Phase 2 study of preoperative radiation with concurrent capecitabine, oxaliplatin, and bevacizumab followed by surgery and postoperative 5-fluorouracil, leucovorin, oxaliplatin (FOLFOX), and bevacizumab in patients with locally advanced rectal cancer: ECOG 3204


  • This manuscript has not been previously published and is not under consideration in the same or substantially similar form in any other peer-reviewed media.



Recent studies have demonstrated the feasibility of combining oxaliplatin with 5-fluorouracil (5-FU) or capecitibine and radiation therapy. The addition of bevacizumab to chemotherapy improves overall survival for metastatic disease. We initiated a phase 2 trial to evaluate preoperative capecitabine, oxaliplatin, and bevacizumab with radiation therapy followed by surgery and postoperative 5-FU, leucovorin, oxaliplatin (FOLFOX) and bevacizumab for locally advanced rectal cancer.


Fifty-seven patients with resectable T3/T4 rectal adenocarcinoma were enrolled. Preoperative treatment was capecitabine (825 mg/m2 twice daily from Monday to Friday), oxaliplatin (50 mg/m2 weekly), bevacizumab (5 mg/kg on days 1, 15, 29), and radiation therapy (50.4 Gy). Surgery was performed by 6 weeks after neoadjuvant therapy. Beginning 8 to 12 weeks after surgery, patients received FOLFOX plus bevacizumab (5 mg/kg) every 2 weeks for 12 cycles.


Fifty-four of 57 enrolled patients were eligible. Forty-nine (91%) patients completed preoperative therapy and underwent surgery. Nine patients (17%; 90% confidence interval, 9%-27%) achieved pathologic complete response. Thirty-two patients (59%) experienced pathologic tumor downstaging, and 53% and 15% of patients experienced worst grade 3 and grade 4 acute toxicity, respectively. Forty-seven percent of patients who underwent surgery experienced a surgical complication.


The primary endpoint of a 30% pathologic complete response rate was not reached; however, the majority of patients experienced pathologic downstaging with this regimen. Increased wound-healing delays and complications may have been related to the addition of bevacizumab, oxaliplatin, or both. Continued observation of these patients will establish the long-term morbidity and efficacy of this combined modality approach. Cancer 2013. © 2012 American Cancer Society.

Colorectal cancer is the third most commonly diagnosed cancer in the United States and the third leading cause of death in males and females.1 Approximately 28% of colorectal cancers arise in the rectum, accounting for 39,870 new cases of rectal cancer diagnosed in the United States in 2011.1 Chemoradiation therapy (CRT) with continuous infusion 5-fluorouracil (CI 5-FU) in addition to surgery is recommended for locally advanced rectal cancer (stage II to III) due to excessive local recurrence rates with surgical resection alone.2, 3 Neoadjuvant 5-FU–based CRT has become the standard of care for stage II to III rectal cancer in the United States due to the findings of improved local control, higher rates of sphincter-sparing procedures, and reduced toxicity versus adjuvant CRT, from the German Rectal Cancer Study Group phase 3 trial.4

Capecitabine (Genentech, San Francisco, Calif) is an oral fluoropyrimidine that was rationally designed as a prodrug to deliver 5-FU predominantly to the tumor tissue. It is rapidly and extensively absorbed as an intact molecule and is then metabolized to 5-FU. Oral capecitabine has been shown to be as effective as CI 5-FU for the treatment of locally advanced rectal cancer when given concurrently with neoadjuvant radiation therapy (RT).5 Oxaliplatin, a third-generation platinum derivative, is the only platinum agent with clinical activity in colorectal cancer, both directly and as a radiosensitizer6; it also has significant additive or synergistic activity with 5-FU.7 At the time of initiation of this trial, the clinical efficacy of the addition of oxaliplatin to neoadjuvant fluoropyrimidine-based CRT was not yet known. Preliminary results from several phase 3 randomized trials have since demonstrated higher toxicity rates without benefit in terms of pathologic complete response (CR) rate with the addition of oxaliplatin to standard neoadjuvant CRT.5, 8

Bevacizumab (Genentech, San Francisco, Calif) is a recombinant humanized monoclonal antibody that acts against circulating vascular endothelial growth factor. The addition of bevacizumab to chemotherapy for metastatic colorectal cancer resulted in a significant 34% relative reduction in risk of death versus chemotherapy alone.9 Bevacizumab has also been shown to have synergistic antitumor effects with RT in preclinical models.10, 11 Several small studies have demonstrated the feasibility of integrating bevacizumab into fluoropyrimidine-based neoadjuvant CRT, with pathologic CR rates ranging from 8% to 32%.12-14

On the basis of these findings, we initiated a multicenter phase 2 trial of preoperative capecitabine, oxaliplatin, and bevacizumab with RT followed by surgery and postoperative 5-FU, leucovorin, oxaliplatin (FOLFOX), and bevacizumab for resectable locally advanced rectal cancer. The purpose of this trial was to evaluate the safety, tolerability, and efficacy of this regimen with a primary endpoint of pathologic complete response.


