ORIGINAL ARTICLES

Chemotherapy and Radiation Therapy Before Transhiatal Esophagectomy for Esophageal Carcinoma Mark B. Orringer, MD, Arlene A. Forastiere, MD, Claudia Perez-Tamayo, MD, Susan Urba, MD, Bonnie J. Takasugi, MD, and Judith Bromberg, PhD Section of Thoracic Surgery, Department of Surgery, Division of Hematology-Oncology, Department of Internal Medicine, and Department of Radiation Oncology, University of Michigan Medical Center, and Department of Biostatistics, School of Public Health, Ann Arbor, Michigan

Recent efforts to improve survival in patients with esophageal carcinoma have combined both systemic and local therapy. From October 1985 to October 1987, 43 patients with local-regional esophageal cancer (adenocarcinoma in 21, squamous cell in 22) were treated with cisplatin, vinblastine, and 5-fluorouracil chemotherapy concurrent with 4,500 cGy radiation therapy for 21 days before transhiatal esophagectomy 3 weeks later. Two patients died of chemotherapyhadiation therapy toxicity. Fortyone completed preoperative chemotherapyhadiation therapy. At operation, 2 patients had incurable metastatic disease; 39 underwent transhiatal esophagectomy. Eleven patients had no residual tumor in the resected specimen for a 27% (11 of 41) pathological complete response rate. Preoperative chemotherapyhadiation therapy resulted in no increased perioperative morbidity as compared with our historical controls. One patient died

postoperatively of an unrecognized brain metastasis (2% operative morbidity). At a median follow-up of 27 months, 20 patients (47%)are alive and clinically diseasefree and 21 have died, 19 from progression of their carcinoma. The median survival time for all 43 patients is 29 months (Kaplan-Meier estimate), and cumulative survival is 72% at 12 months, 60% at 24 months, and 46% at 36 months. All 11 patients with a complete response are alive at a median follow-up of 36 months, and all are disease-free. The 2-year survival of 60% of this group as compared with 32% in our earlier patients treated with transhiatal esophagectomy alone suggests that intensive combined modality therapy improves survival in these patients. A randomized prospective trial is now in progress.

E

enthusiasm for THE was based on a clear perception of decreased postoperative morbidity and mortality for patients requiring resection for esophageal carcinoma. With an overall hospital mortality of approximately 5%, we are able to provide efficient palliation in the majority of our patients. Unfortunately, as we expected, long-term survival after THE, as after standard transthoracic resection, is seldom achieved; our overall 2-year survival is 32%, and our 48-month survival is only 17%. Our desire to achieve long-term survival and even potential cure in these patients has led us, like other investigators, to consider multimodality therapy for esophageal carcinoma. Preoperative chemotherapy in the treatment of localregional esophageal carcinoma was instituted by Kelsen and colleagues [6] at the Sloan-Kettering Memorial Hospital in New York and by Steiger and co-workers [7] at the Wayne State Medical Center in Michigan in the 1970s. Cisplatin drug combinations had a significant local effect on many of these tumors [8, 91. The Wayne State experience with combined preoperative radiation therapy (RT) and chemotherapy [7, 101 stimulated other researchers to evaluate further such combination adjuvant therapy [ll]. We conducted a trial of combined preoperative chemotherapy with cisplatin, vinblastine, and mitoguazone in 36 patients with both adenocarcinomas and epidermoid carcinomas, and 25 of these patients then underwent THE

sophageal cancer remains a lethal disease; 5-year survival rarely exceeds 20% [l, 21. Because approximately 75% of these patients have either local tumor invasion or distant metastases that preclude cure at the time the disease is diagnosed, traditional therapy has been focused on providing expedient palliation of dysphagia with the lowest possible risk. Since 1976, transhiatal esophagectomy (THE) without thoracotomy and a cervical esophagogastric anastomosis have emerged as our preferred method of esophageal resection and reconFor editorial comment see page 345. struction both for benign and malignant disease; the procedure avoids the morbidity of a thoracotomy and the devastating consequences of an intrathoracic esophageal anastomotic leak [3-51. As our experience with THE has increased, a relatively large referral base has been established, and we have now performed nearly 500 of these procedures, approximately 350 for carcinoma. Our initial Presented at the Twenty-fifth Anniversary Meeting of The Society of Thoracic Surgeons, Baltimore, MD, Sep 11-13, 1989. Address reprint requests to Dr Orringer, Section of Thoracic Surgery, University of Michigan Medical Center, 1500 E Medical Center Dr, 2120 TCi0344, Ann Arbor, MI 48109.

