USA Intergroup 113 - CiteSeerX

10 downloads 0 Views 1003KB Size Report
Aug 20, 2007 - David P. Kelsen, Katryn A. Winter, Leonard L. Gunderson, Joanne Mortimer, Norman C. Estes, ...... Weber WA: Use of PET for monitoring cancer.
VOLUME

25



NUMBER

24



AUGUST

20

2007

JOURNAL OF CLINICAL ONCOLOGY

O R I G I N A L

R E P O R T

Long-Term Results of RTOG Trial 8911 (USA Intergroup 113): A Random Assignment Trial Comparison of Chemotherapy Followed by Surgery Compared With Surgery Alone for Esophageal Cancer David P. Kelsen, Katryn A. Winter, Leonard L. Gunderson, Joanne Mortimer, Norman C. Estes, Daniel G. Haller, Jaffer A. Ajani, Walter Kocha, Bruce D. Minsky, Jack A. Roth, and Christopher G. Willett From the Memorial Sloan-Kettering Cancer Center, New York, NY; the Radiation Therapy Oncology Group Statistical Office; the University of Pennsylvania, Philadelphia, PA; Mayo Clinic Cancer Center, Scottsdale, AZ; Washington University, St Louis, MO; the University of Kansas, Lawrence, KS; The University of Texas M.D. Anderson Cancer Center, Houston, TX; Duke University, Durham, NC; and the University of Western Ontario, London, Ontario, Canada. Submitted January 4, 2007; accepted May 31, 2007. Supported by Grants No. CA 21661, CA 62115, and CA 37422 from the National Cancer Institute, Bethesda, MD. Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this article. Address reprint requests to David P. Kelsen, MD, Gastrointestinal Oncology Service, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021; e-mail: kelsend@ mskcc.org. © 2007 by American Society of Clinical Oncology

A

B

S

T

R

A

C

T

Purpose We update Radiation Therapy Oncology Group trial 8911 (USA Intergroup 113), a comparison of chemotherapy plus surgery versus surgery alone for patients with localized esophageal cancer. The relationship between resection type and between tumor response and outcome were also analyzed. Patients and Methods The chemotherapy group received preoperative cisplatin plus fluorouracil. Outcome based on the type of resection (R0, R1, R2, or no resection) was evaluated. The main end point was overall survival. Disease-free survival, relapse pattern, the influence of postoperative treatment, and the relationship between response to preoperative chemotherapy and outcome were also evaluated. Results Two hundred sixteen patients received preoperative chemotherapy, 227 underwent immediate surgery. Fifty-nine percent of surgery only and 63% of chemotherapy plus surgery patients underwent R0 resections (P ⫽ .5137). Patients undergoing less than an R0 resection had an ominous prognosis; 32% of patients with R0 resections were alive and free of disease at 5 years, only 5% of patients undergoing an R1 resection survived for longer than 5 years. The median survival rates for patients with R1, R2, or no resections were not significantly different. While, as initially reported, there was no difference in overall survival for patients receiving perioperative chemotherapy compared with the surgery only group, patients with objective tumor regression after preoperative chemotherapy had improved survival. Conclusion For patients with localized esophageal cancer, whether or not preoperative chemotherapy is administered, only an R0 resection results in substantial long-term survival. Even microscopically positive margins are an ominous prognostic factor. After a R1 resection, postoperative chemoradiotherapy therapy offers the possibility of long-term disease-free survival to a small percentage of patients.

0732-183X/07/2524-3719/$20.00 DOI: 10.1200/JCO.2006.10.4760

J Clin Oncol 25:3719-3725. © 2007 by American Society of Clinical Oncology

INTRODUCTION

Definitive control of the primary site is one of the most important factors in delivering curative therapy for patients with locoregional esophageal cancer. While recent studies indicate that chemoradiotherapy therapy is an alternative, surgery has traditionally been the primary curative treatment. However, because the risk of distant failure is very high, there is intense interest in combining regional treatment of surgery and/or radiation with systemic therapy. We performed a random assignment trial comparing combined-modality therapy using

chemotherapy followed by surgery with surgery alone in patients with either adenocarcinoma or squamous cell cancer of the esophagus. The initial results of this trial did not show a significant improvement in survival for patients receiving cisplatin and fluorouracil chemotherapy followed by surgery versus those receiving surgery alone.1 This report updates those results with long-term follow-up, and in addition, studies the relationship between the completeness of the resection and outcome. We evaluated both patients randomly assigned to surgery alone as well as patients randomly assigned to chemotherapy followed by surgery. Resections were defined using standard criteria: R0 versus R1 versus 3719

