PII S0360-3016(99)00199-6

CLINICAL INVESTIGATION Esophagus

IS COMBINED CHEMOTHERAPY AND RADIATION THERAPY EQUALLYEFFECTIVE AS SURGICAL RESECTION IN LOCALIZED

ESOPHAGEAL CARCINOMA?

ALEXANDER CHAN, M.D., FRCPC,*AND ALFRED WONG, M.D., FRCPC†

Departments of *Radiation Oncology and†Medical Oncology, Tom Baker Cancer Center, Calgary, Alberta, Canada

Purpose: This is a retrospective cohort comparison of combined chemotherapy and radiation versus esophagec-tomy in nonmetastatic esophageal cancers.Methods and Materials: Between 1984 and 1994, 82 patients received concurrent chemotherapy and radiation astheir primary treatment. Their treatment consisted of 50–60 Gy of radiation in 20–30 fractions over 4–6 weeks,concurrent with bolus mitomycin C (8 mg/m2) on day 1, 5-fluorouracil (5-FU) infusion (20 mg/kg/day)1/2leucovorin (20 mg/m2/day) on days 1–4 and 22–25. This group was compared to another cohort of 81 patientswho had esophagectomy. Both groups were restaged according to the 1983 AJCC clinical staging system andthere was more clinical Stage III disease in the chemoradiation group, 30% versus 16%.Results: The complete response rate was 68% after chemoradiation (by clinical assessment) and 83% foresophagectomy (by pathological assessment). At 5 years, the local relapse rate was 59% for chemoradiation and51% for esophagectomy. The 5-year disease-free rate and survival were 23% and 25% for chemoradiation, and21% and 23% for esophagectomy respectively. There was no significant difference in the disease control andsurvival between the two treatments. The pretreatment AJCC clinical stage was a strong prognosticator ofoutcome. The 5-year survival was 55% for Stage I, 16% for Stage II, and 8% for Stage III (p 5 0.00003).Conclusion: Combined chemotherapy and radiation appeared to be as effective as esophagectomy in localizedesophageal cancer. © 1999 Elsevier Science Inc.

Esophageal cancer, Chemotherapy, Radiation, Esophagectomy.

INTRODUCTION

Esophageal carcinoma is a relatively uncommon but deadlydisease. For decades, surgical resection (transthoracic ortranshiatal esophagectomy) has been the standard treatmentfor esophageal cancers. However, radical resection was onlypossible in those patients who could tolerate the surgicalprocedure and in whom all the gross tumor could be re-sected. Many patients were excluded because of their poorgeneral condition and/or their advanced stage of disease atthe time of diagnosis. The latter group was treated withradiation therapy alone with disappointing results (1). In the1980s, several Phase II studies using combined radiationtherapy and chemotherapy (5-fluorouracil [5-FU] infusion,cis-platinum, or mitomycin C) have shown encouragingresults in local disease control and survival (2–8). Thissuperiority of combined radiation therapy and chemother-apy over radiation therapy alone was subsequently con-firmed in Phase III randomized trials (9–11). However,there was no direct comparison (Phase III randomizedstudy) between combined radiation therapy and chemother-apy with surgical resection alone. Several factors have madesuch a comparison difficult: Different staging system

(pathological staging versus clinical staging) and differentmeans of evaluating tumor response to treatment. Thepathological evaluation of an esophagectomy specimen canprovide very accurate information on the stage of the tumorand the completeness of the resection. In contrast, there isalways uncertainty about the evaluation of the tumor stage,particularly the lymph node status, and the response totherapy by clinical assessment. Treatment-related inflam-matory changes can mask residual disease and post-treat-ment benign stricture can be mistaken as persistent disease.

In an attempt to determine whether combined radiationtherapy and chemotherapy is providing similar disease con-trol and survival as esophagectomy, we carried out a 10-year (1984–1994) retrospective review of these two treat-ment strategies for localized esophageal carcinoma in ourinstitution. Comparison was made on the disease-specificsurvival, the disease-free rate, the local failure rate, and thedistant failure rate between these two treatments.

