| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J Thorac Cardiovasc Surg 2005;129:64-72
© 2005 The American Association for Thoracic Surgery
General Thoracic Surgery |
a Division of Thoracic Surgery,
b Department of Surgery, Department of Medical Oncology,
c Department of Pathology,
d Department of Epidemiology,
e Division of Pulmonary and Critical Care Medicine, Department of Medicine, The Johns Hopkins Medical Institutions, Baltimore, Md
Read at the Eighty-fourth Annual Meeting of The American Association for Thoracic Surgery, Toronto, Ontario, Canada, April 25-28, 2004.
Received for publication April 23, 2004; revisions received August 14, 2004; accepted for publication August 31, 2004. * Address for reprints: Malcolm V. Brock, MD, 624 Osler, 600 N. Wolfe St, Baltimore, MD 21287 (E-mail: mabrock{at}jhmi.edu).
 |
Abstract |
|---|
 Top Abstract Patients and methodsResultsDiscussionReferences |
|---|
METHODS: We reviewed our institutional experience of 1415 patients with small cell lung cancer from 1976 to 2002 among whom 82 (6%) underwent surgery with curative intent.
RESULTS: Median age at surgery was 62 years, and small cell lung cancer of mixed morphology represented 14 of 82 (17%). Treatment consisted of surgery alone in 11% of cases (9/82), surgery with neoadjuvant therapy in 22% (18/82), and surgery with adjuvant therapy in 55% (45/82). Prophylactic cranial irradiation was given to 23% (19/82). The 5-year survival of the entire cohort was 42%. The 5-year survival of patients receiving adjuvant chemotherapy (n = 41) was significantly different according to whether patients had received platinum or nonplatinum regimens (68% vs 32.2%, P = .04). Among patients with stage I disease who received adjuvant chemotherapy (n = 24), the 5-year survivals for patients receiving platinum and nonplatinum chemotherapy were 86% and 42%, respectively (P < .02). If patients who received either neoadjuvant or adjuvant therapy (n = 56) were considered, the 5-year survival was significantly better for platinum than for nonplatinum chemotherapy (62% vs 36%, P = .05). The 5-year survival was also better for those undergoing lobectomies (n = 52) than for those with limited resections (n = 15, 50% vs 20%, P = .03). Survival outcomes also differed by gender, with female patients having a 5-year survival advantage over male patients (60% vs 28%, P = .004).
CONCLUSION: These results support a reevaluation of the role of surgery in the multimodality therapy for small cell lung cancer, which currently includes only radiotherapy and chemotherapy.
The role of surgery in SCLC is controversial. Two large, randomized, prospective trials of surgery versus radiotherapy organized by the British Medical Research Council in the 1960s and 1970s reported that surgery and radiotherapy were equally ineffective for limited stage SCLC.4,5 Proponents of surgery argue that the Medical Research Council studies included very few patients with very early stage (T1-2 N0) disease, and that this is precisely the patient population most likely to benefit from surgery.
Since the Medical Research Council studies were performed, more than 3 decades ago, the advent of new, powerful diagnostic tools, such as spiral computed tomography and positron emission tomography, along with traditional mediastinoscopy allow for very limited disease to be more readily identified and to be adequately staged preoperatively. In addition, since the early 1980s, multiagent platinum-based chemotherapeutic regimens have been identified with significant activity in SCLC with, a response rate as high as 80%.6 Radiotherapy options have also increased significantly since that time, with consideration of various radiation port volumes, intensity modulation, and exploration of several alternative doses and schedules.