Eligibility Criteria

Patients ≥ 18 years old with histologically confirmed, nonmetastatic, primary T3 or T4 adenocarcinoma of the rectum within 12 cm of the anal verge by proctoscopic examination were eligible for the study. Patients were staged according to the American Joint Committee on Cancer sixth edition staging system.3 Additional eligibility criteria included T3/T4 staging by CT scan plus endorectal ultrasound or magnetic resonance imaging (MRI), tumor prospectively defined by surgeon as either initially resectable or potentially resectable after preoperative chemoradiation, good general condition allowing major surgery (Eastern Cooperative Oncology Group performance status of 0 to 1), normal liver, renal, and bone marrow function, and a written informed consent. Exclusion criteria included: tumor outside of the true pelvis; high-grade large bowel obstruction (lumen diameter < 1 cm), unless diverting colostomy has been performed; female patient is pregnant or breast-feeding; prior chemotherapy for rectal cancer or any prior pelvic irradiation; history of stroke or transient ischemic attack at any time; myocardial infarction/unstable angina within 12 months of study entry; or any other serious condition or comorbidity that could potentially interfere with administration of trial therapy or follow-up. Patients with clinically significant congestive heart failure, peripheral vascular disease, bleeding diathesis, uncontrolled hypertension ≥ 150/100, or > grade 1 neuropathy were also excluded. Patients must have been disease-free from any other malignancy for ≥ 5 years, with the exception of nonmelanomatous skin cancer or cervical carcinoma in situ. The trial was approved by the local institutional review boards of all participating centers. This trial was registered with (identifier NCT00321685).


Preoperative Chemotherapy

Preoperative chemotherapy consisted of capecitabine (825 mg/m2 every 12 hours, by mouth, 5 days per week during RT), oxaliplatin (50 mg/m2, IV over 2 hours, once per week, days 1, 8, 15, 22, 29 of RT), and bevacizumab (5 mg/kg, IV over 30 to 90 minutes, once every other week, days 1, 15, 29 of RT).

Radiation Therapy

Standard or conformal treatment planning was allowed. Intensity modulated radiation therapy (IMRT) was not allowed. Linear accelerator–based treatment with a minimum photon energy of 4 MV was required. All patients underwent simulation with oral contrast to visualize the small bowel. The gross tumor volume (GTV) was defined as the rectum, perirectal and internal iliac lymph nodes, and the full extent of the presurgical mass. The clinical target volume (CTV) was defined as the GTV with a 2-cm margin, additionally including the presacral space. The initial planning target volume (PTV1) was defined as the CTV with a minimum of 1 cm margin. Beam arrangements could be 3-field (posterior with opposed lateral fields) or 4-field (anterior/posterior and opposed lateral fields). 45 Gy in 25 fractions of 1.8 Gy per day, 5 days per week, over 5 weeks, was delivered to the PTV1. An additional boost was then delivered to the boost PTV, which was defined as the primary tumor and lymph nodes > 1.5 cm with a 2-cm margin. The boost was 5.4 Gy over 3 fractions, for a total of 50.4 Gy given at 1.8 Gy per fraction for 28 fractions, over 5.5 weeks.

Quality assurance for RT consisted of review of the treatment fields, prestudy diagnostic imaging, dose-volume histograms, dose distributions, and portal imaging by the Quality Assurance Review Center (QARC) within 3 days of the start of RT. All linear accelerator units used in this study had their calibration verified by the Radiological Physics Center (RPC).


Surgery was scheduled to occur 6 to 8 weeks after the completion of preoperative CRT. Patients were locally restaged using the same imaging modality that was used before inclusion into the study (CT and endorectal ultrasound or MRI). Restaging prior to surgery also included a chest x-ray or CT of the chest. The choice of operative procedure was at the discretion of the treating surgeon. It was strongly recommended that the entire mesorectum be removed and that a distal rectal margin of ≥ 2 cm be obtained for sphincter-preserving procedures.

