| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Heart, Lung, and Esophageal Surgery Institute, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania
b Division of Hematology-Oncology, Department of Medicine, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania
c University of Pittsburgh Cancer Institute, Pittsburgh, Pennsylvania
d University of Pittsburgh Cancer Institute Biostatistics Facility, Pittsburgh, Pennsylvania
Accepted for publication January 28, 2008.
* Address correspondence to Dr Luketich, Heart, Lung and Esophageal Surgery Institute, Department of Cardiothoracic Surgery, University of Pittsburgh, Suite C-800, 200 Lothrop St, Pittsburgh PA 15213 (Email: luketichjd{at}upmc.edu).
Presented at the Forty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 29–31, 2007.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
Methods: Patients were staged with computed tomography scan (n = 70), endoscopic ultrasonography (n = 63), and laparoscopy with or without thoracoscopy (n = 70). The pretreatment stages were T2N0 (n = 1), T2N1 (n = 15), T3N0 (n = 13), and T3N1 (n = 41). Chemotherapy consisted of 2 or 3 cycles of cisplatin, 5-fluorouracil, and paclitaxel followed by esophagectomy and adjuvant chemotherapy. Patients were monitored for recurrence and survival.
Results: A total of 70 patients were enrolled (66 adenocarcinoma, 4 squamous cell carcinoma; 64 men and 6 women; median age, 60 years). Esophagectomy was performed in 63 patients. Operative mortality was 0%. The median overall survival of the entire group was 27.4 months. Seventeen patients were alive at a median follow-up of 62.8 months (range, 39.1 to 142). Fourteen patients were alive without recurrence at a median follow-up of 79 months (range, 39 to 138). Nodal status was an important predictor of overall survival. Patients who were downstaged experienced a significantly improved median survival of 63.4 months versus 21.5 months and overall survival (p = 0.005).
Conclusions: This prospective study for esophageal carcinoma demonstrates encouraging long-term results. In particular, downstaging of the tumor with preoperative chemotherapy is predictive of better long-term outcome. Our results support the role for perioperative chemotherapy for locally advanced resectable esophageal cancer.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
In the past decade, clinical trials have evaluated the role of neoadjuvant chemotherapy, radiation, or both, for patients with surgically resectable esophageal carcinoma, but have all failed to demonstrate a consistent survival benefit [1–5]. In an effort to improve the efficacy of systemic chemotherapy, we conducted a prospective study to evaluate the regimen of cisplatin, 5-fluorouracil (5-FU), and paclitaxel as perioperative therapy for patients with resectable esophageal carcinoma. Paclitaxel has demonstrated preclinical synergy with both cisplatin and 5-FU [6, 7]. Furthermore, in a phase II study for patients with advanced esophageal carcinoma, a provocative response rate of 48% was noted with the combination of cisplatin, 5-FU, and paclitaxel [8].
This prospective phase II study was conducted using a two-stage study design, evaluating the results of neoadjuvant chemotherapy followed by surgery, and adjuvant chemotherapy. The study was designed for the first cohort of patients to receive two cycles of preoperative chemotherapy and the second cohort to receive three cycles of preoperative chemotherapy with cisplatin, 5-FU, and paclitaxel followed by surgery, and adjuvant chemotherapy. During the study period, a total of 606 patients underwent esophagectomy, and 70 patients were enrolled for the study. The preliminary results of this study have been presented earlier [9], and we now report the long-term results of this study.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
Staging Procedures
All patients underwent a complete history and physical examination. Other staging studies were upper gastrointestinal series including barium swallow, bronchoscopy for midesophageal lesions, radionuclear bone scan, endoscopic ultrasonography (EUS), computed tomography (CT) scan, and minimally invasive staging as described previously [10]. Magnetic resonance imaging was done if EUS was not feasible. Staging with laparoscopy was performed in all patients, and thoracoscopy was done for those with upper and midthoracic tumors, or evidence of thoracic extension of gastroesophageal junction tumors. The post–induction staging investigations included barium esophagram, EUS, chest radiograph, and CT scan of chest and abdomen.