We hypothesized that in the modern era of multimodality cancer treatment, surgery as a technique for maximizing local control would confer a survival advantage for selected patients with SCLC if used with platinum based chemotherapy, optimal delivery of radiotherapy and prophylactic cranial irradiation. We examined our experience during the last 28 years at the Johns Hopkins Hospital to help define the role of surgery in the modern therapy of SCLC.
|
Patients and methods |
|---|
|
TopAbstractPatients and methodsResultsDiscussionReferences |
|---|
We included all operations performed with curative intent and excluded patients if they had (1) resection for pulmonary metastases, (2) operation for diagnostic biopsy, or (3) staging thoracoscopy. Eastern Cooperative Oncology Group performance status data on individual patients were not analyzed.
Patient clinical variables and outcomes
Intraoperative mediastinal lymph node sampling was performed as systematic sampling of levels R2, R4, 7, R8, R9, and R10 on the right side, and 5, 6, 7, L8, and L9 on the left side, if present. Nodes were labeled according to the American Thoracic Society lymph node mapping system. All patients who underwent segmental or greater resections also underwent mediastinal nodal sampling. Among the 10 wedge resections, in only 1 case was nodal sampling performed. For the purposes of this study, all patients underwent staging according to the 1997 TNM classification criteria, which was reported as pathologic staging except in the case of the 9 patients with wedge resection, for whom staging was reported as cTNM because of a lack of adequate mediastinal sampling. Those who underwent resection before 1997 had their disease reclassified according to the new pTNM system. There were 18 patients who received neoadjuvant therapy, but all had their disease classified according to the pTNM staging determined after resection. Definitive data on which patients in the neoadjuvant group underwent preoperative sampling of mediastinal lymph nodes, either by transbronchial fine needle aspiration or by mediastinoscopy, were incomplete (mainly because the preoperative sampling was performed outside our institution) and do not factor into this analysis. Similarly, data regarding objective response rates to neoadjuvant therapy and the Eastern Cooperative Oncology Group performance status data of the patients were incomplete and consequently not analyzed.
Although all patients with SCLC were referred for consideration of adjuvant therapy, 19 patients did not receive any additional postoperative treatment. Specific adjuvant treatment information was unavailable for 10 patients referred back to their primary care physicians without further follow-up. Of the 9 patients who underwent surgery alone, 4 refused any adjuvant therapy and 5 died before completion of the chemotherapy recommended by the medical oncologist.
Adjuvant treatment was recorded as radiotherapy, chemotherapy, or both as appropriate. Because multiple chemotherapeutic drugs were used during the 3 decades of the study period, the chemotherapy regimens were codified as either platinum or nonplatinum. We chose January 1, 1987, as a time point of significance because this date represented the first time in our hospital when platinum-based drugs were in widespread use, and it is conveniently positioned near the halfway point of our study. Although postoperative radiotherapy was given mainly to patients with nodal disease, 6 of 48 patients with stage I disease (12%) were treated with adjuvant radiotherapy. Prophylactic cranial irradiation was performed on 19 patients, including 8 patients with stage I disease.
Patients' data were obtained to their last date of follow-up or until death. Vital status data were provided by the National Death Index of the National Center for Health Statistics. Survival was defined as the interval between date of surgery and date of death or last follow-up. Death included all causes and was not specifically cancer related.
Statistical analysis
The Kaplan-Meier product limit method and the log-rank test were used for survival analysis. The association of factors with time to event end points was estimated with the Cox proportional hazards model for multivariate analysis. Comparisons of continuous and dichotomous variables between groups were performed with the Student t test (2-tailed) and 
2 tests, respectively. All analyses were accomplished with the STATA statistical software package (Stata Corporation, College Station, Tex).
|
Results |
|---|
|
TopAbstractPatients and methodsResultsDiscussionReferences |
|---|
Surgery
All operations were elective. Most patients underwent lobectomy (n = 52, 63%), because that was the preferred operation if the lesion was completely resectable, the general medical condition of the patient was acceptable, and respiratory mechanics and gas exchange were satisfactory. Twelve patients underwent pneumonectomy, and 3 patients underwent bilobectomy. Ten patients underwent only wedge resection. The wedge resection was carried out during the 1970s or 1980s (when this was considered standard practice) in 6 cases, in another 3 cases intraoperative decisions prompted a lesser resection because of concerns about the patient's ability to tolerate a lobectomy, and in 1 case the wedge resection was preferred for undocumented reasons. There was a consistent distribution of operative procedures throughout the study period, with 45 of 82 (55%) before 1987 and 37 of 82 (45%) after 1987 (Table 1). There were no evident changes in the relative frequencies of the various surgical procedures performed before or after 1987.