Postoperative Chemotherapy

Postoperative chemotherapy was administered to all patients who had a macroscopic complete resection, beginning 4-12 weeks after surgery. Postoperative chemotherapy consisted of leucovorin (400 mg/m2, administered intravenously [IV] over 2 hours on day 1), 5-FU (400 mg/m2, IV bolus, day 1), 5-FU (2400 mg/m2, continuous IV infusion over 46 hours, days 1 and 2), oxaliplatin (85 mg/m2, IV infusion over 2 hours, day 1), and bevacizumab (5 mg/kg, IV infusion over 90 minutes, day 1). Cycles were repeated every 2 weeks for a total of 12 (2-week) cycles, except for oxaliplatin, which was administered for 9 cycles only.


Toxicity was graded according to the National Cancer Institute's Common Terminology Criteria for Adverse Events (version 3.0 through June 30, 2011, version 4.0 after June 30, 2011). Patients were assessed weekly during neoadjuvant CRT, prior to surgical resection, and prior to each adjuvant chemotherapy cycle. After completion of all therapy, patients were assessed every 3 months for the first 2 years. Afterward, follow-up was planned for every 6 months up to 7 years, and every year thereafter. Pathologic response was assessed in all patients who underwent surgical resection. A pathologic CR was defined as no evidence of malignancy in the primary or nodal surgical specimens. Acute preoperative toxicity was defined as adverse events that occurred during neoadjuvant chemoradiation or after neoadjuvant therapy, but prior to surgery. Postsurgical complications were defined as adverse events that occurred during the inpatient hospital stay in the immediate postoperative period. Late surgical complications were defined as adverse events that occurred after the initial postsurgical hospital discharge.

Statistical Considerations

The primary endpoint of this study was pathologic CR. A total of 55 eligible patients were planned to be accrued in this trial, but to allow for 5% ineligibility, 58 patients were planned to be entered. If the treatment was indicative of a true pathologic CR rate of 30%, we would consider it a promising regimen for further study. A true pathologic CR rate of less than 15% was considered unpromising. To limit accrual if the treatment was not effective, a 2-stage design was used, allowing early stopping if the true pathologic CR rate was less than 15%. If at least 4 pathologic CRs were observed among the first 23 eligible patients, 34 additional patients (assuming 32 would be eligible) would be entered in the second stage. Five of the first 15 patients (33%) who completed preoperative therapy and surgery demonstrated a pathologic CR after the first stage of accrual, and the decision was made to continue to the second stage.15 If 12 or more pathologic CRs are seen in the 55 eligible patients, the treatment will be considered promising.


Patient Characteristics

A total of 57 patients were enrolled between July 2006 and May 2010 from 11 participating sites. Of the 57 patients, 1 was found ineligible due to T2 disease and 2 patients never started treatment, leaving 54 eligible patients that were included in the analysis. Data updated through November 8, 2011 were analyzed. At the time of analysis, 4 of the 54 eligible patients had data collected that had not been centrally reviewed. The median follow-up period for alive patients was 17 months (range, 2-54 months). The median age was 54 years (range, 26-83 years). Clinical T stage was T3 for 50 patients (93%) and T4 for 4 patients (7%). Clinical nodal stage was NX for 2 patients (4%), N0 for 16 (30%), N1 for 31 (57%), and N2 for 5 (9%). Table 1 summarizes patient demographics.

Table 1. Patient Characteristics
  1. Abbreviations: APR, abdominoperineal resection; LAR, low anterior resection; TME, total mesorectal excision.

Eligible patients54(100)
ECOG performance status  
Clinical T stage  
Clinical N stage  
Surgical procedure  
Median age, y54 (range, 26-83) 

Treatment Characteristics

Forty-nine patients (91%) completed preoperative CRT, all of whom underwent curative surgical resection. Forty-four patients (82%) completed preoperative CRT per protocol or with minor deviations only. The median radiation dose delivered was 50.4 Gy (range, 50-50.4 Gy). The median time period from the last chemotherapy cycle to surgery was 7.6 weeks (range, 5.3-11.9 weeks), and the median time period to initiation of adjuvant chemotherapy after surgery was 9 weeks (range, 6-14 weeks). A total of 67% of patients underwent surgical resection within 8 weeks of completion of neoadjuvant CRT. Five patients (9%) did not undergo surgical resection. Reasons for not undergoing surgery were patient refusal (n = 1), adverse event (n = 2), and patient death during neoadjuvant CRT (n = 2). Of the 49 patients who underwent curative surgery, 25 (51%) began adjuvant chemotherapy. Reasons for not starting adjuvant chemotherapy included wound complication for 13 patients and patient refusal for 8 patients. For the 25 patients who began adjuvant chemotherapy, the median number of cycles completed was 12 (range, 1-12). Sixteen patients (30%) received all 12 cycles and completed the entire treatment per protocol.

