- Email: [email protected]
The role of adjuvant therapy after resection of T1 N1 M0 and T2 N1 M0 non–small cell lung cancer
J
THoRAc CARDIOVASC SURG
91:344-349, 1986
The role of adjuvant therapy after resection of Tl Nl MO and T2 Nl MO non-small cell lung cancer Thirty-four cOiRCutive patients with non-small cell lung cancer plus Nl nodal metastases (eight with Tl Nl MO and 26 with T2 Nl MO) were retrospectively reviewed. Nineteen had adenocarcinoma, 11 had squamous disease, and four had large cell carcinoma. Eleven patients had surgical resection alone (32.3% ~ with a median survival of 13 months.Sevenpatients (20.6%) had resectionfollowed by radiation therapy, with a median survival of 19.2 months. Sixteen patients (47.1%) had resection followed by radiation therapy and chemotherapy, consisting of cyclophosphamide, doxorubicin, methotrexate, and procarbazine. Median survival for the latter group was 45.5 months, significantly greater than for those treated with resection alone (p < 0.005). We did not observe any relationship between survival and age, cell type, number or location of diseased hilar nodes, distance of tumor from the resected bronchial margin, tumor size, the presence or absence of visceral pleural involvement, or the type of resection performed. Resection in combinationwith adjuvant radiation therapy and chemotherapyoffers improved median survival over resection alone in patients with Tl Nl MO and T2 Nl MO non-small ceUlung cancer.
Mark K. Ferguson, M.D. (by invitation), Alex G. Little, M.D. (by invitation), Harvey M. Golomb, M.D. (by invitation), Philip C. Hoffman, M.D. (by invitation), Tom R. DeMeester, M.D.,* Roy Beveridge, M.D. (by invitation), and David B. Skinner, M.D., Chicago, Ill.
Non-small cell lung cancer accompanied by N 1 nodal metastases forms the smallest major subgroup of lung cancer. Nevertheless, it is important, as controversies still exist over its staging, appropriate treatment, and prognosis. An initial review of our experience with T1 Nl MO and T2 Nl MO non-small cell lung cancer at the University of Chicago was presented in 1982 and concluded that surgical resection combined with postoperative chemotherapy, radiation therapy, or both yielded improved survival over surgical resection alone. I Since that report another series has been published that supports the use of operation as a single treatment From the Departments of Surgery and Medicine, University of Chicago Hospitals and Clinics, Chicago, Ill. Read at the Sixty-fifth Annual Meeting of The American Association for Thoracic Surgery, New Orleans, La., April 29-May I, 1985. Address for reprints: Mark K. Ferguson, M.D., The University of Chicago, Department of Surgery, Box 255, 5841 South Maryland Ave., Chicago, Ill. 60637. *Current address: Department of Surgery, Creighton University, 601 North 30th St., Suite 3740, Omaha, Neb.
344
modality for patients with hilar nodal metastases.' Disagreement thus continues over whether resection in combination with chemotherapy, radiation therapy, or both offers improved survival over operation alone in patients with lung cancer complicated by Nl nodal metastases. For these reasons we have updated our data to reassess whether adjuvant therapy is beneficial in treatment of these patients after surgical resection.
Patients and methods From June, 1974, to December, 1984,34 consecutive previously untreated patients with T1 Nl MOor T2 Nl MO non-small cell lung cancer were seen by the University of Chicago Chest Oncology Group. There were 24 men and 10 women aged 38 to 80 years (median 67 years). Initial evaluation included history and physical examination, blood counts, chemistries, and pulmonary function tests. A whole-body tomographic gallium scan was done in 32 of 34 patients, and in recent years computed tomography of the thorax and upper part of the abdomen including the adrenal glands was performed routinely. When indicated by these tests,
Volume 91 Number 3 March. 1986
Non-small cell lung cancer 3 4 5
Table I. Distribution by operation and adjuvant therapy Pneumonectomy
Treatment Resection only Resection + RT Resection + RT
3
+ chemo
Total
100 90 80
Toral >.
5 10
8 2 6
II 7 16
:0
18
16
34
to
11l .0
60
ll:
50
>
40
0
.~
Legend: RT. Radiation therapy. Cherno, Chemotherapy.
:>
rJ)
30 20
Table II. Distribution by cell type and TNM classification
10
Histologic type
Tl Nl MO
T2 Nl MO
I~
Adenocarcinoma Squamous cell carcinoma Large cell carcinoma
7
12
19
I
10
II
044
Total
8
26
70
34
mediastinoscopy was performed to exclude mediastinal nodal involvement. Patients had liver-spleen scan, bone scan, or brain scan as indicated or to confirm abnormalities on gallium scan. All patients underwent thoracotomy with pulmonary resection and mediastinal nodal sampling. Eighteen patients had lobectomy or bilobectomy and 16 had pneumonectomy. Before 1976 no further treatment was provided. Since then we have recommended adjuvant postoperative radiation therapy (refused by three) and chemotherapy (refused by nine). Eleven patients underwent surgical resection alone (32.3%), seven patients underwent resection plus radiation therapy (20.6%), and 16 patients had resection followed by radiation therapy and chemotherapy (47.1%) (Table I). Radiation therapy consisted of 3,000 rads in 15 fractions to the primary tumor site, ipsilateral hilum, and mediastinum. The chemotherapy regimen included cyclophosphamide, 300 mg/rn! intravenously, days 1 and 8; doxorubicin, 20 mg/rn' intravenously, days 1 and 8; methotrexate, 15 mg/rrr' intravenously, days 1 and 8; and procarbazine, 100 mg/m' per os, days 3 through 13 (CAMP). Cycles of CAMP were repeated every 28 days for 1 year or until recurrence was diagnosed. Patients were routinely followed up in our Chest Oncology Clinic with chest x-ray films and periodic gallium scan and/or computed tomographic scan of the thorax. Tumors were evaluated according to size, pleural involvement, histologic type, and location. Nl nodes were dissected from the resected specimen and assigned to one of three levels: intrapulmonary, peribronchial, or hilar. Postoperative staging was performed according to
0 2
3
4
5
6
Years From Diagnosis
Fig. 1. Survival curve for all patients with T1 Nl MO and T2 N 1 MO non-small cell lung cancer. Vertical bars indicate length of follow-up for patients currently alive.
the TNM system.' Survival was calculated from the date of operation until death or until March, 1985. The mean length of follow-up for all patients is 35.8 months with the longest follow-up being 10 years. Statistical analyses were performed by the algorithm of Lee and Desu' and Kaplan-Meier' life-table techniques.
