Esophageal cancer

January 1, 2005
I. Benjamin Paz, MD

Jimmy J. Hwang, MD

Rajesh V. Iyer, MD

Lawrence R. Coia, MD

Although still relatively uncommon in Western countries, esophageal cancer isfatal in the vast majority of cases. In the United States, an estimated 14,520 newcases will be diagnosed in the year 2005, and 13,570 deaths will result from thedisease. This high percentage of deaths rivals that of pancreatic cancer and ismore than four times that of rectal cancer.

Although still relatively uncommon in Western countries, esophageal cancer isfatal in the vast majority of cases. In the United States, an estimated 14,520 newcases will be diagnosed in the year 2005, and 13,570 deaths will result from thedisease. This high percentage of deaths rivals that of pancreatic cancer and ismore than four times that of rectal cancer.The esophagus extends from the cricopharyngeal sphincter to the gastroesophageal(GE) junction and is commonly divided into the cervical, upper tomid-thoracic, and thoracic portions. This can be important, as histology andoptimal treatment approaches may vary considerably based on the site of thecancer. It may not be possible to determine the site of origin if the cancer involvesthe GE junction itself.EpidemiologyGender Esophageal cancer is seven times more common and slightly morelethal in men than in women.Age Adenocarcinoma of the esophagus (now more common in the UnitedStates than the squamous cell type) has a median age at diagnosis of 69 years.The incidence of squamous cell cancer of the esophagus increases with age aswell and peaks in the seventh decade of life.Race The incidence of squamous cell esophageal cancer is three times higherin blacks than in whites, whereas adenocarcinomas are more common in whitemen.Geography Evidence for an association between environment and diet andesophageal cancer comes from the profound differences in incidence observedin different parts of the world. Esophageal cancer occurs at a rate 20-30 timeshigher in China than in the United States. An esophageal "cancer belt" extendsfrom northeast China to the Middle East.Survival Although the overall outlook for patients diagnosed with esophagealcancer has improved in the past 30 years, most patients still present with advanceddisease, and their survival remains poor. One-third to one-half of patientstreated with either chemoradiation therapy or chemoradiation therapyplus surgery are alive at 2 years, without recurrence of esophageal cancer.Disease site The rate of cancer of the distal esophagus is about equal to that ofthe more proximal two-thirds. In general, squamous cell carcinoma is found inthe body of the esophagus, whereas adenocarcinoma predominates in lesionscloser to the GE junction.Etiology and risk factorsCigarettes and alcohol Squamous cell carcinomas of the esophagus have beenassociated with cigarette smoking and/or excessive alcohol intake. Furthermore,cigarette smoking and alcohol appear to act synergistically, producing highrelative risks in heavy users of tobacco and alcohol. Esophageal adenocarcinomais increased twofold in smokers.Diet High-fat, low-protein, and low-calorie diets have been shown to increasethe risk of esophageal cancer. Exposure to nitrosamines has been proposed asa factor in the development of both squamous cell carcinoma and adenocarcinomaof the esophagus.Barrett's esophagus and other factors Gastroesophageal reflux disease(GERD) and Barrett's esophagus (adenomatous metaplasia of the distal esophagus)have been linked to adenocarcinoma of the esophagus. Tylosis, Plummer-Vinson syndrome, history of head and neck cancer, and achalasia have alsobeen associated with a higher-than-normal risk of developing squamous cellcancer of the esophagus.Signs and symptomsBecause symptoms do not alert the patient until the disease is advanced, fewesophageal cancers are diagnosed at an early stage.Dysphagia The most common presenting complaint is dysphagia, which generallyis not noted until the esophageal lumen is narrowed to one-half to onethirdof normal, due to its elasticity.Weight loss is common and has a significant role in prognosis (> 10% of totalbody weight as poor prognosis).Cough that is induced by swallowing is suggestive of local extension into thetrachea with resultant tracheoesophageal fistula.Odynophagia and pain Pain with swallowing (odynophagia) is an ominoussign. Patients who describe pain radiating to the back may well have extraesophagealspread. Supraclavicular or cervical nodal metastases may be appreciatedon examination.Hoarseness may be a sign of recurrent laryngeal nerve involvement due toextraesophageal spread.Metastatic disease may present as malignant pleural effusion or ascites. Bonemetastasis can be identified by pain involving the affected site or by associatedhypercalcemia. The most common metastatic sites are retroperitoneal or celiaclymph nodes.The American College of Surgeons conducted a study utilizing its national cancerdatabase to assess the presentation, stage distribution, and treatment of pa-tients diagnosed with esophageal cancer between 1994 and 1997 (n = 5,044).The most common presenting symptoms were dysphagia (74.0%), weight loss(57.3%), reflux (20.5%), odynophagia (16.6%), and dyspnea (12.1%). Fifty percentof patients had tumors located in the lower third of the esophagus. Squamouscell histology was found in 51.6%, and 42.0% of patients had adenocarcinomas.Barrett's esophagus was found in 39% of those patients with adenocarcinoma.Patients undergoing initial surgical resection had the following stagedistribution: stage I (13.3%), II (34.7%), III (35.7%), and IV (12.3%).DiagnosisIn Western countries, the diagnosis of esophageal cancer is generally made byendoscopic biopsy of the esophagus. In the Far East, cytologic evaluation isfrequently utilized.Endoscopic ultrasonography (EUS) is extremely accurate (> 90%) in establishingthe depth of tumor invasion (T stage) but less accurate (70%-80%) indetermining nodal involvement (N stage) unless combined with fine-needleaspiration (FNA) of the involved nodes (93% accuracy) when nodes greaterthan 5 mm are biopsied. The addition of FNA increases the sensitivity from63% to 93% and the specificity from 81% to 100%. EUS is not reliable in determiningthe extent of response to neoadjuvant treatment.Endoscopy and barium x-rays Endoscopy allows for direct visualization ofabnormalities and directed biopsies. Barium x-rays are less invasive and providea good assessment of the extent of esophageal disease.Bronchoscopy should be performed to detect tracheal invasion in all cases ofesophageal cancer except adenocarcinoma of the distal third of the esophagus.CT scan Once a diagnosis has been established and careful physical examinationand routine blood tests have been performed, a CT scan of the chest, abdomen,and pelvis should be obtained to help assess tumor extent, nodal involvement,and metastatic disease.PET A prospective trial designed to evaluate the utility of PET vs CT and EUSwas performed by obtaining these studies in 48 consecutive patients prior toesophagectomy. PET achieved a 57% sensitivity, 97% specificity, and 86% accuracycompared with CT, which was 99% sensitive, 18% specific, and 78% accurate.In terms of nodal staging, PET was correct in 83% of cases, as comparedwith 60% of cases for CT and 58% for EUS (P = .006). This analysis suggests theimproved accuracy of PET in the staging work-up of patients with esophagealcancer.Numerous studies report the accuracy of PET scanning in determining the presenceof metastatic disease, with sensitivity approaching 90% and specificityover 90%.As PET becomes more widely available, its use will probably become an importantpart of the preoperative evaluation of these patients. In a prospectivetrial of 39 patients with esophageal cancer, PET detected additional sites ofmetastatic disease at the initial evaluation when compared with conventionalimaging. After induction therapy, PET did notadd to the estimation of locoregional resectabilityand did not detect new distant metastases.However, this study suggested thatchanges in [18fluorodeoxyglucose] FDG-PETfollowing induction therapy may predict disease-free and overall survival after inductiontherapy and resection in patients with esophagealcancer. A large prospective national trialwill evaluate the use of PET scan in the treatmentof esophageal cancer.Bone scan A bone scan should be obtainedif the patient has bone pain or an elevated alkalinephosphatase level.Thoracoscopy/laparoscopy Investigators have recently begun to examine therole of surgical staging prior to definitive therapy. These procedures are designedto allow pathologic review of regional lymph nodes and the accurateassessment of extraesophageal tumor spread by direct visualization. A recentlycompleted multi-institution trial (Cancer and Leukemia Group B [CALGB]9380) found these procedures to be feasible in over 70% of patients; they resultedin the upstaging of patients in 38% of cases reviewed. Further investigationsneed to be completed to determine the appropriate use of these tools intreatment algorithms for patients with esophageal cancer.Warning Staging studies should be undertaken only if management wouldchange on the basis of specific findings.Screening and surveillanceHIGH-RISK PATIENTSAdenocarcinoma The role of screening patients with GERD and surveillanceof patients with Barrett's esophagus by upper GI endoscopy remains underinvestigation. In 833 patients studied by endoscopy, there was a 13% incidenceof intestinal metaplasia (Barrett's esophagus). Dysplasia or cancer was seen in31% of patients with long-segment Barrett's esophagus, in 10% of short-segmentBarrett's esophagus, and in 6% of GE-junction intestinal metaplasia.Squamous cell carcinoma Mass screening in the high-risk areas of China andJapan is considered appropriate.PathologyAdenocarcinoma The incidence of esophageal adenocarcinoma involving theGE junction has risen 4%-10% per year since 1976 in the United States andEurope. As a result, adenocarcinoma is now the predominant histologic subtypeof esophageal cancer. The distal one-third of the esophagus is the site oforigin of most adenocarcinomas.