0 1990 by The Society of Thoracic Surgeons

(Ann Thorac Surg 1990;49:348-55)

0003-4975/90/$3.50

Ann Thorac Surg 1990;49:34%55

ORRINGER ET AL MULTIMODALITY THERAPY FOR ESOPHAGEAL CARCINOMA

Table 1. Tumor Location and Cell Type: Multimodality Therapy for Esophageal Cancer“

Criteria for eligibility in this study included the following:

Squamous

Site Cervical

Upper thoracic Midthoracic Lower thoracic Total a

Adenocarcinoma Cell Carcinoma

Total

0 1 20

3 3 10 6

3 (7) 3 (7) 11 (26) 26 (60)

21 (49)

22 (51)

43 (100)

0

349

Numbers in parentheses are percentages

[12]. Substantial toxicity was experienced, and the mean survival time of 14 months did not differ significantly from the 12.5-month mean survival time in our 100 earlier patients treated with THE without chemotherapy [5]. Complete eradication of the tumor was achieved in only 4% of the patients, and in 40% local-regional recurrence developed in follow-up. Radiation therapy was therefore added to the chemotherapy in an effort to improve our results. This report reviews our experience with preoperative chemotherapy (cisplatin, 5-fluorouracil, and vinblastine) administered concurrently with RT before THE in patients with esophageal carcinoma.

Material and Methods Between October 1985 and October 1987,43 patients with esophageal carcinoma were treated with preoperative RT and chemotherapy followed by transhiatal esophagectomy without thoracotomy. This treatment proposal entitled “Concurrent chemotherapy and radiotherapy followed by transhiatal esophagectomy in local-regional esophageal cancer” was described in University of Michigan protocol No. 1885, approved by the University of Michigan Medical School Institutional Review Board on October 1, 1985. Included within this group of patients were 32 men and 11 women with an average age of 62 years (range, 35 to 76 years). There were two black and 41 white patients. Twenty-two (51%) of the patients had squamous cell carcinoma and 21 (49%), adenocarcinoma of the thoracic esophagus or cardia (Table 1). Of the adenocarcinomas, 12 were associated with Barrett’s epithelium in patients with long-standing gastroesophageal reflux. Of the 43 patients, 39 (91%) had experienced weight loss that ranged from 2.3 to 22.5 kg (mean, 7.7 kg). Tumor length varied from 1.5 to 3.0 cm in 11 patients (26%),from 3.5 to 5.0 cm in 8 (17%), from 6.0 to 8.0 cm in 13 (30%), and from 9.0 to 12.0 cm in 8 (17%). Three patients had totally obstructing tumors, the length of which could not be determined preoperatively. According to the 1983 American Joint Committee on Cancer staging criteria (131, 79% of the patients had a clinical T, primary tumor (lesion greater than 5 cm in length, circumferential or obstructing). Regional adenopathy believed to represent metastatic disease was evident on chest and abdominal computed tomographic (CT) scans in 24 (56%)of the patients.

1. Newly diagnosed esophageal cancer without previous treatment 2. Patient believed to be an acceptable operative risk for esophagectomy 3. Tumor judged to be totally resectable (no evidence of distant metastases or tracheobronchial invasion) 4. Disease limited to the esophagus and regional lymph nodes such that it could be encompassed by a single (tolerable) radiation field 5. Karnofsky performance status of at least 60% 6. Adequate bone marrow reserve (white blood cell count greater than 3,500 cells/pL, platelets greater than 100,000/pL) 7. Adequate renal function (creatinine clearance greater than 50 mgimL) 8. Informed consent from the patient. Before entry into the study, all patients underwent the following staging examinations: chest roentgenogram, barium esophagogram, chest and abdominal CT scans, head CT scan, and bone scan. The treatment plan consisted of in-hospital combined RT and triple-drug chemotherapy with cisplatin, vinblastine, and 5-fluorouracil for 21 days, followed by a 3-week rest, and then THE on approximately day 42. Cisplatin (20 mg/m2/day) was administered by continuous intravenous infusion on days 1 through 5 and 17 through 21. To minimize renal toxicity, patients were prehydrated with normal saline solution, the total cisplatin dose was infused over 24 hours, and intravenous hydration was continued for the next four days. Vinblastine (1 mg/ m’iday) was administered by intravenous bolus on days 1 through 4 and 17 through 20. 5-Fluorouracil (300 mg/ m2/day)was administered by continuous infusion for the entire 21-day course of preoperative therapy. A 10-MeV high-energy linear accelerator was used to administer RT. Simulation included the tumor as well as 5-cm longitudinal margins and 2-cm lateral margins. As indicated by pretreatment staging scans, the mediastinum as well as celiac or supraclavicular lymph nodes, or both, were treated using a three- or four-field technique, with patients in the prone position to increase the distance between the spinal cord and the esophagus. Our initial plan called for daily fractions of 250 cGy, five days per week, for a total of 3,750 cGy over 21 days. However, after the first 20 patients (10 with squamous cell carcinoma and 10 with adenocarcinoma), in an effort to further increase the histological complete response rate, we initiated hyperfractionated therapy, administering 150 cGy twice daily, five days per week, to a total dose of 4,500 cGy over 21 days. Spinal cord doses were kept at or below 3,000 to 3,250 cGy. After completing the above 3-week in-hospital course of combined chemotherapy and RT, and after a 3-week rest, the patients underwent THE without thoracotomy as described previously [3, 51. Alimentary continuity was established at the same operation, with the stomach mobilized into the posterior mediastinum in the original

350

ORRINGER ET AL MULTIMODALITY THERAPY FOR ESOPHAGEAL CARCINOMA

esophageal bed and anastomosed to the cervical esophagus. In one patient who had undergone a previous gastric resection for peptic ulcer disease, a short-segment colon interposition was required after resection of the tumor. A pyloromyotomy and feeding jejunostomy were standard. Accessible subcarinal, paraesophageal, and celiac axis lymph nodes were sampled routinely for staging purposes.

