseudopolyp is cut with electrocautery. D, Ligate-and-cut (band-and-cut) technique, which uses a banding device to suction the target lesion, followed by application of a rubber band prior to polypectomy.

Figure 46-12. A, Endoscopic image of a nodular superficially ulcerated lesion at the gastroesophageal junction, which, on biopsy, was an adenocarcinoma in the setting of Barrett’s esophagus. B, An endoscopic mucosal resection (EMR) was performed using the ligate-and-cut technique (see Fig. 46-11D). C, Photomicrograph of the EMR specimen showed adenocarcinoma (Hematoxylin and eosin, ×2). The inset is a magnified view (Hematoxylin and eosin, ×40) of the area of the mucosa indicated by the square in which an intramucosal adenocarcinoma (m1; see Fig. 46-3) could be identified by the presence of fused glands with severe cytological atypia.

(Courtesy Giovanni De Petris, MD, Scottsdale, Ariz.)

To date, the largest series using EMR in the setting of Barrett’s adenocarcinoma reported 100 consecutive patients with early esophageal adenocarcinoma. These patients underwent 144 resections (maximum, 3 resections) using the ligate-and-cut technique; a complete local remission was achieved within 1.9 months (range, 1 to 18 months) in 99 of the 100 patients. During a mean follow-up period of 36.7 months, recurrent or metachronous carcinomas were detected in 11% of the patients, but successful repeat treatment with endoscopic resection was possible in each case. The calculated five-year survival rate was 98%. Minor bleeding, which was easily controlled with endoscopic therapy, occurred in 11% of patients.¹⁵⁸

Other investigators have reported similar success with endoscopic resection of early esophageal cancer, including in patients with squamous cell cancers.^180,181 The ideal lesion for EMR is a solitary, nodular lesion, less than 20 mm in diameter, limited to the mucosa and with a suitable macroscopic appearance (macroscopic types I, IIa, IIb, and IIc by the Paris classification). En-bloc resection is always preferred to piecemeal resection, the latter of which increases the risk of local recurrence in addition to making a histopathologic interpretation of complete resection more difficult.

Endoscopic submucosal dissection (ESD) is becoming popular in Japan and involves a deeper and larger resection of the esophageal wall by dissecting the submucosal connective tissue just beneath the target lesion from the underlying muscle layer using the hook knife or other electrocautery devices. ESD is usually reserved for larger lesions with evidence of some degree of submucosal invasion. ESD also requires high level of expertise on the part of the endoscopist.¹⁸¹ In a comparative trial, ESD using the hook knife had a higher en-bloc and curative resection rate than EMR using a transparent cap, but with a similar complication rate.¹⁸²

Endoscopic resection is often combined with other forms of endoscopic ablative therapy such as photodynamic therapy (PDT), laser ablation, thermal ablation with argon plasma coagulation (APC), or radiofrequency ablation. Combination therapy is particularly applicable for larger lesions, multifocal lesions, or for treatment of flat high-grade dysplasia in patients with Barrett’s esophagus. Experience with EMR combined with other endoscopic ablative therapy is limited, but may be a viable alternative to surgery.¹⁸³ PDT has been studied as single-modality therapy. In a recent report, a complete response could be achieved in all 31 patients with early adenocarcinoma and Barrett’s esophagus treated with 5-aminolevulinic acid–based PDT, but local recurrence occurred in almost one third of patients.¹⁸⁴