Squamous cell carcinomas occur most often in the proximal two-thirds ofthe esophagus. Squamous cell carcinoma is still the most prevalent histologicsubtype worldwide.Other tumor types Other, less frequently seen histologic subtypes includemucoepidermoid carcinoma, small-cell carcinoma, sarcoma, adenoid cysticleiomyosarcoma, and primary lymphoma of the esophagus. Occasionally, metastaticdisease from another site may present as a mass in the esophagus or amass pressing on the esophagus.Metastatic spread The most common sites of metastatic disease are the regionallymph nodes, lungs, liver, bone, adrenal glands, and diaphragm. Adenocarcinomacan also metastasize to the brain.Staging and prognosisBased on data demonstrating that the depth of penetration has important prognosticsignificance, the American Joint Committee on Cancer (AJCC) TNMstaging system for esophageal cancer was changed from a clinical one (1983) toa pathologic one in 2002. Both the clinical and pathologic staging systems areshown in Table 1, as patients may be cured without an operation. Althoughpathologic information obtained from an esophagectomy specimen is of prognosticimportance, postoperative therapy to improve prognosis has not beenrigorously tested. Moreover, recurrence rates for stage I (30%) and stage II(70%) cancers suggest early systemic spread undetected by current noninvasivestaging.

Pathologic information obtained from an esophagectomy specimen is of significantprognostic importance. Immunohistochemical analysis of the initialbiopsy specimen may also have prognostic relevance. Clinical staging has beenshown to be of prognostic importance, particularly in patients managed withprimary radiotherapy or chemoradiation therapy.Histology and grade Neither histology nor grade has been shown to be ofprognostic importance in esophageal carcinoma.Other prognostic factors Patient age, performance status, and degree ofweight loss are of prognostic importance. The prognostic implications of tumor-suppressor genes and oncogenes are an area of active investigation.TreatmentTreatment options for the various disease stages are given in Table 2, alongwith 5-year survival rates.TREATMENT OF LOCALIZED DISEASE
Only 40%-60% of patients with esophageal cancer present with clinically localizeddisease. National Comprehensive Cancer Network (NCCN) guidelinesstate that patients with clinically localized disease may be treated with resectionor chemotherapy plus irradiation (Tables 3 and 4). The overall 5-year survivalrates for either surgery alone or combined chemotherapy and irradiation appearequivalent.Chemoradiation therapy as primary management of localized or locoregionallyconfined esophageal cancer has been shown to be superior to irradiationalone. A series of randomized trials have demonstrated that adjuvant postoperativechemoradiation therapy does not offer a survival advantage to patientswith esophageal cancer. Adequate patient selection, tumor staging, and treat-