|
Survival
There was only 1 postoperative death, that of a patient with a pulmonary embolus at postoperative day 13, giving a 30-day surgical mortality of 1.2%. Data on postoperative complications were available for 62% of patients (51/82). There were 7 nonfatal complications, all involving the cardiorespiratory system, including 1 case of pneumonia, 2 cases of dysrhythmia, and 4 cases of air leakage. The mean follow-up period for the entire study population was 2.6 years. The median survival for all patients was 2.0 years, with overall 1-, 5-, and 10-year survivals of 78%, 42%, and 36%, respectively.
The 5-year survival for patients with stage I SCLC (T1-2 N0) who underwent complete surgical resection was significantly better than for those patients with stage II, stage III, or stage IV disease (58% vs 18%, 23%, and 0%, respectively P < .001; Figure 1). Another subgroup with good long-term survival was the group of 41 patients who received postoperative adjuvant chemotherapy, with overall 1-, 3-, and 5-year survivals of 82%, 50%, and 47%, respectively. By stage, the 5-year survivals for the 41 patients receiving adjuvant chemotherapy were 63%, 25%, and 25% for stages 1, 2, and 3, respectively (Table 2). In particular, the 5-year survival for these 41 patients was significantly different according to whether they had received platinum versus nonplatinum chemotherapy (68.0% vs 32.2%, respectively, P = .04). If only those patients in the series who had stage I disease and received adjuvant chemotherapy are considered (n = 24), the 5-year survival of patients with stage I disease who had platinum adjuvant chemotherapy was 85.7%, versus only 41.7% (P < .02) for those who received nonplatinum chemotherapy (Figure 2).
|
|
|
Because 1987 was the first year of widespread platinum chemotherapy use in our institution, we analyzed our data before and after 1987. This date also represented approximately the middle point of the study. Table 1 illustrates that there were no differences during the study period in the distribution of types of surgical resections that occurred, the proportion of patients who were given neoadjuvant versus adjuvant chemotherapy, and the number of patients who received prophylactic cranial irradiation (19 patients). The only significant difference in treatment observed was in the proportion of patients who received platinum versus nonplatinum chemotherapy. We observed a statistically significant difference in survival between patients treated with adjuvant chemotherapy (n = 41) before and after 1987 (33% vs 71.4%, P < .03).
Survival was also dependent on the type of surgical resection. The 5-year survival of those patients who underwent lobectomy (n = 52) was significantly better than that of patients who underwent segmentectomy or wedge resection (n = 15, 50% vs 20%, P = .03). This difference in survival remained significant even after adjustment for stage. Compared with male patients (n = 47), female patients (n = 35) had a survival advantage. The 1-, 3-, and 5-year survivals for male and female patients were 73%, 35%, and 28%, versus 86%, 63%, and 60%, respectively (P = .004). Median survival for male patients was 16 months (range 1-221 months), versus 39 months for female patients (range 2-216 months, P = .002). There were no deaths observed among the 8 women with T1-2 N0 SCLC who underwent complete surgical resection and received adjuvant platinum chemotherapy. No differences were observed between the sexes in the types of adjuvant chemotherapy treatment received.
Mixed cell variants represented 14 of 82 (17%) of the study cases. Although the number of cases was small, the 5-year survivals of mixed cell variants was 68%, versus 38% for histologically pure SCLC (P = .20).