Surgical Therapy

Although recent studies have reported excellent outcomes with endoscopic therapy for early esophageal cancers, there is a relative paucity of long-term follow-up data. Thus, the role of endoscopic therapy for early esophageal cancer is still a matter of controversy, particularly in the United States. Also, excellent outcomes after endoscopic therapy of early esophageal cancer are typically seen in specialized centers treating a small number of patients; it is unclear whether such excellent outcomes can be achieved in community practice.¹⁸⁵ In a recent study of early esophageal cancer that analyzed data from a large population-based cancer registry, cancer-free survival was similar in patients undergoing endoscopic therapy or surgical resection.¹⁸⁶ Obviously, management decisions for treatment of early esophageal cancer should be individualized. Although younger, otherwise healthy patients with larger mutifocal lesions with submucosal invasion may benefit from surgery, selected older patients with limited disease and multiple comorbidities may be more appropriate candidates for endoscopic therapy. When surgical resection is considered for treatment of early esophageal cancer without submucosal involvement, less morbid surgical techniques such as vagal-sparing esophagectomy, which has fewer postoperative complications, may be considered.¹⁸⁷

Radiation Therapy and Chemotherapy

The role of radiotherapy, with or without chemotherapy, as definitive treatment of early esophageal cancer has been evaluated. In a series of 141 Japanese patients with superficial esophageal cancer who were treated with external beam radiotherapy, the three-year survival rates were 90% and 70% for tumors limited to the mucosa and submucosa, respectively, and were comparable with reported surgical outcomes.¹⁸⁸ It is likely that concurrent chemotherapy may be effective in decreasing local recurrence, particularly in tumors with submucosal invasion. Addition of high-dose brachytherapy may decrease the rate of local recurrence, but at a higher risk of radiation toxicity.¹⁸⁹ Some investigators have reported excellent results using a combination of endoscopic resection followed by chemoradiotherapy or even using chemoradiotherapy after EMR for management of nodal recurrence.^190,191 Until controlled studies are conducted, radiotherapy (with or without concurrent chemotherapy) should be reserved for those patients with early esophageal cancer who cannot undergo endoscopic resection due to contraindications to this procedure (e.g., esophageal varices or severe cervical spinal disease) in whom endoscopic procedures may be more risky.

Surgical Therapy

Different surgical techniques have been used for primary management of esophageal cancer, the choice of which depend on factors such as tumor location, stage, extent of lymphadenectomy, replacement conduit planned, use of neoadjuvant therapy, and the preference of the surgeon. Overall, 30% to 59% of patients with esophageal cancer are candidates for potentially curative resection, with five-year survival rates ranging from 15% to 24%. However, the operative mortality and perioperative morbidity rates (e.g., cardiopulmonary events, infection, anastomotic leaks or strictures) vary from 4% to 10% and 26% to 41%, respectively.¹⁹²

Transhiatal esophagectomy and Ivor-Lewis transthoracic esophagectomy are the most common techniques used in the United States. Modifications of transthoracic esophagectomy, such as transthoracic total esophagectomy with node dissection and cervical esophagogastric anastomosis (the three-field technique) and left thoracoabdominal incision with gastric pull-up into the left chest, are also used frequently. These techniques have their own advantages and disadvantages. For example, a complete mediastinal exploration and lymphadenectomy is not possible with the transhiatal approach. With the transthoracic technique, there is greater likelihood of a troublesome intrathoracic anastomotic leakage and a higher incidence of postoperative bile reflux. In a prospective randomized trial, transhiatal esophagectomy was associated with lower morbidity than transthoracic esophagectomy with extended en-bloc lymphadenectomy.¹⁹³ However, a recent update of this cohort showed a 14% five-year survival advantage in patients with distal esophageal adenocarcinomas if they had been randomized to the transthoracic approach.¹⁹⁴ Despite this finding, an earlier meta-analysis of published trials, and a more recent population-based outcomes study found that transthoracic and transhiatal resections yielded similar long-term survival rates.^195,196 Published data also suggest that patients undergoing esophagectomy in high-volume hospitals may have less perioperative morbidity, lower mortality, and improved early clinical outcomes, although there are conflicting data on the effect of hospital volume on long-term survival.^197,198

