Preoperative chemotherapy for esophageal cancer based, on chemosensitivity testing

Preoperative chemotherapy for esophageal cancer based on chemosensitivity testing To examine the role and significance of in vitro chemosensitivity testing on ceUs cultured from endoscopic biopsy specimens, we evaluated outcomes in 57 patients with squamous ceU carcinoma of the thoracic esophagus. Over 3% years, 57 patients were treated with a single course of preoperative combination chemotherapy based on the results of chemosensitivity testing utilizing eight antineoplastic agents. A radiologically evident response (partial response) was seen in 23 patients (40.4 %) and histologic effects on grade fib tumors were seen in six patients (10.5 %). The median survival time of aU 57 patients was 16.3 months. The median survival time of nine patients who received palliative resection was 4.5 months. The median survival times after curative resection of the 19 patients with a partial response and 29 patients with no response were 27.7 months and 17.3 months, respectively (p < 0.009). The most frequent combination chemotherapy used was cis-dichlorodiammine platinum (II) vindesinej5-ftuorouracil, and a high rate of response was noted after therapy with etoposide and 4' -D-tetraliydropyranyl-doxorubicin. Chemosensitivity testing appears to be useful in selecting preoperative chemotherapeutic agents for patients with esophageal cancer, decreasing tumor recurrence after surgery and resulting in more partial responses than conventional chemotherapy. (J THoRAc CARDIOVASe SURG 1994;108:823-9)

Toshitada Qkuma, MD, Masakazu Yoshioka, MD, Susumu Isechi, MD, Keiichiro Kondo, MD, Yoichi Tabira, MD, Yoshitsugu Torigoe, MD, and Yoshimasa Miyauchi, MD, Kumamoto, Japan

DeSPite advances in surgical technique, esophageal cancer continues to have a poor prognosis, often recurring in lymph nodes or with distant metastases, even if the primary tumor is completely resected. Multimodality treatments, such as radiotherapy and chemotherapy.l' have been used in patients with esophageal cancer but have had no effect on long-term survival. Many trialsv 5 with neoadjuvant combination chemotherapy for esophageal cancer have been performed with cis-dichlorodiammine platinum (II) (CDDP) since this compound was shown to be a potent antitumor agent.? In our hospital, preoperative combination chemotherapy followed by surgery has been performed since 1983.7 This neoadjuvant chemotherapy has improved survival From the First Department of Surgery, Kumamoto University, School of Medicine, Honjo 1-1-1, Kumamoto City, Kumamoto 860, Japan Received for publication Nov. 17, 1993. Accepted for publication July 12, 1994. Address for reprints: Toshitada Okuma, MD, Izumi City Hospital, 520 Myojin-cho, Izumi City, Kagoshima 899-01, Japan. Copyright

w

1994 by Mosby-Year Book, Inc.

0022-5223/94 $3.00

+0

12/1/59162

significantly, particularly in patients with a good radiologic response. To establish the most effective chemotherapeutic regimen, we have used chemosensitivity testing of esophageal cancer cells obtained by endoscopic biopsy before operation since 1989.8 The purpose of this study was to assess the role and significance of neoadjuvant chemotherapy on the basis of the results of chemosensitivity testing. Patients and methods Sensitivity testing. Sensitivity was tested according to the dye exclusion method of Hongo.? Weisenthal, 10 and their associates. In brief, a pan-endoscope and the accompanying biopsy forceps (GIF-XQ 20, Olympus Co. Ltd., Tokyo, Japan) were used to obtain 10 specimens from the primary tumor of each patient with intrathoracic esophageal cancer. After the samples were washed with antibiotics (0.01% piperacillin sodium: Toyama Chemical Industry, Tokyo, Japan; 0.001 % dibecacin sulfate: Meiji Seika Co., Tokyo, Japan), they were minced, incubated with collagenase and trypsin-ethylenediaminetetraacetic acid for 10 minutes, filtered through mesh, and counted. Cells were cultured at a concentration of 2 to 4 X 105jml in RPMI-1640 medium with 10% inactivated autologous serum. Two hundred microliters of the following drugs at the indicated concentration were added to each well of a 96-well

823

The Journal of Thoracic and Cardiovascular Surgery November 1994

8 2 4 Okuma et al.