Kelsen et al

R2 or versus no resection. Outcome was defined as disease-free and overall survival. We also evaluated response to preoperative chemotherapy, failure patterns, and the effect of postoperative treatment on outcome. PATIENTS AND METHODS This study involved investigators from four national cooperative groups: Radiation Therapy Oncology Group (the coordinating group), the Cancer and Leukemia Group B, the Southwest Oncology Group, and the Eastern Cooperative Oncology Group. The primary end point of the study was overall survival. For the purposes of this analysis, the secondary end points included relapse pattern as a function of the type of the resection (R0 v R1 v R2), as well as disease-free survival. Eligibility and pretreatment evaluation have previously been reported in detail.1 In brief, study eligibility requirements included histologically confirmed epidermoid or adenocarcinoma of the esophagus, including the gastroesophageal junction, with or without regional lymph node metastasis (tumor stages I, II, or III, any nodal stage, and no metastasis [M0]) using the TNM classification schema. All patients were at least 18 years of age, had adequate hepatic, renal, and bone marrow function, and could tolerate the planned surgical procedure. Patients with tumors in the cervical esophagus (upper border fewer than 18 cm from the incisor teeth), those with supraclavicular or distant metastasis, or those with T4 tumors were ineligible. Patients could not have received prior therapy for their esophageal cancer. Pretreatment evaluation included a complete medical history and physical examination, computed tomography scan of the chest and abdomen, barium contrast x-ray of the upper gastrointestinal tract, a chest x-ray, and complete blood and platelet counts, and biochemical screening profiles. Bronchoscopy was performed in those patients with tumors in the upper third of the esophagus; endoscopic ultrasonography was encouraged but not required. The randomization scheme has also been previously described.1 Patients were stratified on the basis of extent of weight loss and cell type (epidermoid v adenocarcinoma), and then randomly assigned to either immediate surgery or to receive three cycles of chemotherapy using cisplatin and fluorouracil before surgery. Patients randomly assigned to chemotherapy who had stable or responding disease, and in whom an R0 resection was performed, were to receive two cycles of chemotherapy after resection. Surgery The same surgery was performed whether or not preoperative chemotherapy was given. The following surgical procedures were considered acceptable: an Ivor-Lewis esophagogastrectomy with a high intrathoracic anastomosis above the level of the azygos vein, a subtotal thoracic esophagectomy by means of a left thoracoabdominal incision with an anastomosis above the aortic arch, or a complete thoracic esophagectomy. Either a gastricesophageal anastomosis or colonic interposition with a cervical anastomosis was acceptable. Transhiatal esophagectomy was acceptable only for lesions located below the carina. The proximal and distal margins had to be at least 2 cm below the gross tumor. Assessment of the margin by examination of a frozen section before completion of the surgery was strongly recommended. Resection Type Resections were classified as curative when all gross tumor tissue was removed and microscopical examination revealed all margins to be free of tumor (R0). Resections were considered palliative either when microscopical examination revealed positive margins (R1; a positive margin was defined as tumor tissue at or ⬍ 1 mm from the radial [deep], proximal, or distal margins) or when there was residual local (but not distant) gross disease (R2). Chemotherapy The chemotherapy treatment plan involved cisplatin and fluorouracil. The doses and schedule were reported previously.1 Radiation Therapy Radiation therapy was not part of the pre- or postoperative treatment plan. However, patients who had an R1 or R2 resection or who had local recurrence at a later date could receive radiation therapy at the discretion of the 3720