ment standardization will be required before the optimal therapeutic modalitiesin these patients will be determined.SurgeryPreoperative medical evaluation helps determine the patient's risk of developingpostoperative complications and mortality. In addition to the stagingand nutritional status, it should include an evaluation of the pulmonary, cardiac,renal, and hepatic functions.Extent of surgical resection The extent of resection depends on the locationof the primary tumor, histology of the tumor, and nature of the procedure(palliative vs curative). Retrospective study has reported that superficial mucosallesions may be treated via endoscopic mucosal resection, but those patientswith submucosal invasion require esophagectomy. For tumors of the intrathoracicesophagus (squamous cell carcinomas) and tumors with extensiveBarrett's esophagus (adenocarcinomas), it is necessary to perform a totalesophagectomy with cervical anastomosis to achieve a complete resection. Fordistal lesions of the abdominal esophagus (adenocarcinomas) and cardia, it isoften possible to perform an intrathoracic esophageal anastomosis above theazygus vein, although many surgeons would prefer to perform a totalesophagectomy.The resected esophagus may be replacedwith tubularized stomach in patients withtumors of the intrathoracic esophagus or witha colon interposition in patients with tumorsinvolving the proximal stomach, since suchinvolvement makes this organ unsuitable foresophageal reconstruction. The esophagealreplacement is usually brought up throughthe posterior mediastinum, although theretrosternal route is often used in palliativeprocedures.Patient selection The indications foresophagectomy in esophageal cancer varyfrom center to center within the UnitedStates.Clearly, patients with distant metastases, evidenceof nodal metastases in more than onenodal basin, or tumor extension outside theesophagus (airway, mediastinum, vocal cordparalysis) are candidates for palliativetherapy. Patients with disease limited to theesophagus and no evidence of nodal metastases(stages I and IIA) may be treated with esophagectomy, although thesepatients can also be considered for definitive treatment with chemoradiationtherapy.Method of resection Considerable controversy also exists among surgeonsregarding the method of resection. To date, two randomized studies have comparedtranshiatal esophagectomy (without thoracotomy) with the Ivor-Lewis(transthoracic) esophagectomy (with thoracotomy). These studies failed to showdifferences between the two procedures with regard to operative morbidityand mortality. In a recent randomized trial of 220 patients treated either with atransthoracic or transhiatal esophageal resection, there was a trend toward animprovement in 5-year survival. A meta-analysis failed to show differences in5-year survival rates. Over the past 5 years, successful attempts have been madeto use minimally invasive approaches to esophageal cancer with thoracoscopyand laparoscopy. Although those studies have shown a decrease in morbidityand the minimally invasive approach appears to be oncologically sound fromthe point of view of resection margins, the number of nodes resected is still notcomparable to that of the standard transthoracic approach.The need for pyloric drainage (pyloroplasty) following esophagectomy is anotherarea of debate. A meta-analysis of nine randomized trials that included553 patients showed a trend favoring pyloric drainage in improving gastricemptying and nutritional status, whereas bile reflux was better in the nondrainagegroup. The gastric emptying time evaluated by scintigraphy was twice as longin the nondrainage group than in the pyloric drainage groups.Lymphadenectomy Considerable controversy exists regarding the need forradical lymphadenectomy in esophageal disease. Much of the controversy isdue to the fact that different diseases are being compared.Japanese series include mostly patients with squamous cell carcinomas of theintrathoracic esophagus, with 80% of the tumors located in the proximal andmiddle sections of the esophagus. Americans report combined series, with atleast 40%-50% of patients with adenocarcinomas of the distal esophagus. Skinnerand DeMeester favor en bloc esophagectomy with radical (mediastinal andabdominal) lymphadenectomy, based on 5-year survival rates of 40%-50% inpatients with stage II disease, as compared with rates of 14%-22% in historiccontrols.In a retrospective study, Akiyama found a 28% incidence of cervical nodemetastases in patients with squamous cell carcinomas located in the middleand distal portions of the esophagus, as opposed to 46% in those with tumors ofthe proximal third. Overall survival at 5 years was significantly better in patientswho underwent extended lymphadenectomy (three fields) than in thosewho had conventional lymphadenectomy (two fields); this finding was true inpatients with negative nodes (84% and 55%, respectively) and in those withpositive nodes (43% and 28%, respectively). Extended lymphadenectomy affordedno survival advantage in patients with tumors in the distal third of theesophagus.In a study of 1,000 patients with esophagogastric junction adenocarcinomas,the tumors were classified according to the location of the center of the tumormass in adenocarcinomas of the distal esophagus, cardia, and subcardia. Thetumors located in the cardia and subcardia regions spread primarily to theparagastric and left gastric vessel nodes and did not benefit from extendedesophagectomy.Preoperative chemotherapy
The frequency of metastatic disease as the cause of death in patients with esophagealcancer has resulted in exploration of the early application of systemictherapy in the treatment of esophageal cancer. The first of the two large studieswas intergroup study 113. A total of 440 patients were treated with surgicalresection alone or preceded by three cycles of cisplatin and 5-FU. Objectiveresponses were reported in only 19% of patients receiving chemotherapy. Nodifference in resectability, operative mortality, median survival (14.9 monthswith chemotherapy vs 16.1 months with surgery alone), or 2-year survival (35%vs 37%) was reported.However, the Medical Research Council evaluated 802 patients with resectableesophageal cancer in a similar study. Patients randomized to receive chemotherapywere administered two cycles of cisplatin (80 mg/m2) and 5-FU (1 g/m2/d as a continuous infusion for 4 days). Microscopically complete resectionwas performed more frequently in patients receiving chemotherapy, with nodifference found in postoperative complications or mortality. Moreover, patientsreceiving neoadjuvant chemotherapy had significantly longer mediansurvivals (16.8 months vs 13.3 months) and 2-year survivals (43% vs 34%) thanpatients treated with surgery alone. The reasons for the differences in the outcomesare unclear but may be related to the chemotherapy regimen and scheduleemployed in the intergroup study, patient population, or study design. As aresult, the role of neoadjuvant chemotherapy remains in question but is promising,especially with the potentially more efficacious newer generation of chemotherapyagents.Polee et al have evaluated a biweekly combination of cisplatin and paclitaxel inthis setting in a phase II study, with promising results. Objective responsesoccurred in 59% of 49 patients. No patients had progressive disease. Although71% of patients had severe neutropenia, it was often asymptomatic. Forty-sevenpatients underwent resection subsequently. Complete pathologic responsesoccurred in 14% of patients. The median survival of patients in this study was20 months, but it was 32 months in patients who had disease responsive tochemotherapy. The 3-year survival rate was 32%.Given the uncertainty about the efficacy of preoperative chemotherapy andchemoradiation therapy, some investigators have administered preoperativechemotherapy, followed by chemoradiation therapy, then surgery. The trueutility of this approach will need to be defined by randomized studies, butclearly, the approach is feasible, without a significant increase in toxicity oroperative morbidity. Interestingly, these reports have also demonstrated thatmost patients had significant improvement or resolution of dysphagia with theinduction chemotherapy alone.Radiotherapy
