Surgical treatment for limited small-cell lung cancer

J THORAC CARDIOVASC SURG 1991;101:385-93

Original Communications

Surgical treatment for limited small-cell lung cancer The University of Toronto Lung Oncology Group experience Since 1977, 119 patients with limited small-ceU lung cancer have undergone combined modality therapy including surgery at our institution. Seventy-nine patients (58 male, 21 female; median age 63 years) had surgery first, and 67 of these bad adjuvant chemotherapy. Forty (27 male, 13 female; median age S9 years) had chemotherapy first, and 94 % had a complete or partial response before the operation. Pretreatment staging revealed 69 stage I, 27 stage II, and 23 stage rn tumors. Twenty-six patients required pneumonectomy, 88 lobectomy, and five bad no resection. Four patients had gross and six had microscopic residual disease. Postoperative pathologic examination showed smaU-ceU lung cancer only (n = 95), non-small-ceU lung cancer (n = 3~ mixed (n = 17), and no residual tumor (n = 4~ Postoperative staging revealed 3S stage I, 36 stage II, and 48 stage rna tumors. The median survival of the entire group is III weeks and the projected S-year survival rate is 39%. No survival difference was seen between patients treated with chemotherapy before the operation and those undergoing an initial operation foUowed by chemotherapy (p = 0.7~ The median survival for patients with pathologic stage I disease has not been reached, and the projected S-year survival rate is SI %. This is significantly better than for the patients with stage II (median 82 weeks, p = 0.001) or stage rn (median 83 weeks, p = 0.001) disease, who have projected S-year survival rates of 28% and 19%, respeCtively. Seven of the 12 patients who bad no adjuvant chemotherapy remain aHve at 6 to 48+ months. Sixty-seven patients have died (11 bad no evidence of disease~ Only 10 patients bad a relapse in the primary site alone, seven at the primary and distant sites, and 39 only in distant sites. In summary, resection improves control at the primary site, and a significant proportion of patients with stage I (NO) disease achieve long-term survival and cure with combined modality therapy including surgery. Stage II and rna patients have survival predicticm similar to stage rna non-small-ceU lung carcinoma treated surgically.

FrancesA. Shepherd, MD, FRcpca (by invitation), Robert J. Ginsberg, MD, FRCSC,b Ronald Feld, MD, FRCPCC (by invitation), William K. Evans, MD, FRcpca (by invitation), and Else Johansen" (by invitation), Toronto, Ontario, Canada

From the Departments of Medicine of the Toronto General Hospital" and the Princess Margaret Hospital," the Department of Surgery of the Mount Sinai Hospital.> and the University of Toronto. Supported in part by the Myers and Sylvia Lawrence Oncology Research Fund. Read at the Seventieth Annual Meeting of The American Association for Thoracic Surgery, Toronto, Ontario, Canada, May 7-9, 1990. Addressfor reprints: Frances A. Shepherd, MD, Toronto General Hospital, 200 Elizabeth St., mjlw 2-035, Toronto, ON M5G 2C4, Canada.

12/6/25672

Surgical resectionalone is inadequate therapy for limited small-cell lung cancer (SCLC), because it results in lessthan a 5%long-termsurvival rate.' The poor survival rate of patients in the surgicalarm of the British Medical Research Councilstudy- led most investigators to abandon surgery as the initial treatment for bronchogenic neoplasms of this cell type. At this time, standard therapy for SCLC, evenin the absence of metastatic disease,consists of chemotherapy alone or in combination with local radiation therapy.v' 385

The Journal, of· Thoracic and Cardiovascular Surgery

3 8 6 Shepherd et al.

Table I. Characteristics of 119 patients undergoing surgical treatmentfor limited SCLC Characteristics

Surgery first

Number 79 Sex Male 58 Female 21 Age (yr) Median 63 Range 36-84 Pretreatment clinical stage TI NO 31 T2NO 27 Total stage I 58 7 TI NI T2NI 8 Total stage 2 15 TI N2 I T2N2 I T2NO 3 T3NI I Total stage 3 6

Adjuvant surgery

Total

40

119

27 13

85 34

59 37-74

61 36-84

5 6 II 7 5 12 6 II 0 0 17

36 33 69 14 13 27 7 12 3 I 23

Table II. Summary ofpostoperative chemotherapy given to 79 patients after surgical treatment for SCLC No. of patients CAY One cycle Two cycles Three cycles Four cycles Five cycles Six cycles CA V alternating with etoposide and cisplatin Six cycles CA V plus etoposide Five cycles CA V (five cycles) plus methotrexate, CCNU, and procarbazine (18 mo) Etoposide and cisplatin Six cycles None

4 2

4 4 8 34

4 I 4

2

12

CA Y, Cyclophosphamide, doxorubicin (Adriamycin), and vincristine; CCNU, lomustine.