Radiotherapy

Radiation therapy has been a cornerstone of treatment of esophageal cancer. Modern radiotherapy techniques such as three-dimensional conformal radiotherapy [3D-CRT] and intensity modulated radiotherapy [IMRT] have reduced toxicity from radiotherapy. Radiotherapy alone can result in long-term survival in patients with esophageal cancer, and reports of five-year survival of 21% to 59% have been published in selected patients with early disease. In a recent Chinese trial, 269 patients with resectable squamous esophageal cancer type were randomized to surgery versus late-course accelerated hyperfractionated conformal radiotherapy alone; the five-year survival rates were similar: 36.9% versus 34.7%, respectively.¹⁹⁹

Combined Chemoradiotherapy

More commonly, combined modality treatment with chemotherapy and radiotherapy is used. This approach is based on the argument that the antitumor effect is additive (related to radiation sensitization) and is based on outcome data from randomized control trials. Combined chemoradiotherapy can be definitive therapy or preoperative (neoadjuvant) therapy. Chemoradiotherapy can be used concurrently or administered sequentially.

In a Cochrane review of randomized trials, concomitant (but not sequential) chemoradiotherapy provided a significant reduction in mortality at one and two years. Combined chemoradiotherapy provided an absolute risk reduction for mortality by 9% and 4% after one year and two years, respectively, with a 12% reduction in local recurrence rate.²⁰⁰ However, there was significant toxicity related to such treatment. Thus, it appears that combined concurrent chemoradiotherapy, particularly using cisplatinum-based chemotherapy, is superior to radiotherapy alone in the setting of definitive therapy for locoregional disease (without surgical intervention), but better outcomes come with a cost of higher toxicity. Also, use of higher dosage of radiotherapy did not increase survival, but added to toxicity.²⁰¹

Given the poor long-term results with surgical resection alone, investigators have studied the effect of adding preoperative chemoradiotherapy on long-term survival. This neoadjuvant approach carries the theoretical promise of improved local tumor control and eradication of micrometastasis. Several randomized controlled trials have been published comparing neoadjuvant chemoradiotherapy with surgery alone in patients with localized cancer of the thoracic esophagus or esophagogastric junction. Although it appears that neoadjuvant chemoradiotherapy, particularly administered concurrently, may impart a modest survival advantage, the results from the various trials are not uniform. Two meta-analyses have addressed this issue. The first showed that compared with surgery alone, neoadjuvant chemoradiation followed by surgery improved three-year survival and reduced locoregional cancer recurrence.²⁰² The overall rate of resection attempt was reduced, but when resection was attempted, a higher rate of complete (R0) resection was achieved. There was also a trend toward increased treatment complications with chemoradiotherapy, and benefits were restricted to those undergoing concurrent chemoradiotherapy administration. In a second, more recent meta-analysis, pooled results from 10 randomized trials comparing neoadjuvant chemoradiotherapy with surgery alone in more than 1000 patients showed that the relative risk for all-cause mortality with neoadjuvant chemoradiotherapy versus surgery alone was 0.81 (95% CI 0.70 to 0.93), resulting in a 13% survival difference at two years. Results were similar in patients with squamous cell or adenocarcinoma of the esophagus.²⁰³ In order to consolidate the modest gains achieved by neoadjuvant therapy in this situation, several investigators have attempted intensification of the preoperative regimen by including cycles of induction chemotherapy prior to concurrently administered chemoradiotherapy. Also, other chemotherapeutic drugs with better radiosensitizing potential (such as paclitaxel, carboplatin, and oxaliplatin) have been studied, with some initial encouraging results. However, the strategy of intensification of the preoperative regimen has not been proven in randomized controlled trials. The concern remains whether the increased complications seen with such aggressive regimens will be justified by a corresponding increase in long-term survival.