METHODS AND MATERIALS

A Phase II study of concurrent chemotherapy and re-gional radiation therapy in esophageal cancers was first

Reprint requests to: Dr. Alexander Chan, Department of Radi-ation Oncology, Tom Baker Cancer Center, 1331 - 29th Street,

NW, Calgary, Alberta, T2N 4N2, Canada.Accepted for publication 14 May 1999.

Int. J. Radiation Oncology Biol. Phys., Vol. 45, No. 2, pp. 265–270, 1999Copyright © 1999 Elsevier Science Inc.Printed in the USA. All rights reserved

0360-3016/99/$–see front matter

265

introduced in 1984 in our institution. In the following 10years, 82 patients with localized esophageal carcinoma(without hematogenous metastasis) were treated with thisnonsurgical protocol. In the same period (1984–1994) therewas another cohort of 81 patients with localized esophagealcarcinoma who had esophagectomy. These two groups con-stituted the patient population in our review.

Pretreatment characteristicsIn the combined chemotherapy and radiation group, there

were 55 males and 27 females. Their median age was 60years (range: 41 to 85 years). In the esophagectomy group,there were 63 males and 18 females. The median age was 60years (range: 31 to 83 years). In the chemoradiation group,the tumor location was as follows: eight in the upper esoph-agus (10%), 34 in the middle esophagus (42%), and 40 inthe lower esophagus and the gastroesophageal junction(48%). Sixty-eight patients had squamous cell carcinoma(83%) and 14 patients had adenocarcinoma (17%). In theesophagectomy group, there were three upper esophageal(4%), nine midesophageal (11%), and 69 lower esophagealor gastroesophageal (85%) tumors. Twenty-four patientshad squamous cell carcinoma (30%) and 57 had adenocar-cinoma (70%). The esophagectomy group had a highernumber of distal esophageal cancers because tumors in thislocation were more accessible to a transhiatal esophagealresection. Many surgeons had referred those patients withupper and middle esophageal tumors for combined chemo-therapy and radiation therapy instead of performing a trans-thoracic esophagectomy.

To obtain an equitable comparison between the twogroups, we have retrospectively restaged all 163 patientsusing only their pretreatment clinical information. Pretreat-ment assessment included esophagoscopy and biopsy of theprimary lesion, esophagram, and radiological evaluation ofthe lung (by chest X-ray or CT of the thorax) and the liver(by upper abdominal ultrasonography or CT) for hematog-enous metastasis. The clinical stage was based on the 1983American Joint Committee on Cancers (AJCC) TNM clas-sification (Table 1). All 163 patients had esophagoscopy butonly 54% of the surgical group and 46% of the chemora-diation group had esophagram. CT scan of the thorax andliver was performed in 33 patients in the surgical group and55 patients in the chemoradiation group. We used theesophagoscopy and esophagram findings to determine thelength of the primary tumor (5 cm or less versus greater than5 cm) and the extent of luminal involvement (noncircum-ferential versus circumferential). Extraesophageal extension(i.e., T3) was established by clinical findings of recurrentlaryngeal nerve palsy or phrenic nerve palsy, or broncho-scopic evidence of tracheoesophageal/ bronchoesophagealfistula. We did not up-stage any tumor to a clinical T3 stagebased on CT findings of possible extraluminal extensionalone. Most patients were Nx because there was no patho-logical information about their nodal status. Some patientsin this Nx group actually had CT findings of suspiciousmediastinal or celiac lymphadenopathy. The only disparity

between the two groups was in the N1 disease. More pa-tients in the chemoradiation group had biopsies of suspi-cious lymphadenopathy identified on CT scans. These pa-tients were initially planned for surgical resection. Theywere later referred for chemoradiation after lymph nodemetastasis was diagnosed by mediastinoscopy or at the timeof laparotomy. The clinical stage distribution for bothgroups is given as follow: Nine patients or 11% with StageI disease (T1N0/xM0), 48 patients or 59% with Stage IIdisease (T2N0/xM0), and 25 patients or 30% with Stage III(T3, any N, M0 or any T, N1M0) disease in the chemora-diation group, and 23 (28%) Stage I, 45 (56%) Stage II, and13 (16%) Stage III in the esophagectomy group.

