- Email: [email protected]
Prognostic factors in surgical treatment of solitary brain metastasis after resection of non-small-cell lung cancer
Lung Cancer 24 (1999) 99 – 106
Prognostic factors in surgical treatment of solitary brain metastasis after resection of non-small-cell lung cancer Yukio Saitoh a,*, Takehiko Fujisawa a, Mitsutoshi Shiba a, Shigetoshi Yoshida a, Yasuo Sekine a, Masayuki Baba a, Toshihiko Iizasa a, Motoo Kubota b a
Department of Surgery, Institute of Pulmonary Cancer Research, School of Medicine, Chiba Uni6ersity, 1 -8 -1 Inohana, Chuo-ku, Chiba 260 -8670, Japan b Department of Neurosurgery, School of Medicine, Chiba Uni6ersity, Chiba 260 -8670, Japan Received 23 October 1998; received in revised form 18 March 1999; accepted 26 March 1999
Abstract Patients with brain metastasis after resection of non-small-cell lung cancer usually have poor prognosis. A few such patients, however, survive for long periods after surgical resection of brain metastases. To evaluate the prognostic factors in resection of solitary brain metastasis from non-small-cell lung cancer, we reviewed 24 cases undergoing resection of solitary brain metastasis after resection of the primary site from 1977 to 1993. The patient population consisted of 20 men and four women ranging in age from 40 to 75 years old (average, 57.8 years old). None of the patients had systemic metastasis except in the brain at the time of brain surgery. The overall survival rates were 12.5% at 3 years and 8.3% at 5 years after brain surgery. The longest survival periods were 11.5 years after brain surgery and 15.4 years after lung surgery. The interval between lung and brain surgery (5 360 days vs. \ 360 days), differentiation of primary cancer (poor vs. moderate), size of primary site ( 5 5.0 cm vs. \ 5.0 cm), and operation of primary site (lobectomy vs. pneumonectomy) significantly affected survival as shown by univariate analysis (P B 0.05). Other clinical factors (age, gender, histology, T- and N-status, ‘resectability with curative intent’ of the primary site, location of the brain metastasis and postoperative radiation therapy) did not affect survival. Multivariate analysis using Cox’s proportional hazards model indicated that an interval of more than 360 days between the two surgical procedures (hazard ratio =0.2351, P =0.0136) and lobectomy (hazard ratio= 0.5274, P =0.0416) were independent prognostic factors. In conclusion, patients with solitary brain metastasis from non-small-cell lung cancer without other organ metastasis, in whom relapse in the brain occurred more than 1 year after resection of the primary site and in whom lobectomy was performed, should be treated surgically to maximize the chance of prolonged survival. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Surgery; Brain metastasis; Non-small-cell lung cancer
* Corresponding author. Tel.: +81-43-222-7171; fax: + 81-43-226-2172. 0169-5002/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 9 - 5 0 0 2 ( 9 9 ) 0 0 0 3 4 - 3
100
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
1. Introduction
2. Materials and methods
The brain is a common target of metastasis from non-small-cell lung cancer (NSCLC), and its relapse after lung surgery is correlated with poor prognosis [1]. In pathological examinations of autopsy cases, Knights [2] reported that 53% of pulmonary carcinoma cases had brain metastases. In addition, five of 21 cases showed only intracranial metastases without metastasis in other organs, and three of these 21 cases showed solitary brain lesions. These findings showed that there may be indications for surgical resection in selected cases with brain metastasis from NSCLC. Clinically, brain metastases occur frequently as the first relapse site during the follow-up period after resection of NSCLC [3,4]. Other investigators have reported good results in symptom-free conditions and quality of life after resection of brain metastases, and a small number of these cases have shown good long-term survival [5 – 7]. Recently, two randomized trials have indicated that surgical resection of solitary brain metastasis followed by radiation therapy results in better survival rate after treatment than radiation therapy alone [8,9]. On the other hand, Mintz et al. [10] detected no difference in survival rate between radiation therapy alone and surgery plus radiation therapy. Median survival periods of groups with surgery in these studies were very short, i.e. 40 weeks [8], 12 months [9] and 5.6 months [10], respectively, and the majority of patients had early death, suggesting that not all cases with solitary brain metastasis should be treated with brain surgery. One reason for these unsatisfactory survival results was that the above three studies included many cases with extracranial metastasis at the time of brain surgery. Therefore, in addition to a lack of extracranial metastasis, other clinical factors influencing survival should be investigated prior to scheduling brain surgery. To elucidate prognostic factors for long-term survival and clarify the indications of surgical treatment for solitary brain metastasis, we reviewed the clinical records of cases treated at our institute.