Treatment Plan
Neoadjuvant chemotherapy
The planned chemotherapy regimen consisted of paclitaxel (175 mg/m2 intravenously by 3-hour infusion) and cisplatin (80 mg/m2 intravenously by a 4-hour infusion) on day 1 followed immediately by continuous infusion of 5-FU over 96 hours at a dose of 1,000 mg/m2/day (Fig 1). Treatment cycles were repeated every 4 weeks. Initially, the protocol included two preoperative cycles, and this was increased to three cycles preoperatively after 41 patients were enrolled, followed by two cycles of the same regimen given in the postoperative setting.
|
Postoperative chemotherapy
Two additional cycles of chemotherapy were planned beginning approximately 4 to 6 weeks after surgery or when clinically appropriate in all patients except those with objective evidence of disease progression during the neoadjuvant chemotherapy. All toxicities were graded according to the National Cancer Institute Common Toxicity Criteria, version 2.0.
Patient follow-up
Patients were monitored for recurrence and survival. The follow-up schedule in the clinic was as follows: 1 month after the final cycle of chemotherapy, 3 weeks after surgical resection, then every 3 months for 2 years, followed by every 6 months for an 2 additional years, and then annually. Computed tomography scans were performed at 1-month follow-up, then every 6 months for 2 years, then at the discretion of the physician.
Statistical Design and Analysis
The primary endpoints of the study were to determine the median survival and recurrence-free survival. Kaplan-Meier plots were constructed using Greenwood confidence limits for assessing overall survival and recurrence-free survival. Recurrence-free time was computed as the time from surgical resection to disease recurrence or death among patients who were rendered disease free. Downstaging was determined by either T status or N status downstaging by histopathologic examination. Patients were stratified based on downstaging. The log-rank test was used to analyze differences between the groups. In addition, analysis of individual covariates predictive of survival was performed with the Cox proportional hazards regression method. Multivariate analysis of significant covariates was performed by Cox proportional hazards regression after forward selection.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
|
Tolerance and Toxicity of Chemotherapy
A total of 65 patients (93%) completed all the planned preoperative cycles of chemotherapy; 43 patients (61.4%) completed two cycles of postoperative chemotherapy. In general, chemotherapy was tolerated well, except for 1 death during preoperative chemotherapy. Febrile neutropenia developed in 12 patients. The hematologic and nonhematologic toxicities are summarized in Table 2. A total of 22 patients required a 5-FU dose reduction, and 1 patient required 5-FU dose reduction of 20% and a 50% cisplatin dose reduction secondary to low creatinine clearance. Twelve patients had their treatment delayed by 1 or 2 weeks owing to toxicity. The median duration between completion of chemotherapy and surgery was 36 days (range, 17 to 96; mean, 42.2).
|
Survival
The median overall survival of the entire group was 27.4 months (95% confidence interval [CI]: 17.7 to 36.8). Seventeen of 69 patients were alive at a median follow-up of 62.8 months (minimum and maximum: 39.1 and 142; Fig 2). The estimated 2- and 5-year overall survival rates were 55.1% (95% CI: 44.5 to 68.2) and 28.4% (95% CI: 19.4 to 41.5), respectively. Patients who were downstaged experienced a significantly improved median survival of 63.4 months (n = 18; 95% CI: 5.6–not reached) versus not downstaged (21.5 months; n = 51; 95% CI: 13.5 to 31.1) and overall survival (p = 0.005 by log-rank test; Fig 3).
|
|
|
|
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
Accurate staging is critical for clear stratification of patients, enrollment in clinical trials, assessment of downstaging, and evaluation of results [12]. One of the unique features of this trial included accurate staging with CT scanning, EUS, and minimally invasive staging with laparoscopy and thoracoscopy. This trial was initiated in 1994 and positron emission tomography (PET) was not a requisite. We have, however, reported an increased accuracy of minimally invasive staging in comparison with PET scans [13]. In an analysis of 100 consecutive PET scans from the University of Pittsburgh, Luketich and colleagues [13] reported that although PET scanning was more accurate than CT scanning, its sensitivity was only 69%, specificity was 94%, and overall accuracy was 84% when compared with minimally invasive staging. Similarly, we have reported an increase in accuracy with minimally invasive staging in comparison with EUS in the diagnosis of lymph node metastases [10]. Another unique feature of this trial is that minimally invasive techniques, as described previously [11], were used to perform esophagectomy in many patients. Minimally invasive esophagectomy is currently being evaluated in a multicenter setting (Eastern Cooperative Oncology Group E2202).
The potential benefits of neoadjuvant chemotherapy include earlier treatment of micrometastatic disease and downstaging of the tumor with a possible increase in curative resection. Roth and colleagues [5] randomly allocated 39 patients to either preoperative chemotherapy followed by surgery or to surgery alone. There were no significant differences in survival between these groups, but patients who responded to chemotherapy had a better survival. Kelsen and colleagues [4] reported the results of a prospective randomized study in North America comparing preoperative chemotherapy followed by surgery versus surgery alone in 440 patients. There were no differences in survival in these arms, with the median survival being 16.1 months in the surgical arm and 14.9 months in the treatment arm. The 2-year survival was 37% in the surgery-alone arm and 35% in the treatment arm. The rate of microscopically negative resection was 62% in the treatment arm, and the operative mortality was 6%.