Stat is t ical Methods The BMDPlL program [14] was used to obtain productlimit estimates of survival times and to compare survival distributions. Log-rank (Mantel-Cox) test statistics are reported, although generalized Wilcoxon (Breslow) test statistics also were computed. In both comparisons, results of the generalized Wilcoxon tests were consistent with those of the log-rank tests. Preoperative characteristics of patients with negative pathology at operation were compared with those of patients with positive pathology at operation using Student’s t tests (BMDP3D) for continuous variables and x2 tests for categorical variables (BMDP4F).

Results

Toxicity of Preoperative Therapy Two patients died before undergoing operation of sepsis associated with bone marrow suppression. Hematological toxicity was common. Forty patients (93%) experienced leukopenia to a white blood cell count of 1,900lpL or less; 27 (63%) had febrile neutropenia; and 10 (23%) had thrombocytopenia to a platelet count of 49,OOO/pL or less. One third of the patients required blood transfusions to correct anemia induced by myelosuppression. Only 11 (26%)of the patients tolerated 100% of the planned dose of all three drugs; overall, the group received an average of about 85% of the ideal dose of each. Anorexia and fatigue occurred in all patients. Nausea and vomiting were generally mild. The most serious nonhematological complication was radiation-induced esophagitis and associated odynophagia, which occurred in 37 patients (86%)and necessitated supplemental nutritional support in 34 (79%) (28 [65%] enteral and 6 [14%] total parenteral). Eight (19%) of the patients lost more than 10% of their initial body weight. No patient experienced renal or ototoxicity or peripheral neuropathy. In 2 patients, RT was not completed because of the development of severe radiation esophagitis. The other 41 patients received the scheduled total dose of 3,750 or 4,500 cGy without interruption.

Operat ion Of the 41 patients operated on, 25 (61%) had operation within 1 week of the scheduled day 42 of the treatment plan; delayed recovery from bone marrow suppression or availability of operating time influenced the ultimate date of operation. The patients underwent operation an average of 50 days after initiation of chemotherapy and RT. One patient who attempted suicide at home after completing chemotherapy and RT required psychiatric treatment, and esophagectomy was delayed until day 77.

Ann Thorac Surg 1990;49:34%55

Two of the 41 patients operated on did not undergo an esophagectomy because of the finding of incurable disease at exploration: a hepatic metastasis in 1 and extensive prevertebral fascia1 invasion and involvement of the aortic arch and great vessels with an upper thoracic squamous carcinoma in the other. Thus, the overall operability rate in this series was 95% (41 of 43). Thirty-nine patients underwent THE. In 1 patient who had undergone a previous gastric resection for peptic ulcer disease, a shortsegment colon interposition through a right thoracotomy was required to restore alimentary continuity after esophagectomy. In the other 38 patients, a cervical esophagogastric anastomosis was performed as described previously [15], using a single-layer 4-0 polyglycolic acid suture interrupted technique. In the majority of patients, THE was not believed to be technically more difficult because of increased periesophageal reaction from radiation and chemotherapy. On the contrary, some initially large tumors on barium swallow examination were so reduced in size or eliminated by preoperative therapy that the resection was believed to be technically expedited. Measured intraoperative blood loss in the 38 patients undergoing THE and cervical esophagogastric anastomosis averaged 832 mL (range, 220 to 3,000 mL), not significantly different from that of our previously reported THE patients who had no preoperative chemotherapy or radiation therapy [5]. A chest tube was required in 36 of the 38 standard THE patients because of entry into one or both pleural cavities. Of the 39 patients undergoing esophagectomy, gross total removal of all visible or palpable tumor was believed to have been obtained in all but 1, in whom residual transmurally invasive tumor had to be fractured away from the prevertebral fascia. Unexpected residual disease was documented in 2 other patients. One patient with an upper thoracic squamous cell carcinoma was found on permanent histologic section of the esophagus to have a microscopic focus of carcinoma at the cervical margin. The other had a seizure 48 hours postoperatively, and a brain CT scan showed a metastasis. This patient was treated with high-dose steroids, developed an anastomotic leak, and died on postoperative day 42, the one postoperative death in this series (2% operative mortality). Routine preoperative staging with a brain scan in our patients was instituted after this experience. One additional patient in this group had a postoperative complication, small bowel obstruction at his jejunostomy tube insertion site leading to vomiting and an anastomotic leak that healed spontaneously after reoperation on the abdomen. No patient experienced recurrent laryngeal nerve paresis or paralysis or postoperative chylothorax. The length of postoperative hospitalization averaged 12 days and ranged from 10 to 52 days (the latter in the patient with the small bowel obstruction). Only 5 of the 39 esophagectomy patients were hospitalized longer than 12 days after operation.