investigator. For this reason, a detailed plan of radiation therapy was included in the study protocol. In most cases, for patients undergoing an R1 or R2 resection, chemotherapy was given concurrently with radiation. The treatment plan most commonly used involved fluorouracil given by infusion plus concurrent cisplatin. Chemotherapy as a portion of concurrent treatment with radiation for patients having less than R0 resections was not mandated in the original protocol. Relapse Pattern Patients were followed frequently for up to 5 years from entrance into this study. For R0 and R1 patients, if the patient had recurrence of disease, this was coded as local (at the primary site including the anastomosis), locoregional (including regional lymph node drainage), or distant. For all patients, distant disease was defined as distant lymph node sites, or involvement of distant organs including lung, liver, bone, and subcutaneous tissue. Postoperative chemotherapy whether given as a single modality or concurrently with radiation was not defined in this study, but was left to the discretion of the treating investigator. Statistics Failure to achieve overall survival was defined as death due to any cause, and survival time was measured from the date of randomization to the date of death or last follow-up. Failure for disease-free survival was defined as local, regional, or distant failure or death due to any cause. Disease-free survival time was measured from the date of randomization to the date of first failure or last follow-up. Patients without an R0 or R1 reaction were considered to have experienced local failures at day 1 and therefore have a DFS time of randomization date ⫹ 1. Both overall and disease-free survival were estimated using the Kaplan-Meier method,2 and comparisons were made with the log-rank test.3 The following multivariate analyses were performed using the Cox regression model for both overall and disease-free survival; for both treatment arms combined, these variables were used: dummy variables created for resection using R0 as the reference level, weight loss (⬍ 10% v ⱖ 10%), and histology (squamous v adenocarcinoma), and for the chemotherapy plus surgery arm, the only variables used were induction response (responded v did not respond), weight loss (⬍10% v ⱖ 10%), and histology (squamous v adenocarcinoma).

RESULTS

From August 1990 to December 1995, a total of 467 patients were registered for this study. Since the initial reporting of this study,1 additional information was received that resulted in the case status of three patients randomly assigned to the chemotherapy arm to be changed from ineligible to eligible. Of the 467 patients, 443 were eligible and had adequate follow-up. Two-hundred sixteen patients received chemotherapy followed by surgery, and 227 patients underwent surgery alone. Patient characteristics are presented in Table 1. For this long-term update and the purpose of the analysis of failure pattern on the basis of extent of resection, all data received and processed through July 2003 are included in the analysis. The median possible duration of participation in the study ranges from 7.8 to 13.0 years. Outcome of Surgery Of 227 eligible patients randomly assigned to the surgery arm, 218 underwent surgery; 59% had R0 resections, 15% had R1 resections, and 15% had R2 resections (Appendix Table A1, online only). Among the 213 patients randomly assigned to chemotherapy followed by surgery, 180 underwent surgery; 63% of all eligible chemotherapy and surgery patients had an R0 resection. R1 resections were performed in 4% and R2 were performed in 11% of these patients. JOURNAL OF CLINICAL ONCOLOGY

Chemotherapy and Surgery v Surgery for Esophageal Cancer

No. of Patients Characteristic and Treatment Group

Surgery Chemotherapy ⫹ Surgery

Registered All 234 Epidermoid cancer 108 Adenocarcinoma 126 Eligible, with adequate follow-up data All 227 Epidermoid cancer 106 Adenocarcinoma 121 Sex All Male 187 Female 40 Epidermoid cancer Male 79 Female 27 Adenocarcinoma Male 108 Female 13 Median age, years All 62 ⫾ 9.5 Epidermoid cancer 62 ⫾ 9.5 Adenocarcinoma 62 ⫾ 9.6 Race All White 176 Black 35 Other 16 Epidermoid cancer White 59 Black 33 Other 14 Adenocarcinoma White 117 Black 2 Other 2 Loss of ⬎ 10% of body weight, % of patients All 23 ⫾ 7.1 Epidermoid cancer 31 ⫾ 7.4 Adenocarcinoma 15 ⫾ 6.4

233 112 121 216 101 115

185 31 80 21 105 10 61 ⫾ 9.7 62 ⫾ 9.0 61 ⫾ 10.3

167 36 13 53 35 13 114 1 0 23 ⫾ 7.3 30 ⫾ 8.4 17 ⫾ 6.0

We evaluated failure pattern on the basis of completeness of resection. Appendix Table A2 (online only) presents the first failure pattern in patients undergoing an R0 (potentially curative) resection1; 129 patients undergoing surgery alone had an R0 resection, excluding patients who died in the immediate postoperative period. Locoregional failure only was seen in 21% of patients and locoregional plus distant disease in an additional 8% of patients. Distant disease occurred in 43% of patients. The locoregional failure rate was slightly but not statistically lower in patients undergoing chemotherapy plus surgery who had an R0 resection (126 patients). Locoregional failure only was seen in 19% of chemotherapy and surgery patients, while locoregional plus distant disease was seen in an additional 8% of patients. Distant disease occurred in 41% of patients. Overall then, 29% of patients undergoing surgery alone after an R0 resection and 27% of those receiving chemotherapy plus surgery undergoing an R0 resection had a component of locoregional failure as their first site of failure. For patients undergoing an R1 resection in which the only residual disease was a microscopically positive margin, the failure pattern www.jco.org