Although radiotherapy alone is inferior to chemoradiation therapy in the managementof locoregionally confined esophageal cancer, it may offer palliationto patients with advanced local disease toofrail for chemotherapy.Preoperative radiotherapy has beenshown to be of little value in convertingunresectable cancers into resectable ones orin improving survival. However, it decreasesthe incidence of locoregional recurrence.Postoperative radiotherapy (usually to 50or 60 Gy) can decrease locoregional failurefollowing curative resection but has no effecton survival.Brachytherapy Intraluminal isotope radiotherapy(intracavitary brachytherapy) allowshigh doses of radiation to be delivered to asmall volume of tissue. Retrospective studiessuggest that a brachytherapy boost mayresult in improved rates of local control andsurvival over external-beam radiotherapyalone. This technique can be associated witha high rate of morbidity if not used carefully.A multi-institution prospective trial was conductedby the RTOG to determine the feasibilityand toxicity of chemotherapy, externalbeamirradiation, and esophageal brachytherapyin potentially curable patients withesophageal cancer. Nearly 70% of patientswere able to complete external-beam irradiation,brachytherapy, and at least two cycles of5-FU/cisplatin. The median survival was 11months, and the 1-year survival was 49%.Because of the 12% incidence of fistula formation,the investigators urged caution in theroutine application of brachytherapy as partof a definitive treatment plan.Chemoradiation therapyPreoperative chemoradiation therapy Initialtrials of preoperative chemoradiationtherapy reported unacceptably high operative mortality (~26%). Subsequenttrials reported operative mortality of 4%-11%, median survival as long as 29months, and 5-year survival rates as high as 34%. In general, 25%-30% of patientshave no residual tumor in the resected specimen, and this group tends tohave a higher survival rate than those who have a residual tumor discovered bythe pathologist.The superiority of preoperative chemoradiation therapy over surgery alone inesophageal adenocarcinoma has been demonstratedin a prospective trial. This trialincluded 113 patients with adenocarcinomaof the esophagus. These patients were randomizedto receive either preoperativechemoradiation therapy (two courses of fluorouracil[5-FU] and cisplatin given concurrentlywith 40 Gy of radiotherapy in 15 fractions)or surgery alone. Median survival wasstatistically superior in the combined-modalityarm than in the surgery-alone arm (16 vs11 months). Rates of 3-year survival againstatistically favored the combined-modalityarm (32% vs 6%). Although toxicity was notsevere, the short survival in the surgery controlarm has minimized the impact of theseresults in the United States.A trial performed at the University of Michiganenrolled 100 patients and randomizedthem to undergo surgery alone or preoperative chemoradiation therapy(cisplatin, 20 mg/m2/d), 5-FU (300 mg2/d), and vinblastine (1 mg/m2/d)and radiotherapy (45 Gy/1.5 Gy bid). With a median follow-up of 8.2 years,the 3-year survival was 16% (surgery alone) vs 30% (induction chemoradiationtherapy). This difference did not reach statistical significance, as the study wasdesigned to detect a relatively large increase in median survival from 1 to 2.2years. A meta-analysis of randomized trials comparing neoadjuvantchemoradiation therapy followed by surgery with surgery alone found thatneoadjuvant concurrent chemoradiation therapy improved 3-year survival (oddsratio, 0.66) compared with surgery alone, with a nonsignificant trend towardincreased mortality with neoadjuvant treatment.Newer chemotherapy agents are active and may improve outcome over theseolder trials. A phase II trial of 129 patients employed paclitaxel/carboplatin[Paraplatin]/5-FU with 45 Gy of radiation therapy followed by esophagectomy.A pathologic complete response was seen in 38% of patients, with a mediansurvival of 22 months and a 3-year survival of 41%.Another phase II trial from the University of Michigan administered paclitaxel/cisplatin with 45 Gy of radiation therapy twice daily (1.5 Gy bid). In this study,19% of patients exhibited a pathologic complete response, with a 24-monthmedian survival and a 3-year survival of 34%. Ongoing studies by the RadiationTherapy Oncology Group (RTOG) are employing these newer chemotherapyagents.Primary chemoradiation therapy Patients with locally advanced esophagealcancer (T1-4 N0-1 M0) may be cured with definitive chemoradiation therapy.Randomized trials have demonstrated a survival advantage for chemoradiationtherapy over radiotherapy alone in the treatment of esophageal cancer. In anRTOG randomized trial involving 129 patients with esophageal cancer, irradiation(50 Gy) with concurrent cisplatin and 5-FU provided a significant survivaladvantage (27% vs 0% at 5 years) and improved local control over radiationtherapy alone (64 Gy). Median survival also was significantly better in thecombined-therapy arm than in the irradiation arm (14.1 vs 9.3 months).A recently completed randomized intergroup trial was designed to investigatethe role of high-dose irradiation in conjunction with systemic therapy. Thisstudy compared doses of 50.4 Gy with 64.8 Gy. Both treatment arms of thestudy administered concurrent 5-FU and cisplatin. This trial was stopped afteran interim analysis revealed no statistically significant difference in survivalbetween the two groups. The authors concluded that higher dose radiationtherapy did not offer any survival benefit compared with the 50.4-Gy dose.Patient selection Patients with disease involving the mid to proximal esophagusare excellent candidates for definitive chemoradiation therapy because resectionin this area can be associated with greater morbidity than resection ofmore distal tumors.Most of the trials demonstrating the efficacy of chemoradiation therapy havehad a high proportion of patients with squamous cell cancers. Chemoradiationtherapy has thus become a standard treatment of locoregionally confined squamous cell cancer of the esophagus. It is essential that chemotherapy be givenconcurrently with irradiation when this approach is chosen as primary treatmentfor esophageal cancer. A typical regimen is 50-60 Gy over 5-6 weeks,with cisplatin (75 mg/m2) and 5-FU (1 g/m2/24 h for 4 days) on weeks 1, 5, 8,and 11.The literature also supports offering patients with adenocarcinoma primarysurgery, preoperative chemoradiation therapy, or primary chemoradiationtherapy with surgical salvage if necessary. Entering these patients on protocolswill allow us to further define standard treatment.Sequential preoperative chemotherapy and radiation therapy Only modestbenefits have been found with preoperative chemoradiation therapy todate, with systemic failure continuing to be an important problem. Thus, sequentialtherapy with chemotherapy followed by chemoradiation therapy hasbeen explored.Ajani et al reported a series of 43 patients who received 12 weeks of cisplatinand irinotecan (Camptosar) followed by weekly paclitaxel with infusional5-FU and concurrent radiation therapy (4,500 cGy) and then esophagectomy.Therapy was well tolerated, with no deaths from chemotherapy orchemoradiation therapy, and an operative mortality rate of 5%. Cisplatin andirinotecan induced responses in 37% of patients, and 91% of patients underwentcomplete resection. Pathologic complete responses occurred in 26% ofpatients, and some tumor shrinkage was noted in 63% of patients. With amedian follow-up of more than 30 months, the median disease progressionfreesurvival was 10. 2 months, the median survival was 22.1 months, and the2-year survival was 42%. The patients who had a pathologic response to therapyhad significantly better outcomes than the rest of the study population. However,systemic recurrences remained a prominent cause of failure, with fivepatients experiencing recurrence first in the brain and an additional five patients,in the liver.Esophagectomy following chemoradiation therapy Considerable controversyexists regarding the need for esophagectomy following chemoradiationtherapy. The incidence of residual disease in patients who have a completeclinical response to chemoradiation therapy is 40%-50%. Patients with completeresponse following chemoradiation therapy have the best survival rateswith surgery.TREATMENT OF ADVANCED DISEASE