TreatmentOf the 81 patients who had surgical resection, 50 patients

had a transhiatal esophagectomy and 31 had an esophagealresection through a thoracoabdominal approach, for exam-ple, an Ivor-Lewis esophagectomy. Postoperative radiationof 40 to 60 Gy was given to 19 patients. Fifteen patientsreceived radiation for lymph node metastasis after curativeresections. Four patients received radiation for unclear re-section margins and N1 disease. None of these 19 patientsreceived any chemotherapy concurrent with or subsequentto their radiation therapy.

For the combined chemotherapy and radiation group, thetreatment protocol has gone through two modifications be-tween 1984 to 1994. The initial 23 patients were treatedwith a split-course design. Each treatment course consistedof 25 Gy of radiation in 10 fractions in 2 weeks, concurrentwith a single bolus mitomycin C (8 mg/m2) on day 1, and 96hours of 5-FU infusion (20 mg/kg/day) on days 1 to 4. After

Table 1. The 1983 AJCC-TNM staging system foresophageal cancer

Primary tumor (T)T0 No demonstrable tumorTis Carcinomain situT1 A tumor 5 cm or less in length, produces no

obstruction, and has no circumferentialinvolvement or extraesophageal spread

T2 A tumor more than 5 cm in length, or a tumorthat produces obstruction or that involves theentire circumference. No extraesophagealspread

T3 Any tumor with extraesophageal spread

Regional lymph nodes (N)Nx Regional lymph nodes that are not ordinarily

accessible for evaluation (clinical evaluation)N0 No positive node (surgical evaluation)N1 Positive nodes (surgical evaluation)

Distant metastasis (M)Mx Not assessedM0 No distant metastasisM1 Distant metastasis present

Stage I Tis or T1 N0 or NX M0Stage II T2 N0 or NX M0Stage III Any T3 Any N1 Any M1

266 I. J. Radiation Oncology● Biology ● Physics Volume 45, Number 2, 1999

4 weeks of rest, the same treatment was repeated without themitomycin C on day 1. In the next 33 patients, the 4-weekrest period was deleted. Sixty Gy of radiation was givencontinuously in 30 fractions over 6 weeks, concurrent withbolus mitomycin C (8 mg/m2) on day 1, and 96 hours of5-FU infusion (20 mg/kg/day) on days 1–4 and days 22–25.In the last 26 patients, the 5-FU (20 mg/kg/day) waschanged from a 24-hour daily infusion to a 6-hour dailyinfusion and low-dose leucovorin (20 mg/m2/day) wasadded on days 1–4 and days 22–25 respectively.

Radiotherapy techniqueMost patients were treated using 6 MV photons. A few

patients in 1984–1986 were treated with60Co radiation. Allpatients were treated in a supine position. A barium swallowstudy was performed during simulation to delineate theesophageal tumor. The treatment field encompassed a 5-cmmargin proximal and distal to the primary tumor. The lateralfield borders covered the adjacent mediastinal lymph nodesplus a 1.5-cm margin. For tumors in the lower esophagusand the gastroesophageal junction, the inferior field borderwas extended to include the celiac axis. Cerroband blockswere employed to shield as much adjacent lung, heart, andkidneys as possible. The heart dose was kept below 30 Gyand the spinal cord dose below 42 Gy.

For upper third tumors, the initial treatment was deliveredwith anteroposterior opposing fields encompassing theesophageal tumor, adjacent mediastinal nodes, and bothsupraclavicular fossae. An anterior vertical wedge was em-ployed as a compensator for the contour change between theneck and the upper chest. After 36–38 Gy, the primarytumor and adjacent mediastinum were treated with an an-terior wedged pair to avoid the spinal cord, while thesupraclavicular fossae were treated with separate anteriorphoton or electron fields to a total dose of 44 Gy formicroscopic disease and 50–60 Gy for gross disease.

Tumors in the middle and lower third esophagus weretreated with a 3-fields technique: an anterior and two pos-terior oblique or lateral fields. Patients were lying on aspecial head and arm rest attached to the treatment couch.Their arms were raised above their heads, holding onto thearm-rest. When the inferior treatment border was extendedto included the celiac axis, two lateral fields instead of twoposterior oblique fields were used to reduce the radiationdose to the kidneys.