Results Thirty-four consecutive patients with T1 Nl MO or T2 Nl MO lung cancer were seen and all were evaluable. Fourteen of these patients were new since February, 1981, the closure date of our previous report. Nineteen patients had adenocarcinoma (55.9%), 11 patients had squamous carcinoma (32.3%), and the remaining four patients had large cell carcinoma (11.8%). Eight patients had T1 Nl MO tumors and 26 patients had T2 Nl MOdisease (Table 11).There was a predominance of isolated hilar nodal metastases, and three patients had involvement of nodes from more than one level (Table III). Median overall survival for patients was 19.7 months, and 5 year survival rate was 30.2% (Fig. 1). Median survival by cell type was 19.6 months for adenocarcinoma, 20.5 months for squamous cell carcinoma, and was not reached at 18 months for large cell carcinoma (no statistical differences) (Fig. 2). There were 23 deaths with 11 patients alive at the time of reporting. Two patients died perioperatively, one of sepsis and the other of a cardiac arrhythmia. Fifteen patients died of progressive disease, two of other proved causes without known recurrence, and four of unknown causes. Survival according to treatment was examined.
The Journal of
346 Ferguson et al.
Thoracic and Cardiovascular Surgery
100 100 Adenocarcinoma
90
>-
:0
'"
.0 0
Q:
80
Squamous Cell Carcinoma
i
Large Cell Carcinoma
'--,
70
'"
.0 0
I
60
::J
en
a:
I
IL
50
-
>-
:0
I
,
lii >
I
<;; > 40
's ...
90
I L
_
1'---------
30 20
's ... ::J
en
80
I
Surgery + RT+ Chemo
I
Surgery+ RT
I I
I
L
70 60 50 40 30
I I L
Surgery
l
I
,
L'
I L
I
I
L_ I I
L,
20 10
10
0
0
I 3
I
2
i 4
I
5
I 6
Years From Diagnosis
Fig. 2. Survival curve for patients with T1 Nl MO and T2 N I MO non-small cell lung cancer according to cell type.
.
I
L_ ,
IL
_
2
3
5
4
6
Years From Diagnosis
Fig. 3. Survival curve for patients with TI NI MO and T2 N I MO non-small cell lung cancer according to treatment groups. Chemo, Chemotherapy. RT, Radiation therapy.
Table m. Locations of tumors and nodal metastases Incidence of nodal metastases
Primary tumor Location
No.
Intrapulmonary
Peribronchial
RUL RLL RML LUL LLL Total
34
Hilar
7
1
0
8 I 12 6
3
I
0
0
1
3
2
1
0
9 4
8
3
24
Indeterminate
6 4
3
Legend: RUL, RLL. and RML, Right upper, lower, and middle lobe. LUL and LLL, Left upper and lower lobe.
Patients undergoing resection followed by adjuvant radiation therapy and chemotherapy had an overall median survival of 45.5 months and a 5 year survival rate of 45.9%. This survival period was significantly longer than the median survival in patients who had surgical resection alone (13.0 months excluding two operative deaths, p < 0.005). Patients who underwent resection followed by radiation therapy had a median survival of 19.2 months and a 5 year survival rate of 28.6%, figures which did not differ significantly from either of the other treatment groups (Fig. 3). Patients who received chemotherapy had a median survival of 45.5 months compared to 14.3 months for patients whose treatment did not include chemotherapy (p < 0.005 excluding two operative deaths). We found no relationship between survival and T classification (Fig. 4), absolute tumor size, number or location of involved nodes, distance of tumor from resected bronchial margin, sex of the patient, presence or absence of visceral pleural involvement (Fig. 5), and type of resection performed (Fig. 6).
Fifteen patients had proved recurrences, only two of whom are alive 3 and 6 months after the diagnosis of recurrence. Initial recurrences were local in three patients, with a median time to recurrence of 11.5 months. Distant recurrences were found in 13 patients, with a median time to initial recurrence of 9.0 months. The most frequent site of distant recurrence was the brain, which was affected in six patients, four of whom had adenocarcinoma. Discussion
In a previous publication from this institution we' reported results of treatment of Tl Nl MO and T2 Nl MO non-small cell bronchogenic carcinoma and we highlighted the effects of a combined modality approach. In that report adjuvant radiation therapy and chemotherapy appeared to benefit patients by increasing the survival rate up to 75% at 5 years. This report updates that experience and expands the group originally reported from 20 to 34 patients. Although the previously excellent survival rates for the multimodality
Volume 91
Non-small cell lung cancer
Number 3 March, 1986
100
100 90 80
~
70
.0
60
0..
50
>
40
:::J
30
:ctil
I
>.
:ctil
.~
en
Pleura Negative
I I
Pleura Positive
80
L,
70
I I I
I
L_,
.0
60
0..