Table I. Patient demographics No. of patients Age (yr) Mean Range Sex (n) Male Female Tumor location (n) Upper thoracic Middle thoracic Lowerthoracic pTNM stage (n) I lIA lIB

57

64.4 ± 9.5

39-81 48 9 5 40

12 5 8

III

8 19

IV

17

microplate containing 104 cells/50 JII well and incubated for 48 hours: COOP, 0.2 JIg/ml; 5-fluorouracil (5-FU), 1.0 JIg/ml; vindesine (VDS), 0.005 JIg/ml; peplomycin (PLM), 0.4 JIg/ml; methotrexate (MTX), 0.03 JIg/ml; mitomycin C (MMe), 0.1 JIg/ml; etoposide (VP-16), 0.1 JIg/ml; and 4' -D-tetrahydropyranyl-doxorubicin (THP-adriamycin; THP-ADM, Adria Laboratories, Columbus, Ohio), 0.04 JIg/mI.l I Triplicate wells were tested for each drug. Plates were incubated at 37° Cin 95% air/5% carbon dioxide for 48 hours. Cell sensitivity was evaluated by the dye exclusion method. The absorbance of crystal violet in each well was analyzed by a Titertek Multiskan MC type 340 (Flow Laboratories Ltd., Ayrshire, United Kingdom). A cytotoxicity index (CI) was calculated by the following formula: Anticancer agent wells mean absorbance 00 CI = (l - Control wells mean absorbance) X 1 (%) Drug sensitivity was ranked according to the cytotoxicity index' Patients. Between January 1989 and May 1993,57 patients with squamous cell carcinoma of the thoracic esophagus underwent preoperative chemotherapy. All patients received a single course of preoperative chemotherapy using the three drugs with the highest cytotoxicity indexes as determined by chemosensitivity testing. Three or 4 weeks after completion of the preoperative chemotherapy, subtotal esophagectomy was performed via a thoracoabdominal approach followed by cervical esophagogastrostomy. Three-field dissection of lymph nodes was carried out in 51 (age <75 years) of the 57 patients, as detailed previously. I 2 The dosage and schedule of administration for each drug were as follows: COOP, 70 mg/m 2 per day, was administered on day 1; 5-FU, 500 mg/rn? per day, or PLM, 5 mg/day, was infused continuously from days 2 to 6; VDS, 3 mg/day, THPADM, 30 mg/m2 per day, and MMC, 10 mg/m2 per day were all administered intravenously on day 1; VP-16, 60 mg/day was administered on days 1 and 4; MTX, 10 mg/rn? per day was given on days 2, 4, and 6. Informed consent was obtained from all subjects before treatment. Radiologic and endoscopic evaluation of tumor size was per-

formed within 14 days after the completion of chemotherapy and measured according to the Guidelines for Clinical and Pathologic Studies on Carcinoma of the Esophagus.P After radical surgery, the effect on tumor histology was evaluated according to the classification of Shimosato, Oboshi, and Baba. 14 Staging of the disease was determined according to the pTNM classification of malignant tumors of the International Union Against Cancer (UICC), 1987. 15 Postoperative therapy. A portion of the resected tumor was immediately minced aseptically in the operating room and subjected to chemosensitivity testing. One or two courses of combination chemotherapy based on the results of postoperative chemosensitivity testing were planned for each patient. Thirtyfour (15 patients with a partial response and 19 patients with no response) of the 57 patients were treated with combination chemotherapy after the operation. However, 23 patients did not receive chemotherapy because they refused or postoperative complications precluded therapy. Pneumonia, anastomotic dehiscence, and sepsis accounted for the majority of contraindications to postoperative chemotherapy. In addition, when the cytotoxicity index was determined to be a negative value for the anticancer agents tested, postoperative chemotherapy was not administered. Postoperative radiotherapy consisting of 30 to 45 Gy was given to 10 (3 patients with a partial response and 7 with no response) of 48 patients treated with curative resection when the tumor stage was T3 or T4. Statistical analysis. Rates of survival and Kaplan-Meier survival curves were determined. The survival curves were compared by means of the generalized Wilcoxon test. Other statistical calculations were conducted by means of Fisher's exact test. Significant differences were determined to exist when a p value less than 0.05 (p < 0.05) was noted.