Histopathological Stage Grouping Histopathological assessment of the resected specimens indicated a complete response rate, ie, no residual cancer (TO MO), of 27% (11of 41), or 26% of all 43 patients entered into the study, The final TNM stage grouping for these

Ann Thorac Surg 1990;49348-55

ORRINGER ET AL MULTIMODALITY THERAPY FOR ESOPHAGEAL CARCINOMA

Table 2. Histopathological TNM Stage Grouping After Preoperative Radiation Therapy and Chemotherapy for Esophageal Carcinoma No. of Patients

Stage

11 (27%) 2

0 (TO NO MO) I (T1 NO MO) IIA T2 NO MO T3 NO MO IIB TO N1 MO” TI N1 MO T2 N1 MO I11 T3 N1 MO T4 N1 MO IV (TX N1 M1)

13 7 6 9 2 0 7 5 4 1 1 41

Total

Two patients with no residual primary tumor, but with metastatic squamous cell carcinoma in regional lymph nodes, were staged as IIB.

a

patients based on the 1988 revised American Joint Committee on Cancer classification [16] is shown in Table 2. Thirteen patients (32%) were surgically staged as TO. Lymph nodes sampled for staging purposes were positive for metastatic tumor in 15 patients (37%).

Survival The 2 patients who were operated on but did not undergo esophagectomy received palliative RT postoperatively. All other patients have been followed at regular intervals

351

after operation without undergoing further treatment either until their death or to the present time. With a median follow-up of 27 months, 21 patients have died, 19 of their disease (2 from complications of myelosuppression) and 2 of other causes (myocardial infarction, hepatorenal syndrome). Twenty of the 43 patients (47%) are alive and clinically disease-free. In 15 of the 36 patients in whom ”complete” resections were believed to have been achieved, the carcinomas recurred: locally in 1 patient with squamous cell carcinoma, at a distant site in 9 patients with adenocarcinoma and 3 patients with squamous cell carcinoma, and both locally and at a distant site in 1 patient with adenocarcinoma and 1 with squamous cell carcinoma. The median overall survival for the group is 29 months (Table 3). Cumulative survival is 91% at 6 months, 72% at 1 year, 60% at 2 years, and 46% at 3 years. Although patients with adenocarcinomas appear to have a survival advantage over those with squamous cell carcinoma, the differences in survival distribution are not statistically significant ( p = 0.25). For the purposes of comparison, surgical stage has been divided into two categories (Table 3): “no disease” (patients shown to be TO NO at operation) and ”residual” (20 patients shown to have stage I, IIA, IIB, or 111 disease after treatment) (Table 2). The residual disease category has been further subdivided into patients with local disease confined to the esophagus (stages I and IIA) and patients with regional disease (regional lymph node involvement, stages IIB and 111). The 11 patients who had no residual disease at operation (TO NO) are all still alive and disease-free; 1 underwent resection of a solitary brain metastasis 10 months after esophagectomy. Minimum follow-up time for these patients is 25 months, the median follow-up is 36 months. Their survival is clearly superior to that of patients with

Table 3. Survival After Chemotherapy and Radiation Therapy Combined With Transhiatal Esophagectomy for Esophageal Carcinomaa’b Variable

n

Overall Histological type Adenocarcinoma Squamous cell carcinoma Surgical stagese No disease (TO NO) Residual Local (I and IIA) Regional (IIB and 111)

43 21 22

Median Survival Time (mo)

Cumulative Survival (%) 6 mo

12 mo

24 mo

36 mo

29 (d)

91 (4)

72 (7)

60 (8)

46 (8)

NR

100 (d) 82 (8)

86 (8) 59 (10)

67 (10) 55 (11)

52 (11) 41 (12)

100 (d)

10 (dl

Significance‘

0.25 28 (13)

0.0004 11

NR

30

23 (4)

93 (5) 93 ( 6 )

67 (9) 67 (12)

49 (9) 59 (13)

29 (9) 42 (14)

15 (12)

93 ( 6 )

67 (12)

39 (10)

16 (11)

15

15

100 (d)

100 (10)

Survival calculated from date of diagnosis. Numbers in parentheses are standard errors. MantelLCox (log-rank) test for differences in survival distributions. Could not be calculated. Survival distribution of patients with no disease as compared with that of patients with residual disease. Two patients who died of complications from preoperative therapy are excluded. a

NR

=

not reached;

SE = standard error.

