Seminars in Surgical Oncology 6:255-262 (1990)
Neoadjuvant Treatment in locally Advanced Non-Small Cell Lung Cancer L. PENFIELD FABER, MD, AND PHILIP D. BONOMI, MD From the Departments of Cardiovascular and Thoracic Surgery (L. P. F.) and Medicine (P. D.B.), Rush-Presbyterian-St. Luke's Medical Center, Chicago, Illinois
Any neoadjuvant or preoperative treatment program that would increase survival in clinically advanced non-small cell lung cancer (NSCLC) patients would be of significant benefit. In addition to improving patient survival, the program should be associated with minimal toxicity and surgical mortality. The rationale for any preoperative program is that the ability to resect the advanced cancer will be enhanced and micrometastasis will be eradicated. We have analyzed 323 patients enrolled in various types of neoadjuvant and preoperative studies. Review of this data indicates that cisplatin containing regimens produce relatively high responder rates in locally advanced NSCLC patients, resectability rate can be increased by a neoadjuvant program and chemotherapeutic toxicity and operative mortality are not prohibitive. Survival data frequently includes T3NO-1 patients, but there does appear to be increased survival at 3 and 4 years. These studies remain experimental and continued analysis is necessary before they can be accepted as standard therapy for clinically advanced NSCLC cancer. ~
KEYWORDS:stage I11 non-small cell lung- cancer, preoperative treatment, chemotherapy, radiation
INTRODUCTION Approximately one-third of patients with non-small cell lung cancer (NSCLC) present with locally advanced disease without evidence of distant metastasis. Although subsets of locally advanced patients can have a relatively good prognosis, the overall survival for this group is poor despite the fact that most of them present with a relatively good performance and nutritional status. Surgery or radiation alone have been the most commonly used treatments for locally advanced lung cancer, but long-term results have not been good. Any program of preoperative therapy that would increase the incidence of a complete resection and decrease the occurrence of distant metastasis disease would be of obvious benefit. More than 30 years ago, clinical investigators conducted trials of preoperative radiation [1,2]; and a decade ago, the first trial of preoperative chemotherapy was reported [3]. There are several recent reports of trials evaluating preoperative chemotherapy with or without concurrent thoracic radiation, and they indicate that further investigation of this modality of therapy for locally advanced NSCLC is warranted. 0 1990 Wiley-Liss, Inc.
STAGING AND PROGNOSTIC FACTORS One of the difficulties in designing and evaluating studies in patients with Stage 111NSCLC is the fact that this disease stage comprises a diverse group of patients. This issue has been partially clarified by the new staging system in which the TNM classification has been revised to include T4 and N3 subgroups [4].The newly defined T and N categories have resulted in the subdivision of Stage I11 into IIIa and IIIb. Consideration for resection can be given to patients with Stage IIIa disease because mediastinal lymph node involvement is ipsilateral, while Stage IIIb patients are not surgical candidates because of contralateral and supraclavicular lymph node involvement. The new staging system represents an improvement, but there are still perceived shortcomings when attempting to categorize patients with locally advanced disease. For instance, Stage IIIa includes patients with surgically Address reprint requests to L. Penfield Faber, M.D., Rush-Presbyterian-St. Luke's Medical Center, 1653 West Congress Parkway, Chicago, IL 60612.
256
Faber and Bonomi
pathologically staged T3NO disease which is associated with a 3 0 4 5 % five year survival rate 1451. In stark contrast, Stage IIIa patients whose disease is classified as T3N2 have virtually no chance of surviving 5 years [6]. There are also variations in prognosis in patients who have histologically confirmed N2 disease. Patients who have microsopic nodal metastasis found at the time of surgical resection have a 20-30% five year survival rate [7-91. In contrast, patients in whom nodal metastasis are detected by mediastinoscopy have approximately a 10% chance of surviving 5 years [7]. Lastly, the 5 year survival rate for nodal metastasis which are evident on routine chest x-ray is less than 5% [8]. When Stanley [ 101 conducted statistical analysis of 5,100 lung cancer patients, performance status, previous weight loss, and the extent of disease were shown to be the most important prognostic factors. Kaplan-Meier product limit method was used to estimate survival in relation to potential prognostic factors in 129 Stage 111 NSCLC patients receiving concurrent radiation and chemotherapy (62 had subsequent surgery) at the Rush University Medical Center. The following parameters were assessed in this study: age, sex, performance status, previous weight loss 5%, histology, hemoglobin, serum albumin, histology, race, eligible versus ineligible for surgery, Stage IIIa versus Stage IIIb, subsets of stage defined as IIIa (T3NO) versus IIIa (N2) versus IIIb, and method of mediastinal staging (clinical versus histologic). Surgical eligibility, performance status, and the subgroups of Stage IIIa were found to be the most important prognostic factors (Table I). These analyses show that Stage IIIa is not a homogenous group with the median survival durations for T3NO and N2 being 851 days and 500 days, respectively. All of the above have serious implications for designing clinical trials in patients with locally advanced NSCLC and when evaluating these studies, prognostic
factors, and subgroups of staging must be taken into consideration.
CHEMOTHERAPY: THEORETICAL AND PRACTICAL CONSIDERATIONS
The rationale for neoadjuvant chemotherapy is based on the assumption that chemotherapy-induced regression of the primary lesion will result in more effective local control with subsequent surgery or radiation [ 111. It represents the earliest possible chance to treat micrometastases, possibly resulting in the elimination of the rapidly growing tumor cells, which might not occur at a later time when the number of tumor cells is greater and they also may be less sensitive to chemotherapy because they are in a resting phase or have become drug resistant. The majority of chemotherapy data in NSCLC has been obtained from trials in patients with Stage IV disease. The larger trials have shown that the “most active” regimens contain cisplatin and they produce response rates of approximately 25% in Stage IV patients [12]. Despite discouraging response rates in metastatic disease, there is evidence that treatment with cisplatin containing chemotherapy has a modest but statistically significant beneficial effect on survival in Stage IV NSCLC ~31. There are compelling theoretical and practical reasons for using preoperative systemic therapy in Stage I11 NSCLC. One of the main factors which determine curability is tumor burden, and the cure rate is inversely related to the amount of tumor. Chemotherapy produces higher cure rates in animals injected with smaller number of tumor cells than in animals receiving larger numbers of tumor cells. Similarly, delaying systemic therapy following injection of tumor cells into animals results in lower cure rates because a longer period of tumor growth results in a greater tumor burden [ 141. It is believed that systemic therapy which is marginally effective against gross disease might be more effective when used to treat micrometastases. However, from TABLE I. Prognostic Factors in Stage I11 NSCLC Patients a practical point of view, clinicians have employed the Median rule of 50% when designing trials of adjuvant systemic survival Variable Patients (days) P-value therapy. In short, the use of neoadjuvant chemotherapy should be limited to tumors with recurrence rates greater Surgical eligibility than 50% and for which chemotherapy has produced reEligible 83 663 ,003 sponse rates of greater or equal to 50% in advanced Ineligible 46 385 disease [ 151. Most multi-institution studies have shown Performance status 0-1 response rates of approximately 25% for cisplatin con102 614 ,001 2-3 27 268 taining combination chemotherapy regimens in Stage IV Stage disease [16]. However, recent reports indicate that the IIIa 83 656 ,004 same regimens are capable of producing a response rate IIIb 46 368 at approximately 50% in patients with Stage I11 disease Stage I11 ,004 [17-191. In fact, intensification of the doses in the miT3NO 42 85 1 IIIa 41 500 tomycin, vinblastine, and cisplatin (MVP) regimen has IIIb 46 368 produced response rates of 70% in previously untreated