Multivariate analysis
According to the Cox proportional hazards model of multivariate analysis, if the model was adjusted for the age of the patient at time of surgery, then a patient's stage, sex, date of surgery, and type of surgical resection were all strong, independent predictors of long-term survival (P < .05). If a patient was male, he was 2.25 times more likely to die. If a patient had more advanced disease than T1-2 N0, the risk of death increased 2.8 times. Similarly, surgery before 1987 was associated with poorer survival. Finally, if a patient underwent a lesser pulmonary resection than a lobectomy, the risk of death was 2.5 times greater than for patients who underwent a lobectomy or greater resection. Taken together, male patients with SCLC who before 1987 underwent a segmentectomy or wedge resection followed by nonplatinum chemotherapy had a very poor long-term survival.
|
Discussion |
|---|
|
TopAbstractPatients and methodsResultsDiscussionReferences |
|---|
As observed in older studies, disease that had advanced enough to involve hilar or mediastinal lymph nodes significantly reduced the 5-year survival of any patient.8,15,16,25 Our study achieved similar overall 5-year survivals by stage as previous adjuvant series (Table 2). Our observations confirm that if modern preoperative clinical staging can correctly identify cT1-2 N0 SCLC, and if complete surgical removal of the lesion is possible, platinum-based adjuvant chemotherapy can lead to long-term survival. Moreover, our 5-year survival of 32.2% for nonplatinum chemotherapy compares favorably with the 5-year survivals of many studies performed when that type of chemotherapy was the standard of care.
The multivariate analysis identified patient factors such as stage, date of surgery, surgical procedure, and sex as significantly influencing survival outcome in this study. Multiple studies have shown clinical and pathologic stage to be important in SCLC.26 The date of procedure (before or after 1987) was a surrogate marker for the availability of adjuvant platinum chemotherapy. Administration of platinum chemotherapy appears to have been a substantial factor in improving survival in patients with SCLC amenable to surgical resection. In our series, lobectomy was an effective means of locally controlling SCLC and may represent a better mode than chemotherapy or radiotherapy for completely eradicating a small SCLC lesion.27 We agree with Meyer's commentary on the study of Shields and colleagues8 that surgical resection by itself is inadequate treatment for SCLC of any TNM stage or for a mixed subtype and that adjuvant chemotherapy is essential. Because our data concerning specific sites for local and distant recurrences were limited, we were unable to address any questions concerning local recurrence in this study. Although several of our patients did receive prophylactic cranial irradiation, we also did not address the role of prophylactic cranial irradiation in the inhibition of early development of distant metastases in the brain.
Finally, it is possible that it was not the surgery itself, but rather the selection factors enabling surgery, that played a critical role in producing favorable outcomes. This study was not randomized, and the surgical patients were highly selected. Potential confounding variables, such as performance status of individual patients or renal function, were not assessed. Because this study was a hospital-based, retrospective review spanning more than 2 decades and involving many practitioners and multiple drug combinations and dosages, we did not attempt to analyze individual doses of chemotherapy. It is plausible that patients receiving platinum chemotherapy were given more intensive therapies than those who received nonplatinum regimens. There is considerable evidence, however, to suggest that combinations of chemotherapeutic drugs with platinum, such as etoposide with cisplatin, may even have efficacy in patients with refractory SCLC not responsive to nonplatinum regimens such as cyclophosphamide, doxorubicin, and vincristine.28,29
In summary, our data corroborate the historical literature regarding surgical rates of resectability of SCLC, efficacy of resecting early stage disease with surgery, and the importance of adjuvant chemotherapy. Because we analyzed our data more than a decade after the introduction of platinum chemotherapy, our findings suggest that platinum adjuvant chemotherapy after a lobectomy led to excellent outcomes in patients with early stage disease. These data indicate a need to reconsider the current standard of care for select patients with early stage SCLC and for a cooperative study to reevaluate designing a randomized trial with modern preoperative diagnostic procedures and current adjuvant chemotherapy.