352

Ann Thorac Surg 1990;49:348-55

ORRINGER ET AL MULTIMODALITY THERAPY FOR ESOPHAGEAL CARCINOMA

Table 4. Preoperative Characteristics

Variable M/F Age (yr)

KPS Histology, adenol squamous Location, cervical-mid/ lower Length of tumor (mean, cm)

No Disease (TO NO, n = 11)

Residual Disease (n = 30)

813 62 (SE, 2) 83 (SE, 2) 516 6/5 8.5 (SE, 0.7)

KI'S = Karnofsky performance status;

SE

2317 59 (SE, 2) 87 (SE, 1) 16/14

0.79 0.36 0.10 0.65

11/19

0.30

7.0 (SE, 0.7) =

P

Value

0.24

standard error.

residual disease at operation ( p = 0.004). These latter patients have a median survival time of 23 months, and their cumulative survival at 12, 24, and 36 months is 67%, 49%, and 29%, respectively. There is no statistically significant difference in survival distribution between patients with local disease and those with regional disease ( p = 0.13). To determine whether the superior survival of the stage 0 patients over that of the remaining patients was attrib-

utable to systematic differences in the stage 0 group at entry into the study, we assessed the importance of differences in demographic and clinical characteristics at baseline (Table 4). There was no significant difference between the two groups in age, sex, or performance status or in tumor histology, location, or size.

Comment This report adds to the current wave of enthusiasm for multimodality therapy for esophageal carcinoma, a phenomenon that is likely borne of near-universal frustration with the results of conventional therapy using any single treatment option. The treatment plan described has several relatively unique features: administration of 5fluorouracil as a continuous infusion for 3 weeks, the total dosage of cisplatin, use of vinblastine, RT administered in hyperfractionated (twice daily) dosage, and the standard operative technique-THE without thoracotomy. Since 1984, several published series have reviewed experience with preoperative chemotherapy for esophageal cancer [ll, 17-20], and in nearly half of these, RT has also been administered, generally preoperatively (Table 5). Recent multiinstitutional nonrandomized trials [25,26] have evaluated the Wayne State experience [28] with preoperative cisplatin, 5-fluorouracil, and RT. As with our patients, patients undergoing esophagectomy had a complete re-

Table 5. Preoperative Chemotherapy and Radiation Therapy for Esophageal Carcinoma

Preoperative Drugs

Preoperative Radiation (rads)

No. Completing Preoperative Therapy

No. Operative Undergoing Mortality Esophagectomy (%)

Complete Response Rate (%)

Survival After Resection

Reference

Year

Austin et a1 [21]

1986

DDPI5-FU

3,000

11

11

2.5

36.0

Campbell et a1 [22]

1985

DDP/5-FU

3,000

21

15

20.0

14.0

McFarlane et a1 [23]

1988

3,000

22 (9 SCC, 13 Adeno)

10.5

36.0

Popp et a1 [24]

1986

3,000

21

10

17.0

9.5

21 at 30 mo

Poplin et a1 [25]

1987

DDP/ 5-FU(SCC), DDP/5-FU/ Mito-C (Adeno) DDP/5-FU, 5-FUI Mito-C, 5-FU/DDP/ Vincr DDP/5-FU

3,000

102

55

11.0

18.0

Seydel et a1 (261 Wolfe et a1 [27]

1988 1987

3,000 4,500

29 29 63

27 29 (17 SCC, 12 Adeno)

3.7 0

36.0 16.0

Orringer [Present report]

1990

DDP/5-FU DDP/ VP-l6(SCC), DDP/5-FU (Adeno) DDP/5-FU/ Vinbl

14 at 36 mo NA 44 at 24 mo

3,750 (21 patients), 4,500 (22 patients)

41

39 (20 SCC, 19 Adeno)

2.0

27.0

Aden0 = adenocarcinoma; squamous cell carcinoma;

DDP = cisplatin; 5-FU = 5-fluorouracil; Vinbl = vinblastine; Vincr = vincristine.

Mito-C

=

19 (9 SCC, 10 Adeno)

mitomycin;

(%)

36 at 30 mo 52 at 24 mo 66 at 24 mo, 33 at 36 mo

NA = data not available in reference;

60 at 24 mo, 46 at 36 mo SCC =

Ann Thorac Surg 1990;49:34&55

sponse rate of 18%to 30%, but mean survival times were not improved; in patients undergoing resection mean survival was only 14 months in the Southwest Oncology Group study [25]. This contrasts with a mean survival time of 29 months in our own series. Although few studies of preoperative adjuvant therapy for esophageal cancer have evaluated adenocarcinomas as well as squamous cell carcinomas [23, 271, our results indicate that this treatment regimen is equally effective in both histological types and that the likelihood of having a TO NO response is not significantly different in the two groups. Multimodality therapy has not yet been proven in a prospective randomized trial to result in improved survival in patients with esophageal carcinoma. Clearly, the price for such preoperative treatment is high and is measured in terms of toxicity, the need for prolonged preoperative hospitalization, and its physiological and psychological effects on an already weakened patient who must undergo a major thoracic operation. Despite this, we have not observed an increase in postoperative morbidity or mortality after pretreatment with RT and chemotherapy. Interdisciplinary cooperation and constant dialogue is required throughout the patient’s treatment course. Endoscopic placement of a nasogastric feeding tube may be necessary in patients whose dysphagia worsens during therapy. Operating schedules must be coordinated with patients’ completion of chemotherapy and RT, and flexibility must be maintained if a cancellation is required because of a persistently low white blood cell count. For most surgeons, this treatment is inconvenient and timeconsuming. It represents a departure from the long-held tenet that the “best” treatment for esophageal cancer is an efficient esophagectomy and esophagogastrostomy that returns the patient to his or her family for the few remaining months of life. Yet any of us who have devoted years of professional life to the care of patients with esophageal cancer cannot help but be favorably impressed by the encouraging data that are emerging. Skinner and associates have advocated an en bloc radical resection for potentially curable esophageal carcinomas as a means for improving survival. This extensive and meticulous dissection of the esophagus and surrounding mediastinal soft tissue may be viewed by some as the best, most aggressive ”cancer operation” currently available for esophageal carcinoma. In a preliminary report of 80 radical esophagectomies [29], however, the 3-year survival rate of 24% was not appreciably different from the 22% rate we reported for THE without thoracotomy and no in-continuity lymph node dissection [5]. More recently, Skinner and associates [30] used their operation more selectively in 31 additional patients, one half with squamous cell carcinoma and one half with adenocarcinoma. The overall 1-year and 2-year survival in this group is 65% and 32%, respectively, as compared with our overall 1- and 2-year survival of 72% and 60%, respectively in patients undergoing multimodality therapy. The actuarial 3-year survival in the patients identified by Skinner and associates as those most likely to be benefited by the more radical resection (14 W1 NO patients and 11 W1 N1 patients) was 71% and 55%, respectively.