The review by Shields and associates' of the Veterans Administration Surgical Oncology Group (VASOG) studies of SCLC resulted in two important observations that have led to a reawakening in the interest of surgical treatment for certain patients with limited SCLC. Shields and colleagues reported that patient survival was significantly influenced by TNM staging and that within the

Table III. Preoperative chemotherapy given to 40 patients with SCLC identified prospectively as surgical candidates No. of patients CAY Two cycles Three cycles Four cycles Five cycles Six cycles CA V and etoposide Five cycles Etoposide and cisplatin (VPP) Four cycles Six cycles CA V alternating with VPP Two cycles Response (%) Complete Partial None

3 4

2 18 8

2 I

18 (45%) 20 (50%)

2

broad classic definition of limited disease each TNM stage had a survival pattern that was more favorable than the next advanced stage. Patients with T1 NOtumors had a projected 5-year survival rate of 60%, and this was significantly better than that of other categories (p < 0.00 I). Second, there was a suggestion from the early VASOG adjuvant chemotherapy trials that survival could be prolonged by chemotherapy after complete surgical resection. With current chemotherapy combinations, response is achieved in 80% of patients with SCLC or more.v" Despite this,long-term survival is disappointing, with less than 15% of patients with limited disease alive at 2 years in most series. For that reason, interest has turned, once again, to surgical treatment to determine whether combined modality therapy with aggressive chemotherapy and surgical resection could result in improved survival and a greater potential for cure. For the past 10 years, most patients with SCLC who had undergone initial surgical resection have been treated with adjuvant chemotherapy applied in the same intensive combinations that are standard for patients not treated surgically. In addition, we prospectively identified a cohort of patients with limited local disease without bulky mediastinal node involvement, supraclavicular adenopathy, or pleural effusion, preoperatively believed to have stage II or lIla disease, who were then treated with induction chemotherapy followed by surgical resection. We report here the results of our combined modality treatment program in 119 patients.

Volume 101 Number 3 March 1991

Limited SCLC 3 8 7

Table IV. Summary of surgical procedures

Procedure Lobectomy Pneumonectomy Wedge resection No resection Residual disease None Microscopic Gross Complications

Surgery first

Adjuvant surgery

Total

59 18 2 0

27 8 0 5

86 36 2 5

73 6 0 2*

36 0 4 6t

109 6 4 8

*Bronchopleural fistula and prolonged air leak. tSevere bronchospasm, pulmonary edema, prolonged ventilation, prolonged atelectesis, and arrhythmia (n = 2).

Table V. Postoperative TNM staging for 119 patients who had surgical treatment for limited SCLC Stage

Surgery first

Adjuvant surgery

Total

15 13 28 8 19 27 6 II 3 I 3 24

4 3 7 6 3 9 7 II 0 0 6 25

19 16 35 14 22 36 13 22 3 I 9 48

Tl NO T2NO Total stage I Tl NI T2NI Total stage 2 TIN2 T2~2

DNO DNI T3N2 Total stage 3

Patients and methods One hundred nineteen patients with limited SCLC underwent surgical resection at six University of Toronto teaching hospitals between 1977 and 1987.To confirm limited disease, we conducted staging procedures for all patients. These included a complete blood count with white blood cell differential and platelet counts, serum electrolytes, calcium, liver function tests (including serum glutamic oxaloacetic transaminase, alkaline phosphatase, and bilirubin), and renal function tests (including blood urea nitrogen and creatinine), radionuclide scanning of the bone, radionuclide, ultrasound, or computed tomographic (CT) scanning of the abdomen, radionuclide or CT scanning of the brain, and a bone marrow aspiration and trefine biopsy.Surgical staging was undertaken for all patients, and both clinical and surgical pathologic stages have been reported according to the new International Staging System for Lung Cancer," Seventy-nine patients received surgical resection as their first-line therapy, with adjuvant chemotherapy and radiotherapy being administered postoperatively. These patients underwent operation as their first treatment rather than chemotherapy because a diagnosis of malignancy could not be confirmed