50
I IL __
> 'S
40
I I
~
30
I
0
0
iii
L_
90
T2 Tumors
1__ -
---,
I
T 1 Tumors
-I
iii
en
20
20
10
10
347
0
0
2
3
4
5
2
6
Fig. 4. Survival curve for patients with TI NI MO and T2 NI MO non-small cell lung cancer according to T classification. Median survival for patients with T1 tumors was 20.0 months and for those with T2 lesions, was 20.8 months (p = NS).
treatment group have been tempered in this report by extended follow-up and an expanded patient pool, our conclusions regarding the benefit of adjuvant treatment in these patients remain the same. In 1983 Martini and others' reported results of treatment of 75 patients with non-small cell lung cancer complicated by Nl nodal metastases. Sixty-two of the 75 patients were treated by resection alone, and the combined overall survival rate was 49% at 5 years. These results are significantly better than our overall survival rate of 30% at 5 years and significantly better than the 5 year survival rates for Stage II lung cancer of 7%, 29%, and 31% reported from other institutions." The reasons for such a discrepancy in survival when comparing our patients with those of the Memorial group' are not immediately clear. Both groups had similar distributions of patient age, tumor stage, and type of surgical resection. However, the Memorial group appears to have had a higher incidence of both peripheral carcinoma and squamous cell carcinoma than occurred in our group, whereas their incidence of hilar nodal involvement was considerably lower than ours (37.3% versus 70.6%). In addition, although the value of extended mediastinal lymph node resection in prolonging survival may be unproved, there is little doubt that it can provide improved pathologic staging. There is a high incidence of residual local disease after "curative" resection of early lung cancer, especially with squamous cell carcinomas." It is possible that the Memorial group performed more extensive mediastinal lymph node dissections, eliminating a number of otherwise undetecta-
4
3
6
5
Years From Diagnosis
Years From Diagnosis
Fig. 5. Survival curve for patients with T1 NI MO and T2 N I MO non-small cell lung cancer according to the presence or absence of visceral pleural involvement. Median survival for II patients with "positive" pleura was 19.7 months and for 22 with "negative" pleura, 20.0 months (p = NS). 100 90
>.
.stil .0 0
80 70 60 50
>
40
:::J
30
.~
en
L,
Pneumonectomy
I I I
0..
iii
Lobectomy
II
20
L~
I l
I
L I
L I
L_ I
~
------------,1
I
L
_
10 0
2
3
4
5
6
Years From Diagnosis
Fig. 6. Survival curve for patients with T1 N I MO and T2 N I MO non-small cell lung cancer according to the type of resection. Median survival for lobectomy/bilobectomy patients was 25.9 months and for pneumonectomy patients, 16.0 months (p = NS excluding two operative deaths).
ble cases of Stage III disease from their study group and improving their staging accuracy and the resultant overall survival rates. Finally, unrecognized differences probably exist in our patient populations. The incidence of local recurrence is low in our series. The presence of lymphatic spread, whether or not it is limited to hilar nodes, is usually an indication that local control of disease remains in doubt. For this reason we routinely irradiate the mediastinum and hilus postoperatively in patients with involved nodes. This approach is
The Journal of
348
Ferguson et al.
Thoracic and Cardiovascular Surgery
supported by others who report significant improvement in survival in patients with hilar and mediastinal lymph node metastases when radiation therapy is employed after adequate resection.'? The use of radiation as a single treatment modality, however, is not acceptable in patients who have resectable disease and are physiologically able to undergo operation." Although our local recurrence rate was low, there was a high rate of distant recurrence, primarily in the brain. Two thirds of those patients with metastatic disease to the brain had adenocarcinoma. These findings support previous data reported in patients with non-small cell lung cancer and emphasize the high incidence of cerebral metastases in patients with advanced stages of adenocarcinoma of the lung. This rate is as high as 25% in some series." 12 We interpret these data as supporting the study of prophylactic cranial irradiation in patients with advanced stages of adenocarcinoma. We have been investigating prophylactic cranial irradiation at this institution for the past 3 years. The high incidence of distant metastases underscores the need in these patients for systemic adjuvant therapy. We l3 previously reported that CAMP appeared to prolong survival in Stage III non-small cell bronchogenic carcinoma. Although we recognize the shortcomings of a retrospective nonrandomized review, the results reported herein nevertheless demonstrate an improvement in long-term survival in patients with non-small cell lung cancer and N I nodal metastases treated with resection, radiation therapy, and chemotherapy compared to patients in our series and those of others'" treated with resection alone. These data thus support the use of adjuvant therapy for Tl NI MO and T2 NI MO non-small cell lung cancer. Other reports identify a number of factors that impact on survival in patients with lung cancer, including cell type," proximity of the primary tumor to the hilum, and the presence of visceral pleural involvement.3 We were unable to confirm any of these as risk factors in the present study. The concept that more extended pulmonary resection is indicated for patients with hilar nodal metastases, specifically pneumonectomy rather than lobectomy, also cannot be supported. In patients with non-small cell lung cancer and hilar nodal metastases, we conclude: 1. Surgical therapy is the primary form of treatment. 2. Adjuvant postoperative therapy, particularly chemotherapy, improves survival. 3. The rate of recurrence is high, particularly in the brain. Prophylactic cranial irradiation should be admin-