The goal of esophageal cancer treatment is generally palliative for patients withbulky or extensive retroperitoneal lymph nodes or distant metastatic disease.Therapeutic approaches should temper treatment-related morbidity with theoverall dismal outlook.Local treatment In patients with a good performance status, the combinationof 5-FU/mitomycin (Mutamycin), or 5-FU/cisplatin, and radiotherapy (50 Gy)results in a median survival of 7-9 months. This regimen usually renders patientsfree of dysphagia until death.Photodynamic therapy (PDT) Porfimer sodium (Photofrin) and an argonpumpeddye laser can provide effective palliation of dysphagia in patients withesophageal cancer. A prospective, randomized multicenter trial comparing PDTwith neodymium/yttrium-aluminum-garnet (Nd:YAG) laser therapy in 236patients with advanced esophageal cancer found that improvement of dysphagiawas equivalent with the two treatments.A recent review of 119 patients treated with endoluminal palliation reported asignificant improvement in dysphagia scores and an increased ability to relievestenosis caused by tumor when PDT was used in conjunction with laser therapyand irradiation.Other approaches include external-beam radiotherapy with or without intracavitarybrachytherapy boost, simple dilatation, placement of stents, andlaser recannulization of the esophageal lumen.Palliative resection for esophageal cancer is rarely warranted, although it doesprovide relief from dysphagia in some patients.Chemotherapy Recently published phase I and II studies have demonstratedmoderate response rates to taxanes in esophageal cancer. Taxanes in combinationwith platinum compounds and fluoropyrimidines are being tested in regimenswith irradiation.Although chemotherapy alone may produce an occasional long-term remission,there is no standard regimen for patients with metastatic cancer. Patientswith advanced disease should be encouraged to participate in well-designedtrials exploring novel agents and chemotherapy combinations.CHEMOTHERAPY IN ADVANCED ESOPHAGEAL CANCER
Advances have been made in the treatment of lower GI malignancies withnew chemotherapeutic agents, particularly irinotecan and oxaliplatin(Eloxatin). Progress has been slower, however, for treatment of upper GImalignancies. Given the different etiologies of the predominant forms of esophagealcancer, one problem in assessing the efficacy of chemotherapy in thisdisease is the fact that it remains uncertain whether squamous cell carcinomasand adenocarcinomas respond differently to chemotherapy. To date, bothhistologies appear to respond similarly to chemotherapy.Limited data from randomized studies are available regarding the use of chemotherapyin metastatic esophageal cancer. Most current data come fromphase II studies, making treatment recommendations difficult. In light of thelimited efficacy of current chemotherapeutic agents and available data aboutthe newer agents, no standard therapy for advanced esophageal carcinomapresently exists.The mainstay of therapy for advanced esophageal cancer has been cisplatin,which produces responses in approximately 15%-35% of patients, with a mediansurvival of about 4-6 months. With the addition of 5-FU as a 5-day continuousinfusion (in a manner similar to its use in squamous cell carcinoma ofthe head and neck), a modest increase in response rates to approximately 25%-35% occurs, but with no clear improvement in survival rates.In Britain, the ECF regimen, a combination of epirubicin (Ellence, 50 mg/m2)and cisplatin (60 mg/m2), both repeated every 21 days, with continuous infusionof 5-FU (200 mg/m2/d), is considered to be a standard regimen for advancedesophagogastric cancers. Several phase III studies have been performedand consistently demonstrated objective responses in about 40% of patients,with a median survival of 9 months and 1-year survival of 36%-40%. The mainsevere toxicities of this regimen are neutropenia, in about one-third of patients(32%-36%), lethargy (18%), and nausea and vomiting (11%-17%).Attempts to improve the efficacy of this regimen, which is logistically difficultto administer, are ongoing. Given the accepted role of capecitabine (Xeloda), a5-FU prodrug, in colorectal and breast cancers, as well as oxaliplatin, a studyevaluating the potential roles of these drugs is being performed. Sumpter et alhave reported the preliminary data from their study, with results suggestingsimilar efficacy and toxicity with the different combinations. The final resultswill not be available for several years but are eagerly awaited. A regimen withintravenous chemotherapy on one day combined with oral capecitabine, ratherthan infusional 5-FU, would gain much greater acceptance than ECF if theefficacy and toxicity of the regimens are similar.With the advent of many new chemotherapeutic agents (the taxoids, paclitaxeland docetaxel [Taxotere], irinotecan, and gemcitabine [Gemzar]) with varyingmechanisms of activity, further studies have been conducted, and each of thesedrugs has demonstrated activity, with responses achieved in approximately 15%-30% of patients (Table 5).However, the primary route of investigation for these new agents has beenin combination with cisplatin and/or 5-FU. The results available to datesuggest promising activity, with response rates often around 50% in phaseII studies. Irinotecan (65 mg/m2) and cisplatin (30 mg/m2) administered weeklyfor 4 weeks every 6 weeks have also been active, with responses in 20 of 35patients (57%) and an impressive median survival of 14.6 months.Following the example of the combination of irinotecan and 5-FU/leucovorinin colon cancer, the investigators have explored a modification of that schedule,with therapy administered for 2 weeks, with cycles repeated every 3 weeks.This simple modification has been investigated in 27 patients and was welltolerated, with severe neutropenia in 18% of patients and severe diarrhea in11% of patients.Paclitaxel (180 mg/m2 over 3 hours) and cisplatin (60 mg/m2 over 3 hours)administered every 14 days have been extensively evaluated in Europe andwere reported to produce objective responses in 43% of 51 patients, includingtwo complete responses, and 43% of patients were alive 1 year after initiationof therapy (Table 6).Because of the toxic and logistic difficulties of using cisplatin, carboplatin hasbecome a popular chemotherapeutic drug. Several studies of carboplatin withpaclitaxel have been undertaken in esophageal cancer. El-Rayes et al administeredcarboplatin (area under the curve of 5) with paclitaxel (200 mg/m2over 3 hours) every 3 weeks in 33 chemotherapy-naive patients. Objective