Table VI. A comparison ofpreoperative and postoperative pathology for 119 patients having surgical treatment for SCLC Adjuvant surgery

Total

No tumor SCLC only None-SCLC Mixed

Preoperative 26 23 25 5

0 37 0 3

26 60 25 8

No tumor SCLC only Non-SCLC Mixed

Postoperative 0 65 0 14

4 30 3 3

4 95 3 17

Pathology

Surgery first

Table VII. Site offirst relapse for 119 patients undergoing surgical treatment for limited SCLC

Patients alive Patients deceased Site of relapse None (NED) Primary only Primary plus Brain Bone Liver Other

Surgery first

Adjuvant surgery

Total

32 47

20 20

.52 67

9 4 5 12 9 13 13

2 6 2 4

11 10 7 16 13 21 18

4 8 5

NED, No evidence of disease.

by preoperative investigations (26 patients), non-SCLC disease was identified (30 patients), or clinical staging suggested operable stage I or II lesions (23 patients). Forty patients received induction chemotherapy followed by adjuvant surgical treatment and radiotherapy. Because patients were treated over a IOeyear span, chemotherapy was not uniform. Most patients received cyclophosphamide, I gm/m2, doxorubicin, 50 mg/rrr', and vincristine, 2 mg, all administered intravenously every 3 weeks, usually for six cycles (CAV). More recently, some patients received etoposide (VP-16), 100 mg/m 2, and cisplatin, 25 mg/m', both administered intravenously each day for 3 days and repeated at 3- to 4week intervals (VPP). Five patients had CAV alternating with VPP, two patients had CA V plus VP-16, and four patients had CAV followed by maintenance therapy for 1 year with methotrexate, lomustine (CCNU), and procarbazine. Postoperative radiotherapy was administered to the tumor bed and mediastinum after completion of both surgical therapy and chemotherapy for most patients with stage II or III disease. The total dose ranged from 2500 cGy in 10 fractions to 3500 cGy in 20 fractions. Prophylactic cranial irradiation was given to the whole brain as 2000 cGy in five fractions at the completion of treatment.

The Journal of

388

Shepherd et al.

Thoracic and Cardiovascular

Suroerv

110 100 90

ro>

.~

:J (J)

'0 .~

:0 as

.0

e

Il.

80

Median Survival: 111 weeks

70 60 50 40 30 20 10 0

0

2 3 Time in Years

4

5

Fig. 1. Survival of 119 patients with SCLC treated with combined modality therapy including surgery.

All patients have had follow up for at least I year, and survival time has been calculated from the date of diagnosis until the date of death or most recent follow-up. Actuarial survival curves were prepared by means of the Kaplan-Meier method," and survival data were compared by the two-tailed generalized Wilcoxon-Gehan test. 8

Results The characteristics of the 119 patients are shown in Table I. There were 85 male and 34 female patients; the median age was 61 years (range 36 to 84). Pretreatment clinical staging revealed 69 patients with stage I (NO) disease, 27 patients with stage II (Nl) disease, and 23 patients with stage III (N2) disease. Preoperative pathologic or cytologic examination revealed pure SCLC in 60 patients. Twenty-five patients were thought initially to have non-SCLC tumors, and mixed SCLC and non-SCLC disease was found in eight. For 26 patients, malignancy could not be confirmed by any of the preoperative diagnostic procedures. Twelve patients who underwent an initial operation did not receive any adjuvant treatment, one because of advanced age and another because of severe alcoholic hepatitis and liver failure postoperatively. The 10 remaining patients all had stage I tumors and were not referred for chemotherapy at the discretion of their surgeons. The chemotherapy protocols and the number of cycles administered to the remaining 67 patients are shown in Table II. The preoperative chemotherapy given to 40 patients is shown in Table III. After induction chemotherapy, 18 patients (45%) had a complete clinical response and 20 (50%) a partial response. Only two patients had no response to their preoperative treatment.