Preoperative CRT Toxicity

Fifty-five patients were included in the toxicity analysis: 54 eligible patients and 1 patient with a T2 tumor who underwent study therapy, but was later found to be ineligible. Two patients (4%) experienced grade 5 events during preoperative CRT, one of which was secondary to aspiration and attributed to study therapy. The other patient died during the third week of neoadjuvant CRT, which was reported as not associated with study therapy and definitely related to rectal cancer. Overall, there were 29 patients (53%) with worst grade 3 toxicity and 8 patients (15%) with worst grade 4 toxicity during preoperative therapy. A total of 16 (29%) and 4 (7%) patients experienced grade 3 and grade 4 hematologic toxicity, respectively. The most common hematologic grade 3/4 toxicities were neutropenia (18%), leukopenia (15%), and lymphopenia (15%). Table 2 summarizes hematologic toxicity. A total of 25 (45%) and 7 (13%) patients experienced grade 3 and grade 4 nonhematologic toxicity, respectively. The most common nonhematologic grade 3/4 toxicities were diarrhea (13%), fatigue (15%), and rectal pain (16%). Table 3 summarizes the most common nonhematologic toxicity.

Table 2. Highest Hematologic Toxicity for Each Patient During Chemoradiation (n = 55)
CharacteristicGrade 2Grade 3Grade 4
Worst degree8164
Table 3. Highest Nonhematologic Toxicity for Each Patient During Chemoradiation (n = 55)
CharacteristicGrade 2Grade 3Grade 4Grade 5
  1. Abbreviations: CNS, central nervous system; NOS, not otherwise specified.

Weight loss22
CNS hemorrhage1
Sensory neuropathy24
Abdominal pain331
Rectal pain572
Deep vein thrombosis2
Death NOS1
Worst degree182352

Postoperative Complications

The median time interval from last bevacizumab administration to surgery was 7.6 weeks (range, 5.3-11.9 weeks). Forty-nine patients underwent surgical resection and were included in the postsurgical toxicity analysis. A total of 9 patients (18%) experienced postsurgical complications, including wound infection (n = 8), fascial dehiscence (n = 5), intra-abdominal abscess (n = 1), fistula (n = 1), bowel obstruction (n = 1), and thrombosis/embolism (n = 1). Three patients (6%) had postsurgical complications that required surgical intervention, consisting of intra-abdominal abscess drainage (1 patient), perirectal abscess drainage (1 patient), and small bowel obstruction (1 patient). Late surgical complications occurred in 23 patients (47%). These consisted of wound infection (n = 23), wound/fascial dehiscence (n = 12), bowel obstruction/ileus (n = 5), intra-abdominal abscess (n = 2), and anastomotic leak (n = 1). Five patients experienced both postsurgical and late complications, indicating the development of a complication that persisted from the immediate postsurgical period to beyond the initial hospital discharge.

Treatment Response

Nine patients achieved pathologic CR (17%; 90% confidence interval: 9%-27%). Eleven patients (20%) were ypT0. Fourteen patients (26%) and 29 patients (54%) were ypT0-T1 and ypT0-T2, respectively. Thirty-two patients (59%) experienced pathologic primary tumor downstaging. Thirty-two patients (59%) and 42 patients (78%) were ypN0 and ypN0-N1, respectively. Of the 32 patients with clinical nodal involvement who underwent surgery, 19 (59%) were ypN0 and 21 (66%) experienced pathologic nodal downstaging. Positive surgical margins occurred in 6 of the 49 patients who underwent surgical resection (12%).

With a median follow-up period for alive patients of 17 months (range, 2-54 months), there have been 2 recurrence events. One patient recurred in the liver 11 months after initial surgery, subsequently underwent hepatic metastectomy, and is still alive. One patient recurred in the pelvic lymph nodes 17 months after initial surgery and is still alive. Overall survival and relapse-free survival data have not yet matured and will be analyzed at a later date.