ORRINGER ET AL MULTIMODALITY THERAPY FOR ESOPHAGEAL CARCINOMA

353

These results, although admirable, do not approach the 3-year survival of 100% in our TO NO patients. The logical explanation for the difference is that the latter group has had better local-regional control of their disease with multimodality therapy than can be provided by radical resection of the esophagus. This adds further support to the concept that biological behavior of the tumor, rather than magnitude of resection, determines survival in patients with esophageal carcinoma. That one quarter of the patients we treated had no residual tumor in their resected esophagus or adjacent lymph nodes (TO NO) and all of them are still alive and tumor-free with a median follow-up of 36 months cannot be a chance occurrence. The tumor-free survival of these patients appears to be a function of the preoperative therapy they received. If the apparent value of multimodality therapy in these patients is confirmed with prospective trials now underway, the challenge will then be to identify the predictive factors in the patient with esophageal carcinoma that indicate that a favorable response to such treatment is likely. Alternative treatment, perhaps immunotherapy, could then be offered to patients for whom expenditure of resources required for multidisciplinary therapy should be avoided. The suggestion that esophagectomy be omitted in the treatment plan of the patient undergoing chemotherapy and RT for local-regional disease [31] is premature and unjustified. Unless an effective noninvasive means of detecting residual disease in the esophagus can be found, esophageal resection not only provides the ultimate means for assessing the effectiveness of therapy, but also offers the best means of providing local disease control and potential prolongation of life. At present, there is simply no method of clinical staging that insures the pathological findings, and as new variations of preoperative therapy for esophageal cancer emerge (as they surely will), careful histological assessment of the resected specimen will remain the ”gold standard’ for documenting effectiveness of therapy. The ultimate proof of the value of multidisciplinary therapy in the treatment of esophageal cancer awaits the results of a controlled, randomized study now underway. Nonetheless, these preliminary results add further support to the growing notion that the natural history of esophageal carcinoma can be altered, and that long-term survival, not just palliation, may be achievable in many patients with this disease.

References 1. Ellis FH. Carcinoma of the esophagus. Cancer 1983;33:

264-80. 2. Anonymous. Cancer facts and figures. New York: American Cancer Society, 1988. 3. Orringer MB, Sloan H. Esophagectomy without thoracotomy. J Thorac Cardiovasc Surg 1978;76:643-54. 4. Orringer MB, Orringer JS. Transhiatal esophagectomy without thoracotomy-a dangerous operation? J Thorac Cardiovasc Surg 1983;85:72-80. 5. Orringer MB. Transhiatal esophagectomy without thoracotomy for carcinoma of the thoracic esophagus. Ann Surg 1984;200:282-7. 6. Coonley CJ, Bains M, Hilaris B, Chapman R, Kelsen DP.