Neoadjuvant Treatment
257
Stage 111 NSCLC patients [20,21]. Therefore, it appears 130 patients were evaluated for a program of combined that combination chemotherapy is capable of producing chemotherapy and radiation. There were 85 that were response rates of greater than or equal to 50% in Stage I11 considered eligible for surgery and there were 45 not patients (Table 11). considered candidates for surgical intervention because of insufficient cardiopulmonary reserve (16), contralatSURGICAL NEOADJUVANT STUDIES era1 and supraclavicularlymph node metastasis (15), and Radiation therapy can be used with chemotherapy ei- extensive mediastinal disease as evidenced by vena caval ther to enhance the effect of the alternate method or to obstruction or tracheal and carinal involvement (14). treat separate components of the malignancy, i.e., local There were two different groups of drugs used and the disease and microscopic metastatic disease. Radiation first included cisplatin and 5-fluorouracil (5-FU), while and chemotherapy work similarly to affect biological tu- etoposide (VP-16) was added to the second group. Split mor cell destruction, and these modes of therapy can be course radiation was administered every other week for 5 used in combination for the following purposes: 1) to days in the first group and every third week for 5 days in increase the effect of the other method in the instance of the second group. Total radiation dose was 4,000 cGy. bulky localized disease, 2) to accelerate response, and 3) This series included 18 patients with T3NO disease, of to increase the effect of the other method in the instance whom 9 had chest wall invasion, 6 demonstrated mediof marginally responsive tumors. Disadvantages of such astinal involvement with obliteration of adjacent fascial a program of treatment include toxicity of the treatment planes on the computed tomographic scan, and 3 had program resulting in morbidity or death, delay in surgical main stem bronchial disease less than 2 cm from the therapy to control local disease, an increase in operative carina. There were 5 patients with T4NO-2 disease. complications and mortality, and a failure to achieve There were 5 patients that failed to receive the planned improved survival rates over that of surgery or radiation four cycles because of the development of metastasis during therapy (3), nausea and vomiting (l), and death alone. Seven groups of investigators have reported results of related to treatment (1). This latter patient had a white non-randomized surgical neoadjuvant trials, which have blood cell count of 100/ml and death was attributed to evaluated approximately 30 or more Stage I11 NSCLC sepsis. Two other patients died of documented myocarpatients. Analysis and interpretation of these studies are dial infarction during therapy. A clinical response was limited by the fact that three have been reported in ab- achieved in 70% (60/85) of the patients. Of the group of stract form only. All of the regimens included cisplatin, 85 patients, there were 62 that had a thoracotomy. The but the dose and schedule of this agent varied consider- reasons thoracotomy was not performed were treatment ably and a variety of other chemotherapeutic agents were related death (l), myocardial infarction during treatment included in these treatments. In addition, preoperative (2), disease progression on treatment (7), and patient treatment consisted of chemotherapy alone in two stud- refusal (13). In the group that refused surgery, there is ies, concurrent chemotherapy and radiation in 4 studies one patient that is alive and free of disease at 62 months. and sequential chemotherapy and radiation in one study. In this series, it is important to note that 97% had a The dose of radiation has ranged from 3,000 to 4,000 pulmonary resection accomplished (60/62) and complete cGy and the schedules have varied from continuous to resection (tumor free margins) was accomplished in 94% (58/62). The operative mortality was 5% (3/62) and comsplit course (Table 111). plications included 4 patients with bronchial fistula. The The Rush study [22] is currently the largest series and median survival of the 85 patients eligible for surgery as of July 1, 1989, is 22.1 months and the proportion surTABLE 11. Response Related to Stage in NSCLC* viving at 4 years is 36% (Fig. 1). The median survival of Response the 62 patients having a thoracotomy is 31.3 months Stage Regimen rate, % Investigator (Fig. 2). Martini [23] reported on 41 patients who were evaluIV MVP 28 Ruckdeschel et al. [6] Wagner et al. [191 111 MVP 53 ated in a surgical neoadjuvant program. Patients elimiIII MVP' 69 Gralla et al. [20] nated from consideration included those with superior Burkes et al. [21] 111 MVP" 69 vena caval syndrome, T4 lesions, pleural effusions, and Ruckdeschel et al. [16] IV CAP 23 those unable to tolerate resection. All of the patients had Fram et al. [18] 111 CAP 47 confirmation of N2 disease either by mediastinoscopy Longeval and Klastensky [17] IV EP 28 Longeval and Klastensky [17] (31) or at the time of thoracotomy (10). There were no 111 EP 56 T3NO patients included in this series. The chemothera*MVP = mitomycin vinblastine, cisplatin; CAP = cyclophosphapeutic program was vigorous in that cisplatin at a dose of mide, adriamycin, cisplatin; EP = etoposide-Cisplatin. 120 mg/M2 was given along with mitomycin and a vinca "High dose MVP.
258
Faber and Bonomi
TABLE 111. Regimen Used in Surgical Neoadjuvant NSCLC Trials No. of pts.
Treatment
Cycles
Faber et al. [221
56
4
Faber et al. [221
29
Martini et al. [231
41
Weiden and Piantadosi [24]
53
Skarin et al. [25]
41
Eagan et al. [26]
39
Burkes et al. [21]
35
Strauss et al. [27]
29
Cisplatin 60 mg/m2 IV day 1, 5FU 800/m2/day on days 1-5 as continuous 120 h infusion; radiation-200 cGy daily on days 1-5; repeated every 14 days Cisplatin 60 mg/m21V day 1, 5FU 800 mg/m2/day on days 1-4 or 96 h infusion; radiation-200 cGy daily on days 1-5; repeated every 21 days Cisplatin 120 mg/m2 on days 1, 29, and 71; vindesine (3mg/m2) or vinblastine (4.5 mg/m2); then every other week mitomycin 8 mg/m2 only on days 1, 29, 71 Cisplatin 75 mg/m2 on day 1, 5FU 1 gm/m2 on days 1-4 as 96 h continuous infusion; radiation-3,000 cGy on days 1-19, 200 cGy fractions Cisplatin 50 mg/m2 day 1, q 28 days cyclophosphamide 500 mg/m2 day, doxorubicin 5 mg/m2 day 1; radiation-3,W cGy over 3 weeks starting 3 weeks after second course of chemotherapy Cisplatin 60 mg/m2 day 1, cyclophosphamide 600 mg/m2 day, doxorubicin 40 mg/m2 day repeated every 28 days, radiation-300 cGy daily on days 29-33 and 57-61; total dose 3,000 cGy Cisplatin 120 mg/m2 days 1 & 29, mitomycin 8 mg/m2 days 1 & 29, vindesine 3 mg/m2 days 1, 8, 15, 22, 29, 43, 57 Cisplatin 100 mg/m2 day 1, vinblastine 3 mg/m2 days 1 & 3, 5FU 30 mglkglday on days 1-3 as a continuous infusion, repeated every 28 days; radiation-3,W cGy days 1-21
Investigator
Eligible for Surgery 85 Pts Median 22.1 Mos
4
2-3
2 2
3
2
2
Surgery 62 Pts Median 31.3 Mos
z
F
0.6
CC
0
-
F
0.6
IT
-
, 0
1
a
0.4
-
0.2
-
a: a
0.2 r
0
0
0
250
500
750
1000
1250
1500
DAYS
0
250
500
750
1000
1250
DAYS
Fig. 1. Kaplan-Meier survival estimate of 85 Stage 111 NSCLC patients eligible for preoperative chemotherapy and radiation.
Fig. 2. Kaplan-Meier survival estimate for 62 patients undergoing thoracotomy after chemotherapy and radiation.