In this prospective multi-institutional phase 2 trial, neoadjuvant therapy consisting of capecitabine, oxaliplatin, and bevacizumab with RT was administered prior to curative surgical resection. This was to be followed by adjuvant chemotherapy consisting of 5-FU, leucovorin, oxaliplatin, and bevacizumab. The addition of oxaliplatin and bevacizumab to standard neoadjuvant CRT did not achieve the expected pathologic CR rate of 30%. Neoadjuvant therapy intensification demonstrated significant acute toxicity with 2 preoperative deaths (1 of which was attributed to study therapy), 53% of patients with worst grade 3 toxicity, and 15% of patients with worst grade 4 toxicity. Late surgical complications were also common, with 47% of patients who underwent surgery experiencing some form of surgical complication, most commonly wound infection and wound/fascial dehiscence. Furthermore, 50% of all patients began adjuvant chemotherapy, with only 30% receiving all treatment as per protocol.

This trial was designed and initiated prior to the release of results from several phase 3 studies investigating the addition of oxaliplatin to fluoropyrimidine-based neoadjuvant CRT for locally advanced rectal cancer. The results were recently reported for NSABP (National Surgical Adjuvant Breast and Bowel Project) R-04, a 4-arm trial comparing CI 5-FU–based CRT with or without oxaliplatin and capecitabine-based CRT with or without oxaliplatin for clinical stage II and III rectal cancer.5 There were no differences in outcome between the oxaliplatin and no-oxaliplatin arms, with pathologic CR rates of 21% and 19%, respectively (P = .46). However, there was a significant increase in grade 3/4 diarrhea with the addition of oxaliplatin, with rates of 15% and 7%, respectively (P < .001). In another phase 3 trial, the study by Gerard et al examined capecitabine-based CRT to 45 Gy (Cape45) versus capecitabine and oxaliplatin-based CRT to 50 Gy (CapeOx50).8 There was no difference in the primary outcome of pathologic CR, with rates of 14% and 19%, respectively (P = .09). However, the rate of grade 3/4 toxicity was significantly increased in the CapeOx50 arm (25% versus 11%, P < .001). Furthermore, Aschele et al reported the results of the STAR-01 (Studio Terapia Adiuvante Retto) trial, a comparison of CI 5-FU CRT and the same regimen with weekly oxaliplatin.16 Again, the addition of oxaliplatin significantly increased toxicity (grade 3/4 toxicity 8% versus 24%, P < .001) without affecting local tumor response. The preponderance of evidence supports the conclusion that the addition of oxaliplatin to fluoropyrimidine-based CRT increases toxicity without a benefit in local tumor response or pathologic CR rates.

Bevacizumab has been associated with increased postoperative complication risk after major surgery when added to neoadjuvant CRT for several disease sites.17-19 Willet et al reported on a phase 2 trial examining the addition of bevacizumab to standard CI 5-FU-based CRT for locally advanced rectal cancer.13 Five of 32 patients (16%) experienced pathologic CR, and the 5-year actuarial local control and overall survival rates were both 100%. Surgery occurred 7 to 10 weeks after the completion of neoadjuvant therapy, and the regimen containing bevacizumab appeared to be tolerable and safe. Postoperative complications included anastomotic leak with presacral abscess requiring drainage (n = 1), vaginal tear with presacral hematoma and abscess requiring drainage (n = 1), pelvic hematoma (n = 1), delayed healing of perineal incision (n = 2), ileus (n = 2), neurogenic bladder (n = 1), perforated ileostomy/stent–related (n = 1), pulmonary embolus (n = 1), and wound infection (n = 3). A similar trial was performed by Crane et al, where patients with stage II or III rectal cancer were treated with neoadjuvant bevacizumab, capecitabine, and RT.14 Eight of 25 patients (32%) experienced pathologic CR, with a 2-year actuarial local recurrence rate of 6%. There were 3 major wound complications (12%) that required surgical intervention: coloanal anastomotic dehiscence requiring completion abdominoperineal resection (APR) (n = 1) and perineal wound dehiscences after initial APR (n = 2). There were also 5 minor complications (20%) that were managed without surgical intervention, yielding a total surgical complication rate of 32%. However, complication rate did differ by type of surgery, with 4 of 6 patients (67%) who underwent APR developing a complication versus 2 of 10 patients (20%) who underwent low anterior resection. There was no detectable association between the time from last bevacizumab dose to surgery and surgical complication rate. Neither trial was able to specifically attribute the observed surgical complications to the addition of bevacizumab, with other possible explanations including patient selection, differences in surgical technique, and chance alone.