354

Ann Thorac Surg 1990;49:34655

ORRINGER ET AL MULTIMODALITY THERAPY FOR ESOPHAGEAL CARCINOMA

Cisplatin and bleomycin in the treatment of esophageal carcinoma: a final report. Cancer 1984;54:2351-5. 7. Steiger Z, Franklin R, Wilson RF, et al. Eradication and palliation of squamous cell carcinoma of the esophagus with chemotherapy, radiotherapy, and surgical therapy. J Thorac Cardiovasc Surg 1981;82:713-9. 8. Kelsen D, Fein R, Cooney C, Heelan R, Bains M. Cisplatin, vindesine, and mitoguazone in the treatment of esophageal cancer. Cancer Treat Rep 1986;70:255-9. 9. Kelsen D, Hilaris 8, Coonley C, et al. Cisplatin, vindesine, and bleomycin chemotherapy of local-regional and advanced esophageal carcinoma. Am J Med 1983;75;64!5-52. 10. Leichman L, Steiger Z, Seydel HG, et al. Preoperative chemotherapy and radiation therapy for patients with cancer of the esophagus: a potentially curative approach. J Clin Oncol 1984;2:7%9. 11. Carey RW, Hilgenberg AD, Wilkins EW, Choi NC, Mathisen DJ, Grillo H. Preoperative chemotherapy followed by surgery with possible postoperative radiotherapy in squamous cell carcinoma of the esophagus. Evaluation of the chemotherapy component. J Clin Oncol 1986;4:697-701. 12. Forastiere AA, Gennis M, Orringer MB, Agha FP. Cisplatin, vinblastine, and mitoguazone chemotherapy for epidermoid and adenocarcinoma of the esophagus. J Clin Oncol 1987; 8:1143-9. 13. American Joint Committee on Cancer. In: Beahrs OH, Myers MH, eds. Manual for staging of cancer. 2nd ed. Philadelphia: JB Lippincott, 1983:61-72. 14. Dixon WJ, ed. BMDP statistical software manual. Berkeley: University of California Press, 1988. 15. Orringer MB, Stirling MC. Cervical esophagogastric anastomosis for benign disease-functional results. J Thorac Cardiovasc Surg 1988;96:887-93. 16. American Joint Committee on Cancer. In: Beahrs OH, Heuson DE, Hutter RV, Myers MH, eds. Manual for staging of cancer. 3rd ed. Philadelphia: JB Lippincott, 1988:63-65. 17. Hdgenberg AD, Carey RW, Wilkins WE, Choi NC, Mathisen DJ, Grill0 HC. Preoperative chemotherapy, surgical resection, and selective postoperative therapy for squamous cell carcinoma of the esophagus. Ann Thorac Surg 1988;45: 357-63. 18. Miller JI, McIntyre B, Hatcher CR Jr. Combined treatment approach in surgical management of carcinoma of the esophagus: a preliminary report. Ann Thorac Surg 1985;40289-93. 19. Roth JA, Pass HI, Flanagan MM, Graeber GM, Rosenberg JC,

20.

21. 22. 23.

24. 25.

26.

27. 28.

29. 30. 31.

Steinberg S. Randomized clinical trial of preoperative and postoperative adjuvant chemotherapy with cisplatin, vindesine, and bleomycin for carcinoma of the esophagus. J Thorac Cardiovasc Surg 1988;96:242-8. Shields TW, Rosen ST, Hellerstein SM, Tsang T, Ujiki GT, Kies MS. Multimodality approach to treatment of carcinoma of the esophagus. Arch Surg 1984;119:55%62. Austin JC, Postier RG, Elkins RC. Treatment of esophageal cancer: the continued need for surgical resection. Am J Surg 1986;152:592-6. Campbell WR, Taylor SA, Pierce GE, Hermrelk AS, Thomas JH. Therapeutic alternatives in patients with esophageal cancer. Am J Surg 1985;150665-8. McFarlane SD, Hill LD, Jolly PC, Kozarek RA, Anderson RP. Improved results of surgical treatment for esophageal and gastroesophageal junction carcinomas after preoperative combined chemotherapy and radiation. J Thorac Cardiovasc Surg 1988;95:41!5-22. Popp MB, Hawley D, Reising J, et al. Improved survival in squamous esophageal cancer. Arch Surg 1986;121:1330-5. Poplin E, Fleming T, Leichman L, et al. Combined therapies for squamous cell carcinoma of the esophagus, a Southwest Oncology Group study (SWOG-8037). J Clin Oncol 1987; 5:622-8. Seydel HG, Leichman L, Byhardt R, et al. Preoperative radiation and chemotherapy for localized squamous cell carcinoma of the esophagus: a RTOG study. Int J Radiat Oncol Biol Phys 1988;14:33-5. Wolfe WG, Burton GV, Seigler HF, Crocker IR, Vaughn AL. Early results with combined modality therapy for carcinoma of the esophagus. Ann Surg 1987;205:563-71. Leichman L, Steiger Z, Seydel HG, Vaitkevicius VK. Combined preoperative chemotherapy and radiation therapy for cancer of the esophagus: the Wayne State University, Southwest Oncology Group, and Radiation Therapy Oncology Group experience. Semin Oncol 1984;11:17&?-85. Skinner DB. En bloc resection for neoplasms of the esophagus and cardia. J Thorac Cardiovasc Surg 1983;85:59-71. Skinner DB, Ferguson MK, Soriano A, Little AG, Staszak VM. Selection of operation for esophageal cancer based on staging. Ann Surg 1986;204:391401. Poplin E, Leichman L, Seydel HG, Steiger L, Fleming C. SWOG 8073: combined therapy for squamous cell carcinoma of the esophagus (SCCE) [Abstract]. Proc Am SOCClin Oncol 1986;5:80.