alkoid. There was only one drug related death secondary to the toxicity of this chemotherapeutic regimen. A clinical response of 73% (30/41) was achieved and there was only one patient who had disease progress under therapy. Of the 41 patients placed in the program, 34 had a thoracotomy and the reasons for not proceeding with surgery were varied. These included tumor progression, death, medical contraindications and a subjective deci-
sion by the surgeon that resection could not be accomplished. There were 30 patients that had a partial or complete response to chemotherapy and 28 underwent surgery. Of these 28 patients, 21 had a complete resection, although a precise definition of complete was not given. It is interesting to note that in 3 patients, residual tumor could not be palpated in the lung and only mediastinal lymph nodes were removed and these were neg-
Neoadjuvant Treatment
ative in two of these three patients. The operative mortality was 3% (1/34). Martini was more selective in the patients who had minimal or no response to chemotherapy, as only 6/11 had a thoracotomy and a complete resection was accomplished in 3. An additional variant was added to this series in that all patients with viable tumor in the resected specimen received an additional two cycles of chemotherapy, 21 patients received postoperative radiation, and it was given in 11 following complete resection and in 10 after partial or no resection. 1251seeds were also used in 4 patients who had a partial resection. The tumor was sterilized in 20% (8/41) of the entire series. Reported survival at 3 years was 34% with a median survival of 20 months. Weiden and Piantadosi [24] evaluated 76 technically unresectable patients with advanced NSCLC. The chemotherapeutic regimen consisted of cisplatin and 5-FU. Radiation of 3,000 cGy was given on days 1 through 19 of the preoperative program. Only the responders were considered for surgery and 53 patients have been evaluated. The staging was both clinical and surgical, as 28 had biopsy proven N2 disease, 13 had N2 disease documented by CT scan, and 12 were apparently T3NO patients. There were two cardiac “events” that were apparent deaths related to chemotherapeutic toxicity. This program had a somewhat lower clinical response rate of 57% (30/53). Thoracotomy was carried out in 60% (32/ 53) and a complete resection was accomplished in 5 of 8 non-responders and in 17 of 24 responders. Definition of a complete resection was not provided. It is assumed that there was no operative mortality in this series, as none was mentioned. Of the 32 patients who had a thoracotomy, 7 had no viable tumor in their resected specimen. The median survival of the entire patient group (53 or 76?) is 10.5 months. Skarin [25] described 41 marginally resectable patients treated with a neoadjuvant chemotherapeutic program consisting of CAP followed by 3,000 cGy given in 3 weeks. Surgical histologic staging was accomplished in 27 patients and 14 were staged radiologically. There were 4 patients that progressed during treatment and 37 underwent thoracotomy. There were 31 patients that had N2 disease and 10 who had T3NO-1 disease. The resection rate of 97% (36/37) is impressive, and only one patient could not be resected due to extensive hilar fibrosis; this further substantiates the difficult technical problems associated with programs of preoperative therapy. Viable tumor was found in all of the 36 resected specimens. The bronchopleural fistula rate was 8.3% (31 36) and the operative mortality was 5% (2/37). It is somewhat difficult to evaluate the preoperative aspects of such a program as 27 patients also received approxi-
259
mately 2,500 cGy of postoperative radiation therapy and 23 patients received from I to 6 courses of postoperative CAP therapy. There were 7 patients who also received prophylactic cranial radiation. There was no statistical difference in survival comparing stage or responders versus non-responders after initial radiation and chemotherapy. The median disease free survival from the date of resection in 36 patients was 13.1 months and the median survival was 32.3 months, with a 1 year survival rate of 75%. Eagan et al. [26] described 39 patients with limited Stage I11 NSCLC who were treated preoperatively with 3 cycles of chemotherapy (CAP) at 4 week intervals, and during the second and third cycles, a total of 3,000 cGy of radiation was given. Surgical histologic staging was accomplished in 33 of 39 patients. Thirty-seven were staged as having N2 disease and 2 as T3N1 disease. The authors were stringent in their criteria for surgery, as patients were only eligible for thoracotomy if they had tumor regression. There were 19 patients who actually came to surgery and 13 were described as having complete resections , while 6 had incomplete resections. Definition of complete or incomplete resection was not given but reference is made that a complete resection indicated that no gross or microscopic tumor was left behind. This would appear to be a somewhat low complete resection rate, especially when those with no disease regression were eliminated from surgical consideration. This careful selection resulted in no surgical deaths and no bronchopleural fistula, which is an excellent result. The median survival for the 39 patients was 11.O months and the median survival for the 19 patients undergoing surgery was 15.0 months. Improved survival was noted among the patients who had a complete resection. Burkes et al. [21] reported 35 patients with mediastinoscopy proven Stage 111 N2 NSCLC who received two cycles of neoadjuvant chemotherapy. Patients who had a response to chemotherapy had a thoracotomy and postoperative chemotherapy. There were 21 patients who had surgery and 17 had a complete resection and 4 an incomplete resection. There was no evidence of tumor in 8.6% (3/21). The operative mortality was 10% (2121). The toxicity of the chemotherapy program which consisted of mitomycin, vindesine, and cisplatin was severe in that there were 4 septic deaths and one patient developed a mitomycin lung complication. The median survival had not been reached at the time of the report, although a 68% response rate was felt to be effective. Strauss et al. [27] reported on 29 Stage 111 NSCLC patients who received preoperative chemotherapy and radiation of 3,000 cGy. All patients had staging by mediastinsocopy and 24 had N2 disease and 8 had T3NO-1 disease. Resection was accomplished in 18 patients and
260
Faber and Bonomi
there was no tumor in the operative specimen in 2 of the 18. The operative mortality was 5% (1/22). There were 2 deaths related to the chemotherapy program. There were 18 patients reported to be in “remission” and 6 patients were alive with disease. These authors felt that their results were encouraging and suggested a phase I11 randomized trial employing their program of therapy. Detailed analysis of these neoadjuvant studies reveals variability among chemotherapeutic drugs used and their dosages, use and amount of radiation, staging, selection of patients for surgery, and use of postoperative adjunctive therapy. These variables make it somewhat difficult to analyze what is being accomplished, but some common themes do appear. The chemotherapeutic regimens are associated with some toxicity and reported mortality varies from 0% to 11%. The cumulative data indicate a preoperative mortality rate of 3.4% (1 1/323). Not all reports provide a description of hematologic, gastrointestinal, renal, and neurotoxicity and a detailed description in follow-up studies would be helpful. If significant improvements in survival can be achieved, this incidence of toxicity appears acceptable. Operative mortality varies from 0% to 10% with the combined mortality being 4.6% (9/195). There is undoubtedly a lack of uniformity among clinicians regarding the decision as to which patients can tolerate a major pulmonary resection, but the overall mortality of 4.6%in these series is most acceptable when dealing with clinically advanced NSCLC. Surgeons must be aware that the technical aspects of any resection following neoadjuvant therapy can be extremely difficult due to fibrosis and obliteration of normal vascular planes. Postoperative morbidity is also increased as evidenced by a postoperative bronchopleural fistula rate of 6.5% (4/62) in our own series. A complete resection is more frequently achieved when preoperative therapy is utilized and it varies from 68% to 97% in these reported series. These results compare favorably to previous reports of surgery alone or of surgery following radiation in this group of patients. Potential differences in staging and in patient characteristics make it difficult to compare results from earlier surgical series to results observed in the groups of patients who have received neoadjuvant treatment. However, the relatively high resectability rates observed in these neoadjuvant trials suggest that this treatment approach is increasing the resectability rate in Stage IIIa NSCLC patients. Complete resection is an important factor in achieving long-term survival. A significant observation is that no evidence of tumor was found in the resected specimen in up to 20% of the total patients entered into the trials and this percentage