The surgical procedures for the two study groups are shown in Table IV. Of the 79 patients undergoing initial surgical resection, 18 underwent complete pneumonectomy, 59 required only a lobectomy, and two patients had a wedge resections. All had complete lymph node staging. The operation was believed to have been complete for 73 patients (including the two who had only wedge resections). No patient had gross residual disease, but six had microscopic residual disease (diseased resection margins). Significant surgical complications included a prolonged air leak and a bronchopleural fistula in one patient each, but there were no surgical deaths. Among the 40 patients who underwent thoracotomy after induction chemotherapy, eight required a pneumonectomy and 27 a lobectomy (Table IV). All had radicallymph node dissection. In five patients resection was not done at thoracotomy. Four were believed to have unresectable disease, and in one patient no residual tumor could be identified pathologically, and the surgeon was unwilling to subject the patient to a complete pneumonectomy. Postoperative complications in this cohort included severe bronchospasm, prolonged atelectasis, and pulmonary edema in one patient each, supraventricular cardiac arrhythmias in two patients, and the need for assisted ventilation for 6 weeks in one patient. There were no postoperative deaths. The postoperative pathologic TNM staging is shown in Table V. Only 35 patients were found to have stage I (NO) disease. Thirty-six patients had stage II disease, and 48 patients had stage IlIa disease. No residual tumor could be identified pathologically in four patients (10%) who had undergone neoadjuvant chemotherapy (Table VI).

Volume 101

Limited SCLC 3 8 9

Number 3 March 1991

110 100 90

~

80

'0

60

::J

C/)

I

70

50

40 30 20

10

==35) - - > p = 0;001 > _0001 - - > p - 0958 P - . =36) 48) - -.

Stage I (n Stage II (n Stage III (n

°0

3

2

4

5

Time in Years

Fig. 2. A comparison of survival by pathologic stage for 119 patients who underwent surgery for limited SCLC.

110 100

90 (ij

>

.~

::J

C/)

'0 ~

:c III

e a.

.0

80 70

Stage I (n = 69) - Stage II (n = 27) -----. Stage III (n = 23)··············

60 50 40 30

20 10

°0

5

Time in Years

Fig. 3. A comparison of survival by pretreatment clinical stage for 119 patients who underwent surgery for limited SCLC.

Ninety-fivepatients had pure SCLC and 17 patients had combined SCLC and non-SCLC. Three patients had pure non-SCLC at resection. All of these patients had had SCLC pretreatment, with an excellent response to chemotherapy. We believe they had mixed tumors with complete disappearance of the small cell component. After the completion of chemotherapy and the operation, 68 patients received radiation to the tumor bed and mediastinum, and 85 patients received prophylactic cranial irradiation. The overall survival of the 119 patients is shown in Fig. 1. The median survival is 111 weeks, and the

projected 5-year survival rate is 39%. Survival by pathologic stage is shown in Fig. 2. It can be seen that patients with stage I (NO) disease had significantly longer survival (median not yet reached) than those with stage II or stage III disease (median survival 82 and 83 weeks, respectively). It is of interest, though, that a similar statistically significant difference could not be shown when survival was analyzed according to preoperative clinical stage (Fig. 3). No difference in survival could be detected between patients undergoing an initial operation and those having the operation after induction chemotherapy (Fig. 4).

The Journal of Thoracic and Cardiovascular

3 90 Shepherd et al.

Surgery

110 100 90 Cii >

80

(j)

~

70

'0

60

.~

.~ :0 <1l .D

0

It

50

=

Surgery First (n 79) ••••••.••. ) Surgery After Chemotherapy (n = 40) _ _

'"

'!._·~·:·····";,

u

p = 0.756

••

.......·.. l ...

40

!..••••••••••••••

L

~

.

30 20 10 00

2

3

4

5

Time in Years

Fig. 4. A comparison of the survival of patients with SCLC who had surgery first or surgery after chemotherapy.

Fifty-two patients remain alive and 67 patients have died (Table VII). Eleven patients (9.2%) died of nonmalignantcauseswithoutevidence of recurrentdisease. Only 10patients(8.3%)had a relapseat the primarysitealone. This includes the four patients with unresectable disease that progressed locally. Seven other patients (5.9%) had a relapse concurrently at the primary and distant sites. Onlyoneof the patients withmicroscopic residualdisease remainsaliveand diseasefree;but, of interest,relapsewas systemic rather than localfor all patients.Sevenof the 12 patients who did not have postoperative chemotherapy remain alive(twowith recurrence)6 to 48+ monthsafter diagnosis. One died with recurrent disease and the rest died of nonmalignantcauses. Discussion