Results The clinical characteristics of the 57 patients enrolled are outlined in Table I. All patients tolerated the preoperative chemotherapy without severe toxicity, and all received subtotal esophagectomy. Of the 57 patients, 44 (77.2%) had stage lIB, III, or IV disease. The numbers of patients with a given tumor stage (pT) were as follows: Tl, 10; T2, 10; T3, 26, and T4, 11. Resection was performed for palliation in 9 of the 57 patients and with curative intent in 48 patients. Response rate and histologic efficacy. Tumor size was evaluated by radiologic study of the esophagus, because it afforded more precise sizing than endoscopic studies. Thirty-four of the 57 patients had a reduction in tumor size ofless than 50% (no response), and 23 patients (40.4%) had a reduction of more than 50% (partial response) (Table 11). The numbers of patients whose disease was classified as grade I, IIa, and lIb by histologically apparent chemotherapeutic efficacy were 29, 22, and 6, respectively. No patient had grade III or IV disease (see Table 11). All 6 patients with grade lIb disease had a reduction in tumor size of more than 50%.

The Journal of Thoracic and Cardiovascular Surgery Volume 108, Number 5

Okuma et al.

Table II. Radiologic response rate and histologic effect Radiologic effect Reduction in tumor size <50% (NR) ~50% (PR) Histologic effect Grade I Grade IIa Grade IIb Grade III

Table III. Tumor stage distribution of the patients with curative resections

NR 34 23 (40.4%)

29 22 (38.6%) 6 (10.5%)

o

NR: No response (either no change or progressive disease, with a reduction in tumor size of less than 50%); PR, partial response (tumor reduction rate of greater than 50%).

Survival. The median follow-up period of these 57 patients was 37 months (range 3 to 56 months), and the median survival time was 16.3 months. Nine patients who underwent palliative resection died of cancer recurrence within 489 days of the operation. The median survival time was 4.5 months. Of the 48 patients who received curative resections, 19 had preoperative tumor reduction of more than 50% (partial response group) after chemotherapy, whereas the other 29 patients had a tumor reduction of less than 50% (no response group). The distribution of tumor stage for each group is seen in Table III. Fewer patients with stage IV disease were present in the partial response group than in the no response group; however, the overall stage distribution was not significantly different between groups. The median survival times of the partial response group and the no response group were 27.7 months and 17.3 months, respectively. A significant difference was seen in the Kaplan-Meier survival curves between the partial response group and the no response group (p < 0.009). One-, 2-, 3-, and 4-year survivals of patients in the partial response group were 94.7%, 94.7%, 60.9%, and 60.9%, respectively, whereas those of patients in the no response group were 72.0%, 42.8%, 32.1%, and 32.1%, respectively (Fig. 1). Cause of death. Only one patient in the partial response group who received curative resection died of local recurrence. He was 39 years old and had a tumor stageofpT4 Nl MO (T: 12cm by radiologic assessment). The azygos vein, thoracic duct, pericardium, and lung were invaded by the tumor and removed with the esophagus. He was treated with two courses of chemotherapy based on the results of chemosensitivity testing of neoplastic cells from the resected specimen, as well as radiotherapy, 45 Gy, after the operation. He died of local recurrence 1051 days after the operation. Three other

825

PR*

pTNM stage

(n = 29)

(n = /9)

I IIA IIB III

2 5 3 8 II

3 3

IV

3

7 3

NR, No response; PR, Partial response.

"No significant difference versus NR.