was analyzed for all patients, as well as histologic subtype (adenocarcinoma v squamous cell cancer). Data was reviewed for 34 patients (data on one patient were not available) undergoing surgery alone who had an R1 resection; only nine patients receiving chemotherapy and surgery had an R1 resection. Approximately one half of all patients had adenocarcinoma, and one half had squamous cell cancers. The distribution by histologic subtype for the group of patients undergoing an R1 resection was similar to that of all patients entered onto the study. For R1 patients undergoing surgery alone, locoregional failure occurred first in 41% of patients, and 32% of patients had distant failure only as first site of failure. Both locoregional and distant failure were higher in patients with a histology of adenocarcinoma, as opposed to those with squamous histology. The number of patients having an R1 resection who first received chemotherapy was small (nine patients). Four of nine patients experienced distant failure first. No clinical recurrence was seen in a total of nine R1 patients (21%; two underwent surgery after induction chemotherapy). All of these patients received postoperative chemoradiotherapy therapy. Among surgery-only patients, 18 of 34 patients received postoperative radiation therapy usually given with chemotherapy. Seven of these patients (21%) were long-term disease-free survivors. Among the nine surgery-only patients who had local failure only, three had received postoperative chemoradiotherapy therapy. Data regarding the schedule of chemotherapy plus concurrent radiation therapy given to patients who had an R1 (or an R2) resection were not required to be reported in this study. When available, these data indicated that radiation therapy was usually given in a total dose of 45 to 50 Gy. The chemotherapy regimens used usually included cisplatin and fluorouracil. Disease-Free and Overall Survival Shown in Figures 1 and 2 and Appendix Figures A1 and A2 (online only) are disease-free and overall survival for patients by study arm and by extent of resection. As can be seen, only patients who underwent an R0 resection had a substantial chance of long-term disease-free survival. With a minimum follow-up of 1.1 years and a

100 Chemo + surgery Surgery alone

Total Dead MST 216 184 1.3 227 190 1.3

75

Alive (%)

Table 1. Patient Characteristics

50

25

0

1

2

3

4

5

6

7

8

9

Time Since Random Assignment (years) Patients at risk Chemo + surgery 216 Surgery alone 217

130 135

77 83

52 60

45 41

42 45

31 38

29 30

27 24

18 15

Fig 1. Overall survival by study arm for chemotherapy followed by surgery versus surgery only. MST, median survival time. 3721

Kelsen et al

Total Dead MST 270 205 2.2 43 41 1.0 57 56 0.6 73 72 0.24

100 R0 R1 R2 Not resected

Alive (%)

75

50

25

0

1

2

3

4

5

6

7

8

9

Time Since Random Assignment (years) Patients at risk R0 R1 R2 Not resected

270 43 57 73

210 21 14 20

142 10 4 2

103 5 1 2

90 3 1 2

83 2 1 1

68 1 0 0

58 1 0 0

50 1 0 0

33 0 0 0

Fig 2. Overall survival by type of resection. R0, complete resection with negative microscopical margins; R1, all gross disease resected but microscopical examination reveals positive margins; R2, gross residual visible tumor; MST, median survival time.

median follow-up of 8.8 years, a total of 69 patients (16%) remain alive. The median duration of survival for patients undergoing an R0 resection who were still alive at the time of analysis was 8.9 years. The median duration of survival for patients undergoing an R1 resection who were still alive at the time of analysis was 7 years. The sole surviving R2 resection was alive at 5.8 years, and the surviving nonresected patient was alive at 1.7 years. The difference in survival for patients undergoing an R0 resection was highly significantly superior to those of patients undergoing R1, R2, or no resection (P ⬍ .0001). Three-year survival for patients undergoing an R0 resection was 39%. For patients undergoing an R1 resection, it was 12%; and for R2 and for no resection it was 4% each. There was no significant difference in the percentage of patients surviving for 3 years for R1 versus R2 versus no resection. Multivariate analyses results for overall and disease-free survival are presented in Table 2 and Appendix Table A3 (online only), respectively. For overall survival in both treatment arms combined, after