The initial optimism engendered by the success of combination chemotherapy in the 1970s has been tempered in the 1980s by the observation that cure still remains an elusive goal for the majority of patients with SCLC. Despitethe introductionof newchemotherapeutic agents and innovative manipulations of drug doseand schedule, there has been no significant improvement in either overallsurvival or cure rate overthe past 10 years, and lessthan 20%of patients with limiteddisease survive more than 2 years. An analysis of the site of first relapse demonstrates that, even for patients who have achieved clinical complete response, the primary tumor bed and hilar or mediastinallymph nodeareas are the mostfrequentsinglesites of failure." Despite the addition of thoracic irradiation,

relapsecontinues to occurat the primarysite in onequarter to one third of patients.10, II In 1983,wel 2 reportedour early experience with combined modalitytherapy including surgeryand notedthat relapse at the primary site occurred in only two of 35 patients who had undergone surgical resection. Similar resultswere reportedby Comisand hiscolleagues, 13 who identified no local relapses after surgical treatment in their small feasibility study of 22 patients. . It is recognized that surgery alone is inadequate therapy for the majorityof patientswithSCLC. Nonetheless, almost all surgical seriesfrom the pre-combination chemotherapy era reported a small percentage of patients who achieved 5-year survival with surgical treatment alone." This observation led Shields and associates' to review the VASOG resultsof surgicaltherapy for SCLC in an attempt to identify a subgroup of patients with limited disease who might benefit from surgical resection. They made the important observation that the TNM staging, which has long been recognized to have prognostic importance for non-SCLC, could also define prognostic subgroups for patients undergoing surgeryfor SCLC. Perhapsofevengreater importancewasShields'observationthat there wasa suggestion of prolongation of survival in the patients who had received adjuvant chemotherapy in these early VASOG trials. There were insufficient patientsin thesetrialsfor the resultsto achieve statistical significance, and the chemotherapy they received would certainly be considered inadequate by today's standards. Nonetheless, these optimistic results,

Volume 101 Number 3 March 1991

and the suggestion of better control at the primary site after operation, provided the encouragement necessary for our groupand others to undertake prospective studies of combined modality therapy including surgery in patients with limited SCLC. With the current standard chemotherapy regimens, complete clinicalremission can beachieved in almost50% of patients with limited disease. On the basis of the suggestion of prolonged survival in patients receiving singleagent treatment after surgicaltreatment in the VASOG studies, we initiatedour studiesof more aggressive adjuvantcombination chemotherapyfor all patientsundergoinginitialoperations for SCLC with the expectationthat this treatment would improve overall survival and increase the frequency of cure. Our early promising results with this treatment'? led us to initiate a second study of surgery as adjuvant therapy in more advanced limited disease (clinical stagesII and IlIa) after induction of remission with chemotherapy. In this report we have reviewed our experience from bothofthesetrials.The mediansurvival of 111weeks and projected 5-yearsurvival rate of 39%are both encouragingand are clearlysuperiorto the survival seenin patients whohave had an operationwithout chemotherapy. I, 2, 14 Althoughwe speculatethat the favorable survival in our series is in part due to surgery, it is not possible to state this ~ith certainty.The survival of patients in single-arm trials of surgery cannot be compared directly with trials of standard chemotherapyand radiotherapy, becauseall surgical seriesrepresenta selectsubgroupof limiteddisease. Patients with known adverse prognostic factors, such as supraclavicular adenopathy, bulky mediastinal involvement, and pleuraleffusions, havebeen specifically excluded. Our analysis of survival by final pathologic stage supports Shields' earlier observation that the TNM staging systemis of prognostic importance in SCLC, as well as non-SCLC. The median survival of our patients with stage I disease has not yet been reached. Although the median survival rates of patients with stage II and III disease wereconsiderably lessthan thoseof patients with stageI disease (p = 0.001),their projected5-yearsurvival rates of 30%and 28%must also be viewed as encouraging, In fact, they do not appear to be significantly different from the survival rates of patients with a similar stage of non-SCLC after operation. Similar results have been reported by other groups, bothfor patientsundergoing an initialoperationfollowed by chemotherapy'> and for those undergoing adjuvant surgery after inductionof remission by chemotherapy.l" Karrer and associates'? reported a 78% 2-year survival rate after surgery followed by chemotherapy for 47