DISCUSSION DR F. HENRY ELLIS, JR (Burlington, MA): I congratulate Dr Omnger and his associates for their contribution to the serious therapeutic dilemma that faces all of us who have tried to care for patients with cancer of the esophagus or cardia. It is quite obvious that current modalities of therapy leave much to be desired, and it is hoped that studies such as this will provide us with new insights that will improve our results. To evaluate studies such as this, we must enumerate some of the pros and cons of neoadjuvant therapy as compared with operation alone. As reported in the literature, about a third of patients so treated, whether with chemotherapy alone or with chemotherapy and radiation therapy, will show a complete response. It is said that this form of treatment provides better local control of the disease, and those that are responders show evidence of increased longevity. Are there any disadvantages? Neoadjuvant therapy certainly requires a longer period of treatment in which the patient is

under the care of a physician or is hospitalized, sometimes as long as 4 months. The incidence of toxicity is substantial, and as yet, there has been no evidence of any increase in 5-year survival rate. Now how do the results of Orringer and associates compare with what has been reported in the literature? They show approximately the same percentage of complete responders, close to a third. They do not mention much about local control of the disease, but longevity is somewhat increased, a median survival time of 29 months, which is slightly higher than in our patients with operation alone. Forty-six percent were alive at 3 years, though to be sure, this is a cumulative figure and not an actuarial one. How this figure relates to a historical control group that had a poorer survival is difficult to say. There are some negative features to the study. The period of therapy is prolonged (2 months in this series), though it is somewhat less than other reports of preoperative chemotherapy

Ann Thorac Surg 1990;49:34%55

and radiation therapy. Toxicity was considerable, with a 4.7% mortality, double the surgical mortality, and only one quarter of the patients were able to complete the adjuvant therapy as originally designed. Whether long-term survival is better than after operation alone has not been answered by this study. Results of the proposed prospective randomized study may provide the answer. For the sake of comparison, let me briefly relate our results with operation alone. 82.3% of our patients were operable and their resectability rate was close to 90%, with a 30-day mortality rate of 2.3%. The 5-year survival for curative resections was 23.3% and close to 34% as 3 years, lower to be sure than the 46% Orringer and associates report but their figures are cumulative, not actuarial, and are difficult to compare. In conclusion, I would like to ask Dr Orringer two questions: Why do you not proceed with palliative resection in the face of a metastatic deposit in the liver? I have found a doubling of the expected survival rate in such patients with excellent palliation of dysphagia. Finally, as it is quite clear from this study and from others that the "complete" responders do much better than nonresponders, what steps are you taking to determine some way to identify those patients who will be complete responders before initiating this form of treatment for everybody, two thirds of whom will probably not benefit from it? DR ORRINGER: As usual, Dr Ellis has targeted the key questions here. I think we are entering a new and exciting time in the treatment of this disease. We have seldom before been able to talk about such a group of patients with esophageal cancer in which 27% are alive with no residual disease that can be detected at 3 years. It truly seems that the natural history of this disease can be altered in certain patients. But what is different about those patients who have responded to preoperative neoadjuvant therapy and those who have not? This is where the new areas of cancer investigation are going to play a role, and this information, I believe, will change our basic approach to these patients. For example, we may soon be routinely assessing preoperative biopsy specimens for deoxyribonucleic acid content of the tumor cells with flow cytometry or evaluating other characteristics of the tumors at a cellular level using immunological tools. The results

ORRINGER ET AL MULTIMODALITY THERAPY FOR ESOPHAGEAL CARCINOMA

355

of the present study have stimulated us to begin a prospective randomized trial, and there is a need to utilize other diagnostic modalities such as esophageal ultrasound to help us differentiate those patients who are suitable for the type of multimodality therapy we have described from those in whom this treatment is ineffective and some alternative therapy is indicated. Patients who have spread of their cancer beyond the confines of the esophagus, to the liver or distant spread, are staged according to the current TNM staging as having stage IV esophageal cancer. In our experience, the majority of those patients live 6 months or less, and we are hard pressed to recommend major resectional therapy for patients with such advanced disease. So I do not share Dr Ellis' enthusiasm for going ahead with resection in the face of a hepatic metastasis. I do not feel that the effort is justified by the small reward gained and the increased morbidity such patients with advanced disease often experience. Multimodality therapy of esophageal cancer appears to be a "new frontier" area, and I hope to bring you follow-up on a prospective, randomized study, which really must now be done. I want to emphasize that our report today represents but another uncontrolled, unrandomized study, the conclusions of which do not justify establishing multimodality therapy for esophageal carcinoma as "state of the art" in the treatment of this disease. Perhaps one other word of caution is indicated here. A number of investigators have favored multiinstitutional studies of multimodality therapy for esophageal cancer, as few institutions can accumulate enough patients quickly to reach useful conclusions about the efficacy of specific therapy. But even in talking with some of my colleagues here it becomes clear that there is a tremendous difference in radiation therapy techniques from one institution to another. Our radiation therapists, for example, utilize a very innovative three-dimensional CT dosimetry planning system that allows precise delivery of radiation with much less scatter and toxicity. This may explain why our patients have not experienced an increase in pulmonary toxicity after preoperative irradiation and chemotherapy, which some of our colleagues have seen in their patients. Radiation therapy, chemotherapy, and surgical techniques vary widely between institutions. So we have to be careful in generalizing from single papers that this is the way to go for all. The answer will come from the badly needed prospective, randomized studies that are currently under way.