Few trials have compared neoadjuvant chemotherapy (without radiation) with surgery alone. In the same meta-analysis referred to earlier, neoadjuvant chemotherapy was favored over surgery alone, albeit with a lower reduction in mortality compared with neoadjuvant chemoradiotherapy.²⁰³

Several issues remain unanswered in the management of locally advanced esophageal cancer and esophagogastric junctional cancers. It is not known if there is any difference in outcome in patients treated with primary chemoradiotherapy or with neoadjuvant chemoradiotherapy followed by surgery. Experience from few controlled studies suggests that patients treated with neoadjuvant chemoradiotherapy followed by surgery had better locoregional control and needed fewer palliative procedures, but had similar overall survival and quality of life compared with those treated with primary (definitive) therapy.^204,205 It is important to note that these conclusions are based on studies in which most patients had squamous cell cancer; in patients with adenocarcinoma, the case for the necessity of surgery after chemoradiotherapy is stronger. Salvage esophagectomy in patients with recurrence after primary (definitive) chemoradiotherapy may be a feasible option in a minority of patients, but such a strategy is plagued by two problems: first, difficulty in identifying recurrent disease in these patients and second, the increased risk of perioperative complications in patients undergoing salvage esophagectomy compared with primary esophagectomy or a planned esophagectomy after combined modality treatment.

Metastatic Cancer Chemotherapy

Similar to other advanced gastrointestinal cancers, systemic chemotherapy has a role in palliative treatment in patients with advanced, unresectable esophageal cancer, particularly those with distant metastatic disease. Although a large volume of published information exists in this area evaluating a large number of chemotherapy regimens, there is no convincing evidence that one particular regimen is superior to others. Based on randomized trials, it appears that a combination of cisplatinum, 5-fluorouracil (5-FU), and either epirubicin (an anthracyclic antineoplastic agent) or docitaxel is appropriate for first-line therapy. Capecitabine, the orally administered fluoropyrimidine, has attracted attention as a replacement agent for 5-FU infusion, because it eliminates the need for central venous access and an ambulatory infusion pump. Recent evidence suggests that capecitabine and oxaliplatin are as effective as 5-FU and cisplatin.²⁰⁶ Irinotecan-containing regimens have also been studied recently in advanced esophageal and esophagogastric junctional cancers.

Biological agents such as cetuximab (a monoclonal antibody against the EGF receptor), erlotinib (an orally active tyrosine kinase inhibitor), and bevacizumab (which targets VEGF) have been studied in a preliminary fashion as single agents or parts of combination chemotherapeutic regimens. In the recent Southwest Oncology Group (SWOG) trial, patients with unresectable or metastatic esophagogastric junctional adenocarcinoma had a reasonable response to erlotinib, with a very attractive toxicity profile.²⁰⁷ Experience with second-line chemotherapy in patients who fail initial chemotherapy regimens has been generally disappointing.

Endoscopic Palliative Treatment

Palliation of dysphagia, restoration or preservation of swallowing, and maintenance of nutritional status are the priority concern in planning palliative therapy and are the most important determinants of quality of life. Although systemic chemotherapy and radiation therapy (external beam or brachytherapy) can provide relief of dysphagia in the majority of patients, such relief is usually temporary. Therefore, most patients with unresectable esophageal or esophagogastric junctional cancers eventually become candidates for endoscopic palliative therapy for sustained relief of dysphagia. Endoscopic therapy for palliation of dysphagia can be divided into two categories: (1) techniques that ablate neoplastic tissue, and (2) techniques such as dilation and stent placement that displace neoplastic tissue.

Endoscopic tissue ablative techniques include contact thermal, noncontact thermal, cytotoxic injection, and photodynamic therapies. The electrosurgical tumor probe (BICAP tumor probe, ACMI Circon, Santa Barbara, Calif) was often used in the past for thermal contact ablation of bulky esophageal tumors and could establish esophageal patency, with improvement in dysphagia, in up to 80% of patients. However, this technique has been largely abandoned because of an unacceptable rate of life-threatening complications, such as tracheoesophageal fistula, esophageal perforation, and delayed hemorrhage in up to 20% of patients. It is interesting to note this contact thermal ablation device is the precursor of the radiofrequency ablation balloon device, which has recently become a popular form of endoscopic therapy for patient with dysplastic Barrett’s esophagus.