Evaluation of responseIn the esophagectomy group the tumor response was

determined by pathological evaluation. Complete responsewas defined as a complete resection with negative resectionmargins. Those patients with either positive resection mar-gins or macroscopic residual disease were considered aspartial responders. In the chemoradiation group, the tumorresponse was determined by clinical assessment: esopha-goscopy, biopsy, CT scan, and/or esophagram at 6–8 weeksafter treatment. Further follow-up esophagoscopy and/oresophagram were performed at 3 months and 6 months as

indicated. Complete responders were those patients whohad no visible residual disease in their follow-up esopha-goscopy and a negative biopsy if one was taken, and therewas no disease recurrence in the following 6 months. Partialresponders were those patients who had significant tumorreduction and there was no disease progression in the fol-lowing 3 months. Those patients who did not have post-treatment evaluation due to various reasons, for example,developing metastatic disease, were considered as havingfailed at the primary site.

Statistical methodSurvival was calculated from the date of diagnosis using

the Kaplan-Meier method. Comparison of the disease-spe-cific survival, the disease-free rate, the local failure rate, andthe distant failure rate was performed using the log-ranktest.

RESULTS

In the chemoradiation group, the complete response ratewas 68% (56/82) and the partial response rate was 11%(9/82). Ten patients who did not have post-treatment endo-scopic evaluation were considered as local failures. In theesophagectomy group, the complete response rate was 83%(67/81) and the partial response rate was 16% (13/81).Postoperative radiation therapy appeared to reduce the localrecurrence rate after esophagectomy, especially for thosepatients with positive resection margins. Among the 67patients with complete response, local esophageal relapseoccurred in 19 of the 52 patients who did not receivepostoperative radiation, and 2 of the 15 patients who hadpostoperative radiation. In the subgroup with partial re-sponse, all 9 patients without postoperative radiation and 2of the 4 patients who had postoperative radiation developedlocal recurrence in the esophagus. The 5-year actuarial localfailure rate was 51% for the esophagectomy group and 59%for the chemoradiation group (Fig. 1), and the differencewas not statistically significant (p 5 0.27).

There are 22 long-term survivors with 11 patients in eachgroup. They have been followed from 4 years to 13 years.All 22 survivors are free of disease. The remaining 141patients had died either of their disease or other nonmalig-nant causes. There was no difference in the disease-specificsurvival between esophagectomy and chemoradiation (Fig.2). The 5-year survival was 23% for the esophagectomygroup and 25% for the chemoradiation group. The mediansurvival was 16 months for esophagectomy and 15 monthsfor chemoradiation. The 5-year disease-free rate was 21%and 23% respectively (Fig. 3). Of the 82 patients who hadchemoradiation, 24 patients (29%) had no evidence of dis-ease, 26 patients (32%) had local failures alone, 20 patients(24%) had distant failures alone, and 12 patients (15%) hadboth local and distant failures. For the 81 patients who hadesophagectomy, 22 patients (27%) had no evidence of dis-ease, 8 patients (10%) had local recurrences alone, 28patients (35%) had distant failures alone, and 23 patients

267Chemo-RT in esophageal Ca● A. CHAN et al.

(28%) had both local and distant failures. The 5-year distantfailure rate was 72% for esophagectomy and 55% for che-moradiation (Fig. 4,p 5 0.028).

When we looked at the outcome according to the tumorhistology, we did not find any significant difference in boththe disease-specific survival (p 5 0.81) and the disease-freerate (p 5 0.82) between squamous cell carcinoma andadenocarcinoma. For those patients with squamous cellcarcinoma, the 5-year disease-specific survival was 26% forboth esophagectomy and chemoradiation. For the subgroupwith adenocarcinoma, the 5-year disease-specific survivalwas 19% after esophagectomy and 22% after chemoradiation.

Of the 32 clinical Stage I (1983 AJCC clinical stagingsystem) patients who had either esophagectomy or chemo-radiation, 19 patients (59%) had no evidence of disease(NED). In comparison, the NED rate was 22% for clinicalStage II (20/93) and 18% for clinical Stage III (7/38). Thedifference was statistically significant with a two-tailedp-

value of 0.000067. The 5-year disease-specific survival was55% for clinical Stage I, 16% for clinical Stage II, and 8%for clinical Stage III (Fig. 5,p 5 0.00003).