patients in the partial response group died: one committed suicide 108 days after the operation and two died of unrelated causes without recurrence. Eleven patients in the no response group who received curative resection died of recurrent cancer, mainly distant metastases to the lung, bone, and brain (Table IV). One patient died of sepsis within 30 days of the operation; autopsy revealed that sepsis resulted from necrotic pancreatitis unrelated to the chemotherapy. Drug selection. Six of the 57 patients had negative cytotoxicity indexes for all eight drugs, and they received conventional chemotherapy of COOP/VOS/5-FU used in our hospital before 1989. Twenty-four patterns of combination chemotherapy were determined by sensitivity testing for the other 51 patients. The content of the three-drug chemotherapy regimens for the 51 patients is shown in Table V. The most frequently used combination was COOP/VOS/5-FU, administered to seven patients. The second most common combination was COOP/PLM/THP-AOM, administered to four patients. COOP/PLM/MTX, COOP/5-FU/THP-AOM, COOP/5-FU/VP-16, COOP/5-FU/MTX, and 5-FU/ MMC/THP-AOM were administered to three patients, respectively. Other combinations were administered to one or two patients. Especially, the combinations of THP-AOM/VP-16/5-FU and VOS/MMC/PLM are rare regimens for esophageal cancer, and two of three patients who were treated with the regimens had a reduction in tumor size of less than 50%. The most frequently selected drug was COOP administered to 41 of the 51 patients (80.4%). 5-FU, PLM, and THP-AOM were used for 24,21, and 16 patients, respectively. MMC was administered to 14 patients and resulted in the lowest response rate (28.6%) (Table VI). Discussion Lymph node metastases are found in the early stages of esophageal cancer, because the esophagus has rich lymphatic networks above and below the muscularis

826

The Journal of Thoracic and Cardiovascular Surgery November 1994

Okuma et al.

19

14

11

5

29

20

8

5

3

100

PR group ( 0=19 )

5

NR group ( 0=29 )

oL-----.----...----..,-----,---2

3

Years after surgery

Fig. 1. Kaplan-Meier survival curvesfor the partial response and no response groupsof patientsreceiving curative resection, p < 0.009 (generalized Wilcoxon test). PR, Partial response assessed radiologically; NR, no response (no change, or progressive disease) assessed radiologically. The small number at the top of the figure is the number of patients still being followed up at the indicated point along the survival curve.

Table IV. Cause of death in the NR and PR groups* Cause of death

NR

Cancer recurrence Other disease

In = 29) II 6

PRln=19)

I 3

NR. No response; PR, partial response.

*All patients received curative surgical

resection.

mucosae. Specifically, 50.0% of the patients in this study with esophageal cancer invasion to the submucosa by Tl tumors had lymph node metastases. Therefore we believe esophageal cancer with invasion to the submucosa is an advanced cancer, whereas other Tl tumors represent early cancer. In addition, esophageal cancer rapidly penetrates the esophageal wall, invading not only neighboring structures such as the aorta and respiratory organs, but also often resulting in distant metastases. As a result, esophageal cancer is one of the gastrointestinal malignant diseases with a typically dismal outcome after surgery. Various multimodality approaches have been used after surgery to improve this poor prognosis, but none has been successful to date. Neoadjuvant combination chemotherapy for esophageal cancer has been studied by Kelsen and coworkers" since 1978. Neoadjuvant chemotherapy has several theoretical advantages. First, laboratory studies have dem-

onstrated superiority of preoperative administration of chemotherapy in animal models. 16, 17 Second, administration of anticancer drugs when metastatic tumor growth is low may minimize the probability of spontaneous drug resistance.P Finally, preoperative chemotherapy allows in vivo assessment of the effectiveness of a drug or drug combination in a given patient by analyzing tissue obtained at operation. As a result of these theoretical and practical considerations, we have performed preoperative chemotherapy since 1983, and we detail the efficacy of neoadjuvant chemotherapy in this report. The survival of patients with a partial response to preoperative chemotherapy by radiologic assessment was significantly longer than that of patients with no change or progressive disease in a prior study.l" In that study, the overall survival improved, but neoadjuvant chemotherapy with CDDP/VDS/PLM or CDDP/ VDS/5-FU was not always associated with a good outcome as assessed by radiologic and histologic criteria. In fact, empiric neoadjuvant chemotherapy was used for 62 patients but resulted in a greater than 50% decrease in tumor size in only nine of the 62 patients (14.5%), and histologic efficacy of grade lib disease was seen in four of the 62 patients. Four-year survival was 27.4%. The concept of tumor cell "individuality" is derived from the observation that the same drug is not always

The Journal of Thoracic and Cardiovascular Surgery Volume 108, Number 5

Okuma et al.