istered to patients with advanced stages of adenocarcinoma of the lung in investigative programs. 4. Cell type, tumor size, number of nodes involved, involvement of visceral pleura, and type of resection have no proven effect on survival. We wish to acknowledge the assistance of Margaret Johnson in data analysis, Consuelo Skosey and Cynthia YoungStrickland in patient care, and Cathy Ward in manuscript preparation. REFERENCES Newman SB, OeMeester TR, Golomb HM, Hoffman PC, Little AG, Raghavan Y: Treatment of modified Stage II (Tl Nl MO, T2 Nl MO) non-small cell bronchogenic carcinoma. J THORAC CARDlOVASC SURG 86:180-185, 1983 2 Martini N, Flehinger BJ, Nagasaki F, Hart B: Prognostic significance of Nl disease in carcinoma of the lung. J THORAC CARDIOVASC SURG 86:646-653, 1983 3 American Joint Committee for Cancer Staging and EndResults Reporting. Manual for Staging of Lung Cancer, Chicago, 1979, Whiting Press 4 Lee E, Desu M: A computer program for comapring k samples with right-censored data. Comput Programs Biomed 2:315-321, 1972 5 Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 6 Kirsh MM, Rotman H, Argenta L, Bove E, Cimmino V, Tashian J, Ferguson P, Sloan H: Carcinoma of the lung. Results of treatment over ten years. Ann Thorac Surg 21:371-377, 1976 7 Vincent RG, Takita H, Lane WW, Gutierrez AC, Pickren JW: Surgical therapy of lung cancer. J THORAC CARDlOVASC SURG 71:581-591, 1976 8 Mountain CF: Assessment of the role of surgery for control of lung cancer. Ann Thorac Surg 24:365-373, • 1977 9 Matthews MJ, Kanhouwa S, Pickren J, Robinette D: Frequency of residual and metastatic tumor in patients undergoing curative surgical resection for lung cancer. Cancer Chemother Rep 4:63-67, 1973 10 Green N, Kurohara SS, George FW, Crews QE Jr: Postresection irradiation for primary lung cancer. Radiology 116:405-407, 1975 II Cooper JO, Pearson FG, Todd TRJ, Patterson GA, Ginsberg RJ, Basiuk J, Blair Y, Cass W: Radiotherapy alone for patients with operable carcinoma of the lung. Chest 87:289-292, 1985 12 Bitran JO, Golomb HM, OeMeester TR, Desser RK, Hoffman P, Cohen L, Green M, Sovik C: Combined modality therapy for Stage III non-small cell bronchogenic carcinoma (NSBC). Proc Am Soc Clin Oncol 21:446, 1980
Volume 91 Number 3 March, 1986
13 Bitran JD, Desser RK, DeMeester TR, Shapiro CM, Billings A, Rubenstein L, Evans R, Colman M, Rao Y, Griem M, Golomb HM: Combined modality therapy for Stage III MO non-oat cell bronchogenic carcinoma. Cancer Treat Rep 62:327-332, 1978 14 Shields TW, Vee J, Conn JH, Robinette CD: Relationship of cell type and lymph node metastasis to survival after resection of bronchial carcinoma. Ann Thorac Surg 20:501-510,1975
Discussion DR. E. CARMACK HOLMES Los Angeles, Calif
It is always gratifying to hear a report suggesting that adjuvant therapy is beneficial in patients with resected lung cancer, because there have been so many negative reports over the years. Although it is gratifying to hear such a report, I do have problems with data. It is difficult to draw conclusions based on 34 patients in a three-arm study. Statisticians tell us that to generate meaningful data, a two-arm study usually requires 60 patients in each arm. I would ask the authors to comment on this in their summation. I would like to share some data generated by the National Lung Cancer Study Group. One study is somewhat analogous to the one presented by Dr. Ferguson's group. Patients with Stage II and Stage III adenocarcinoma and large cell undifferentiated carcinoma were operated on, stratified, and then randomized to receive immunotherapy or a combination chemotherapy. At thoracotomy these patients underwent very careful intraoperative lymph node sampling for staging purposes. There is a significant difference in favor of the chemotherapy group. I would like to emphasize that this study includes 130 patients with 60 to 65 patients in each of the arms. Survival with disease, disease-free survival, and overall survival are all prolonged significantly. A second study is one that involves patients with squamous cell carcinoma of the lung. As you recall, the previous study evaluated only adenocarcinoma and large cell undifferentiated carcinoma. These are patients with Stage II and Stage III squamous carcinoma of the lung, most of whom had diseased mediastinal nodes. They were operated on, carefully staged intraoperatively, and then randomized to receive radiation therapy, 5,000 rads postoperatively, or no further treatment. There is absolutely no difference between the patients receiving postoperative irradiation and those undergoing operation alone. Of interest, however,is that the patients who did receive radiation therapy had a much lower incidence of local
Non-small cell lung cancer 3 49
recurrence than the patients receiving no further therapy (p = 0.001). Thus, postoperative radiation therapy in these patients does not prolong survival but does diminish local recurrence. I would like to pose two questions to the authors. First, there was- a dramatic difference in the survival of patients with adenocarcinoma and squamous carcinoma in the subset T2 N I that they presented. I find it difficult to accept lumping these two cell types into one study, as the authors did. Finally, I would again question the authors regarding conclusions that impact on reclassification and restaging of lung cancer, conclusions regarding the effect on prognosis of T2 N I disease and pleural involvement based on a series of only 34 patients. DR. FERGUSON(Cwsm~ I would like to thank Dr. Holmes for his insightful comments. As he points out, the number of patients that we have analyzed is relatively small. Unfortunately, patients with isolated hilar nodal metastases are uncommon, and it is difficult to generate large numbers of such patients for study. I would emphasize that in analyzing our various treatment groups, we found quite significant statistical differences despite the small numbers in each group. Thus, while it is nice to have an "ideal" number of patients in each study group, one cannot really ignore such significant differences when they exist. On the other hand, one cannot become too free in interpreting differences between small numbers of patients. While there is frequently said to be a difference in survival between patients with adenocarcinoma and those with squamous cell carcinoma, we were unable to demonstrate such a difference. Such variations in survival between cell types, when they exist, are actually much less significant than the variations in survival among stages. We group these cell types together because in the end we find there are more similarities between them than there are differences. Dr. Holmes has presented some interesting data generated by the Lung Cancer Study Group, which support many of the conclusionsfrom our paper. First, the use of adjuvant radiation therapy in patients with NI nodal metastases can limit the incidence of local recurrence. Second, systemic chemotherapy in patients with advanced stages of lung cancer can prolong survival when compared to results of surgical treatment alone. Finally, I should reemphasize the importance of complete intraoperative staging, which includes mediastinal nodal dissection. Only in this way can adequate pathologic staging be obtained for appropriate assignment to treatment groups and assessment of prognosis.