DISCUSSION

Although there is no randomized study that directly com-pared surgery alone versus chemoradiation alone in local-ized esophageal carcinoma, many Phase II studies usingcombined chemotherapy and radiation therapy (2–8) havereported survival and loco-regional control that were com-parable to surgical resection. The present review attemptedto compare the treatment outcome between these two man-agement strategies within the same institution. As a retro-spective review, we cannot avoid the inherent problem ofpatient selection bias. Patients referred for chemoradiationgenerally had either worse physical condition or more ad-vanced disease. Many patients were considered not fit for a

Fig. 1. Actuarial local failure rate after chemoradiation (82 pa-tients) and esophagectomy (81 patients).

Fig. 2. Actuarial disease-specific survival after chemoradiation (82patients) and esophagectomy (81 patients).

Fig. 3. Actuarial disease-free rate after chemoradiation (82 pa-tients) and esophagectomy (81 patients).

Fig. 4. Actuarial distant failure rate after chemoradiation (82patients) and esophagectomy (81 patients).

268 I. J. Radiation Oncology● Biology ● Physics Volume 45, Number 2, 1999

radical resection that involved a thoracotomy. There wererelatively fewer patients with early disease (clinical Stage I)in the chemoradiation group (11%) as compared to theesophagectomy group (28%). Another potential bias waspostoperative adjuvant therapy after esophagectomy. Nine-teen patients in the esophagectomy group (23%) had re-ceived postoperative radiation therapy.

We have found no difference in both the disease-specificsurvival and the disease-free rate between these two treat-ments. Our treatment results (a median survival of 15–16months and a 5-year survival of 21–23%) were fairly com-parable to the published results after surgical resection orchemoradiation. In a literature review of 130 papers (76,911patients) on the surgical management of esophageal cancerby Muller et al. (12), the mean 3-year and 5-year survivalrate after esophagectomy was 25% and 20%. Coiaet al. (7)reported a 5-year disease-specific survival of 30% for clin-ical Stage I and II esophageal carcinoma treated with 5-FUinfusion, mitomycin, and 60 Gy of radiation. In the RTOG85-01 trial (10, 11), the median survival was 14 months andthe 5-year survival was 30% after combined cisplatin, 5-FUinfusion, and 50 Gy of radiation.

In our review, the complete response rate was higher afteresophagectomy, 82%, than chemoradiation, 68%. This dif-ference may be misleading because of the difference inevaluating tumor response after treatment. Pathologicalevaluation definitely gave a more accurate assessment of theresponse than clinical evaluation. In the chemoradiationgroup, 12% of the patients who did not have post-treatmentevaluation were considered as local failures. This mighthave resulted in an underestimation of the actual responserate after chemoradiation. The 5-year local failure rate wasfairly comparable, 51% after esophagectomy and 59% afterchemoradiation. The distant metastatic rate was lower in thechemoradiation group (55%) than in the esophagectomygroup (72%). It is possible that concurrent chemotherapymight have reduced the incidence of distant metastasis.

However, this difference must be viewed with caution be-cause of the potential selection bias in a retrospective re-view. More early deaths from uncontrolled local diseasebefore the occurrence of distant metastases in the chemora-diation group is another plausible but less likely explanationsince the actuarial local failure rates were very similar forboth treatment groups (Fig. 1).