2.1. Patients From August 1977, through June 1993, 1049 patients underwent resection of NSCLC at our hospital. Of these, 30 underwent combined lung surgery for the primary site and brain surgery for brain metastases. Four cases had brain surgery prior to lung surgery because of their neurological symptoms and two cases had multiple brain metastases. Twenty-four patients (2.3%), excluding the above six patients, in whom the solitary brain metastases were surgically treated during the follow-up period after lung surgery, were analyzed in this study. The resected brain tumors were confirmed pathologically to be metastases from NSCLC. These patients included 20 men and four women ranging in age from 40 to 75 years old (average, 57.8 years old).
2.2. Histological examination Histology and T, N classification were confirmed according to the criteria reported by Mountain [11] and the general rule for clinical and pathological recording for lung carcinoma, edited by the Japan Lung Cancer Society [12]. Cellular differentiation was graded as well, moderately or poorly differentiated for those patients with adenocarcinoma and squamous cell carcinoma. Large cell carcinomas were labeled as undifferentiated [13].
2.3. Statistical analysis Clinical data obtained from medical records were classified into two groups and used as variables (Table 1). Relations between clinical variables were tested by Fisher’s exact test or t-test. Survival time was measured from the date of brain surgery until death or the most-recent date of follow-up for surviving patients. Survival rate was calculated by the Kaplan–Meier method and comparisons of survival were estimated by generalized Wilcoxon test [14] for univariate analysis. Cox’s proportional hazards model (forward stepwise procedure) [15] was used for multivariate
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
analysis. Significance was defined as P B 0.05. All these analyses were performed with the SPSS software package (SPSS version 6.1 for Windows, SPSS, Chicago, IL).
3. Results
3.1. O6erall sur6i6al Of the 24 cases examined, two patients survived for 11.5 and 9.7 years following brain surgery,
101
respectively. Of the remaining 22 patients, two died due to operation-related complications within 30 days after brain surgery (8%) and 20 died due to the cancer. The causes of operationrelated death were pulmonary infarction on the seventh day and cerebral bleeding on the twentyfirst day after brain surgery. Among the cases of cancer-related death, 16 patients died within a year after brain surgery because of systemic recurrence (15 cases) and multiple brain metastases (one case), while four survived for over 1 year. The overall survival rates were 25% at 1 year,
Table 1 Clinical characteristics of 24 patients underwent surgery for brain metastasis and results of univariate analysis for prognostic factors
Gender Age (year; 57.8 9 10.3)b Interval between two operations (days; 396 9 327) Primary lung tumor Histology Differentiation
Location
T-factor Size (cm; 5.4 9 2.7) N-factor Resectability Operation Brain metastasis Location Resectability Radiation
a b
Generalized Wilcoxon test. Mean9 SD.
n
Median survival time (days)
P valuea
Male Female ]60 B60 5360 \360
20 4 12 12 12 12
203 173 216 117 117 292
0.5328
Adenocarcinoma Others Moderate Poor, undifferentiated Right Left Upper lobe Others T1 or T2 T3 55.0 \5.0 N0 N1 or N2 Complete resection Incomplete resection Lobectomy Pneumonectomy
15 9 11 13
224 203 292 134
9 15 18 6 15 9 15 9 11 13 20 4 18 6
117 254 203 173 224 203 239 102 267 189 203 134 239 75
Supratentorial Infratentorial Complete resection Incomplete resection With Without
16 8 20 4 10 14
224 173 216 189 203 173
0.1923 0.0069
0.7894 0.0152
0.063 0.7203 0.9057 0.023 0.4178 0.6386 0.0057
0.4206 0.9399 0.4374
102
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
3.3. Brain metastasis
Fig. 1. Actuarial survival according to the types of pulmonary resection. Cases undergoing lobectomy had a significantly prolonged survival.
12.5% at 3 years and 8.3% at 5 years. The median survival period was 6.8 months.