Laser Therapy

Laser photoablation using neodymium:yttrium-aluminum-garnet (Nd:YAG), potassium titanyl phosphate (KTP), and argon lasers has been used extensively in the palliation of malignant dysphagia associated with esophageal cancers (Fig. 46-13). The bulk of the experience has been attained with the Nd:YAG laser in the noncontact mode. The endoscopist usually passes the laser fiber through the accessory channel of the endoscope and with the tip of the fiber placed about 10 mm from the tumor tissue, a combination of tissue vaporization and coagulation necrosis is achieved with a typical Nd:YAG laser setting of 40 to 100 W, used either in a pulsed or continuous mode. The ablation is typically started at the distal margin of the tumor, moving proximally in a circumferential manner. However, if there is complete luminal obstruction, the laser ablation is started at the proximal extent of the tumor in order to try to carefully bore through the obstructive tumor tissue. In the majority of patients, more than one session is required to establish reasonable luminal patency, and the procedure is typically repeated at 48- to 72-hour intervals. Chest pain, odynophagia, low-grade fever, and leukocytosis may occur after the procedure. Follow-up endoscopy with a “touch-up” session in three to four weeks is often required, depending on the response to laser ablation. Reasonable improvement in dysphagia can be achieved in up to 70% of patients, which typically lasts three to six weeks and rarely beyond 12 months in absence of retreatment. Complication rates are approximately 4% in experienced hands, with a 1% procedural mortality rate. Polypoid, fleshy, noncircumferential lesions in the body of the esophagus are easier to treat with laser ablation. Anastomotic recurrences and tumor overgrowth in a previously placed endoprosthesis can also be treated by laser photoablation.

Figure 46-13. Laser therapy of esophageal cancer. A focal fungating adenocarcinoma (A) can be debulked for palliation of dysphagia with a noncontact laser (B), resulting in tumor coagulative necrosis and vaporization, with reconstitution of the esophageal lumen (C).

Photodynamic Therapy

PDT has a role in palliation of dysphagia in patients with advanced esophageal cancer (Fig. 46-14). Comparative trials suggest that PDT may be better than endoscopic laser ablation in terms of fewer perforations and better tumor response, particularly in proximal tumors and long tumors. Photosensitivity is a significant problem in up to 20% of patients. PDT has also been used for ablation of tumor in-growth after stent placement.

Figure 46-14. A, Photodynamic therapy of high-grade dysplasia in Barrett’s esophagus. This patient was not a candidate for esophagectomy. B, Forty-eight hours after photodynamic therapy there is intense, circumferential, superficial tissue destruction with sharply demarcated borders delineating the proximal extent of laser light exposure.

Esophageal Stent Placement

Use of esophageal stents for palliation of malignant dysphagia has increased rapidly and compares favorably with other endoscopic palliative techniques. Self-expanding covered metallic stents are the most effective palliative therapy for tracheoesophageal fistula. However, palliative stent placement in patients with esophageal cancer is not always followed by improvement of nutritional parameters.²⁰⁸

The rigid plastic esophageal stents have been replaced by self-expanding metal stents (SEMS) because of their familiarity, ease of placement, improved palliation of dysphagia, and better safety profile. The covered SEMS are the only effective practical treatment available for management of esophagorespiratory fistula (Fig. 46-15). SEMS are typically placed as an outpatient procedure under conscious sedation. The patient is placed in left lateral or supine position and an initial endoscopic examination is carefully performed to estimate the lesion’s location, length, configuration, luminal diameter, and relationship to the upper esophageal sphincter and esophagogastric junction. A recent contrast esophagogram is often helpful to plan the placement of the SEMS. Esophageal dilation is often performed prior to stent placement, but may not be always recommended or required. Most endoscopists use fluoroscopic guidance for placement, although one study questioned the utility of such a practice.²⁰⁹ Accurate marking of the tumor margins in relation to the distal and proximal end of the SEMS, and consideration of the stent foreshortening are critical to optimal placement. A post-placement barium contrast study should be performed routinely in patients with esophagorespiratory fistula to document effective sealing of the fistula. Most patients can be started on oral intake immediately after the endoscopic procedure. Patients should have individualized dietary instruction to prevent stent occlusion by food impaction. Antiemetic and antitussive drugs may be needed if the patient experiences significant postprocedure retching, coughing, or hiccups, both to relieve those symptoms and prevent stent dislodgement.