An article by Dipetrillo et al described a single-institution trial of neoadjuvant bevacizumab and oxaliplatin with fluoropyrimidine-based CRT for stage II to III rectal cancer.20 Patients on this trial received 1 month of induction bevacizumab with modified infusional 5-fluorouracil, leucovorin, and oxaliplatin (FOLFOX) regimen followed by concurrent bevacizumab, oxaliplatin, CI 5-FU, and RT. The oxaliplatin dose had to be reduced from 50 mg/m2 to 40 mg/m2 weekly due to gastrointestinal toxicity. The trial was terminated early due to excessive toxicity after 26 of a planned 29 patients were enrolled. Five of 25 patients (20%) achieved a pathologic CR. Grade 3/4 toxicity during neoadjuvant CRT occurred in 19 of 25 patients (76%), consisting primarily of diarrhea (n = 16), pain (n = 11), and neutropenia (n = 6). Nine patients (36%) developed postoperative wound complications including infection (n = 4), delayed healing (n = 3), leak/abscess (n = 2), sterile fluid collection (n = 2), ischemic colonic reservoir (n = 1), and fistula (n = 1). Two patients (12%) developed deep venous thrombosis. Only 25% of all patients were able to complete all treatment as per protocol. The results of Dipetrillo et al are similar to this trial, wherein 67% of patients experienced grade 3/4 acute toxicity, only 30% of patients completed all therapy as per protocol, and a pathologic CR rate of 17% was achieved. It is important to note that the trials assessing the addition of oxaliplatin and/or bevacizumab to neoadjuvant chemoradiation have typically only reported tumor downstaging rates and toxicity to date. The more important endpoints of overall survival, relapse-free survival, local recurrence, and distant recurrence are yet immature for most trials and are pending longer follow-up periods. The validity of pathologic CR as a surrogate for long-term oncologic outcome has not been firmly established in this disease site, and conflicting reports as to the prognostic significance of pathologic CR exist in the literature.21-23 Pathologic CR rates reflect response to neoadjuvant therapy only; however, long-term outcome may also be influenced by the intensification of adjuvant therapy in trials such as this.

Nine (18%) and 23 (47%) of the 49 patients who underwent surgery in this trial experienced postsurgical and late surgical complications, respectively. Five patients experienced both postsurgical and late complications, indicating the development of a complication that persisted from the immediate postsurgical period to beyond the initial hospital discharge. This event rate is slightly higher than the combined surgical complication rate of 48% reported by Dipetrillo et al; however, there are 3 important differences between these 2 studies: 1) The phase 2 study described here is a multi-institutional trial with 11 participating centers compared with a single-institution experience of Dipetrillo et al; 2) the concurrent chemotherapy regimen differed between trials, with capecitabine instead of 5-FU and all patients receiving 50 mg/m2 oxaliplatin weekly in this trial; and 3) the median time interval from last bevacizumab to surgery was shorter in this trial (median 7.6 weeks versus 8.9 weeks). The interval from the last dose of bevacizumab to surgery has been proposed as a possible risk factor for wound healing complications.

In summary, the primary endpoint of this trial was not reached, although the majority of patients experienced pathologic downstaging with the addition of bevacizumab and oxaliplatin to capecitabine and RT for locally advanced rectal cancer. Increased wound-healing delays and complications may have been related to the addition of bevacizumab, oxaliplatin, or both. However, we were not able to attribute toxicity to a specific agent with certainty. Relapse-free survival and overall survival data have not yet matured and will be reported in a future publication. Continued observation of these patients will establish the long-term morbidity and efficacy of this combined modality approach.


This study was conducted by the Eastern Cooperative Oncology Group (Robert L. Comis, MD) and supported in part by Public Health Service grants CA23318, CA66636, CA21115, CA27525, CA15488, CA49957, and CA17145 and from the National Cancer Institute, National Institutes of Health, and the Department of Health and Human Services.


All authors have read and approved the manuscript. To the best of our knowledge, no conflict of interest, financial or other, exists except: Steven J Cohen has received honoraria from Genentech; Halla Nimeiri has received honoraria from Genentech; Al Bowen Benson has been a consultant/advisor for Sanofi and Genentech, and has received research funding from Genentech.