3.2. Primary site Histological types of primary lung tumors were adenocarcinoma in 15 cases (63%), squamous cell carcinoma in eight (33%), and large cell carcinoma in one case (4%). Furthermore, adenocarcinoma and squamous cell carcinoma were graded as moderately differentiated in 11 cases (46%) and poorly differentiated in 12 cases (50%). The types of pulmonary resection consisted of lobectomy in 18 cases and pneumonectomy in six. Twenty patients underwent complete resection of lung tumor and four underwent incomplete resection (one in the bronchial stump and three in the chest wall). Three cases with chest wall resection received radiation therapy of the chest wall following surgery. On pathological examination, the lymph node status was classified as N0 in 11 cases, N1 in three and N2 in ten, and the primary tumor status was classified as T1 in four cases, T2 in 11 and T3 in nine. Before lung surgery, the results of Gallium whole-body scan and bone scan indicated no extrathoracic spread of the tumor, and no clinical symptoms of brain metastasis were detected in any of the cases. Use of brain CT was introduced for routine preoperative evaluation in 1982 (19 cases). At the time of lung surgery, six cases had stage I disease, two stage II and 16 stage IIIA.
In all cases, neurological symptoms were detected and brain CT or brain MRI (ten cases since 1990) showed solitary mass lesions. Before brain surgery, chest CT showed no intrathoracic recurrence and clinically there were no signs of metastasis in other organs. The interval from lung surgery to brain surgery in 24 cases ranged from 49 to 1339 days with a mean of 396 days. Location of the brain tumor included the supratentorial region in 16 and infratentorial region in eight. All of the former and four of the latter had complete removal of the brain tumor. Four cases with infratentorial lesions underwent incomplete resection. After brain surgery, five cases received localized radiation therapy (50 Gy) and five were treated with both whole-brain radiation (30 Gy) and localized radiation (ranging from 10 to 20 Gy) therapy. No cases had systemic chemotherapy after brain surgery. Two cases showed intracranial relapse without systemic metastasis. One underwent reoperation for a supratentorial lesion 2 months after the initial treatment, and another received radiosurgery using a gamma knife for a cerebellar lesion 3 months after the first brain surgery.
3.4. Uni6ariate analysis We analyzed 15 variables to evaluate their influence on survival: i.e. age, gender, interval between lung and brain surgery, nine variables related to primary lung tumor and three variables related to brain metastasis (Table 1). Univariate analysis indicated that lobectomy (P= 0.0057, Fig. 1), an interval between the two operations longer than 360 days (P= 0.0069, Fig. 2), moderate differentiation (P=0.0152), and primary lesion less than 5 cm in maximum diameter (P=0.023) were significantly good prognostic factors for survival.
3.5. Multi6ariate analysis Six variables showing differences with PB 0.2 on univariate analysis were selected for multivariate analysis. Multivariate analysis using Cox’s proportional hazards model revealed two signifi-
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
103
4. Discussion
Fig. 2. Actuarial survival according to interval between lung and brain surgery. Cases with interval more than 360 days had a significantly prolonged survival.
cant variables (Table 2). The interval between the two operations was an independent factor with a hazard ratio of 0.2351 (P= 0.0136). The type of pulmonary resection had a hazard ratio of 0.5274 (P= 0.0416), indicating that this was another independent factor.
3.6. Clinical factors associated with inter6al between the two operations and the type of pulmonary resection The interval between the two operations was significantly associated with histological type (P= 0.018) and differentiation of the lung tumor (P= 0.003). Adenocarcinomas and moderately differentiated tumors had longer intervals between the two operations (Table 3). The type of pulmonary resection was significantly associated with differentiation of lung tumor (P= 0.0168) and tumor size (lobectomy; 4.691.9 cm, pneumonectomy; 7.99 3.2 cm, P =0.049). All of the pneumonectomy cases had poorly differentiated carcinomas.