adjusting for weight loss and histology, the hazard ratios (HR) for an increased risk of death for R1, R2, and R3 patients relative to R0 patients were 2.42, 4.18, and 4.45, respectively. In patients assigned to chemotherapy and surgery, not responding to chemotherapy (HR, 2.83; 95% CI, 1.84 to 4.35; P ⬍ .0001) and ⱖ 10% weight loss (HR, 1.47; 95% CI, 1.09 to 1.98; P ⫽ .0109) were associated with increased risk, while adenocarcinoma histology (HR, 0.59; 95% CI, 0.44 to 0.80) was associated with decreased risk. Survival for Patients Responding to Chemotherapy Shown in Figure 3 and Appendix Figure A3 (online only) are the overall and disease-free survival distribution curves for all patients randomly assigned to receive preoperative chemotherapy on the basis of objective response to treatment as assessed by barium esophagram, as well as those assigned to surgery alone. As noted in Methods, repeat barium esophagrams were performed after induction chemotherapy and before surgery. Response to treatment was assessed using previously published criteria for complete and partial radiographic response.4 Only 19% of patients randomly assigned to chemotherapy had major objective regressions (7% complete radiographic regression and 12% partial radiographic regression). The remaining patients had minor improvement, no change, or progression of disease. As shown in Figure 5, responding patients had a highly significantly improved survival compared with nonresponding patients, or with those assigned to surgery alone. The survival distribution curves for patients not responding to preoperative chemotherapy are not significantly different than those undergoing surgery alone. DISCUSSION

Definitive control of the primary tumor plays a crucial role in delivering curative therapy to patients with newly diagnosed esophageal cancers. The results of Intergroup 113 as well as other studies have indicated that salvage therapy for patients with recurrent disease is almost always palliative in nature, and that survival after recurrence is brief. Because the prognosis for esophageal cancer is so poor, some physicians have felt that even a palliative resection is worth performing, as this will improve outcome compared with no surgery. The results of this study do not support this view. Patients with even microscopical residual disease have a prognosis no better than that of

Table 2. Overall Survival Multivariate Analyses Model Both arms combined R0 R1 R2 R3 Weight loss Histology Chemotherapy ⫹ surgery Induction response Weight loss Histology

Comparison

Hazard Ratio

95% CI

P*

Reference level ⫽ 1 if R1, 0 otherwise ⫽ 1 if R2, 0 otherwise ⫽ 1 if R3, 0 otherwise ⬍ 10% v ⱖ 10% Squamous v adenocarcinoma

1.00 2.42 4.18 4.45 1.07 0.95

1.71 to 3.41 3.05 to 5.72 3.33 to 5.94 0.86 to 1.32 0.77 to 1.16

— ⬍ .0001 ⬍ .0001 ⬍ .0001 .5407 .5931

Responded v did not respond ⬍ 10% v ⱖ 10% Squamous v adenocarcinoma

2.83 1.47 0.59

1.84 to 4.35 1.09 to 1.98 0.44 to 0.80

⬍ .0001 .0109 .0006

ⴱ 2

␹ test.

3722

JOURNAL OF CLINICAL ONCOLOGY

Chemotherapy and Surgery v Surgery for Esophageal Cancer

Total Dead MST CT/SX: responded 39 26 3.0 CT/SX: did not respond 177 158 1.1 Surgery only 227 190 1.3

100

Alive (%)

75

50

25

1

0

2

3

4

5

6

7

8

9

Time Since Random Assignment (years) Patients at risk CT/SX: responded CT/SX: did not respond Surgery only

39 177 227

34 96 135

25 52 83

19 33 60

18 27 51

17 25 45

13 18 38

12 17 30

12 15 24

9 9 15

Fig 3. Overall survival distribution curves on the basis of response to preoperative chemotherapy. CT/SX, preoperative chemotherapy followed by surgery; MST, median survival time.