would be even higher if only the resections are tabulated. Multivariate analyses of our own series reveals that sterilization of the tumor is the most important treatment related prognostic factor for long-term survival. This observation suggests that the rate of histologic complete remission may be an early indication of the effectiveness of a particular neoadjuvant treatment and that the histologic complete remission rate may be useful for selecting a promising regimen for evaluation in a randomized trial. Comparing survival data in reported trials reveals that median survival varies from 10.5 months to 32 months with some of the most mature studies showing 3 and 4 year survival rates of 34% and 40%, respectively. When compared to previous studies of surgery alone, it appears that some survival benefit may have been achieved. A lack of uniformity in some phases of these studies would indicate a cautious approach regarding this acceptance as standard therapy. Staging has not been uniform and was accomplished utiiizing both the CT scan to evaluate size of mediastinal nodes and surgical biopsy in 5 of the 7 studies. Surgical staging was used in all of the patients in two of the series, but the absence of positive histologic mediastinal nodes did not eliminate these patients in one of the studies. The lack of 100% histologic staging may permit some patients with enlarged mediastinal lymph nodes to be included in a neoadjuvant program when, indeed, these nodes reflect only inflammatory changes. Several of these studies were designed to include all Stage 111 NSCLC patients and therefore included varying numbers of T3NO-1 patients. It has been firmly established that the T3NO patient has an improved survival over those with N2 disease, and combining these patients into one series enhances survival data. However, Skarin et al. [25] used the log rank test and reported no major difference in survival when comparing the TN stages. T3NO-1 patients should not necessarily be eliminated from these types of studies, as it has been clearly documented that tumor size is diminished and resectability rates are increased. If survival can be increased in this group of patients, then they are candidates for neoadjuvant therapy. It is unclear as to what is a complete resection. Some view a resection as being incomplete when the highest mediastinal node is positive for carcinoma, while others view an incomplete resection as when surgical margins are positive or gross tumor remains. There needs to be developed a uniformity in the definition of complete and incomplete resection so that all surgeons have a clear understanding of what is being accomplished surgically. This is evidenced by the fact that reported complete resections varied from 68% to 97% in these series. There is a lack of uniformity in the selection of patients for resection. Progression of disease and inability
Neoadjuvant Treatment
to tolerate thoractomy are obvious indicators of inoperability. Stable disease or minimal response may result in surgeons’ subjective refusal to operate. We would suggest that all patients be selected as surgical candidates prior to the initiation of the neoadjunctive program and if therapy is completed without evidence of metastatic disease, they should be operated upon. Radiologic response is not a particularly good guideline to use to determine whether surgery is indicated or not. In our own series, there were some discrepancies between clinical response and pathologic findings as is evidenced by 4 patients with a radiologic complete response who had residual gross or microscopic cancer in the resected specimen and no residual tumor was found in the surgical specimen in four patients who had no evidence of clinical response. Conversely, Martini et al. [23]achieved a “complete resection” in 14 of 22 partial responders. Surgical resection should always be included in any therapeutic program of this type, as clinical response may be an unreliable predictor of the subsequent course of the patient. Should these patients receive postoperative adjunctive therapy? Skarin et al. [25]used postoperative radiation, postoperative chemotherapy, and prophylactic cranial radiation. Martini et al. [23]added postoperative chemotherapy if there was viable tumor in the specimen and 2 1/34also received postoperative external beam radiation. 1251 seeds were implanted if the resection was incomplete. There were 11 patients who received postoperative radiation despite the fact that they were defined as having a complete resection. In contrast, we did not use postoperative therapy of any type in any of our 62 patients who had a thoracotomy . Postoperative treatment certainly may skew survival data, and clinical investigators should attempt to agree on its use in any series of this type. Conclusions that can be reached about neoadjuvant programs are that the use of cisplatin achieves relatively high response rates, resectability rates are increased, morbidity and mortality are acceptable, the tumor can be sterilized in some cases, and the design of future studies should be more uniform. Einhorn [ 151 has stated that the evaluation of patient survival “can only be accomplished by a well designed, stratified, random, prospective phase I11 study.” We agree completely with this recommendation. However, do the results from the recently reported phase I1 trials of neoadjuvant therapy warrant evaluation in a randomized study? Our bias is that only treatment regimens associated with a twofold increase in median survival should be considered for a phase I11 trial. Therefore, for Stage 111-N2NSCLC patients, the median survival should increase from 12 to 24 months. So far, no chemotherapy or concurrent chemotherapy-radiation regimen has produced these results.
261
REFERENCES 1. Bromley LL, Szur L: Combined radiotherapy and resection for carcinoma of the bronchus, experience with 66 patients. Lancet 2:937-947, 1955. 2. Bloedom FG: Rationale and benefit of pre-operative irradiation in lung cancer. JAMA 196:128-129, 1966. 3. Takita H, Hollinshead AC, Rizzo DJ, et al.: Treatment of inoperable lung carcinoma. A combined modality approach. Ann Thorac Surg 28:363-368, 1979. 4. Mountain CF: A new international staging system for lung cancer. Chest [Supp1]89:225-233, 1986. 5 . Trastek VF, Pairolero PC, Piehler JM, et al.: En bloc (non-chest wall) resection for bronchogenic carcinoma with parietal fixation. J Thorac Cardiovasc Surg 87:352-358, 1984. 6. Frytak S, Eagan RT, Sawamurak K, et al.: Treatment of limited stage 111 non-small cell carcinoma of the lung. Cancer Invest 6:193-207, 1988. 7. Pearson FG: Radical surgery for N, disease. Chest [Supp1]89: 339-349. 1986. 8. Martini N, Flehinger BJ, Zaman MB, et al.: Results of resection in non-oat cell carcinoma of the lung with mediastinal lymph node metastases. Ann Surg 198:386-396, 1984. 9. Naruke T, Goya T, Tsuchiya R, et al.: The importance of surgery to non-small cell carcinoma of the lung with mediastinal lymph node metastases. Ann Thorac Surg 46:603-610, 1988. 10. Stanley KE: Prognostic factors for survival in patients with inoperable lung cancer. JNCI 65:25-32, 1980. 11. Frei E, 111: What’s in a name-neoadjuvant. JNC! 80:1088, 1089, 1988. 12. Bonomi P: Brief overview of combination chemotherapy in nonsmall cell lung cancer. Semin Oncol 13(Suppl 3):98-91, 1986. 13. Rapp E, Pater JL, Willon A, et al.: Chemotherapy can prolong survival in patients with advanced non-small cell lung cancer: Report of a Canadian multicenter randomized trial. J Clin Oncol 61633-664, 1988. 14. Devita VT, Jr: Principles of Chemotherapy. In DeVita VT, Jr, Hellman G, Rosenberg SA (eds): “Cancer: Principles and Practice of Oncology,” 3rd Edition. Philadelphia: JB Lippincott Publishers, 1989, 276-300. 15. Einhom LH: Neoadjuvant therapy of stage 111 non-small cell lung cancer. Ann Thorac Surg 46:362-365, 1988. 16. Ruckdeschel JC, Finkelstein DM, Ettinger DS, et al.: Chemotherapy for metastatic non-small cell bronchogenic carcinoma. EST 2575, generation V-a randomized comparison of four cisplatin containing regimens. J Clin Oncol 3:72-79, 1985. 17. Longeval E, Klastensky J: Combination chemotherapy with cisplatin and etoposide in bronchogenic squamous cell and adenocarcinoma-a study for the EORTC lung cancer working party (Belgium). Cancer 50:2751-2756, 1982. 18. Fram R, Skaren A, Balikian J, et a!.: Combination chemotherapy followed by radiation therapy in patients with regional stage I11 unresectable non-small cell lung cancer. Cancer Treat Rep 690~587-590, 1985. 19. Wagner H, Ruckdeschel J, Bonomi P, et al.: Treatment of locally advanced non-small cell lung cancer (NSCLC) with mitomycin C, vinblastine, and cisplatin (MVP) followed by radiation therapy: An ECOG pilot study. Proc Am SOCClin Oncol 4:183, 1985. 20. Gralla RJ,Kris MG, Martini N, et al.: A neoadjuvant trial in stage IIIa non-small cell lung cancer in patients with clinically apparent mediastinal node involvement with MVP chemotherapy, (mitovinca alkaloid + cisplatin) Lung Cancer [Suppll4: mycin A161, 1988. 21. Burkes R, Ginsbert R, Shepherd F, et al.: Neoadjuvant trialwith MVP (Mitomycin C + vindesine + cisplatin) chemotherapy for stage 111 (Tl-31, N2,MO) unresectable non-small cell lung cancer. Proc Am SOCClin Oncol 8:221, 1989. 22. Faber LP, Kittle CF, Warren WH, et al.: Preoperative chemotherapy and irradiation for stage 111 non-small cell lung cancer. Ann Thorac Surp. 47:669-677, 1989. 23. Martini N, KrisM, Gralla R, et al.: The effects of preoperative
+
262
Faber and Bonomi
chemotherapy on the resectability of non-small cell lung carcinoma with mediastinal lymph node metastases (N2 MO). Ann Thorac Surg 45:37&379, 1988. 24. Weiden P, Piantadosi S: Preoperative chemoradiotherapy in stage IH non-small cell lung cancer (NSCLC): A phase I1 study of the lung study group (LCSG). Proc Asco 7:197, 1988. 25. Skarin A, Jochelson M, Sheldon T, et al.: Neoadjuvant chemotherapy in marginally resectable stage 111 M, non-small cell lung cancer: Long term follow-up in 41 patients. J Surg Oncol40:266 274, 1989.
26. Eagan R , Ruud C, Lee R, et al.: Pilot study of induction therapy with Cyclophosphamide, Doxorubicin, and Cisplatin (CAP) and chest irradiation prior to thoracotomy in initially inoperable stage 111 M, non-small cell lung cancer. Cancer Treat Rep 71:895-900, 1987. 27. Strauss G , Sherman L, Mathisen D, et al.: Concurrent chemotherapy (CT) and radiotherapy (RT) followed by surgery (S) in marginally resectable stage IIIA non-small cell carcinoma of the lung (NSCLC): A cancer and leukemia group B study. Proc ASCO 7:203, 1988.