More recently, the Medical Research Council reported the results of a similar randomized study using cisplatin and 5-flurouracil followed by resection versus resection alone [2]. Resection was microscopically negative in 60% of the treatment arm versus 54% of the surgery-alone arm. In contrast to the North American trial, there was a significant improvement in the survival, with a 16.8-month median survival in the chemotherapy arm versus 13.3 months in the surgery arm; and the 2-year survival was 43% and 34%, respectively, at a median follow-up of 37 months. The impact of downstaging was not reported in these studies. The reasons for the difference in results between these trials are not clear. The North American trial patients were staged better with CT scanning, which was required, as opposed to the Medical Research Council study, which did not require CT scanning for staging; this may have resulted in an imbalance in the study population. Finally, the results of the Medical Research Council Adjuvant Gastric Infusional Chemotherapy trial utilizing a strategy similar to our study with preoperative chemotherapy, surgery, and adjuvant chemotherapy favoring the treatment arm has recently been published [14]. However, only about a quarter of the patients had cancers of the distal esophagus or gastroesophageal junction, and the majority had gastric cancer.
Our current trial enrolled patients who had locally advanced disease, with 80% of patients having preoperative pathology proven N1 disease and 99% of patients with either a T3 or N1 lesion. Our results, with a median survival of 27 months, compare well with the studies discussed above. Similarly, the rate of microscopically negative margins of 81% and operative mortality of 0% compare well with the studies above.
Another approach to the treatment of esophageal cancer involves the addition of preoperative radiation to chemotherapy [1]. Forastiere and colleagues [15] reported the results in 43 patients treated with chemoradiation followed by transhiatal esophagectomy. In that series, 56% of patients were thought to have nodal metastases preoperatively, and 8 patients (19%) had T1 disease. The median survival was 29 months, and the 5-year survival was 34%. Randomized trials using this approach have not, however, shown a consistent benefit [1]. Urba and colleagues [16] subsequently reported the results of a randomized trial comparing preoperative chemoradiation followed by surgery versus surgery alone and did not find a significant survival benefit, with a median survival of 16.9 months in the treatment arm versus 17.6 months in the surgery arm. The only trial that showed a benefit was reported by Walsh and colleagues [17] evaluating preoperative chemotherapy with 4,000 Gy radiation. With a median survival of 16 months versus 11 months and a 3-year survival of 32% versus 6%, chemoradiation was associated with survival benefit [17]. However, this study had a short follow-up of 1.5 years and did not include CT scanning for staging, thus allowing for imbalances between the treatment groups. Furthermore, the 3-year survival probability with surgery alone was 6%, which is lower than expected [17].
Recently, Reynolds and colleagues [18] from the same institution reported an increase in complications with preoperative chemoradiation in 102 of a prospective, nonrandomized cohort of 200 patients, when compared with the 98 patients treated with surgery alone. There was an increase in rates of sepsis, respiratory failure, and acute respiratory distress syndrome in the multimodal group with no significant difference in survival. Therefore, their results lend support to the view that an effective local control with surgical resection may be adequate and that adding a second local treatment such as radiotherapy preoperatively may increase toxicity. In our study, the pattern of recurrence was predominantly distal and that points to the need for more effective chemotherapy. A recent meta-analysis, with its limitations, has, however, shown a survival benefit for both neoadjuvant chemoradiotherapy and chemotherapy alone followed by surgery [19].
Impact of Downstaging
The overall efficacy of combining preoperative radiation with chemotherapy is controversial; however, patients who have a complete pathologic response (15% to 40%) have an improved survival [15, 20]. However, that has not translated into better survival for the entire cohort. The rate of complete pathologic response after neoadjuvant chemotherapy alone has been lower than that reported for combined chemoradiation preoperatively [21]. The large randomized studies of Kelsen and colleagues [4] and the Medical Research Council study do not provide information on the rate of downstaging. Rice and colleagues [22] have estimated that about 25% to 30% of patients are N downstaged, and downstaging N status is a strong predictor of survival. In our study, multivariate analysis showed that N status was an important factor in overall survival. In our series, both downstaging T or N status resulted in improved survival, which is similar to the study by Korst and colleagues [21]. However, the rate of downstaging was lower than that reported in the retrospective study by Korst and colleagues [21]. The main differences between these studies is that in the Korst study, only the surgical cohort was included in the assessment, and the rate of downstaging may have been overestimated. Second, EUS alone was utilized in the assessment of nodal metastases, with its limitations, and in the current study all patients were staged pathologically by minimally invasive staging, which results in a more accurate assessment of downstaging. Further, other reports of downstaging take into account tumor size and not the depth of invasion, which corresponds to T status [21].