Figure 46-15. Stent therapy of an esophageal fistula. A, This patient with a circumferential esophageal carcinoma, previously treated with chemoradiotherapy, developed an esophagomediastinal fistula, seen inferiorly. B, Placement of a covered self-expanding metallic stent achieved long-term symptomatic palliation.

A variety of esophageal SEMSs are commercially available, most of which are covered and available in different lengths and diameters. The Wallstent II (Microvasive, Boston Scientific, Inc., Natick, Mass) consists of a bilayer chromium alloy tubular mesh coated with a polyurethane sleeve between two mesh tubes. The stent can be recaptured and repositioned when up to 50% of the stent has been deployed. The Wallstent is delivered on an introducer catheter over a previously placed guidewire and deployed by withdrawing a translucent outer sheath under fluoroscopic guidance. The Flamingo Wallstent is a popular modification of this covered stent, with an exaggerated proximal flange to reduce the risk of distal stent migration.

The Ultraflex stent (Microvasive, Boston Scientific, Inc., Natick, Mass) consists of a knitted nitinol wire tube. The stent is constrained on the introducer catheter by a spiral retention suture. Unraveling the suture deploys the stent. One unique advantage of this stent is that models are available with either distal or proximal release systems, allowing precise deployment. The delivery device is 5 to 7 mm in diameter and allows insertion of the endoscope alongside the stent delivery device for concomitant endoscopic and fluoroscopic observation during placement and deployment.

The Z-stent (Cook Medical, Inc., Winston-Salem, N.C.) is made of stainless steel wires shaped in a Z configuration and releases without shortening to a diameter of 18 mm. The version preferred by many endoscopists, because of a lower risk of stent migration, has uncovered segments in the proximal and distal sections. A model designed with a latex “wind-sock” extending from the distal opening of the stent (Dua stent) and intended to prevent gastroesophageal reflux has been marketed for esophagogastric junction and distal esophageal cancers, tumors in which the stent must bridge the esophagogastric junction. Another commercially available antireflux stent (FerX-Ella; Dr Karel Volenec, Ella CS, Hradec Králové, Czech Republic) has been used to minimize reflux of gastric contents. Clinical experience with these antireflux stents is limited, and randomized studies with small numbers of patients failed to show a difference in reflux symptoms or global quality of life.^210,211 Another introduced version of the Z-stent (Evolution Controlled-Release Stent, Cook Medical, Inc., Winston-Salem, N.C.) allows more controlled stepwise release and better ability to recapture and reposition the stent.

A covered hybrid esophageal metallic stent has been marketed (Allimax-E, Alveolus, Charlotte, N.C.). Some experts believe that covered metal stents are more likely to migrate compared with the bare metal uncovered stent. A double-layered stent, the Niti-S stent (Taewong Medical, Seoul, South Korea), with an inner polyurethane layer (to prevent tumor in-growth) and an outer uncovered nitinol wire tube (to allow the mesh of the stent to embed itself in the esophageal wall to minimize stent migration) has been marketed; initial experience with this stent has been promising.²¹²

The Polyflex (Boston Scientific, Inc., Natick, Mass) is a completely coated self-expanding nonmetallic (plastic) stent introduced for palliation of malignant and benign strictures. Advantages are its lower cost and potential of removability. Published series have shown the Polyflex stent to be safe and effective in relieving malignant dysphagia. In a randomized prospective trial of 101 patients with unresectable esophageal cancer, there was no difference between the Polyflex and Ultraflex stents in palliating dysphagia. However, Polyflex stents had a higher rate of complications such as tumor overgrowth, hyperplastic granulomatous reaction, food bolus impaction, and stent migration and may be more technically difficult to place.²¹³