patients in whom resection is not attempted or in those with obvious gross residual cancer. The prognostic implications of an R1 resection in esophageal cancer have been reported by other investigators. Kato et al5 reported the results of a retrospective review of 463 patients with squamous cell carcinoma of the esophagus who were treated at the National Cancer Center in Tokyo, Japan. A transthoracic esophagectomy was performed. The minimum follow-up was 3 years. Sixteen patients had R1 resections; 13 of 16 recurred. They attempted reresection in several patients, two of whom survived after re-operation. While the authors concluded that local failure was not an important component of recurrence, 13 (82%) of 16 patients developed recurrent disease, indicating the ominous prognosis of an R1 resection. Law et al6 reported the results of a retrospective review of a group of 604 patients undergoing resection for an esophageal cancer. Fortyfive patients (7.5%) had positive resection margins. Local failure occurred in 10% of those with histologically positive margins and 4.9% of those with histologically negative margins. The median duration of survival was significantly less for those with positive margins than for those with negative margins (8.8 months v 15 months; P ⫽ .007). Recurrence pattern for patients undergoing complete resection has also been recently reviewed. Sanz et al7 performed a retrospective review of 53 patients who had undergone potentially curative resections. Thirty-four of these patients experienced recurrence. For three patients, local failure at the esophageal margin was the first site of failure. Twenty patients (59%) had distant failure, and 23 patients (68%) had intrathoracic recurrence as part of their recurrence pattern. In summary, our study and retrospective reviews described herein indicate that outcome as measured by disease-free and overall survival strongly depends on performance of a complete resection including negative microscopical margins. The presence of even microscopic residual disease is a strong predictor of poor outcome. The fact that the only long-term R1 resection survivors in our study received postoperative chemoradiotherapy therapy suggests, but does not prove, that curative treatment is possible even in patients with residual microscopic disease. It is not clear from the retrospective reviews www.jco.org

described earlier whether postoperative treatment of any type was given to patients with R1 resections, and whether the few long-term survivors seen in those studies also received additional treatment. While preoperative chemotherapy decreased the incidence of R1 resections in our study, overall survival was not improved. Two additional random assignment trials using preoperative (neoadjuvant) cisplatin and fluorouracil-containing chemotherapy followed by surgery have been compared with surgery alone. In these two trials, an improvement in outcome was reported compared with surgery only. Investigators from the Medical Research Council (MRC) compared preoperative chemotherapy with surgery alone in patients with localized esophageal cancers.8 In their study, 802 patients were randomly assigned to receive two cycles of cisplatin and fluorouracil chemotherapy followed by surgery or to surgery alone. The median age was 63 years, three fourths of patients were men, and two thirds had adenocarcinomas. Complete resections were performed in 78% of chemotherapy plus surgery patients compared with 70% of surgery-only patients (P ⬍ .001). Operative mortality, while higher than in our study (Intergroup 113), was equal between the two arms (10% and 11%, respectively). In an intention-to-treat analysis, overall survival was better for patients receiving chemotherapy before surgery (HR, 0.78; P ⫽ .003). The median survival was 17.2 months for patients receiving chemotherapy and 13.3 months for patients undergoing surgery alone. At 2 years, 43% of chemotherapy and surgery patients and 34% of surgery-only patients were alive. In a second study, Cunningham et al9 reported the results of the Medical Research Council Adjuvant Gastric Cancer Infusional Chemotherapy (MAGIC) trial. This study involves patients with gastric and lower esophageal cancer. Cisplatin, fluorouracil, and epirubicin chemotherapy was given for three cycles before and three cycles after surgery. Patients in the control arm underwent surgery alone; 503 patients were randomly assigned. Three fourths had gastric cancer, and the remainder had a gastroesophageal junction tumor. With a median follow-up of longer than 3 years, the survival HR was 0.75 in favor of chemotherapy (95% CI, 0.6 to 0.93). The 5-year survival rates were 36% for those patients receiving chemotherapy followed by surgery versus 23% for those undergoing surgery alone. They concluded that perioperative chemotherapy improved outcome whether measured as resectability, progression-free survival, or overall survival for patients with gastroesophageal junction and gastric cancer. The reason for the difference in survival outcome between these studies is not yet clear, as all three involved cisplatin and fluorouracilbased chemotherapy. In fact, less chemotherapy was given in the MRC trial. However, the subgroup analysis in the current study indicates that patients who have objective response to neoadjuvant chemotherapy do have a significantly improved survival compared with nonresponding patients or to all patients randomly assigned to surgery These data suggest that effective chemotherapy will have a positive effect on survival for those patients whose tumors respond to the agents. The larger sized MAGIC and MRC trials may have shown this difference. The identification of newer chemotherapeutic regimens that are more effective in esophageal cancer is a high priority. Recently, two new active agents have been identified for gastroesophageal cancer patients with advanced disease. These are the taxanes (docetaxel or paclitaxel) and irinotecan. These agents have been combined with other drugs, including cisplatin and fluorouracil, for patients with 3723