The impact of downstaging despite the lack of a pathologic complete response was striking. Patients who were downstaged had a significantly improved median survival of 63 months. Downstaging did not appear to be a significant prognostic variable in the multivariate model including nodal status, only because those patients who were initially node negative and not downstaged had the longest survival. Interestingly, in this series, even patients who were not downstaged experienced a median survival of 21 months, which compares favorably even with the treatment arm (preoperative chemotherapy group) of the MRC trial. In addition to accurate staging, factors that may have played a role in the outcome include a multidisciplinary approach and a high hospital volume. Despite the limitations of this being a single-institution, nonrandomized study, the robust survival results lend credence to continued evaluation of novel preoperative systemic therapy options for patients with esophageal carcinoma.
In conclusion, this prospective phase II trial for esophageal carcinoma using a neoadjuvant strategy with cisplatinum, 5- FU, and paclitaxel followed by surgery and adjuvant chemotherapy showed encouraging long-term results. In particular, downstaging of the tumor was predictive of better long-term outcome. Issues that merit further investigation include optimizing the chemotherapy regimen to maximize efficacy while decreasing toxicity, investigation of newer chemotherapy and molecularly targeted agents, and patient selection for optimal systemic therapy based on validated predictive models.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
I do have several questions and comments regarding this study. Despite the findings of the MRC, as you know, there is a negative study from the North American Intergroup presented by Kelson. The Kelson study was a better balance of the kind of cancers that we see in the United States, as well as, I think, a better balance in terms of the staging that was stratified for that study versus the MRC. Having said that, this was done as a very well-controlled phase II trial, but that raises several questions. An intent-to-treat analysis would require that not only you look at your survival statistics with all comers but also the morbidity and the mortality, and I think, therefore, when you look at the perioperative outcome, the picture is just a little bit cloudier. This is not a standard of therapy when you look at the final results as opposed to the 2-year survival. You didn't tell us, but on your graph the 5-year survival was roughly about 20%, if I'm close, and although you had some excellent median numbers, when you follow these patients out past 2 years, what you're finding is that many more of these patients are dying over time, probably of metastatic disease.
That was similar to several of the other prospective randomized trials, including even chemo and radiation, such as the Michigan trial. There are at least two studies now that did include chemotherapy and radiation with randomization that showed a 5-year survival benefit that was actually measured, not actuarial, and that survival was 40%, so roughly double.
The question I'm left with is, I have no doubt that it's easier to do an esophagectomy after chemotherapy alone, but if the ultimate outcome in curing esophageal cancer is long-term survival, that we're following patients now with 5-year survival, that we have data from chemorads versus radiation with 10-year survivals, we should be looking at the long-term outcome, and although I think that these are excellent perioperative results in terms of the complications, I'm worried that we're losing some of the survival benefit. Basically you're ending up with half of that survival. So perhaps you could explain to us in response to that question what your proposed follow-up after this Phase II is, or how do you think perhaps adding radiation into your current proposal would allow for better long-term outcome? Again, I think it was an excellent presentation.
DR PENNATHUR: Thank you, Dr Krasna, for your comments, and I think your contributions to staging in esophageal cancer as well as to multimodality therapy are well known.
In terms of overall survival, the 5-year overall survival in this cohort of patients was about 30%. I think that the issue of adding radiation to chemotherapy preoperatively is something we do not use routinely and certainly is something which needs to be investigated. The only randomized trial which was positive was the study by Walsh and colleagues, which was positive in part because the 3-year survival in the surgery group was 6%, and every other study, including the large study by Burmeister and colleagues, was negative in terms of survival by using neoadjuvant chemotherapy and radiation. Now, the Cancer and Leukemia Group B (CALGB) study which you presented last year had very encouraging results and certainly lends some consideration to using this.