Discussion
Dr W. Roy Smythe (Temple, Tex). Thank you for the opportunity to comment on this article. This is a thoughtful retrospective evaluation of a select group of patients with SCLC undergoing resection with curative intent during 26 years at a single institution. The stated hypothesis was to determine whether surgery conferred a survival advantage in the modern era of more effective neoadjuvant and adjuvant therapies, including, importantly, platinum chemotherapy.
I first have some comments. Even though the title of the article is "Surgical Resection of Limited Disease Small Cell Lung Cancer in the New Era of Platinum Chemotherapy: Its Time Has Come," I think it's important to say that what I think is really meant is early stage, rather than limited disease. I think this confusion of terms may be one of the reasons that the results of the British Research Council have been overgeneralized and overapplied, and many people have not operated on patients with early stage disease. The American Joint Committee on Cancer Staging actually defines limited disease as stage I through IIIB, as compared with NSCLC. So we're really talking about stage I and stage II disease.
The second comment is that the hypothesis proposed obviously can't be answered without a randomized trial, although I think there are some important caveats to be taken from this study because we don't have a lot of these patients for analysis. It is also important to note that the American College of Chest Physicians' evidence-based guidelines published in 2002 do suggest that it is reasonable to resect limited disease SCLC defined as early stage, stages I or II, and many of us have been practicing this for quite some time.
I have three questions for you, Dr Brock. First, you talked about the use of prophylactic cranial irradiation, but it appears that you lumped prophylactic cranial irradiation with thoracic irradiation with regard to evaluating its effectiveness or its impact on survival. We know that prophylactic cranial irradiation decreases the rate of central nervous system metastasis by about 50% and has a small but real effect on survival. Did you look at this as a stand-alone variable? I think 19 of your patients received prophylactic cranial irradiation, and the remainder did not.
Dr Brock. Thank you very much, Dr Smythe, for those insightful comments. I fully agree with both of your comments. I think that our medical colleagues do speak a different language from us when it comes to the staging of SCLC. This is partly because our medical colleagues do not have the advantage of routinely referring to the actual pathologic specimen with accompanying lymph nodes and, instead, have to rely on clinical and radiologic impressions for staging. These clinical descriptors are less precise and confuse the issue. We initially struggled in deciding which nomenclature to use in the manuscript, but since this is a surgical forum, we have used the pTNM staging system in the final version of the submitted manuscript.
Let's go to your question regarding prophylactic cranial irradiation and the thoracic irradiation. Our 5-year survival for our 19 patients with prophylactic cranial irradiation was approximately 60%. Our problem really was that we didn't have the cause of death for many of these patients. Therefore we don't know how many of these patients really had recurrences due to their brain metastases.
Dr Smythe. For my second question, do you know the pattern of recurrence? I guess what you're saying is that you don't know the pattern of recurrence in these patients.
Dr Brock. We are trying to secure these data points for all 1400 patients with SCLC in our database. As of right now, we don't have this information.
Dr Smythe. For my last question, you say that the reason the post-1987 group does better is that it's a surrogate for use of cisplatin-based chemotherapy, but isn't it also possible that it's a surrogate for better staging? I would assume that between 1976 and 1987 you didn't uniformly get computed tomographic scans, at least of good quality, and certainly not positron emission tomographic scanning. So couldn't that also just be a staging bias or a surrogate for better staging?
Dr Brock. Yes, I agree with that comment. This is definitely another interpretation of the data. Platinum-based chemotherapy and better staging tools probably both contribute to better survival, and I think that's the premise of our discussion here. We would like to use both of these modern tools in a randomized trial.
Dr William H. Warren (Chicago, Ill). Congratulations on a fine article. This study covers quite a few years, and we are all aware that there can be occasional cases in which the diagnosis of SCLC is made on the basis of a fine-needle aspirate, but the final diagnosis, made on the basis of the resected specimen, is different. For the purposes of this study, was the surgically resected material reviewed, including those specimens before 1987?