Kelsen et al

gastric and gastroesophageal cancers.10,11 Additional studies are focusing on promising targeted therapy plus chemotherapy combinations. The identification of predictive markers that can guide therapy in the individual patient before treatment begins is an area of intense investigation. Proof that an analysis of a predictive marker in tissue will allow choice of therapy that is likely to be beneficial to the individual patient comes initially from studies in women patients with breast cancer. Analysis of HER-2 allows a decision to use or not use trastuzumab.12 Analysis of estrogen or progesterone receptor protein has long guided a choice to use or not use hormone therapy. Preliminary data from other tumors indicate the possibility that a similar analysis of pretreatment tumor tissue may be useful. For example, mutations at exons 11 or 9 in the kit gene in patients with gastrointestinal stromal tumors may eventually guide a decision to use imatinib.13 Preliminary data from patients with non–small-cell lung cancer suggest that mutations in the epidermal growth factor receptor tyrosine kinase are associated with responses to epidermal growth factor receptor tyrosine kinase inhibitors.14More recently, Mellinghoff and colleagues15 reported that mutations affect responses to epidermal growth factor receptor tyrosine kinase inhibitors in glioblastoma multiforma. Preliminary data from a single institution study in esophageal cancer suggest that a gene expression analysis may predict survival for patients receiving neoadjuvant chemotherapy followed by esophagectomy, when regimens containing cisplatin and fluorouracil are used.16 In addition to pretreatment predictive markers, functional imaging using [18F]fluorodeoxyglucose positron emission tomography scans may allow a rapid assessment of efficacy in patients for whom therapy has already been started. Preliminary data from several centers indicate that in gastroesophageal cancers, a marked drop in subtraction uptake value is associated with clinical benefit in patients receiving preoperative systemic therapy.17,18 A strategy of continuing effective therapy in patients with marked improvement in [18F]fluorodeoxyglucose positron emission tomography scans as early as day 15 versus either switching to an alternative chemotherapy approach or aborting preoperative therapy and proceeding directly to surgery are being studied.19 Chemoradiotherapy followed by surgery has also been extensively studied. Preoperative chemoradiotherapy appears to increase the R0 resection rate, although studies reported to date have not demonstrated a clear survival advantage. In some of these trials, all patients received chemoradiotherapy and then were randomly asREFERENCES 1. Kelsen DP, Ginsberg R, Pajak T, et al: Chemotherapy followed by surgery compared with surgery alone for localized esophageal cancer. N Engl J Med 339:1979-1984, 1998 2. Kaplan EL, Meier P: Non-parametric estimation from incomplete observation. J Am Stat 53:457481, 1958 3. Mantel N: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chem Rep 5:163-170, 1966 4. Kelsen DP, Heelan R, Coonley C, et al: Clinical and pathological evaluation of response to chemotherapy in patients with esophageal carcinoma. Am J Clin Oncol 6:539-546, 1983 5. Kato H, Tachimor Y, Watanabe H, et al: Anastomic recurrence of oesophageal squamous cell 3724

signed to either surgery or to complete radiation therapy. In another study, all patients received chemoradiotherapy; responding patients were then randomly assigned to surgery or to complete chemoradiotherapy. There was no significant difference in survival for those receiving surgery versus those completing chemoradiotherapy alone.20,21 In summary, for patients undergoing surgery for esophageal cancer, complete resection with negative microscopical margins is an important goal. Objective response to preoperative therapy is associated with better survival. More effective systemic regimens are being developed, and show promise for improving outcome for patients responding to such treatment. AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment: N/A Leadership: N/A Consultant: Joanne Mortimer, Astra Zeneca Stock: N/A Honoraria: Norman C. Estes, GlaxoSmithKline; Christopher G. Willett, Genentech Research Funds: N/A Testimony: N/A Other: N/A