Neoadjuvant Treatment in locally Advanced Non-Small Cell Lung Cancer L. PENFIELD FABER, MD, AND PHILIP D. BONOMI, MD From the Departments of Cardiovascular and Thoracic Surgery (L. P. F.) and Medicine (P. D.B.), Rush-Presbyterian-St. Luke's Medical Center, Chicago, Illinois
Any neoadjuvant or preoperative treatment program that would increase survival in clinically advanced non-small cell lung cancer (NSCLC) patients would be of significant benefit. In addition to improving patient survival, the program should be associated with minimal toxicity and surgical mortality. The rationale for any preoperative program is that the ability to resect the advanced cancer will be enhanced and micrometastasis will be eradicated. We have analyzed 323 patients enrolled in various types of neoadjuvant and preoperative studies. Review of this data indicates that cisplatin containing regimens produce relatively high responder rates in locally advanced NSCLC patients, resectability rate can be increased by a neoadjuvant program and chemotherapeutic toxicity and operative mortality are not prohibitive. Survival data frequently includes T3NO-1 patients, but there does appear to be increased survival at 3 and 4 years. These studies remain experimental and continued analysis is necessary before they can be accepted as standard therapy for clinically advanced NSCLC cancer. ~
KEYWORDS:stage I11 non-small cell lung- cancer, preoperative treatment, chemotherapy, radiation
INTRODUCTION Approximately one-third of patients with non-small cell lung cancer (NSCLC) present with locally advanced disease without evidence of distant metastasis. Although subsets of locally advanced patients can have a relatively good prognosis, the overall survival for this group is poor despite the fact that most of them present with a relatively good performance and nutritional status. Surgery or radiation alone have been the most commonly used treatments for locally advanced lung cancer, but long-term results have not been good. Any program of preoperative therapy that would increase the incidence of a complete resection and decrease the occurrence of distant metastasis disease would be of obvious benefit. More than 30 years ago, clinical investigators conducted trials of preoperative radiation [1,2]; and a decade ago, the first trial of preoperative chemotherapy was reported [3]. There are several recent reports of trials evaluating preoperative chemotherapy with or without concurrent thoracic radiation, and they indicate that further investigation of this modality of therapy for locally advanced NSCLC is warranted. 0 1990 Wiley-Liss, Inc.
STAGING AND PROGNOSTIC FACTORS One of the difficulties in designing and evaluating studies in patients with Stage 111NSCLC is the fact that this disease stage comprises a diverse group of patients. This issue has been partially clarified by the new staging system in which the TNM classification has been revised to include T4 and N3 subgroups [4].The newly defined T and N categories have resulted in the subdivision of Stage I11 into IIIa and IIIb. Consideration for resection can be given to patients with Stage IIIa disease because mediastinal lymph node involvement is ipsilateral, while Stage IIIb patients are not surgical candidates because of contralateral and supraclavicular lymph node involvement. The new staging system represents an improvement, but there are still perceived shortcomings when attempting to categorize patients with locally advanced disease. For instance, Stage IIIa includes patients with surgically Address reprint requests to L. Penfield Faber, M.D., Rush-Presbyterian-St. Luke's Medical Center, 1653 West Congress Parkway, Chicago, IL 60612.
256
Faber and Bonomi
pathologically staged T3NO disease which is associated with a 3 0 4 5 % five year survival rate 1451. In stark contrast, Stage IIIa patients whose disease is classified as T3N2 have virtually no chance of surviving 5 years [6]. There are also variations in prognosis in patients who have histologically confirmed N2 disease. Patients who have microsopic nodal metastasis found at the time of surgical resection have a 20-30% five year survival rate [7-91. In contrast, patients in whom nodal metastasis are detected by mediastinoscopy have approximately a 10% chance of surviving 5 years [7]. Lastly, the 5 year survival rate for nodal metastasis which are evident on routine chest x-ray is less than 5% [8]. When Stanley [ 101 conducted statistical analysis of 5,100 lung cancer patients, performance status, previous weight loss, and the extent of disease were shown to be the most important prognostic factors. Kaplan-Meier product limit method was used to estimate survival in relation to potential prognostic factors in 129 Stage 111 NSCLC patients receiving concurrent radiation and chemotherapy (62 had subsequent surgery) at the Rush University Medical Center. The following parameters were assessed in this study: age, sex, performance status, previous weight loss 5%, histology, hemoglobin, serum albumin, histology, race, eligible versus ineligible for surgery, Stage IIIa versus Stage IIIb, subsets of stage defined as IIIa (T3NO) versus IIIa (N2) versus IIIb, and method of mediastinal staging (clinical versus histologic). Surgical eligibility, performance status, and the subgroups of Stage IIIa were found to be the most important prognostic factors (Table I). These analyses show that Stage IIIa is not a homogenous group with the median survival durations for T3NO and N2 being 851 days and 500 days, respectively. All of the above have serious implications for designing clinical trials in patients with locally advanced NSCLC and when evaluating these studies, prognostic
factors, and subgroups of staging must be taken into consideration.
CHEMOTHERAPY: THEORETICAL AND PRACTICAL CONSIDERATIONS
The rationale for neoadjuvant chemotherapy is based on the assumption that chemotherapy-induced regression of the primary lesion will result in more effective local control with subsequent surgery or radiation [ 111. It represents the earliest possible chance to treat micrometastases, possibly resulting in the elimination of the rapidly growing tumor cells, which might not occur at a later time when the number of tumor cells is greater and they also may be less sensitive to chemotherapy because they are in a resting phase or have become drug resistant. The majority of chemotherapy data in NSCLC has been obtained from trials in patients with Stage IV disease. The larger trials have shown that the “most active” regimens contain cisplatin and they produce response rates of approximately 25% in Stage IV patients [12]. Despite discouraging response rates in metastatic disease, there is evidence that treatment with cisplatin containing chemotherapy has a modest but statistically significant beneficial effect on survival in Stage IV NSCLC ~31. There are compelling theoretical and practical reasons for using preoperative systemic therapy in Stage I11 NSCLC. One of the main factors which determine curability is tumor burden, and the cure rate is inversely related to the amount of tumor. Chemotherapy produces higher cure rates in animals injected with smaller number of tumor cells than in animals receiving larger numbers of tumor cells. Similarly, delaying systemic therapy following injection of tumor cells into animals results in lower cure rates because a longer period of tumor growth results in a greater tumor burden [ 141. It is believed that systemic therapy which is marginally effective against gross disease might be more effective when used to treat micrometastases. However, from TABLE I. Prognostic Factors in Stage I11 NSCLC Patients a practical point of view, clinicians have employed the Median rule of 50% when designing trials of adjuvant systemic survival Variable Patients (days) P-value therapy. In short, the use of neoadjuvant chemotherapy should be limited to tumors with recurrence rates greater Surgical eligibility than 50% and for which chemotherapy has produced reEligible 83 663 ,003 sponse rates of greater or equal to 50% in advanced Ineligible 46 385 disease [ 151. Most multi-institution studies have shown Performance status 0-1 response rates of approximately 25% for cisplatin con102 614 ,001 2-3 27 268 taining combination chemotherapy regimens in Stage IV Stage disease [16]. However, recent reports indicate that the IIIa 83 656 ,004 same regimens are capable of producing a response rate IIIb 46 368 at approximately 50% in patients with Stage I11 disease Stage I11 ,004 [17-191. In fact, intensification of the doses in the miT3NO 42 85 1 IIIa 41 500 tomycin, vinblastine, and cisplatin (MVP) regimen has IIIb 46 368 produced response rates of 70% in previously untreated