THoRAc CARDIOVASC SURG
91:344-349, 1986
The role of adjuvant therapy after resection of Tl Nl MO and T2 Nl MO non-small cell lung cancer Thirty-four cOiRCutive patients with non-small cell lung cancer plus Nl nodal metastases (eight with Tl Nl MO and 26 with T2 Nl MO) were retrospectively reviewed. Nineteen had adenocarcinoma, 11 had squamous disease, and four had large cell carcinoma. Eleven patients had surgical resection alone (32.3% ~ with a median survival of 13 months.Sevenpatients (20.6%) had resectionfollowed by radiation therapy, with a median survival of 19.2 months. Sixteen patients (47.1%) had resection followed by radiation therapy and chemotherapy, consisting of cyclophosphamide, doxorubicin, methotrexate, and procarbazine. Median survival for the latter group was 45.5 months, significantly greater than for those treated with resection alone (p < 0.005). We did not observe any relationship between survival and age, cell type, number or location of diseased hilar nodes, distance of tumor from the resected bronchial margin, tumor size, the presence or absence of visceral pleural involvement, or the type of resection performed. Resection in combinationwith adjuvant radiation therapy and chemotherapyoffers improved median survival over resection alone in patients with Tl Nl MO and T2 Nl MO non-small ceUlung cancer.
Mark K. Ferguson, M.D. (by invitation), Alex G. Little, M.D. (by invitation), Harvey M. Golomb, M.D. (by invitation), Philip C. Hoffman, M.D. (by invitation), Tom R. DeMeester, M.D.,* Roy Beveridge, M.D. (by invitation), and David B. Skinner, M.D., Chicago, Ill.
Non-small cell lung cancer accompanied by N 1 nodal metastases forms the smallest major subgroup of lung cancer. Nevertheless, it is important, as controversies still exist over its staging, appropriate treatment, and prognosis. An initial review of our experience with T1 Nl MO and T2 Nl MO non-small cell lung cancer at the University of Chicago was presented in 1982 and concluded that surgical resection combined with postoperative chemotherapy, radiation therapy, or both yielded improved survival over surgical resection alone. I Since that report another series has been published that supports the use of operation as a single treatment From the Departments of Surgery and Medicine, University of Chicago Hospitals and Clinics, Chicago, Ill. Read at the Sixty-fifth Annual Meeting of The American Association for Thoracic Surgery, New Orleans, La., April 29-May I, 1985. Address for reprints: Mark K. Ferguson, M.D., The University of Chicago, Department of Surgery, Box 255, 5841 South Maryland Ave., Chicago, Ill. 60637. *Current address: Department of Surgery, Creighton University, 601 North 30th St., Suite 3740, Omaha, Neb.
344
modality for patients with hilar nodal metastases.' Disagreement thus continues over whether resection in combination with chemotherapy, radiation therapy, or both offers improved survival over operation alone in patients with lung cancer complicated by Nl nodal metastases. For these reasons we have updated our data to reassess whether adjuvant therapy is beneficial in treatment of these patients after surgical resection.
Patients and methods From June, 1974, to December, 1984,34 consecutive previously untreated patients with T1 Nl MOor T2 Nl MO non-small cell lung cancer were seen by the University of Chicago Chest Oncology Group. There were 24 men and 10 women aged 38 to 80 years (median 67 years). Initial evaluation included history and physical examination, blood counts, chemistries, and pulmonary function tests. A whole-body tomographic gallium scan was done in 32 of 34 patients, and in recent years computed tomography of the thorax and upper part of the abdomen including the adrenal glands was performed routinely. When indicated by these tests,
Volume 91 Number 3 March. 1986
Non-small cell lung cancer 3 4 5
Table I. Distribution by operation and adjuvant therapy Pneumonectomy
Treatment Resection only Resection + RT Resection + RT
3
+ chemo
Total
100 90 80
Toral >.
5 10
8 2 6
II 7 16
:0
18
16
34
to
11l .0
60
ll:
50
>
40
0
.~
Legend: RT. Radiation therapy. Cherno, Chemotherapy.
:>
rJ)
30 20
Table II. Distribution by cell type and TNM classification
10
Histologic type
Tl Nl MO
T2 Nl MO
I~
Adenocarcinoma Squamous cell carcinoma Large cell carcinoma
7
12
19
I
10
II
044
Total
8
26
70
34
mediastinoscopy was performed to exclude mediastinal nodal involvement. Patients had liver-spleen scan, bone scan, or brain scan as indicated or to confirm abnormalities on gallium scan. All patients underwent thoracotomy with pulmonary resection and mediastinal nodal sampling. Eighteen patients had lobectomy or bilobectomy and 16 had pneumonectomy. Before 1976 no further treatment was provided. Since then we have recommended adjuvant postoperative radiation therapy (refused by three) and chemotherapy (refused by nine). Eleven patients underwent surgical resection alone (32.3%), seven patients underwent resection plus radiation therapy (20.6%), and 16 patients had resection followed by radiation therapy and chemotherapy (47.1%) (Table I). Radiation therapy consisted of 3,000 rads in 15 fractions to the primary tumor site, ipsilateral hilum, and mediastinum. The chemotherapy regimen included cyclophosphamide, 300 mg/rn! intravenously, days 1 and 8; doxorubicin, 20 mg/rn' intravenously, days 1 and 8; methotrexate, 15 mg/rrr' intravenously, days 1 and 8; and procarbazine, 100 mg/m' per os, days 3 through 13 (CAMP). Cycles of CAMP were repeated every 28 days for 1 year or until recurrence was diagnosed. Patients were routinely followed up in our Chest Oncology Clinic with chest x-ray films and periodic gallium scan and/or computed tomographic scan of the thorax. Tumors were evaluated according to size, pleural involvement, histologic type, and location. Nl nodes were dissected from the resected specimen and assigned to one of three levels: intrapulmonary, peribronchial, or hilar. Postoperative staging was performed according to
0 2
3
4
5
6
Years From Diagnosis
Fig. 1. Survival curve for all patients with T1 Nl MO and T2 N 1 MO non-small cell lung cancer. Vertical bars indicate length of follow-up for patients currently alive.
the TNM system.' Survival was calculated from the date of operation until death or until March, 1985. The mean length of follow-up for all patients is 35.8 months with the longest follow-up being 10 years. Statistical analyses were performed by the algorithm of Lee and Desu' and Kaplan-Meier' life-table techniques.