responses were reported in 45% of patients, with a median survival of 9 monthsand 1-year and 2-year survival of 43% and 17%.Polee et al explored a weekly schedule of these drugs in a phase I study. Withtherapy administered for 3 consecutive weeks, followed by a 1-week break, adose of carboplatin (area under the curve of 4) with paclitaxel (100 mg/m2)was recommended for further investigation. Responses were noted in half ofthe 40 patients, with a median survival of 11 months and 1-year survival of46%. Both of these combinations were well tolerated, with the primary toxicityof myelosuppression.The antimetabolite gemcitabine has also been evaluated in combination withcisplatin in esophageal cancer. A Southwest Oncology Group Study reportedby Urba combined 1,000 mg/m2 of gemcitabine on days 1, 8, and 15 with100 mg/m2 of cisplatin on day 15 in 64 patients. Approximately one-quarterof these patients had received prior chemotherapy. Therapy was well tolerated,with severe neutropenia occurring in 31% of patients. The median survivalof patients treated on this study was 7.3 months, and the 1-year survivalwas 20%. However, the heterogeneity of the patient population makes theefficacy of the therapy somewhat difficult to assess.Kroep et al tested a somewhat different schedule of these drugs, based onpreclinical data suggesting that cytotoxicity is higher when cisplatin is administered24 hours before gemcitabine. Thus, cisplatin (50 mg/m2) was given ondays 1 and 8, and gemcitabine (800 mg/m2 over 30 minutes), on days 2, 9,and 16, with cycles repeated every 28 days. Objective responses were notedin 39% of 36 patients, with stable disease in 17 patients, and a median survivalof 9.8 months. Aside from myelosuppression, this combination was also welltolerated.Oxaliplatin may also have a role in the treatment of esophageal cancer, both asa radiosensitizer and an agent in advanced disease. A preliminary report dem-