Neoadjuvant Treatment
257
Stage 111 NSCLC patients [20,21]. Therefore, it appears 130 patients were evaluated for a program of combined that combination chemotherapy is capable of producing chemotherapy and radiation. There were 85 that were response rates of greater than or equal to 50% in Stage I11 considered eligible for surgery and there were 45 not patients (Table 11). considered candidates for surgical intervention because of insufficient cardiopulmonary reserve (16), contralatSURGICAL NEOADJUVANT STUDIES era1 and supraclavicularlymph node metastasis (15), and Radiation therapy can be used with chemotherapy ei- extensive mediastinal disease as evidenced by vena caval ther to enhance the effect of the alternate method or to obstruction or tracheal and carinal involvement (14). treat separate components of the malignancy, i.e., local There were two different groups of drugs used and the disease and microscopic metastatic disease. Radiation first included cisplatin and 5-fluorouracil (5-FU), while and chemotherapy work similarly to affect biological tu- etoposide (VP-16) was added to the second group. Split mor cell destruction, and these modes of therapy can be course radiation was administered every other week for 5 used in combination for the following purposes: 1) to days in the first group and every third week for 5 days in increase the effect of the other method in the instance of the second group. Total radiation dose was 4,000 cGy. bulky localized disease, 2) to accelerate response, and 3) This series included 18 patients with T3NO disease, of to increase the effect of the other method in the instance whom 9 had chest wall invasion, 6 demonstrated mediof marginally responsive tumors. Disadvantages of such astinal involvement with obliteration of adjacent fascial a program of treatment include toxicity of the treatment planes on the computed tomographic scan, and 3 had program resulting in morbidity or death, delay in surgical main stem bronchial disease less than 2 cm from the therapy to control local disease, an increase in operative carina. There were 5 patients with T4NO-2 disease. complications and mortality, and a failure to achieve There were 5 patients that failed to receive the planned improved survival rates over that of surgery or radiation four cycles because of the development of metastasis during therapy (3), nausea and vomiting (l), and death alone. Seven groups of investigators have reported results of related to treatment (1). This latter patient had a white non-randomized surgical neoadjuvant trials, which have blood cell count of 100/ml and death was attributed to evaluated approximately 30 or more Stage I11 NSCLC sepsis. Two other patients died of documented myocarpatients. Analysis and interpretation of these studies are dial infarction during therapy. A clinical response was limited by the fact that three have been reported in ab- achieved in 70% (60/85) of the patients. Of the group of stract form only. All of the regimens included cisplatin, 85 patients, there were 62 that had a thoracotomy. The but the dose and schedule of this agent varied consider- reasons thoracotomy was not performed were treatment ably and a variety of other chemotherapeutic agents were related death (l), myocardial infarction during treatment included in these treatments. In addition, preoperative (2), disease progression on treatment (7), and patient treatment consisted of chemotherapy alone in two stud- refusal (13). In the group that refused surgery, there is ies, concurrent chemotherapy and radiation in 4 studies one patient that is alive and free of disease at 62 months. and sequential chemotherapy and radiation in one study. In this series, it is important to note that 97% had a The dose of radiation has ranged from 3,000 to 4,000 pulmonary resection accomplished (60/62) and complete cGy and the schedules have varied from continuous to resection (tumor free margins) was accomplished in 94% (58/62). The operative mortality was 5% (3/62) and comsplit course (Table 111). plications included 4 patients with bronchial fistula. The The Rush study [22] is currently the largest series and median survival of the 85 patients eligible for surgery as of July 1, 1989, is 22.1 months and the proportion surTABLE 11. Response Related to Stage in NSCLC* viving at 4 years is 36% (Fig. 1). The median survival of Response the 62 patients having a thoracotomy is 31.3 months Stage Regimen rate, % Investigator (Fig. 2). Martini [23] reported on 41 patients who were evaluIV MVP 28 Ruckdeschel et al. [6] Wagner et al. [191 111 MVP 53 ated in a surgical neoadjuvant program. Patients elimiIII MVP' 69 Gralla et al. [20] nated from consideration included those with superior Burkes et al. [21] 111 MVP" 69 vena caval syndrome, T4 lesions, pleural effusions, and Ruckdeschel et al. [16] IV CAP 23 those unable to tolerate resection. All of the patients had Fram et al. [18] 111 CAP 47 confirmation of N2 disease either by mediastinoscopy Longeval and Klastensky [17] IV EP 28 Longeval and Klastensky [17] (31) or at the time of thoracotomy (10). There were no 111 EP 56 T3NO patients included in this series. The chemothera*MVP = mitomycin vinblastine, cisplatin; CAP = cyclophosphapeutic program was vigorous in that cisplatin at a dose of mide, adriamycin, cisplatin; EP = etoposide-Cisplatin. 120 mg/M2 was given along with mitomycin and a vinca "High dose MVP.
258
Faber and Bonomi
TABLE 111. Regimen Used in Surgical Neoadjuvant NSCLC Trials No. of pts.
Treatment
Cycles
Faber et al. [221
56
4
Faber et al. [221
29
Martini et al. [231
41
Weiden and Piantadosi [24]
53
Skarin et al. [25]
41
Eagan et al. [26]
39
Burkes et al. [21]
35
Strauss et al. [27]
29
Cisplatin 60 mg/m2 IV day 1, 5FU 800/m2/day on days 1-5 as continuous 120 h infusion; radiation-200 cGy daily on days 1-5; repeated every 14 days Cisplatin 60 mg/m21V day 1, 5FU 800 mg/m2/day on days 1-4 or 96 h infusion; radiation-200 cGy daily on days 1-5; repeated every 21 days Cisplatin 120 mg/m2 on days 1, 29, and 71; vindesine (3mg/m2) or vinblastine (4.5 mg/m2); then every other week mitomycin 8 mg/m2 only on days 1, 29, 71 Cisplatin 75 mg/m2 on day 1, 5FU 1 gm/m2 on days 1-4 as 96 h continuous infusion; radiation-3,000 cGy on days 1-19, 200 cGy fractions Cisplatin 50 mg/m2 day 1, q 28 days cyclophosphamide 500 mg/m2 day, doxorubicin 5 mg/m2 day 1; radiation-3,W cGy over 3 weeks starting 3 weeks after second course of chemotherapy Cisplatin 60 mg/m2 day 1, cyclophosphamide 600 mg/m2 day, doxorubicin 40 mg/m2 day repeated every 28 days, radiation-300 cGy daily on days 29-33 and 57-61; total dose 3,000 cGy Cisplatin 120 mg/m2 days 1 & 29, mitomycin 8 mg/m2 days 1 & 29, vindesine 3 mg/m2 days 1, 8, 15, 22, 29, 43, 57 Cisplatin 100 mg/m2 day 1, vinblastine 3 mg/m2 days 1 & 3, 5FU 30 mglkglday on days 1-3 as a continuous infusion, repeated every 28 days; radiation-3,W cGy days 1-21
Investigator
Eligible for Surgery 85 Pts Median 22.1 Mos
4
2-3
2 2
3
2
2
Surgery 62 Pts Median 31.3 Mos
z
F
0.6
CC
0
-
F
0.6
IT
-
, 0
1
a
0.4
-
0.2
-
a: a
0.2 r
0
0
0
250
500
750
1000
1250
1500
DAYS
0
250
500
750
1000
1250
DAYS
Fig. 1. Kaplan-Meier survival estimate of 85 Stage 111 NSCLC patients eligible for preoperative chemotherapy and radiation.
Fig. 2. Kaplan-Meier survival estimate for 62 patients undergoing thoracotomy after chemotherapy and radiation.