Results Thirty-four consecutive patients with T1 Nl MO or T2 Nl MO lung cancer were seen and all were evaluable. Fourteen of these patients were new since February, 1981, the closure date of our previous report. Nineteen patients had adenocarcinoma (55.9%), 11 patients had squamous carcinoma (32.3%), and the remaining four patients had large cell carcinoma (11.8%). Eight patients had T1 Nl MO tumors and 26 patients had T2 Nl MOdisease (Table 11).There was a predominance of isolated hilar nodal metastases, and three patients had involvement of nodes from more than one level (Table III). Median overall survival for patients was 19.7 months, and 5 year survival rate was 30.2% (Fig. 1). Median survival by cell type was 19.6 months for adenocarcinoma, 20.5 months for squamous cell carcinoma, and was not reached at 18 months for large cell carcinoma (no statistical differences) (Fig. 2). There were 23 deaths with 11 patients alive at the time of reporting. Two patients died perioperatively, one of sepsis and the other of a cardiac arrhythmia. Fifteen patients died of progressive disease, two of other proved causes without known recurrence, and four of unknown causes. Survival according to treatment was examined.
The Journal of
346 Ferguson et al.
Thoracic and Cardiovascular Surgery
100 100 Adenocarcinoma
90
>-
:0
'"
.0 0
Q:
80
Squamous Cell Carcinoma
i
Large Cell Carcinoma
'--,
70
'"
.0 0
I
60
::J
en
a:
I
IL
50
-
>-
:0
I
,
lii >
I
<;; > 40
's ...
90
I L
_
1'---------
30 20
's ... ::J
en
80
I
Surgery + RT+ Chemo
I
Surgery+ RT
I I
I
L
70 60 50 40 30
I I L
Surgery
l
I
,
L'
I L
I
I
L_ I I
L,
20 10
10
0
0
I 3
I
2
i 4
I
5
I 6
Years From Diagnosis
Fig. 2. Survival curve for patients with T1 Nl MO and T2 N I MO non-small cell lung cancer according to cell type.
.
I
L_ ,
IL
_
2
3
5
4
6
Years From Diagnosis
Fig. 3. Survival curve for patients with TI NI MO and T2 N I MO non-small cell lung cancer according to treatment groups. Chemo, Chemotherapy. RT, Radiation therapy.
Table m. Locations of tumors and nodal metastases Incidence of nodal metastases
Primary tumor Location
No.
Intrapulmonary
Peribronchial
RUL RLL RML LUL LLL Total
34
Hilar
7
1
0
8 I 12 6
3
I
0
0
1
3
2
1
0
9 4
8
3
24
Indeterminate
6 4
3
Legend: RUL, RLL. and RML, Right upper, lower, and middle lobe. LUL and LLL, Left upper and lower lobe.
Patients undergoing resection followed by adjuvant radiation therapy and chemotherapy had an overall median survival of 45.5 months and a 5 year survival rate of 45.9%. This survival period was significantly longer than the median survival in patients who had surgical resection alone (13.0 months excluding two operative deaths, p < 0.005). Patients who underwent resection followed by radiation therapy had a median survival of 19.2 months and a 5 year survival rate of 28.6%, figures which did not differ significantly from either of the other treatment groups (Fig. 3). Patients who received chemotherapy had a median survival of 45.5 months compared to 14.3 months for patients whose treatment did not include chemotherapy (p < 0.005 excluding two operative deaths). We found no relationship between survival and T classification (Fig. 4), absolute tumor size, number or location of involved nodes, distance of tumor from resected bronchial margin, sex of the patient, presence or absence of visceral pleural involvement (Fig. 5), and type of resection performed (Fig. 6).
Fifteen patients had proved recurrences, only two of whom are alive 3 and 6 months after the diagnosis of recurrence. Initial recurrences were local in three patients, with a median time to recurrence of 11.5 months. Distant recurrences were found in 13 patients, with a median time to initial recurrence of 9.0 months. The most frequent site of distant recurrence was the brain, which was affected in six patients, four of whom had adenocarcinoma. Discussion
In a previous publication from this institution we' reported results of treatment of Tl Nl MO and T2 Nl MO non-small cell bronchogenic carcinoma and we highlighted the effects of a combined modality approach. In that report adjuvant radiation therapy and chemotherapy appeared to benefit patients by increasing the survival rate up to 75% at 5 years. This report updates that experience and expands the group originally reported from 20 to 34 patients. Although the previously excellent survival rates for the multimodality
Volume 91
Non-small cell lung cancer
Number 3 March, 1986
100
100 90 80
~
70
.0
60
0..
50
>
40
:::J
30
:ctil
I
>.
:ctil
.~
en
Pleura Negative
I I
Pleura Positive
80
L,
70
I I I
I
L_,
.0
60
0..