Since more than half of the patients had failed at theprimary site, neither esophagectomy nor chemoradiationalone was optimal for primary disease control. A combina-tion of preoperative chemoradiation followed by esophagec-tomy appeared to be the next logical step to explore. SeveralPhase II studies using such combination (13–18) haveshown encouraging results in both the survival and thedisease control. However, others have found either no ben-efit with the addition of surgical resection after chemora-diation (19), or improvement only in the local control, butnot in distant metastases or survival (20). Similarly, threerecent randomized studies on surgery versus preoperativechemoradiation plus surgery also reported conflicting re-sults. Walshet al. (21) showed survival improvement withpreoperative 5-FU, cisplatin, 40 Gy of radiation, and sur-gery in esophageal adenocarcinoma. Despite a prolongationin the disease-free survival and the local disease-free sur-vival, Bossetet al. (22) did not find any overall survivalbenefit with the addition of preoperative cisplatin and 18.5Gy of radiation to esophagectomy in squamous cell carci-noma of the esophagus. In the most recent update of a trialthat compared preoperative cisplatin, vinblastine, 5-FU, andradiation versus surgery alone, Urbaet al. (23) reported animprovement in survival at 3 years, 32% versus 15%. Giventhe degree of treatment morbidity associated with preoper-ative chemoradiation and esophagectomy, more studies areneeded to confirm its efficacy and to define the appropriatepatient population that will benefit from such intensivetreatment.

Because of the imbalance in the tumor histology betweenthe two treatment groups in our review, a subgroup analysisaccording to the tumor histology was performed. We did notfind any difference in both the survival and treatment out-come between squamous cell carcinoma and adenocarci-noma. Other investigators have reported similar observa-tions. Coiaet al. (7), in an analysis of 90 patients treated bychemoradiation, found no survival difference between thesetwo histologies. In another review of the surgical results inesophageal cancers, Mu¨ller et al. (12) could not find anyprognostic correlation between either the tumor histology orthe tumor location and the long-term outcome after esoph-ageal resection. Therefore, it is unlikely that we have misseda significant difference between these two treatment regi-ments because of this imbalance in the tumor histology.

We have found the 1983 AJCC clinical staging system auseful prognosticator. There was good correlation betweenthe survival and the disease-free rate with the pretreatmentstage. For Stage I disease, the 5-year survival was 55%, and59% of the patients was free of recurrence after eitheresophagectomy or combined chemotherapy and radiation.

Fig. 5. Actuarial disease-specific survival according to the 1983AJCC clinical stage: Stage I (32 patients), Stage II (93 patients),and Stage III (38 patients).

269Chemo-RT in esophageal Ca● A. CHAN et al.

In comparison, the 5-year survival was only 8–16% in Stage IIand III, and the disease-free rate was 18–22%. Such a clinicalstaging system will be useful in future preoperative neoadju-vant studies in pretreatment stratification of patients.

In conclusion, we have not found any significant differ-ence in the outcome between primary combined chemother-apy and radiation versus surgical resection in localizedesophageal cancers. However, the result from either treat-ment is far from optimal and there is an ongoing need to find

more effective treatment for this disease. Combining che-motherapy, radiation, and surgery appears to be the nextlogical step, particularly for patients who have a morefavorable prognosis, for example, clinical Stage I esopha-geal cancer. Currently, we are conducting a Phase II studyof preoperative chemotherapy with 5-FU infusion, leucov-orin, and mitomycin C, and 50 Gy of regional radiation,followed by esophagectomy for patients with localizedesophageal cancers in our institution.

REFERENCES

1. Earlam R, Cunha-Melo JR. Oesophageal squamous cell car-cinoma: II. A critical review of radiotherapy.Br J Surg1980;67:457–461.

2. Keane TJ, Harwood AR, Elhakim T,et al. Radical radiationtherapy with 5-fluorouracil infusion and mitomycin-C for oe-sophageal squamous carcinoma.Radiother Oncol1985;4:205–210.

3. Richmond J, Seydel HG, Bae Y,et al. Comparison of threetreatment strategies for esophageal cancer within a singleinstitution. Int J Radiat Oncol Biol Phys1987;13:1617–1620.

4. Herskovic A, Leichman L, Lattin P,et al. Chemo/radiationwith or without surgery in the thoracic esophagus: The WayneState experience.Int J Radiat Oncol Biol Phys1988;15:655–662.

5. Chan A, Wong A, Arthur K. Concomitant 5-fluorouracil in-fusion, mitomycin-C, and radical radiation therapy in esoph-ageal squamous cell carcinoma.Int J Radiat Oncol Biol Phys1989;16:59–65.

6. John MJ, Flam MS, Mowry PA,et al. Radiotherapy alone andchemoradiation for nonmetastatic esophageal carcinoma: A crit-ical review of chemoradiation.Cancer1989;63:2397–2403.