Limited SCLC 391

patients without any nodal involvement. When only N 1 nodeswere involved, 64% of 44 patients were alive at 2 years,comparedwith only 34%of 29 patients with mediastinal (N2) node involvement. SCLC poses specialdifficulties with respect to clinical staging. Most lesions are central, and the differentiation between NOand Nl may be difficult, evenwith CT scanning.Furthermore,micrometastaticdiseasewithinlymph nodesiscommonand occultmediastinalinvolvement frequently can be missed even if mediastinoscopy is performed. We l 8 have reported previously that only 35%of clinicalTNM stagingcorrelated with the final pathologicstaging in our adjuvant surgicaltrial. For patients who have had preoperative or "neoadjuvant" chemotherapy, pathologic staging at thoracotomyalmost always results in a more advanced stage.'? In this study, survival was significantly longer for patients with pathologic stage I disease than for those with pathologic stages II and III. On the other hand, though, no survival differences between stages could be demonstrated for patients in whomstagingwasdoneclinically. We wishto emphasize, therefore, that in prospective randomizedtrials of neoadjuvant chemotherapyfollowed by surgicaltherapy versus standard chemotherapy and radiation, the survival of individual TNM subgroups between the surgical and medicaltreatment arms shouldnot becompared,because one is staged pathologically and one clinically. Survival may be comparedonly as a whole, for TNM subgroups withina singlearm, or for clinically determined preoperative TNM subgroups between arms. TNM stage, rather than "limited" or "extensive," is alsoof prognostic importancefor patients who are treated with chemotherapyand radiotherapywithoutsurgical treatment. In a retrospective analysis of 247 patients with limited SCLC treated in several University of Toronto Lung Oncology Group nonsurgical trials, we20 analyzed the importanceof clinical TNM staging. In a subgroup of patients in clinicalstages I and II without mediastinal node involvement, the projected 5-year survival rate was 23% versus 10% for the patients with mediastinal node involvement. It may evenbe advisable, therefore, to consider stratification by clinical TNM stage in all future studies of limited SCLC. The timingof surgicalintervention remainscontroversial.In our series, no survival difference wasseenbetween patients who underwent an initial operation followed by chemotherapy and those who underwent the operation after inductionof remission by chemotherapy. It should be noted, though, that 70% of the patients who had surgery firsthad pathologic stage I and II tumors, compared with only40% whohad preoperative chemotherapy. We do not recommend surgical resection as the primary

392 Shepherd et al.

management for the majority of patients with SCLC. Patients with clinical stage I lesions frequently have micrometastatic nodal involvement, and true stage I (NO) disease probably exists for less than 20% of patients with resectable limited disease. For that reason, in other than small peripheral tumors, we would recommend initial treatment with combination chemotherapy followed by surgical resection for responding patients. In several single-arm prospective studies, approximately half of the patients identified prospectively as surgical candidates were eligible for resection after remission induction.t': 22 In our own prospective study, 79% of patients were eligible for resection, although only 54% of patients actually underwent resection because of randomization to the nonsurgical arm of another study protocol. 22 The objective of surgical resection in the therapy of SCLC is to achieve better control at the primary site than has been possible with chemotherapy and local radiotherapy. The results of our pilot study suggested that this was the case in patients in clinical stages I and II, both for patients treated initially with operation or those undergoing operation as adjuvant treatment. The results of this trial are similar. Only 8.3% of patients had a relapse or progression at the primary site alone, and this included the four patients who had unresectable disease at thoracotomy. In 5.9% relapse occurred concurrently at the primary and distant sites. Therefore disease was controlled in the thorax in 86% of patients by this combined modality approach. In summary, combined modality therapy including surgical treatment is feasible. The patients who appear to benefit most from such combined modality treatment are those with stage I (NO) tumors, and nodal involvement at any level appears to confer a poorer prognosis. The final role of surgery must await the results of prospective randomized trials currently underway by the large cooperative study groups.P REFERENCES 1. Mountain CF. Clinical biology of small cell lung cancer: relationship to surgical therapy. Semin Oncol 1978;5: 272-9. 2. Fox W, Scadding JG. Medical Research Council comparative trial of surgery and radiotherapy for primary treatment of small-eelled or oat-celled carcinoma of bronchus: ten year follow-up. Lancet 1973;2:63-5. 3. Livingston RB. Current chemotherapy of small cell lung cancer. Chest 1986;89:2585-625. 4. Morstyn G, Ihde DC, Lichter RS, et al. Small cell lung cancer 1973-1983:Early progress and recent obstacles. Int J Radiat Oncol Bioi Phys 1984;10:515-39. 5. Shields TW, Higgins GA, Matthews MJ, Keehn RJ.