Table V. Content of the combined chemotherapy

Table VI. Drug selection and response rate in 51

regimens for 51 patients

patients

Regimen

CDDP jVDSjTHP-ADM CDDPjTHP-ADMjVP-16 CDDPjMMCjMTX CDDPjVDSjMTX CDDPjVDSjPLM CDDPjVDSj5-FU CDDPjPLMjMTX CDDPjPLMjVP-16 CDDPjPLMjTHP-ADM CDDPj5-FU jTHP-ADM CDDPj5-FUjMTX CDDPjTHP-ADMjMTX CDDPj5-FUjVP-16 CDDPjMMCjVP-16 CDDPjPLMjMMC CDDP!5-FUjMMC CDDPj5-FUjPLM CDDPjPLM 5-FUjTHP-ADMjVP-16 5-FUjMMCjTHP-ADM MMCjVDSjPLM MMCjPLMjMTX 5-FUjPLMjMTX MMC/.vP-16jPLM

No. of cases

Patients with susceptible tumor selected by sensitivity testing

I

2 I I

2 7

3 2 4

3 3

2 3 I 2 2 I I

I 3 2 2 I I

CDDP, Cis-diamminedichloroplatinum (II); VDS, vindesine sulfate; THP-ADM, THP-Adriamycin (4' -D-tetrahydropyranyl-doxorubicin); VP-/6, etoposide; MMC, mitomycin C; MTX, methotrexate; PLM, peplomycin; 5-FU, 5-fluorouraeil.

effective in different patients with the same disease, histopathologic characteristics, and stage of disease.P We 8 have attempted to use chemosensitivity testing of esophageal cancer cells from endoscopic biopsy specimens, as previously reported. The dye exclusion method used in this study was introduced by Hongo and associates? in 1986 as a modification of the method ofWeisenthal and colleagues. 10 Hongo's group defined drug sensitivity as a cytotoxicity indexes greater than 60%. The overall cytotoxicity indexes generally was low in this study because endoscopic specimens contain not only cancer cells but also fibroblasts, lymphocytes, tissue debris, and other contaminants.f We have no data to indicate appropriate in vivo drug levels that correlate with in vitro sensitivity tests; thus no specific break point for percent cytotoxicity indexes was used to exclude a specific drug in this trial. We therefore used the three agents with the highest cytotoxicity indexes determined by the results of chemosensitivity testing, the first trial to do so. Although we have no specificdata to support it, our general impression is that a cytotoxicity index of I0% or more is necessary to achieve a good clinical result.

827

Patients showing PR

Drug

No.

%

No.

%

CDDP 5-FU Peplomycin THP-ADM Mitomycin C Methotrexate Vindesine Etoposide

41 24 21 16 14 13 13 10

80.4 47.1 41.2 31.4 27.5 25.5 25.5 19.6

20 9 10 8 4 6 6 5

46.3 33.3 47.6 50.0 28.6 46.2 46.2 50.0

PR, Partial response; CDDP, cis-diarnminedichloroplatinum (II); 5-FU, 5-fluorouracil; THP-ADM, 4' -D-tetrahydropyranyl-doxorubicin.

In this study, six patients whose cytotoxicity indexes was determined to be a negative value received conventional chemotherapy, resulting in only one partial response. This finding highlights the utility of sensitivity testing. So that a negative cytotoxicity index can be avoided, as many anticancer agents as possible must be tested. Although a partial response was radiologically demonstrated in four of the nine patients who received palliative resections, all nine died of cancer. Their deaths demonstrate that palliative resection is rarely associated with long-term survival, even if a partial response to chemotherapy is achieved. Terz and coworkers' have reported that 13 of 17 patients treated with preoperative CDDP/5-FU chemotherapy and radiotherapy had complete clinical responses; however, the treatment did not result in increased survival when compared with that of historical controls. Among their 17 patients, 10 were treated with transhiatal esophagectomy; thus lymph node dissection of the mediastinum may not have been performed. In our study, lymph node metastases were seen in 44 of 57 patients (77 .2%). As a basic concept, we believe operative control of local tumor spread and regional lymph node metastases is necessary to cure patients with esophageal carcinorna.'? Therefore, three-field dissection of lymph nodes was used for our patients with thoracic esophageal cancer. Kelsen/! has reported response rates to various combinations of drugs to be 15% with CDDP/bleomycin, 53% with CDDP/bleomycin/VDS, and 41% with CDDP/ VDS/guanylhydrazone. Schlag and coworkers-i have reported a response rate of 44% with the combination of CDDP/VDS/bleomycin, and Roth and colleagues-' have reported a 47% response rate using CDDP/VDS/ bleomycin. All these authors used two courses of preoperative chemotherapy. Schields and associates/" have