There have been many reports concerning the factors influencing survival after combined resection of NSCLC and brain metastasis, but the results of these studies have been controversial. Cases with these combined resections are grouped into those with synchronous and metachronous presentation, or simultaneous and sequential presentation [16], based on the time of diagnosis of lung tumor and brain metastasis. In general, the former cases had resection of brain metastasis prior to resection of lung tumor, because of their neurological symptoms. The above two groups should be analyzed independently to accurately elucidate the prognostic factors. In this study of patients undergoing lung and brain surgery, those in whom brain metastases were not detected at the time of lung surgery were selected and analyzed. The results of this study provided useful information for deciding whether brain surgery should be performed when the first recurrent site of NSCLC is the brain during follow-up after lung surgery. The 3-year survival rate of 12.5%, 5-year survival rate of 8.3%, and the median survival of 6.8 months after brain surgery reported in this study are unsatisfactory from the viewpoint of prolonged survival. Two of the 24 cases examined showed long-term survival (11.5 and 9.7 years) with no recurrence and little neurological deficits following brain surgery. On the other hand, ten of the 24 patients (42%) died within 6 months and 18 (75%) died within the first year. From these observations, it is obvious that the elucidation of prognostic factors and the selection of candidates for brain surgery are of major clinical importance. Previous studies showed that age and gender [17,18], histological type of tumor [19], N-status
Table 2 Results of multivariate analysis for six prognostic variables shown to be significant (PB0.2) by univariate analysis Variables
Interval between two operations (5360 vs. \360) Type of operation (lobectomy vs. pneumonectomy)
P value
0.0136 0.0416
Hazard ratio
0.2351 0.5274
95% C.I. Lower
Upper
0.0744 0.2851
0.7427 0.9759
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
104
Table 3 Clinical factors associated with interval between the two operations Interval between two operations (days)
P valuea
Gender Male Female
385 9349b 451 9 199
N.S.
Age (year) ]60 B60
3319442 4429385
N.S.
514 9350
0.018
Histology Adenocarcinoma Others Differentiation Moderate Poor, undifferentiated
199 9148 599 9326 224 9214
0.003
Location Right Left
385 9277 402 9362
N.S.
Upper lobe Others
342 9336 558 9255
N.S.
T-factor T1 or T2 T3
425 9249 347 9 440
N.S.
Size (cm) 55.0 \5.0
481 9 363 2539 196
N.S.
N-factor N0 N1 or N2
380 9 224 409 9 402
N.S.
Resectability Complete resec- 428 9 344 tion Incomplete re234 9 161 section Operation Lobectomy Pneumonectomy a b
451 9 355 230 9 136
N.S.
N.S.
t-test. Mean9 SD.
[20 – 22] and complete or incomplete resection of the primary site [23], and supratentorial or infratentorial location of brain metastasis [17] are
significant prognostic factors. In our experience, cases with metachronous presentation after lung surgery, solitary brain lesion and no extracranial recurrence, the clinical factors described above as significant prognostic factors failed to show significant differences in survival after brain surgery. However, the number of cases examined in this study was too small to conclusively exclude the influence of the above factors on survival, especially the resectability of the primary site and the location of brain metastasis. In this study, multivariate analysis indicated that interval between brain and lung surgery was a significant prognostic factor. Mussi et al. [21] and Macchiarini et al. [24] reported similar results concerning the interval between the two operations using univariate analysis. It is important to clarify what the interval between the two operations actually reflects. In our study, this interval was associated with histological type and differentiation of the tumor, but showed no relation with local extent of primary lesion indicated by Tstatus, N-status, tumor size or ‘resectability with curative intent’. In reports of cases with long-term survival after brain surgery, the majority of tumors were histologically classified as adenocarcinomas [6]. Furthermore, in cases with recurrent brain metastasis after brain surgery [25] and in patients treated with radiation therapy alone [26], the histology of adenocarcinoma was also associated with longer survival than other types of tumor. Adenocarcinoma has a tendency to develop brain metastasis [27], but it is well known that this type of tumor shows biological heterogeneity. It seems that this interval reflects the biological characteristics of the tumor, and a longer interval between brain and lung surgery indicates low-grade malignancy of the tumor. Another significant prognostic factor selected in this study was the type of pulmonary resection. There have been a few previous reports concerning this factor. Mussi et al. [21] reported that patients who underwent lobectomy had longer survival than those who underwent pneumonectomy. Our results were consistent with these observations. Magilligan et al. [28] reported that wedge resection of small peripheral tumors showed better results than lobectomy or pneu-
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
monectomy, but this study included cases of synchronous presentation with brain surgery performed before lung surgery. Usually, a major determinant of the type of pulmonary resection is local extent of the primary site. However, in this study this factor was associated with differentiation of lung tumor and primary tumor size. Based on these results, it seems that the biological characteristics of the tumor have the most important effect on long-term survival, and the two prognostic factors selected by multivariate analysis were considered to be good clinical indicators of the biological characteristics of the tumor. Kiss et al. [29] reported that DNA ploidy level in brain metastasis is related to patient survival. It is possible that better prognostic factors will be determined by biological investigation of carcinoma. In this study, ten patients received radiation therapy after brain surgery. There was no significant difference in survival between cases with and without radiation therapy. Since these cases were not randomized, the efficacy of postoperative radiation therapy could not be evaluated in this study. In conclusion, patients with solitary brain metastasis from non-small-cell lung cancer without metastasis to other organs, in whom relapse in the brain occurred more than 1 year after resection of the primary site and in whom the lung operation was lobectomy, should be treated surgically, with a good chance of prolonged survival.