The Journal of Thoracic and Cardiovascular Surgery

Surgical resectionin the management of small cellcarcinoma of the lung. J THORAC CARDIOVASC SURG 1982;84: 481-8. 6. Mountain CF. A new International Staging System for Lung Cancer. Chest 1986;89:2235-55. 7. Kaplan EL, Meier P. Non-parametric estimation from incompleteobservations. JAm StatAssoc 1958;53:457-81. 8. Gehan EA. A generalized Wilcoxon test for comparing arbitrarily singly-sensored samples. Biometrika 1965;52: 203-23. 9. Elliot JA, Osterlind K, Hirsch FR, Hansen HH. Metastatic patterns in small-eelliung cancer: correlation of autopsy findings with clinical parameters in 537 patients. J Clin Oncol 1987;5:246-54. 10. Perez CA, Einhorn L, Oldham RK, et al. Randomized trial of radiotherapy to the thorax in limited small-cellcarcinoma of the lung treated with multi-agent chemotherapy and elective brain irradiation: a preliminary report. J Clin Oncol 1984;2:1200-8. 11. Kies MS, Myra JG, Crowley Ll, et al. Multi-modality therapy for limited small-cell lung cancer: a randomized study of induction combination chemotherapy with or without thoracic irradiation in complete responders and with wide-field versus reduced-field radiation in partial responders; a Southwest Oncology Group Study. J Clin Oncol 1987;5:592-600. 12. Shepherd FA, Ginsberg RJ, Evans WK, et al. Reductionin local recurrence and improved survival in surgically treated patients with small cell lung cancer. J THORAC CARDIaVASC SURG 1983;89:498-506. 13. Comis R, Meyer J, Ginsberg S, et al. The current resultsof chemotherapy (CTH) plus adjuvant surgery (AS) in limited small cell anaplastic lung cancer (SCALC). Proc Am Soc Clin OncoI1982;1:147 (Abstract C-571). 14. Shields TW. Surgery of small cell lung cancer. Chest 1986;89:264s-7s. 15. Meyer JA, Gullo Ll, Ikins PM, et al. Adverse prognostic effect of N2 disease in treated small cell carcinoma of the lung. J THORAC CARDIOVASC SURG 1984;88:495-501. 16. Comis R, Meyer J, Ginsberg S, et al. The impact ofTNM stage on results with chemotherapy (CT) and adjuvant surgery (AS) in small cell lung cancer (SCLC). Proc Am Soc Clin OncoI1984;3:226 (Abstract C-884). 17. Karrer K, Denck H, Karnicka-Mlofkowska H, et al. Multi-modality treatment after surgery for cure of small cell lung cancer (SCLC). Lung Cancer 1988;4:AI53 (Abstract 8.1.01). 18. Shepherd FA, Evans WK, Feld R, et al. Adjuvant chemotherapy following surgical resection for small-cellcarcinoma of the lung. J Clin OncoI1988;6:832-8. 19. Shepherd FA, Ginsberg RJ, Patterson GA, Evans WK, Feld R. A prospectivestudy of adjuvant surgical resection after chemotherapy for limited small cell lung cancer: a University of Toronto Lung Oncology Group study. J THORAC CARDIOVASC SURG 1989;97:177-86. 20. Shepherd, FA, Ginserg R, Evans WK, et al. "Very limited"

Volume 101 Number 3 March 1991

small cell lung cancer (SCLC): results of non-surgical treatment. Proc Am Soc Clin Oncol 1984;3:223 (Abstract C-870). 21. Baker RR, Ettinger DS, Ruckdeschel JD, eta!. The role of surgery in the management of selected patients with small cell carcinoma of the lung. J Clin Oncol 1987;5:696-702. 22. Lad T. The role of surgery in small cell lung cancer (SCLC): an intergroup prospective randomized tria!' Lung Cancer 1988;4:A83 (Abstract 5.13).