There is little comparative information on the utility of different types of SEMS. A randomized trial comparing Flamingo, Ultraflex, and Z-stents failed to show superiority of one stent over another.²¹⁴ In another randomized study comparing Ultraflex, Niti-S, and Polyflex stents, all three stents adequately palliated malignant dysphagia, although Polyflex stents had a higher rate of migration.²¹⁵ The decision to place a particular type of stent should be individualized and is typically determined by assessment factors such as tumor length, location, and configuration; the patient’s performance status and overall prognosis; plans for further palliative therapy, such as chemoradiotherapy; and local expertise.

Available information reports good results with SEMS in palliating dysphagia and patients with esophagorespiratory fistula, with 70% to 100% reported success rates in sealing the fistula.²¹⁶ Complications occur in 30% to 40% of patients with SEMS, although most are minor. Common complications include tumor in-growth or tumor overgrowth (5% to 20%), stent migration (10%), chest pain, procedure-related perforation, food bolus impaction, bleeding, foreign body sensation, and reflux esophagitis. Also, aspiration of gastric contents is commonly seen in patients with a stent placed across the esophagogastric junction.

Brachytherapy

Brachytherapy administered intraluminally by a nasoesophageal applicator has gained attention for palliation of malignant dysphagia. In a randomized trial, single-dose brachytherapy was more effective than metal stent placement for palliation of dysphagia in patients with unresectable esophageal cancer, although dysphagia improved more rapidly in patients who were treated with the Ultraflex SEMS. The brachytherapy group also had fewer complications than the stent group, and costs were similar.²¹⁷ A prognostic score based on age, gender, WHO performance score, tumor length, and presence of metastasis has been suggested to identify patients with a poor prognosis in whom stent placement is at least equivalent to brachytherapy.²¹⁸ In a Markov model analysis using current levels of reimbursements, brachytherapy (and not SEMS) was shown to be the most cost-effective intervention for palliation of malignant dysphagia.²¹⁹ In a randomized trial of 53 patients with advanced esophageal cancer from China, a SEMS loaded with radioactive iodine (¹²⁵I) seeds yielded better palliation of malignant dysphagia and better survival than a conventional SEMS.²²⁰

Nutritional Therapy

Most patients with esophageal cancer are nutritionally compromised and need specific nutritional management for improvement of functional status before and after surgery, during chemoradiotherapy, and as an adjunct to other palliative measures. Success of endoscopic therapy in palliation of dysphagia does not necessarily translate into improved nutritional status, and many of these patients need establishment of enteral access by endoscopic, surgical, or radiologic procedures for maximizing nutritional intake. A surgical feeding jejunostomy should be performed in all patients undergoing esophagectomy; percutaneous endoscopic gastrostomy (PEG) is contraindicated in those patients who may be candidates for gastric pull-up surgery. In many patients with esophagectomy, or those with a SEMS placed across the esophagogastric junction, feeding the patient through a PEG tube may lead to recurrent aspiration, and in these patients a PEG with a jejunal feeding tube extension or even direct percutaneous endoscopic jejunostomy may be preferred. Some patients with markedly compromised nutritional status may benefit for short-term parenteral nutrition, especially during the perioperative period (see Chapters 4 and 5).

Analgesics

Many patients with terminal advanced esophageal cancer may suffer from pain from tumor ulceration or local invasion or from metastatic disease and need narcotic analgesic medications for control of pain. A pain medicine or palliative care consultation may be appropriate in such cases.

Multidisciplinary Care

It is clear that appropriate management of patients with esophageal cancer needs input from many providers from different specialties. A multidisciplinary team approach with a tailored treatment plan for each patient with esophageal cancer has been shown to result in improved staging, lower operative mortality, and improved five-year survival when compared with a group of patients managed by surgeons who were working independently.²²¹