50
I IL __
> 'S
40
I I
~
30
I
0
0
iii
L_
90
T2 Tumors
1__ -
---,
I
T 1 Tumors
-I
iii
en
20
20
10
10
347
0
0
2
3
4
5
2
6
Fig. 4. Survival curve for patients with TI NI MO and T2 NI MO non-small cell lung cancer according to T classification. Median survival for patients with T1 tumors was 20.0 months and for those with T2 lesions, was 20.8 months (p = NS).
treatment group have been tempered in this report by extended follow-up and an expanded patient pool, our conclusions regarding the benefit of adjuvant treatment in these patients remain the same. In 1983 Martini and others' reported results of treatment of 75 patients with non-small cell lung cancer complicated by Nl nodal metastases. Sixty-two of the 75 patients were treated by resection alone, and the combined overall survival rate was 49% at 5 years. These results are significantly better than our overall survival rate of 30% at 5 years and significantly better than the 5 year survival rates for Stage II lung cancer of 7%, 29%, and 31% reported from other institutions." The reasons for such a discrepancy in survival when comparing our patients with those of the Memorial group' are not immediately clear. Both groups had similar distributions of patient age, tumor stage, and type of surgical resection. However, the Memorial group appears to have had a higher incidence of both peripheral carcinoma and squamous cell carcinoma than occurred in our group, whereas their incidence of hilar nodal involvement was considerably lower than ours (37.3% versus 70.6%). In addition, although the value of extended mediastinal lymph node resection in prolonging survival may be unproved, there is little doubt that it can provide improved pathologic staging. There is a high incidence of residual local disease after "curative" resection of early lung cancer, especially with squamous cell carcinomas." It is possible that the Memorial group performed more extensive mediastinal lymph node dissections, eliminating a number of otherwise undetecta-
4
3
6
5
Years From Diagnosis
Years From Diagnosis
Fig. 5. Survival curve for patients with T1 NI MO and T2 N I MO non-small cell lung cancer according to the presence or absence of visceral pleural involvement. Median survival for II patients with "positive" pleura was 19.7 months and for 22 with "negative" pleura, 20.0 months (p = NS). 100 90
>.
.stil .0 0
80 70 60 50
>
40
:::J
30
.~
en
L,
Pneumonectomy
I I I
0..
iii
Lobectomy
II
20
L~
I l
I
L I
L I
L_ I
~
------------,1
I
L
_
10 0
2
3
4
5
6
Years From Diagnosis
Fig. 6. Survival curve for patients with T1 N I MO and T2 N I MO non-small cell lung cancer according to the type of resection. Median survival for lobectomy/bilobectomy patients was 25.9 months and for pneumonectomy patients, 16.0 months (p = NS excluding two operative deaths).
ble cases of Stage III disease from their study group and improving their staging accuracy and the resultant overall survival rates. Finally, unrecognized differences probably exist in our patient populations. The incidence of local recurrence is low in our series. The presence of lymphatic spread, whether or not it is limited to hilar nodes, is usually an indication that local control of disease remains in doubt. For this reason we routinely irradiate the mediastinum and hilus postoperatively in patients with involved nodes. This approach is
The Journal of
348
Ferguson et al.
Thoracic and Cardiovascular Surgery
supported by others who report significant improvement in survival in patients with hilar and mediastinal lymph node metastases when radiation therapy is employed after adequate resection.'? The use of radiation as a single treatment modality, however, is not acceptable in patients who have resectable disease and are physiologically able to undergo operation." Although our local recurrence rate was low, there was a high rate of distant recurrence, primarily in the brain. Two thirds of those patients with metastatic disease to the brain had adenocarcinoma. These findings support previous data reported in patients with non-small cell lung cancer and emphasize the high incidence of cerebral metastases in patients with advanced stages of adenocarcinoma of the lung. This rate is as high as 25% in some series." 12 We interpret these data as supporting the study of prophylactic cranial irradiation in patients with advanced stages of adenocarcinoma. We have been investigating prophylactic cranial irradiation at this institution for the past 3 years. The high incidence of distant metastases underscores the need in these patients for systemic adjuvant therapy. We l3 previously reported that CAMP appeared to prolong survival in Stage III non-small cell bronchogenic carcinoma. Although we recognize the shortcomings of a retrospective nonrandomized review, the results reported herein nevertheless demonstrate an improvement in long-term survival in patients with non-small cell lung cancer and N I nodal metastases treated with resection, radiation therapy, and chemotherapy compared to patients in our series and those of others'" treated with resection alone. These data thus support the use of adjuvant therapy for Tl NI MO and T2 NI MO non-small cell lung cancer. Other reports identify a number of factors that impact on survival in patients with lung cancer, including cell type," proximity of the primary tumor to the hilum, and the presence of visceral pleural involvement.3 We were unable to confirm any of these as risk factors in the present study. The concept that more extended pulmonary resection is indicated for patients with hilar nodal metastases, specifically pneumonectomy rather than lobectomy, also cannot be supported. In patients with non-small cell lung cancer and hilar nodal metastases, we conclude: 1. Surgical therapy is the primary form of treatment. 2. Adjuvant postoperative therapy, particularly chemotherapy, improves survival. 3. The rate of recurrence is high, particularly in the brain. Prophylactic cranial irradiation should be admin-