alkoid. There was only one drug related death secondary to the toxicity of this chemotherapeutic regimen. A clinical response of 73% (30/41) was achieved and there was only one patient who had disease progress under therapy. Of the 41 patients placed in the program, 34 had a thoracotomy and the reasons for not proceeding with surgery were varied. These included tumor progression, death, medical contraindications and a subjective deci-
sion by the surgeon that resection could not be accomplished. There were 30 patients that had a partial or complete response to chemotherapy and 28 underwent surgery. Of these 28 patients, 21 had a complete resection, although a precise definition of complete was not given. It is interesting to note that in 3 patients, residual tumor could not be palpated in the lung and only mediastinal lymph nodes were removed and these were neg-
Neoadjuvant Treatment
ative in two of these three patients. The operative mortality was 3% (1/34). Martini was more selective in the patients who had minimal or no response to chemotherapy, as only 6/11 had a thoracotomy and a complete resection was accomplished in 3. An additional variant was added to this series in that all patients with viable tumor in the resected specimen received an additional two cycles of chemotherapy, 21 patients received postoperative radiation, and it was given in 11 following complete resection and in 10 after partial or no resection. 1251seeds were also used in 4 patients who had a partial resection. The tumor was sterilized in 20% (8/41) of the entire series. Reported survival at 3 years was 34% with a median survival of 20 months. Weiden and Piantadosi [24] evaluated 76 technically unresectable patients with advanced NSCLC. The chemotherapeutic regimen consisted of cisplatin and 5-FU. Radiation of 3,000 cGy was given on days 1 through 19 of the preoperative program. Only the responders were considered for surgery and 53 patients have been evaluated. The staging was both clinical and surgical, as 28 had biopsy proven N2 disease, 13 had N2 disease documented by CT scan, and 12 were apparently T3NO patients. There were two cardiac “events” that were apparent deaths related to chemotherapeutic toxicity. This program had a somewhat lower clinical response rate of 57% (30/53). Thoracotomy was carried out in 60% (32/ 53) and a complete resection was accomplished in 5 of 8 non-responders and in 17 of 24 responders. Definition of a complete resection was not provided. It is assumed that there was no operative mortality in this series, as none was mentioned. Of the 32 patients who had a thoracotomy, 7 had no viable tumor in their resected specimen. The median survival of the entire patient group (53 or 76?) is 10.5 months. Skarin [25] described 41 marginally resectable patients treated with a neoadjuvant chemotherapeutic program consisting of CAP followed by 3,000 cGy given in 3 weeks. Surgical histologic staging was accomplished in 27 patients and 14 were staged radiologically. There were 4 patients that progressed during treatment and 37 underwent thoracotomy. There were 31 patients that had N2 disease and 10 who had T3NO-1 disease. The resection rate of 97% (36/37) is impressive, and only one patient could not be resected due to extensive hilar fibrosis; this further substantiates the difficult technical problems associated with programs of preoperative therapy. Viable tumor was found in all of the 36 resected specimens. The bronchopleural fistula rate was 8.3% (31 36) and the operative mortality was 5% (2/37). It is somewhat difficult to evaluate the preoperative aspects of such a program as 27 patients also received approxi-
259
mately 2,500 cGy of postoperative radiation therapy and 23 patients received from I to 6 courses of postoperative CAP therapy. There were 7 patients who also received prophylactic cranial radiation. There was no statistical difference in survival comparing stage or responders versus non-responders after initial radiation and chemotherapy. The median disease free survival from the date of resection in 36 patients was 13.1 months and the median survival was 32.3 months, with a 1 year survival rate of 75%. Eagan et al. [26] described 39 patients with limited Stage I11 NSCLC who were treated preoperatively with 3 cycles of chemotherapy (CAP) at 4 week intervals, and during the second and third cycles, a total of 3,000 cGy of radiation was given. Surgical histologic staging was accomplished in 33 of 39 patients. Thirty-seven were staged as having N2 disease and 2 as T3N1 disease. The authors were stringent in their criteria for surgery, as patients were only eligible for thoracotomy if they had tumor regression. There were 19 patients who actually came to surgery and 13 were described as having complete resections , while 6 had incomplete resections. Definition of complete or incomplete resection was not given but reference is made that a complete resection indicated that no gross or microscopic tumor was left behind. This would appear to be a somewhat low complete resection rate, especially when those with no disease regression were eliminated from surgical consideration. This careful selection resulted in no surgical deaths and no bronchopleural fistula, which is an excellent result. The median survival for the 39 patients was 11.O months and the median survival for the 19 patients undergoing surgery was 15.0 months. Improved survival was noted among the patients who had a complete resection. Burkes et al. [21] reported 35 patients with mediastinoscopy proven Stage 111 N2 NSCLC who received two cycles of neoadjuvant chemotherapy. Patients who had a response to chemotherapy had a thoracotomy and postoperative chemotherapy. There were 21 patients who had surgery and 17 had a complete resection and 4 an incomplete resection. There was no evidence of tumor in 8.6% (3/21). The operative mortality was 10% (2121). The toxicity of the chemotherapy program which consisted of mitomycin, vindesine, and cisplatin was severe in that there were 4 septic deaths and one patient developed a mitomycin lung complication. The median survival had not been reached at the time of the report, although a 68% response rate was felt to be effective. Strauss et al. [27] reported on 29 Stage 111 NSCLC patients who received preoperative chemotherapy and radiation of 3,000 cGy. All patients had staging by mediastinsocopy and 24 had N2 disease and 8 had T3NO-1 disease. Resection was accomplished in 18 patients and
260
Faber and Bonomi
there was no tumor in the operative specimen in 2 of the 18. The operative mortality was 5% (1/22). There were 2 deaths related to the chemotherapy program. There were 18 patients reported to be in “remission” and 6 patients were alive with disease. These authors felt that their results were encouraging and suggested a phase I11 randomized trial employing their program of therapy. Detailed analysis of these neoadjuvant studies reveals variability among chemotherapeutic drugs used and their dosages, use and amount of radiation, staging, selection of patients for surgery, and use of postoperative adjunctive therapy. These variables make it somewhat difficult to analyze what is being accomplished, but some common themes do appear. The chemotherapeutic regimens are associated with some toxicity and reported mortality varies from 0% to 11%. The cumulative data indicate a preoperative mortality rate of 3.4% (1 1/323). Not all reports provide a description of hematologic, gastrointestinal, renal, and neurotoxicity and a detailed description in follow-up studies would be helpful. If significant improvements in survival can be achieved, this incidence of toxicity appears acceptable. Operative mortality varies from 0% to 10% with the combined mortality being 4.6% (9/195). There is undoubtedly a lack of uniformity among clinicians regarding the decision as to which patients can tolerate a major pulmonary resection, but the overall mortality of 4.6%in these series is most acceptable when dealing with clinically advanced NSCLC. Surgeons must be aware that the technical aspects of any resection following neoadjuvant therapy can be extremely difficult due to fibrosis and obliteration of normal vascular planes. Postoperative morbidity is also increased as evidenced by a postoperative bronchopleural fistula rate of 6.5% (4/62) in our own series. A complete resection is more frequently achieved when preoperative therapy is utilized and it varies from 68% to 97% in these reported series. These results compare favorably to previous reports of surgery alone or of surgery following radiation in this group of patients. Potential differences in staging and in patient characteristics make it difficult to compare results from earlier surgical series to results observed in the groups of patients who have received neoadjuvant treatment. However, the relatively high resectability rates observed in these neoadjuvant trials suggest that this treatment approach is increasing the resectability rate in Stage IIIa NSCLC patients. Complete resection is an important factor in achieving long-term survival. A significant observation is that no evidence of tumor was found in the resected specimen in up to 20% of the total patients entered into the trials and this percentage