7. Coia LR, Engstrom PF, Paul AR,et al. Long-term results ofinfusional 5-FU, mitomycin C, and radiation as primary man-agement of esophageal carcinoma.Int J Radiat Oncol BiolPhys1991;20:29–36.

8. Seitz JF, Giovannini M, Padaut-Cesana J,et al. Inoperablenonmetastatic squamous cell carcinoma of the esophagusmanaged by concomitant chemotherapy (5-fluorouracil andcisplatin) and radiation therapy.Cancer1990;66:214–219.

9. Sischy B, Ryan L, Haller D,et al. Interim report of EST 1282phase III protocol for the evaluation of combined modalities inthe treatment of patients with carcinoma of the esophagus,stage I and II. (Abstr.)Proc Am Soc Clin Oncol1990;9:105.

10. Herskovic A, Martz K, Al-Sarraf M,et al. Combined chemo-therapy and radiotherapy compared with radiotherapy alone inpatients with cancer of the esophagus.N Engl J Med1992;326:1593–1598.

11. Al-Sarraf M, Martz K, Herskovic A,et al. Superiority ofchemo-radiotherapy vs. radiotherapy in patients with esopha-geal cancer: Final report of an intergroup randomized and con-firmed study (Abstr.).Proc Am Soc Clin Oncol1996;15:464.

12. Muller JM, Erasmi H, Stelzner M,et al. Surgical therapy ofoesophageal carcinoma.Br J Surg1990;77:845–857.

13. Poplin E, Fleming T, Leichman L,et al. Combined therapiesfor squamous-cell carcinoma of the esophagus: A SouthwestOncology Group Study (SWOG-8037).J Clin Oncol1987;5:622–628.

14. Seydel HG, Leichman L, Byhardt R,et al. Preoperative radi-ation and chemotherapy for localized squamous cell carci-noma of the esophagus: A RTOG study.Int J Radiat OncolBiol Phys1988;14:33–35.

15. Naunheim KS, Petruska PJ, Roy TS,et al. Preoperative che-motherapy and radiotherapy for esophageal carcinoma.J Tho-rac Cardiovasc Surg1992;5:887–895.

16. Forastiere A, Orringer MB, Perez-Tamayo C,et al. Preoper-ative chemoradiation followed by transhiatal esophagectomyfor carcinoma of the esophagus: Final report.J Clin Oncol1993;11:1118–1123.

17. Sauter ER, Coia LR, Keller SM. Preoperative high-dose radi-ation and chemotherapy in adenocarcinoma of the esophagusand esophagogastric junction.Ann Surg Oncol1994;1:5–10.

18. Stahl M, Wilke H, Fink U,et al. Combined preoperativechemotherapy and radiotherapy in patients with locally ad-vanced esophageal cancer: Interim analysis of a phase II trial.J Clin Oncol1996;14:829–837.

19. Gill PG, Denham JW, Jamieson GG,et al. Patterns of treat-ment failure and prognostic factors associated with the treat-ment of esophageal carcinoma with chemotherapy and radio-therapy either as sole treatment or followed by surgery.J ClinOncol 1992;10:1037–1043.

20. Kavanagh B, Anscher M, Leopold K,et al.Patterns of failurefollowing combined modality therapy for esophageal cancer,1984–1990.Int J Radiat Oncol Biol Phys1992;24:633–642.

21. Walsh TN, Noonan N, Hollywood D,et al. A comparison ofmultimodal therapy and surgery for esophageal adenocarci-noma.N Engl J Med1996;335:462–467.

22. Bosset JF, Gignoux M, Triboulet JP,et al.Chemoradiotherapyfollowed by surgery compared with surgery alone in squa-mous-cell cancer of the esophagus.N Engl J Med1997;337:161–167.

23. Urba S, Orringer M, Turrisi A,et al. A randomized trialcomparing surgery (S) to preoperative concomitant chemora-diation plus surgery in patients (pts) with resectable esopha-geal cancer (CA) (Abstr.).Proc Am Soc Clin Oncol1997;16:A983.

270 I. J. Radiation Oncology● Biology ● Physics Volume 45, Number 2, 1999