istered to patients with advanced stages of adenocarcinoma of the lung in investigative programs. 4. Cell type, tumor size, number of nodes involved, involvement of visceral pleura, and type of resection have no proven effect on survival. We wish to acknowledge the assistance of Margaret Johnson in data analysis, Consuelo Skosey and Cynthia YoungStrickland in patient care, and Cathy Ward in manuscript preparation. REFERENCES Newman SB, OeMeester TR, Golomb HM, Hoffman PC, Little AG, Raghavan Y: Treatment of modified Stage II (Tl Nl MO, T2 Nl MO) non-small cell bronchogenic carcinoma. J THORAC CARDlOVASC SURG 86:180-185, 1983 2 Martini N, Flehinger BJ, Nagasaki F, Hart B: Prognostic significance of Nl disease in carcinoma of the lung. J THORAC CARDIOVASC SURG 86:646-653, 1983 3 American Joint Committee for Cancer Staging and EndResults Reporting. Manual for Staging of Lung Cancer, Chicago, 1979, Whiting Press 4 Lee E, Desu M: A computer program for comapring k samples with right-censored data. Comput Programs Biomed 2:315-321, 1972 5 Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 6 Kirsh MM, Rotman H, Argenta L, Bove E, Cimmino V, Tashian J, Ferguson P, Sloan H: Carcinoma of the lung. Results of treatment over ten years. Ann Thorac Surg 21:371-377, 1976 7 Vincent RG, Takita H, Lane WW, Gutierrez AC, Pickren JW: Surgical therapy of lung cancer. J THORAC CARDlOVASC SURG 71:581-591, 1976 8 Mountain CF: Assessment of the role of surgery for control of lung cancer. Ann Thorac Surg 24:365-373, • 1977 9 Matthews MJ, Kanhouwa S, Pickren J, Robinette D: Frequency of residual and metastatic tumor in patients undergoing curative surgical resection for lung cancer. Cancer Chemother Rep 4:63-67, 1973 10 Green N, Kurohara SS, George FW, Crews QE Jr: Postresection irradiation for primary lung cancer. Radiology 116:405-407, 1975 II Cooper JO, Pearson FG, Todd TRJ, Patterson GA, Ginsberg RJ, Basiuk J, Blair Y, Cass W: Radiotherapy alone for patients with operable carcinoma of the lung. Chest 87:289-292, 1985 12 Bitran JO, Golomb HM, OeMeester TR, Desser RK, Hoffman P, Cohen L, Green M, Sovik C: Combined modality therapy for Stage III non-small cell bronchogenic carcinoma (NSBC). Proc Am Soc Clin Oncol 21:446, 1980
Volume 91 Number 3 March, 1986
13 Bitran JD, Desser RK, DeMeester TR, Shapiro CM, Billings A, Rubenstein L, Evans R, Colman M, Rao Y, Griem M, Golomb HM: Combined modality therapy for Stage III MO non-oat cell bronchogenic carcinoma. Cancer Treat Rep 62:327-332, 1978 14 Shields TW, Vee J, Conn JH, Robinette CD: Relationship of cell type and lymph node metastasis to survival after resection of bronchial carcinoma. Ann Thorac Surg 20:501-510,1975
Discussion DR. E. CARMACK HOLMES Los Angeles, Calif
It is always gratifying to hear a report suggesting that adjuvant therapy is beneficial in patients with resected lung cancer, because there have been so many negative reports over the years. Although it is gratifying to hear such a report, I do have problems with data. It is difficult to draw conclusions based on 34 patients in a three-arm study. Statisticians tell us that to generate meaningful data, a two-arm study usually requires 60 patients in each arm. I would ask the authors to comment on this in their summation. I would like to share some data generated by the National Lung Cancer Study Group. One study is somewhat analogous to the one presented by Dr. Ferguson's group. Patients with Stage II and Stage III adenocarcinoma and large cell undifferentiated carcinoma were operated on, stratified, and then randomized to receive immunotherapy or a combination chemotherapy. At thoracotomy these patients underwent very careful intraoperative lymph node sampling for staging purposes. There is a significant difference in favor of the chemotherapy group. I would like to emphasize that this study includes 130 patients with 60 to 65 patients in each of the arms. Survival with disease, disease-free survival, and overall survival are all prolonged significantly. A second study is one that involves patients with squamous cell carcinoma of the lung. As you recall, the previous study evaluated only adenocarcinoma and large cell undifferentiated carcinoma. These are patients with Stage II and Stage III squamous carcinoma of the lung, most of whom had diseased mediastinal nodes. They were operated on, carefully staged intraoperatively, and then randomized to receive radiation therapy, 5,000 rads postoperatively, or no further treatment. There is absolutely no difference between the patients receiving postoperative irradiation and those undergoing operation alone. Of interest, however,is that the patients who did receive radiation therapy had a much lower incidence of local
Non-small cell lung cancer 3 49
recurrence than the patients receiving no further therapy (p = 0.001). Thus, postoperative radiation therapy in these patients does not prolong survival but does diminish local recurrence. I would like to pose two questions to the authors. First, there was- a dramatic difference in the survival of patients with adenocarcinoma and squamous carcinoma in the subset T2 N I that they presented. I find it difficult to accept lumping these two cell types into one study, as the authors did. Finally, I would again question the authors regarding conclusions that impact on reclassification and restaging of lung cancer, conclusions regarding the effect on prognosis of T2 N I disease and pleural involvement based on a series of only 34 patients. DR. FERGUSON(Cwsm~ I would like to thank Dr. Holmes for his insightful comments. As he points out, the number of patients that we have analyzed is relatively small. Unfortunately, patients with isolated hilar nodal metastases are uncommon, and it is difficult to generate large numbers of such patients for study. I would emphasize that in analyzing our various treatment groups, we found quite significant statistical differences despite the small numbers in each group. Thus, while it is nice to have an "ideal" number of patients in each study group, one cannot really ignore such significant differences when they exist. On the other hand, one cannot become too free in interpreting differences between small numbers of patients. While there is frequently said to be a difference in survival between patients with adenocarcinoma and those with squamous cell carcinoma, we were unable to demonstrate such a difference. Such variations in survival between cell types, when they exist, are actually much less significant than the variations in survival among stages. We group these cell types together because in the end we find there are more similarities between them than there are differences. Dr. Holmes has presented some interesting data generated by the Lung Cancer Study Group, which support many of the conclusionsfrom our paper. First, the use of adjuvant radiation therapy in patients with NI nodal metastases can limit the incidence of local recurrence. Second, systemic chemotherapy in patients with advanced stages of lung cancer can prolong survival when compared to results of surgical treatment alone. Finally, I should reemphasize the importance of complete intraoperative staging, which includes mediastinal nodal dissection. Only in this way can adequate pathologic staging be obtained for appropriate assignment to treatment groups and assessment of prognosis.