Dr Brock. That's a great question. We had concerns that we might even have had some carcinoids that were misclassified as SCLC. After we initially identified the patients, the very next thing we did was to have an independent review of all available tissues by one of our lung pathologists. After re-reviewing about 90% of the cases, he concluded that all the cases were correctly classified either as SCLC or as a mixed variant (about 17% mixed variants). Remarkably, he did not find any difference between his own impression and what was reported back even in the early time period of the study.
Dr Warren. Were any of them sterilized as a result of preoperative chemoradiotherapy?
Dr Brock. I did not look at how many had complete response. We had 18 patients who had induction chemotherapy. I don't know the answer to that question.
Dr Steven J. Mentzer (Boston, Mass). I have just a quick comment related to that histologic diagnosis. Because of the small number of available patients, it may not be realistic to pursue a randomized trial to look at the outcomes of a patient population like this. If we can do genomic profiling and correlate the biologic behavior, it may be possible to do a nonrandomized trial. Correlating molecular features and pattern of failure, which is what Dr Smythe and you alluded to earlier, may provide an opportunity to study this question without the expense of a randomized trial.
Dr Brock. I agree that that's a very good idea.
Dr Alec Patterson (St. Louis, Mo). We're talking about a very unusual lesion, I believe. I'm a little skeptical that you would ever be able to mount or complete a randomized trial.
Dr Brock. I think that it's a very uncommon lesion and I think that various institutions have found that they don't have this lesion very often in their records. We were surprised that we had even 82. It would take multiple institutions to do this and it would take quite a while, but I think it could be done.
The other thing is that we found that the majority of these were patients in whom the surgeon thought it was non–small cell lung cancer and only after resection did he find out that it was SCLC. I think that does happen, but you are right, it's not a common, everyday event.
Dr Thomas K. Waddell (Toronto, Ontario, Canada). It needs to be said that there has been a randomized trial for this situation of limited disease, although the criticism was always that it wasn't what you're talking about, very limited disease. Maybe some of the survivors of that era could comment. It was a negative trial. There was no difference in the group treated with surgery.
I would also comment, to echo what Alec said, for the node-negative cases, where most of the benefit or most of the very good result at least was seen, you accumulated 43 patients in 26 years. So I don't see that we can be changing our practice right now, except with what Steve just said, which is to use more modern science to follow very selected patients. But clearly there are a great many patients with clinically limited disease who have been treated by our medical oncology colleagues. And it's worth noting that there has been tremendous progress, not only with platinum but subsequently with hyperfractionated radiation schedules and perhaps Current Procedural Terminology code 11. So the regimen that you propose is perhaps not the current state of the art in nonsurgical oncology either.
Dr Patterson. Did you want a comment from the survivors of the era, or survivors of the trial?
Dr Waddell. You would do fine.
Dr Patterson. Tom was mentioning the Frances Shepard trial. It was a very small experience, as he said, and most of those patients had IIIA disease. I mean those were N2 cases.
Dr David J. Sugarbaker (Boston, Mass). I wonder how many of these you knew about preoperatively. Because usually you're resecting a solitary nodule, and you did a lobectomy, and after frozen section, the pathologist says it was a malignant tumor. Maybe you had done a mediastinoscopy and the results were negative, so you went ahead and did a lobectomy. Most of us would say in that case to give adjuvant therapy on the basis of the data that are available, much of which are old. Unless we get some sort of cooperative ability to get a large number of cases together, isn't that what we're going to be stuck with, doing the lobectomy and giving adjuvant therapy?
Dr Brock. Yes. I agree 100%. That's exactly why we're proposing to get people together, maybe in an American College of Surgeons Oncology Group trial or similar, where a large number of institutions can come together and design a trial so that we're not "stuck" with an ineffective or out-dated therapy.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractPatients and methodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