onstrated objective responses in 48% of 29 evaluable patients treated withoxaliplatin, 5-FU, and leucovorin according to the FOLFOX 4 schedule(oxaliplatin, 85 mg/m2 on day 1; leucovorin, 500 mg/m2 over 2 hours on days1 and 2; and 5-FU, 400 mg/m2 bolus, then 600 mg/m2 over 22 hours on days 1and 2, repeated every 14 days).The primary toxicity of these regimens is severe neutropenia, occurring in about40%-70% of patients. Severe diarrhea, nausea, and vomiting occur in ~10%-15% of patients in many studies. Fatigue and asthenia also were significantside effects with both therapies.With improving toxicity profiles and modest improvements in therapeuticoutcomes, second-line therapy for advanced esophageal cancer is also increasinglybeing explored. Muro et al treated 28 Japanese patients with squamouscell carcinoma of the esophagus with docetaxel (70 mg/m2) every 3 weeks;these patients had previously received cisplatin and 5-FU. As expected, severeneutropenia was the dominant toxicity (88%, including nine episodes offebrile neutropenia), with severe anorexia, fatigue, and anemia also reported.Objective responses were noted in 16% of these patients.In addition, Lordick et al determined that irinotecan (55 mg/m2) and docetaxel(25 mg/m2) given on days 1, 8, and 15, with cycles repeated every 28 days,were tolerable, with severe asthenia in 21% of 24 patients and severe diarrheain 13%. However, only three partial responses (13%) and eight patients withstable disease were noted, with a resultant median survival of 26 weeks. Althoughthese studies suggest the feasibility of second-line cytotoxic chemo-therapy in esophageal cancer, the significanttoxicity and limited objective response ratewarrant its use only with caution and preferablyon a clinical study.Novel agents are also being actively investigated,with great hope for the future. Thecombination of paclitaxel with the cyclin-dependentkinase inhibitor flavopiridol hasdemonstrated promising activity in patientswith esophageal cancer. A phase I study ofthis combination reported one complete, onepartial, and one minor response in seven patientstreated.In addition to flavopiridol and the epidermalgrowth factor receptor antagonists, it isanticipated that other targeted therapies,such as vascular endothelial growth factorantagonists, will be evaluated in this disease.Such avenues of exploration, in addition toearly diagnosis and therapy for early-stagedisease, are the most likely path toward significantimprovements in the therapy ofesophageal cancer.