Discussion Dr. John A. Meyer (Syracuse. NiY]. The abstract states: "Seven of the 12 patients who had no adjuvant chemotherapy remain alive at 6 to 48+ months." I think Dr. Shepherd would agree that 6 months is minimal follow-up and that more of the patients who did not receive adjuvant treatment will surely have a relapse. Some readers may carry away the message that patients with SCLC do just about as well after surgical resection alone as patients do after surgery for non-SCLC. Therefore, it may be thought, adjuvant chemotherapy probably is not necessary, and anyway, it is a bother. The disease is most notable for early and widespread dissemination. I ask Dr. Shepherd whether, in the present state of our knowledge, we can afford to dispense with adjuvant chemotherapy after surgical resection of SCLC irrespective of clinical stage. . Dr. Shepherd. Almost all patients who remain alive and who do not have adjuvant therapy have early stage TI NO disease. Of the seven who are alive, one of them has had a relapse and is in remission after chemotherapy. I think it would be unwise to recommend no further chemotherapy after surgical treatment for SCLC. That would be a retrograde step. All of the studies before the introduction of chemotherapy really suggest that it is virtually impossible to cure SCLC with surgery alone. What should be done about the small, isolated, peripheral TI NO nodule? Do those patients really need chemotherapy? I am not sure that anyone will ever be able to answer that question because there are so few patients who present in that way with SCLC. Certainly, some of the patients in our group who have done well without chemotherapy had small TI NO primary tumors. However, they represent such a minority of patients with SCLC that it would beunfair to generalize about not adding chemotherapy from an assessment of that subgroup of patients. Dr. Thomas W. Shields (Chicago. Ill.). I'd like to strongly support Dr. Shepherd's and the Toronto group's conclusions. I've been associated with the International Society of Chemotherapy Small Cell program in Europe and in Asia. To date there are 142 patients who have had an initial operation followed by randomized chemotherapy and an additional 16 patients from the Innsbrook group. We have a total of 162 patients who have had an initial operation followed by adequate chemotherapy. Our results at 4 years are exactly the same as Dr. Shepherd's. The 4-year survival rate for patients with NO disease is 61%, and patients with both N I and N2 disease have a 35%survival rate. It is distressing that the prognosis for Nl and N2 disease is equally poor. I am not sure of the reason for that. Among the things that we have learned are that about a third

Limited SCLC 3 9 3

of the patients need to have their stage of disease upgraded at the time of operation, despite relatively intensive preoperative evaluation. Mediastinoscopy has a significant degree of error in analyzing the mediastinal nodes in patients with SCLC. However, I am strongly supportive of the need for postoperative chemotherapy for the patient with a tumor of unknown cell type that turns out to be SCLC. If the diagnosis is made preoperatively, I think only those patients with stage I or stage II disease should undergo exploration and resection, if possible, and receive chemotherapy. Patients with N2 disease should not be operated on electively. However, chemotherapy has permitted a small subset of patients (maybe around 5% of patients small cell cancer, with SCLC) to be operated on with satisfactory long-term survival, and we should continue to operate. As you know from Dr. Shepherd's group, anywhere from 5% to 14% of these patients will have a non-small cell component to their disease. If the patient is given preoperative chemotherapy, this component rises almost to 30%. Thus that subset harbors a significant tumor burden that is not sensitive to the chemotherapy. In these early stages, the addition of surgery plus chemotherapy, either before or after the operation, is an ideal way to handle this very small but select group of patients. Dr. Ginsberg. I would like to emphasize a few points. The survival rate for TIN I and T2 N I adenocarcinoma is only 30% to 40%. Thus these groups of patients with NO and N I small cell disease fare almost as well as those with non-SCLC. Another thing Dr. Shephard did not mention concerns patients who were treated without surgery. Although recurrence was frequent in the primary site, the survival rate in the N I and N2 groups was not too dissimilar to the survival rate we are seeing with surgery. We really do not know what medical oncologists can do with this very limited disease, because we do not have any kind of data base for that. When Dr. Shields says that surgery should not be used in N2 disease, he is presuming that our medical oncologists can treat the patients just as well. I'm not saying they can't, but we don't know that. There is an unpublished trial from the Southwest Oncology Group that is claiming as good results as those that we have presented for stage II and stage III disease. If that is true, then it is possible that surgery should not be recommended in the more advanced stages of resectable disease. Until we know that, surgery should continue to play an important role in early stage SCLC. Dr. Shepherd. The rationale for surgery in SCLC is basically to improve control at the primary site of the tumor. We know that approximately one third of patients with limited disease will have a relapse only at the primary site and half of them at some time will have a relapse at the primary site. That is why we focused on surgery in an attempt to improve control at the primary site. Indeed, in our series we were able to reduce relapse there to less than 10% of resectable patients. Nonsurgical studies are also focusing on improving control at the primary site. Routine radiotherapy has reduced relapse rate at the primary site, but on the whole it has not improved survival over the long term. It may be, though, that with more aggressive radiotherapy, using fractionated dosing schemes, or the use of radiotherapy given concurrently with chemotherapy, it may be possible to achieve the same results that we have achieved with surgery. Only randomized trials will tell.