AUTHOR CONTRIBUTIONS Conception and design: David P. Kelsen, Norman C. Estes, Daniel G. Haller, Jack A. Roth, Christopher G. Willett Administrative support: Christopher G. Willett Provision of study materials or patients: David P. Kelsen, Norman C. Estes, Daniel G. Haller, Jack A. Roth Collection and assembly of data: David P. Kelsen, Kathryn A. Winter Data analysis and interpretation: David P. Kelsen, Kathryn A. Winter, Daniel G. Haller, Jack A. Roth Manuscript writing: David P. Kelsen, Kathryn A. Winter, Leonard L. Gunderson, Daniel G. Haller, Jaffer A. Ajani, Walter Kocha, Bruce D. Minsky Final approval of manuscript: David P. Kelsen, Leonard L. Gunderson, Norman C. Estes, Daniel G. Haller, Jack A. Roth, Christopher G. Willett Other: Joanne Mortimer [Editing]

carcinoma after transthoracic oesophagectomy. Eur J Surg 164:759-764, 1998 6. Law S, Arcilla C, Chu KM, et al: The significance of histologically infiltrated resection margin after esophagectomy for esophageal cancer. Am J Surg 176:286-290, 1998 7. Sanz L, Gonzalez JJ, Miyar A, et al: Pattern of recurrence after esophageal resection for cancer. Hepato-Gastroenterology 46:2393-2397, 1999 8. Medical Research Council Oesophageal Cancer Working P: Surgical resection with or without preoperative chemotherapy in oesophageal: A randomized controlled trial. Lancet 359:1727-1733, 2002 9. Cunningham D, Allum WH, Stenning SP, et al: Perioperative chemotherapy versus surgery alone for resectable gastroesophageal cancer. N Engl J Med 355:11-20, 2006 10. Van Cutsem E, Moiseyenko VM, Tjulandin S, et al: Phase III study of docetaxel and cisplatin plus

fluorouracil compared with cisplatin and fluorouracil as first-line therapy for advanced gastric cancer: A report of the V325 Study Group. J Clin Oncol 24: 4991-4997, 2006 11. Dank M, Zaluski J, Barone C, et al: Randomized phase III trial of irinotecan (CPT-11) ⫹ 5FU/ folinic acid (FA) vs CDDP ⫹ 5Fu in 1st line advanced gastric cancer patients. J Clin Oncol 23:308s, 2005 (abstr 4003) 12. Baselga J: Herceptin alone or in combination with chemotherapy in the treatment of HER2positive metastatic breast cancer: Pivotal trials. Oncology 61:14-21, 2001 (suppl 2) 13. Heinrich MC, Corless CL, Demetri GD, et al: Kinase mutations and imatinib response in patients with metastatic gastrointestinal stromal tumor. J Clin Oncol 21:4342-4349, 2003 14. Lynch TJ, Bell DW, Sordella R, et al: Activating mutations in the epidermal growth factor receptor JOURNAL OF CLINICAL ONCOLOGY

Chemotherapy and Surgery v Surgery for Esophageal Cancer

underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350:2129-2139, 2004 15. Mellinghoff IK, Wang MY, Vivanco I, et al: Molecular determinants of the response of glioblastomas to EGFR kinase inhibitors. N Engl J Med 353:2012-2024, 2005 16. Luthra R, Wu T-T, Lutra MG, et al: Gene expression profiling of localized esophageal carcinomas: Association with pathologic response to preoperative chemoradiation. J Clin Oncol 24:259-267, 2006

17. Ott K, Fink U, Becker K, et al: Prediction of response to preoperative chemotherapy in gastric carcinoma by metabolic imaging: Results of a prospective trial. J Clin Oncol 21:4604-4610, 2003 18. Weber WA: Use of PET for monitoring cancer therapy and for predicting outcome. J Nucl Med 46:983-995, 2005 19. Lordick F: Response-guided treatment in locally advanced adenocarcinoma of the esophagogastric junction (AEG): Survival results and cost analysis of the MUNICON trial. Presented at the American Society of

Clinical Oncology Gastrointestinal Cancer Symposium January 19-21, 2007, Orlando, FL 20. Stahl M, Stuschke M, Lehmann N, et al: Chemoradiation with and without surgery in patients with locally advanced squamous cell carcinoma of the esophagus. J Clin Oncol 23:2310-2317, 2005 21. Bedenne L, Michel P, Bouche O, et al: Chemoradiation followed by surgery compared with chemoradiation alone in squamous cancer of the esophagus: FFCD 9102. J Clin Oncol 25:1160-1168, 2007

■ ■ ■

Acknowledgment We gratefully acknowledge the contributions of the late Robert Ginsberg, MD, to the original Intergroup 113 trial, information that forms the basis of this report. Appendix The Appendix is included in the full-text version of this article, available online at www.jco.org. It is not included in the PDF version (via Adobe® Reader®).

www.jco.org

3725