The Journal of Thoracic and Cardiovascular Surgery November 1994

8 2 8 Okuma et al.

treated 17 patients with three cycles of CDDP/5-FU chemotherapy, and a 76% response rate was achieved. In the present study, we treated 57 patients with a single course of chemotherapy and achieved a 40.4% response rate. Although the response rate was not as high as that noted by other investigators, our patients had a significantly better response rate than that achieved by blind neoadjuvant chemotherapy used in our hospital before 1989 (p = 0.0015). Also, six patients had a reduction in tumor size of greater than 40% (minor improvement) and might have achieved a partial response if two or three courses of preoperative chemotherapy had been used. Moreover, the dose of each drug used in our study was lower than doses used in many of the referenced reports. The reason for this practice was that if progressive disease occurred despite several courses of preoperative chemotherapy, severe toxicity might preclude aggressive surgical resection, and radical surgery is believed to be the mainstay of therapy for patients with esophageal cancer. In general, histologic evidence of efficacy does not parallel clinical response.' In our study only six of 57 patients had histologic evidence of a grade lIb response. No histologically complete responses were seen, perhaps because of the moderate doses we used. Other authors have reported a median survival ranging from 9 to 18 months and a 2-year survival of between 15% and 30%.23,25-27 In our study, the median survival of the 57 patients was 16.3 months, and 2- and 4-year survivals were 50.9% and 35.2%, respectively. Also, the survival curve of the partial response group with curative resection was significantly better than that of the no response group (p < 0.009). Although this study was not a randomized trial, the number of patients who received postoperative chemotherapy and radiotherapy was not different between the partial response group and the no response group. Many patients died of distant metastases in the no response group. The results demonstrate that preoperative chemotherapy may decrease tumor recurrences in patients who achieve a partial response. Sensitivity testing revealed 24 patterns of drug combinations to be of potential value. The most frequently used combination was CDDP/VDS/5-FU. A susceptibility rate of 50.0% was noted for both THP-ADM and VP-16. MMC had a low susceptibility rate (27.5%) and resulted in the lowest response rate, 28.6%. MMC may not be a suitable drug for treatment of esophageal cancer. Although 5-FU had a relatively high susceptibility rate, the response rate was low. Higher doses will be considered for 5-FU therapy in the future. In addition, sensitivity testing to additional anticancer agents should be used to

decrease the number of patients with negative cytotoxicity indexes. Conclusion

Chemosensitivity testing appears to be useful in selecting preoperative chemotherapeutic agents for patients with esophageal cancer, decreasing tumor recurrence after surgery, and achieving a partial response in more patients than does empiric chemotherapy. We thank Miss H. Taniguchi for her technical assistance. REFERENCES I. Goldminc M, Maddern G, Le Prise E, Meunier B, Campion JP, Launois B. Oesophagectomy by a transhiatal approach or thoracotomy: a prospective randomized trial. Br J Surg 1993;80:367-70. 2. Schlag P. Randomisierte Studie zur praoperativen Chemotherapie beim Plattenepithelcarcinoma des Oesophagus. CAO-Studiengruppe Oesophaguscarcinom. Chirurgie 1992;63:709-14. 3. Terz JJ, Leong LA, Lipsett JA, Wagman LD. Preoperative chemotherapy and radiotherapy for cancer of the esophagus. Surgery 1993;114:71-5. 4. Kelsen DP, Cvitkovic E, Bains M, et al. Cis-dichlorodiammineplatinum (II) and bleomycin in the treatment of esophageal carcinoma. Cancer Treat Rep 1978;62:

1041-6. 5. 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. 6. Rosenberg B, Vancanp L, Trosko JE, Mansour VH. Platinum compounds: a new class of potent antitumor agents. Nature 1969;222:385-6. 7. Okuma T, Torigoe Y, Hongo H, Okamura K, Miyauchi Y. Combined therapy with cisplatin and peplomycin for esophageal cancer surgery. Jpn J Gastroenterol Surg 1984; 17:1973-9. 8. Kondo K, Okuma T, Yoshioka M, Trigoe Y, Miyauchi Y, Katsuki T. Preoperative in vitro chemosensitivity test of esophageal cancer with endoscopic specimens. Cancer 1993;71:661-6. 9. Hongo T, Mizuno Y, Haraguchi S, Yoshida T. A new anticancer drug sensitivity test using the microplate culture and surviving tumor cell staining method. Jpn J Cancer Chemother 1986;13:247-54. 10. Weisenthal LM, Marsden JA, Dill PL, Macaluso CK. A novel dye exclusion method for testing in vitro chemosensitivity of human tumor. Cancer Res 1983;43:749-57. 11. Kanzawa F. Update of a human tumor clonogenicassay in carcinoma of the lung. Jpn J Cancer Chemother 1985; 12:1552-9. 12. Okuma T, Kaneko H, Yoshioka M, Torigoe Y, Miyauchi Y. Prognosis in esophageal carcinoma with cervicallymph node metastases. Surgery 1993;114:513-8.

The Journal of Thoracic and Cardiovascular Surgery Volume 108, Number 5

13. Japanese Society for Esophageal Disease. Guidelines for clinical and pathologic studies on carcinoma of the esophagus. 8th ed. Tokyo: Kanehara & Co., Ltd, 1992: 59-93. 14. Shimosato Y, Oboshi S, Baba K. Histological evaluation of effects of radiotherapy and chemotherapy for carcinomas. Jpn J Clin OncoI1971;1:19-35. 15. Hermanek P, Sobin LH. VICC International Union Against Cancer, TNM classificationof malignant tumours. 4th ed. London: Springer-Verlag, 1987:40-2. 16. Fisher B, Gunduz N, Saffer E. Influence of the interval between primary tumor removal and chemotherapy on kinetics and growth of metastases. Cancer Res 1983; 43:1488-92. 17. Pendergrast WJ, Drake WP, Mardinary MR. A proper sequence for the treatment of B16 melanoma: chemotherapy, surgery, and immunotherapy. J Natl Cancer Inst 1976;57:539-44. 18. Goldie JI1, Coldman AJ. The genetic origin of drug resistance in neoplasms: implications for systemic therapy. Cancer Res 1984;44:3643-53. 19. Okuma T, Torigoe Y, KondoK, Okamura K, Miyauchi Y. The outcome of extended nodal dissection and combination chemotherapy for resected cases of thoracic esophageal cancer. In: Skinner DB, ed. Proceedings of the Fourth World Congress of the International Society for Disease of the Esophagus; 1989 Sept. 6-8. Chicago: International Society for Disease of the Esophagus, 1989:116.

Okuma et al.

829

20. Hamburger AW, Salmon SE. Primary bioassay of human myeloma stem cells. J Clin Invest 1977;60:846-54. 21. Kelsen DP. Preoperative chemotherapy in esophageal C(l1'cinoma. World J Surg 1987;11:433-8. 22. Schlag P, Herrmann R, Raeth V, Lehner B, Schwarz V, Herfarth C. Preoperative (neoadjuvant) chemotherapy in squamous cell cancer of the esophagus. Recent Results Cancer Res 1988;110:14-20. 23. Roth JA, Pass HI, Flanagan MM, Graber GM, Roserberg JC, 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. 24. Shields TW, Rose ST, Hellerstein SM, Tat-kin T, Vjiki G, Kies M. Multimodality approach to treatment of carcinoma of the esophagus. Arch Surg 1984;119:558-62. 25. Poplin E, Fleming T, Leichman L. Combined therapies for squamous-cell carcinoma of the esophagus, a Southwest Oncology Group Study (SWOG-8037). J Clin Onco11987; 5:622-8. 26. Seydel HG, Leichman L, Byhardt R. Preoperative radiation and chemotherapy for localized squamous cell carcinoma of the esophagus: AR TOG study. Int J Radiat Oncol Bioi Phys 1988;14:33-5. 27. John MJ, Flam MS, Mowry PA. Radiotherapy alone and chemoradiation for nonmetastatic esophageal carcinoma: a critical review of chemoradiation. Cancer 1989;63:2397403.