References [1] Markesbery WR, Brooks WH, Gupta GD, Young AB. Treatment for patients with cerebral metastases. Arch Neurol 1978;35:754 –6. [2] Knights EM Jr. Metastatic tumors of the brain and their relation to primary and secondary pulmonary cancer. Cancer 1954;7:259 – 65. [3] Figlin RA, Piantadosi S, Feld R. Intracranial recurrence of carcinoma after complete surgical resection of stage,, and non-small-cell lung cancer. New Engl J Med 1988;318:1300– 5. [4] Yano T, Yokoyama H, Inoue T, Asoh H, Tayama K, Takai E, Ichinose Y. The first site of recurrence after complete resection in non-small-cell carcinoma of the lung. Comparison between pN0 disease and pN2 disease. J Thorac Cardiovasc Surg 1994;108:680–3.
105
[5] Sarma DP, Weilbaecher TG. Long-term survival after brain metastasis from lung cancer. Cancer 1986;58:1366 – 70. [6] Shahidi H, Kvale PA. Long-term survival following surgical treatment of solitary brain metastasis in non-small cell lung cancer. Chest 1996;109:271 – 6. [7] Kelly K, Bunn PA Jr. Is it time to reevaluate our approach to the treatment of brain metastases in patients with non-small cell lung cancer? Lung Cancer 1998;20:85 – 91. [8] Patchell RA, Tibbs PA, Walsh JW, Dempsey RJ, Maruyama Y, Kryscio RJ, Markesbery WR, Macdonald JS, Young B. A randomized trial of surgery in the treatment of single metastases to the brain. New Engl J Med 1990;322:494 – 500. [9] Vecht CJ, Haaxma-Reiche H, Noordijk EM, Padberg GW, Voormolen JH, Hoekstra FH, Tans JT, Lambooij N, Metsaars JA, Wattendorff AR, et al. Treatment of single brain metastasis: radiotherapy alone or combined with neurosurgery? Ann Neurol 1993;33:583 – 90. [10] Mintz AH, Kestle J, Rathbone MP, Gaspar L, Hugenholtz H, Fisher B, Duncan G, Skingley P, Foster G, Levine M. A randomized trial to assess the efficacy of surgery in addition to radiotherapy in patients with a single cerebral metastasis. Cancer 1996;78:1470 – 6. [11] Mountain CF. A new international staging system for lung cancer. Chest 1986;89:225S – 33. [12] The Japan Lung Cancer Society. General Rule for Clinical and Pathological Record for Lung Cancer, 4th ed. Tokyo: Kanehara Press, 1995:24 – 30. [13] Robbins SL, Cotran RS. Differentiation and anaplasia. In: Robbins SL, Cotran RS, editors. Pathologic Basis of Disease, 2nd ed. Philadelphia, PA: Saunders, 1979:146 – 50. [14] Gehan EA. A generalized two-sample Wilcoxon test for doubly censored data. Biometrika 1965;52:650 – 3. [15] Cox DR. Regression models and life tables. J R Stat Soc B 1972;34:187 – 220. [16] Vecht CJ. Clinical management of brain metastasis. J Neurol 1998;245:127 – 31. [17] Wronski M, Arbit E, Burt M, Galicich JH. Survival after surgical treatment of brain metastases from lung cancer: a follow-up study of 231 patients treated between 1976 and 1991. J Neurosurg 1995;83:605 – 16. [18] Noordijk EM, Vecht CJ, Haaxma-Reiche H, Padberg GW, Voormolen JH, Hoekstra FH, Tans JT, Lambooij N, Metsaars JA, Wattendorff AR, et al. The choice of treatment of single brain metastasis should be based on extracranial tumor activity and age. Int J Radiat Oncol Biol Phys 1994;29:711 – 7. [19] Nakagawa H, Miyawaki Y, Fujita T, Kubo S, Tokiyoshi K, Tsuruzono K, Kodama K, Higashiyama M, Doi O, Hayakawa T. Surgical treatment of brain metastases of lung cancer: retrospective analysis of 89 cases. J Neurol Neurosurg Psychiatry 1994;57:950 – 6. [20] Torre M, Quaini E, Chiesa G, Ravini M, Soresi E, Belloni PA. Synchronous brain metastasis from lung can-