Ajani JA, Walsh G, Komaki R, et al:

Preoperativeinduction of CPT-111 and cisplatin chemotherapyfollowed by chemoradiotherapy in patients withlocoregional carcinoma of the esophagus or gastroesophageal junction. Cancer 100:2347–2354, 2004.

Armanios MY, Xu R, Forastiere A, et al:

Phase II adjuvant chemotherapy for resectedadenocarcinoma of the esophagus, gastroesophageal (GE) junction and cardia (E8296):A trial of the Eastern Cooperative Group (abstract). Proc Am Soc Clin Oncol 22:296,2003.

Bains MS, Stojadinovic A, Minsky B, et al:

A phase II trial of preoperative combinedmodalitytherapy for localized esophageal carcinoma: Initial results. J Thorac CardiovascSurg 124:270–277, 2002.

Coia LR, Minsky BD, Berkey BA, et al:

Outcome of patients receiving radiation forcancer of the esophagus: Results of the 1992-1994 Patterns of Care Study. J Clin Oncol18:455–462, 2000.

DeYoung CM, Suntharalingam M, Line BR, et al:

The ability of whole body FDG18PET imaging to predict pathologic response to induction chemotherapy in locally advancedesophageal cancer: A prospective phase II trial. Proc ASTRO 57:165–166, 2003.

Downey RJ, Akhurst T, Ilson D, et al:

Whole body 18FDG-PET and the response ofesophageal cancer to induction therapy: Results of a prospective trial. J Clin Oncol21:428–432, 2003.

Fujita H, Sueyoshi S, Yamana H, et al:

Optimum treatment strategy for superficialesophageal cancer: Endoscopic mucosal resection versus radical esophagectomy. WorldJ Surg 25:424-431, 2001.

Gerson LB, Triadafilopoulos G:

Screening for esophageal adenocarcinomas: An evidence-based approach. Am J Med 113:499–505, 2002.

Hulscher JB, van Sandick JW, de Boer AG, et al:

Extended transthoracic resectioncompared with limited transhiatal resection for adenocarcinoma of the esophagus. NEngl J Med 347:1662-1669, 2002.

Kaklamanos IG, Walker GR, Ferry K, et al:

Neoadjuvant treatment for resectablecancer of the esophagus and the gastroesophageal junction: A meta-analysis of randomizedclinical trials. Ann Surg Oncol 10:754–761, 2003.

Kato H, Miyazaki T, Nakajima M, et al:

Sentinel lymph nodes with technetium-99mcolloidal rhenium sulfide in patients with esophageal carcinoma. Cancer 98:932–939,2003.

Kinkel K, Lu Y, Both M, et al:

Detection of hepatic metastases from cancers of thegastrointestinal tract by using noninvasive imaging methods (US, CT, MR imaging,PET): A meta-analysis. Radiology 224:748–756, 2002.

Krasna MJ, Reed CE, Nedzweicki D, et al:

CALGB 9380: A prospective trial of thefeasibility of thoracoscopy/laparoscopy in staging esophageal cancer. Ann Thorac Surg71:1073–1079, 2001.

Kroep JR, Pinedo HM, Giaccone G, et al:

Phase II study of cisplatin precedinggemcitabine in patients with advanced oesophageal cancer. Ann Oncol 15:230–235,2004.

Medical Research Council Esophageal Cancer Working Group:

Surgical resectionwith or without preoperative chemotherapy in esophageal cancer: A randomized controlledtrial. Lancet 359:1727–1733, 2002.

Meluch AA, Greco FA, Gray JR, et al:

Preoperative therapy with concurrent paclitaxel/carboplatin/infusional 5-FU and radiation therapy in locoregional esophageal cancer:Final results of a Minnie Pearl Cancer Research Network phase II trial. Cancer J 9:251–260, 2003.

Muro K, Hamaguchi T, Ohtsu A, et al:

A phase II study of single-agent docetaxel inpatients with metastatic esopahgeal cancer. Ann Oncol 15:955–959, 2004.

Nguyen NT, Roberts P, Follette DM, et al:

Thoracoscopic and laparoscopicesophagectomy for benign and malignant disease: Lessons learned from 46 consecutiveprocedures. J Am Coll Surg 197:902–913, 2003.

Polee MB, Sparreboom A, Eskens FA, et al:

A phase I and pharmacokinetic study ofweekly paclitaxel and carboplatin in patients with metastatic esophageal cancer. ClinCancer Res 10:1928–1934, 2004.

Polee M, Tilanus HW, Eskens FA, et al:

Phase II study of neo-adjuvant chemotherapywith paclitaxel and cisplatin given every 2 weeks for patients with a resectable squamouscell carcinoma of the esophagus. Ann Oncol 14:1253–1257, 2003.

Stahl M, Wilke H, Walz MK, et al:

Randomized phase III trial in locally advancedsquamous cell carcinoma (SCC) of the esophagus: Chemoradiation with and withoutsurgery (abstract). Proc Am Soc Clin Oncol 22: 250, 2003.

Sumpter KA, Harper-Wynne C, Cunningham D, et al:

Randomised, multicenterphase III study comparing capecitabine with fluorouracil and oxaliplatin with cisplatinin patients with advanced oesophagogastric cancer: Confirmation of dose escalation(abstract). Proc Am Soc Clin Oncol 22:257, 2003.