One of the key issues as far as adding a second local therapy, that is radiation to surgical resection is: Is there any potential harm in doing that? Now, from the same institution where Walsh presented the study, Dr Reynolds has published recently in the Journal of Thoracic and Cardiovascular Surgery a study wherein he showed that adding radiation to the chemotherapy was associated with an increase in the complications, such as acute respiratory distress syndrome and pneumonia and so forth. So there is a concern there in terms of adding radiation.
The second issue is what is the local recurrence in these patients? In our series I didn't show the data because I was trying to keep within the time, but we had 39 recurrences which were distant and 3.5% of local only recurrences. So there is the issue about adding a second local therapy when surgery provides a good local clearance. The addition of neoadjuvant chemotherapy is perhaps important, because most of the recurrences were distant, but I agree with you that this is something which we need to look at closely, particularly after some of the excellent results which you have presented. In the future, perhaps trials with targeted therapies should be investigated to see if these agents are beneficial in improving long-term outcome.
DR KRASNA: Just to follow up, one of the surrogate markers that we have that we've heard a lot about over the last few days is pathologic complete response. That was one number that we didn't hear, and I wonder if you could share that with us, because pathologic response has been shown in most of the studies to correlate with improvement in survival, and obviously the potential benefit of adding chemotherapy and radiation together is achieving a higher pathologic response rate, not only with better local control. But, as you said, if you don't have a high pathologic response rate, I'm not sure that adding chemotherapy is going to make up for it. What was your CPR rate?
DR PENNATHUR: We did not have any microscopic complete pathologic response. The pathologic response is in the eye of the beholder, I think to some degree, in the sense that if the research pathologist looks at 1,000 sections as opposed to 10 sections, one may not find a complete response, but, having said that, I think the pathologic complete response is clearly more in the chemoradiation arm. But our thought is that having a complete response locally alone may not translate into long-term survival for the entire cohort because many of these recurrences occur distally. I think if most of these recurrences occur locally, then that would make us think that perhaps there is some value to adding preoperative radiation. I think this is something which we need to investigate clearly a little bit further. In this series, although the complete response rate was low, downstaging after neoadjuvant chemotherapy was associated with an improved survival. Thus downstaging may also be considered a surrogate marker, and in our study was associated with an improved survival.
DR DANIEL L. MILLER (Atlanta, GA): I agree with you, because we presented our series at the Southern Thoracic where we compared chemotherapy only versus chemoradiation, and we only found a 3% complete pathologic response. You did not present that, but now we know there was none.
Now, you said you had a significant downstaging. Was that downstaging related to the tumor or was it related to nodal metastasis? I think when you look at survival and pathologic response, you have to take both of those into consideration to see if you're going to have the most benefit, because if you're cutting out radiation, is it the nodes that are staying positive or vice versa. I think you have to look at that.
DR PENNATHUR: Thank you, Dr Miller, for your comments. Downstaging was defined as either T downstaging or N downstaging, and, as you know, these patients were all pathologically staged as N1 preoperatively, not assumed N1 by an endoscopic ultrasound characteristics such as size and echogenicity. So either T or N downstaging or downstaging of both of them qualified as a downstaging, and was associated with an improved survival. So what we found was any downstaging matters in terms of ultimate survival benefit. Thank you for your comments.
DR SETH D. FORCE (Atlanta, GA): That was a great study trying to clear up still a problem that is pretty foggy for all of us doing esophageal surgery.
My two questions are, number one, 61% of your patients got adjuvant chemotherapy also. Do you have any data comparing those to the patients who received only neoadjuvant?
The second question is more just your thoughts on the fact that 10% of your population, essentially weeded out during the chemotherapy, either died or had advanced disease and didn't get operated on. If you had not given neoadjuvant therapy, they would have had surgery and probably progressed and, you know, as part of the long-term results, just improved because you had a weeding-out phase of the chemotherapy where you didn't operate on some of the patients with more aggressive cancers.
DR PENNATHUR: Thank you for your questions. Yes, there was one mortality during chemotherapy, and I think that there were 4 patients with progressive disease. In terms of weeding out some of the patients, these are the results of all patients who were enrolled and followed-up in the study whether they went on to surgery or not. The results presented were therefore not improved because we excluded patients who did not undergo surgery. I'm sorry, what was the other question?
DR FORCE: 61% of your patients had adjuvant therapy also.
DR PENNATHUR: About two-thirds of the patients had two cycles of chemotherapy postoperatively. Several of them had at least one cycle postoperatively. The numbers were small and we didn't analyze them separately.
Thank you for your comments and I want to thank my colleagues at the University of Pittsburgh for their work and the Society for the privilege of presenting this paper.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