would be even higher if only the resections are tabulated. Multivariate analyses of our own series reveals that sterilization of the tumor is the most important treatment related prognostic factor for long-term survival. This observation suggests that the rate of histologic complete remission may be an early indication of the effectiveness of a particular neoadjuvant treatment and that the histologic complete remission rate may be useful for selecting a promising regimen for evaluation in a randomized trial. Comparing survival data in reported trials reveals that median survival varies from 10.5 months to 32 months with some of the most mature studies showing 3 and 4 year survival rates of 34% and 40%, respectively. When compared to previous studies of surgery alone, it appears that some survival benefit may have been achieved. A lack of uniformity in some phases of these studies would indicate a cautious approach regarding this acceptance as standard therapy. Staging has not been uniform and was accomplished utiiizing both the CT scan to evaluate size of mediastinal nodes and surgical biopsy in 5 of the 7 studies. Surgical staging was used in all of the patients in two of the series, but the absence of positive histologic mediastinal nodes did not eliminate these patients in one of the studies. The lack of 100% histologic staging may permit some patients with enlarged mediastinal lymph nodes to be included in a neoadjuvant program when, indeed, these nodes reflect only inflammatory changes. Several of these studies were designed to include all Stage 111 NSCLC patients and therefore included varying numbers of T3NO-1 patients. It has been firmly established that the T3NO patient has an improved survival over those with N2 disease, and combining these patients into one series enhances survival data. However, Skarin et al. [25] used the log rank test and reported no major difference in survival when comparing the TN stages. T3NO-1 patients should not necessarily be eliminated from these types of studies, as it has been clearly documented that tumor size is diminished and resectability rates are increased. If survival can be increased in this group of patients, then they are candidates for neoadjuvant therapy. It is unclear as to what is a complete resection. Some view a resection as being incomplete when the highest mediastinal node is positive for carcinoma, while others view an incomplete resection as when surgical margins are positive or gross tumor remains. There needs to be developed a uniformity in the definition of complete and incomplete resection so that all surgeons have a clear understanding of what is being accomplished surgically. This is evidenced by the fact that reported complete resections varied from 68% to 97% in these series. There is a lack of uniformity in the selection of patients for resection. Progression of disease and inability
Neoadjuvant Treatment
to tolerate thoractomy are obvious indicators of inoperability. Stable disease or minimal response may result in surgeons’ subjective refusal to operate. We would suggest that all patients be selected as surgical candidates prior to the initiation of the neoadjunctive program and if therapy is completed without evidence of metastatic disease, they should be operated upon. Radiologic response is not a particularly good guideline to use to determine whether surgery is indicated or not. In our own series, there were some discrepancies between clinical response and pathologic findings as is evidenced by 4 patients with a radiologic complete response who had residual gross or microscopic cancer in the resected specimen and no residual tumor was found in the surgical specimen in four patients who had no evidence of clinical response. Conversely, Martini et al. [23]achieved a “complete resection” in 14 of 22 partial responders. Surgical resection should always be included in any therapeutic program of this type, as clinical response may be an unreliable predictor of the subsequent course of the patient. Should these patients receive postoperative adjunctive therapy? Skarin et al. [25]used postoperative radiation, postoperative chemotherapy, and prophylactic cranial radiation. Martini et al. [23]added postoperative chemotherapy if there was viable tumor in the specimen and 2 1/34also received postoperative external beam radiation. 1251 seeds were implanted if the resection was incomplete. There were 11 patients who received postoperative radiation despite the fact that they were defined as having a complete resection. In contrast, we did not use postoperative therapy of any type in any of our 62 patients who had a thoracotomy . Postoperative treatment certainly may skew survival data, and clinical investigators should attempt to agree on its use in any series of this type. Conclusions that can be reached about neoadjuvant programs are that the use of cisplatin achieves relatively high response rates, resectability rates are increased, morbidity and mortality are acceptable, the tumor can be sterilized in some cases, and the design of future studies should be more uniform. Einhorn [ 151 has stated that the evaluation of patient survival “can only be accomplished by a well designed, stratified, random, prospective phase I11 study.” We agree completely with this recommendation. However, do the results from the recently reported phase I1 trials of neoadjuvant therapy warrant evaluation in a randomized study? Our bias is that only treatment regimens associated with a twofold increase in median survival should be considered for a phase I11 trial. Therefore, for Stage 111-N2NSCLC patients, the median survival should increase from 12 to 24 months. So far, no chemotherapy or concurrent chemotherapy-radiation regimen has produced these results.
261
REFERENCES 1. Bromley LL, Szur L: Combined radiotherapy and resection for carcinoma of the bronchus, experience with 66 patients. Lancet 2:937-947, 1955. 2. Bloedom FG: Rationale and benefit of pre-operative irradiation in lung cancer. JAMA 196:128-129, 1966. 3. Takita H, Hollinshead AC, Rizzo DJ, et al.: Treatment of inoperable lung carcinoma. A combined modality approach. Ann Thorac Surg 28:363-368, 1979. 4. Mountain CF: A new international staging system for lung cancer. Chest [Supp1]89:225-233, 1986. 5 . Trastek VF, Pairolero PC, Piehler JM, et al.: En bloc (non-chest wall) resection for bronchogenic carcinoma with parietal fixation. J Thorac Cardiovasc Surg 87:352-358, 1984. 6. Frytak S, Eagan RT, Sawamurak K, et al.: Treatment of limited stage 111 non-small cell carcinoma of the lung. Cancer Invest 6:193-207, 1988. 7. Pearson FG: Radical surgery for N, disease. Chest [Supp1]89: 339-349. 1986. 8. Martini N, Flehinger BJ, Zaman MB, et al.: Results of resection in non-oat cell carcinoma of the lung with mediastinal lymph node metastases. Ann Surg 198:386-396, 1984. 9. Naruke T, Goya T, Tsuchiya R, et al.: The importance of surgery to non-small cell carcinoma of the lung with mediastinal lymph node metastases. Ann Thorac Surg 46:603-610, 1988. 10. Stanley KE: Prognostic factors for survival in patients with inoperable lung cancer. JNCI 65:25-32, 1980. 11. Frei E, 111: What’s in a name-neoadjuvant. JNC! 80:1088, 1089, 1988. 12. Bonomi P: Brief overview of combination chemotherapy in nonsmall cell lung cancer. Semin Oncol 13(Suppl 3):98-91, 1986. 13. Rapp E, Pater JL, Willon A, et al.: Chemotherapy can prolong survival in patients with advanced non-small cell lung cancer: Report of a Canadian multicenter randomized trial. J Clin Oncol 61633-664, 1988. 14. Devita VT, Jr: Principles of Chemotherapy. In DeVita VT, Jr, Hellman G, Rosenberg SA (eds): “Cancer: Principles and Practice of Oncology,” 3rd Edition. Philadelphia: JB Lippincott Publishers, 1989, 276-300. 15. Einhom LH: Neoadjuvant therapy of stage 111 non-small cell lung cancer. Ann Thorac Surg 46:362-365, 1988. 16. Ruckdeschel JC, Finkelstein DM, Ettinger DS, et al.: Chemotherapy for metastatic non-small cell bronchogenic carcinoma. EST 2575, generation V-a randomized comparison of four cisplatin containing regimens. J Clin Oncol 3:72-79, 1985. 17. Longeval E, Klastensky J: Combination chemotherapy with cisplatin and etoposide in bronchogenic squamous cell and adenocarcinoma-a study for the EORTC lung cancer working party (Belgium). Cancer 50:2751-2756, 1982. 18. Fram R, Skaren A, Balikian J, et a!.: Combination chemotherapy followed by radiation therapy in patients with regional stage I11 unresectable non-small cell lung cancer. Cancer Treat Rep 690~587-590, 1985. 19. Wagner H, Ruckdeschel J, Bonomi P, et al.: Treatment of locally advanced non-small cell lung cancer (NSCLC) with mitomycin C, vinblastine, and cisplatin (MVP) followed by radiation therapy: An ECOG pilot study. Proc Am SOCClin Oncol 4:183, 1985. 20. Gralla RJ,Kris MG, Martini N, et al.: A neoadjuvant trial in stage IIIa non-small cell lung cancer in patients with clinically apparent mediastinal node involvement with MVP chemotherapy, (mitovinca alkaloid + cisplatin) Lung Cancer [Suppll4: mycin A161, 1988. 21. Burkes R, Ginsbert R, Shepherd F, et al.: Neoadjuvant trialwith MVP (Mitomycin C + vindesine + cisplatin) chemotherapy for stage 111 (Tl-31, N2,MO) unresectable non-small cell lung cancer. Proc Am SOCClin Oncol 8:221, 1989. 22. Faber LP, Kittle CF, Warren WH, et al.: Preoperative chemotherapy and irradiation for stage 111 non-small cell lung cancer. Ann Thorac Surp. 47:669-677, 1989. 23. Martini N, KrisM, Gralla R, et al.: The effects of preoperative
+
262
Faber and Bonomi
chemotherapy on the resectability of non-small cell lung carcinoma with mediastinal lymph node metastases (N2 MO). Ann Thorac Surg 45:37&379, 1988. 24. Weiden P, Piantadosi S: Preoperative chemoradiotherapy in stage IH non-small cell lung cancer (NSCLC): A phase I1 study of the lung study group (LCSG). Proc Asco 7:197, 1988. 25. Skarin A, Jochelson M, Sheldon T, et al.: Neoadjuvant chemotherapy in marginally resectable stage 111 M, non-small cell lung cancer: Long term follow-up in 41 patients. J Surg Oncol40:266 274, 1989.
26. Eagan R , Ruud C, Lee R, et al.: Pilot study of induction therapy with Cyclophosphamide, Doxorubicin, and Cisplatin (CAP) and chest irradiation prior to thoracotomy in initially inoperable stage 111 M, non-small cell lung cancer. Cancer Treat Rep 71:895-900, 1987. 27. Strauss G , Sherman L, Mathisen D, et al.: Concurrent chemotherapy (CT) and radiotherapy (RT) followed by surgery (S) in marginally resectable stage IIIA non-small cell carcinoma of the lung (NSCLC): A cancer and leukemia group B study. Proc ASCO 7:203, 1988.