106
[21]
[22]
[23]
[24]
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106 cer. Result of surgical treatment in combined resection. J Thorac Cardiovasc Surg 1988;95:994–7. Mussi A, Pistolesi M, Lucchi M, Janni A, Chella A, Parenti G, Rossi G, Angeletti CA. Resection of single brain metastasis in non-small-cell lung cancer: prognostic factors. J Thorac Cardiovasc Surg 1996;112:146–53. Harpole D, Amos A, Alexander E, Loeffler J, Swanson S, DeCamp M, Mentzer S, Skarin A, Richards W, Black P, Sugarbaker D. Stage of the primary is important when treating isolated brain metastases from lung cancer. Proc ASCO 1996;15:382. Burt M, Wronski M, Arbit E, Galicich JH. Resection of brain metastases from non-small-cell lung carcinoma. Results of therapy. J Thorac Cardiovasc Surg 1992;103:399– 411. Macchiarini P, Buonaguidi R, Hardin M, Mussi A, Angeletti CA. Results and prognostic factors of surgery in the management of non-small cell lung cancer with solitary brain metastasis. Cancer 1991;68:300–4.
.
[25] Arbit E, Wronski M, Burt M, Galicich JH. The treatment of patients with recurrent brain metastasis. Cancer 1995;76:765 – 73. [26] Bergqvist M, Brattstrom D, Bennmarker H, Wagenius G, Riska H, Brodin O. Irradiation of brain metastases from lung cancer: a retrospective study. Lung Cancer 1998;20:57 – 63. [27] Komaki R, Cox JD, Stark R. Frequency of brain metastasis in adenocarcinoma and large cell carcinoma of the lung: correlation with survival. Int J Radiat Oncol Biol Phys 1983;9:1467 – 70. [28] Magilligan DJ Jr., Duvernoy C, Malik G, Lewis JW Jr., Knighton R, Ausman JI. Surgical approach to lung cancer with solitary cerebral metastasis: twenty-five years’ experience. Ann Thorac Surg 1986;42:360 – 4. [29] Kiss R, Rorive S, Camby I, Pasteels JL, Brotchi J, Salmon I. DNA ploidy level assessments in 83 human brain metastases. Relationship to the survival of 35 patients. J Cancer Res Clin Oncol 1996;122:127 – 31.
Prognostic factors in surgical treatment of solitary brain metastasis after resection of non-small-cell lung cancer Yukio Saitoh a,*, Takehiko Fujisawa a, Mitsutoshi Shiba a, Shigetoshi Yoshida a, Yasuo Sekine a, Masayuki Baba a, Toshihiko Iizasa a, Motoo Kubota b a
Department of Surgery, Institute of Pulmonary Cancer Research, School of Medicine, Chiba Uni6ersity, 1 -8 -1 Inohana, Chuo-ku, Chiba 260 -8670, Japan b Department of Neurosurgery, School of Medicine, Chiba Uni6ersity, Chiba 260 -8670, Japan Received 23 October 1998; received in revised form 18 March 1999; accepted 26 March 1999
Abstract Patients with brain metastasis after resection of non-small-cell lung cancer usually have poor prognosis. A few such patients, however, survive for long periods after surgical resection of brain metastases. To evaluate the prognostic factors in resection of solitary brain metastasis from non-small-cell lung cancer, we reviewed 24 cases undergoing resection of solitary brain metastasis after resection of the primary site from 1977 to 1993. The patient population consisted of 20 men and four women ranging in age from 40 to 75 years old (average, 57.8 years old). None of the patients had systemic metastasis except in the brain at the time of brain surgery. The overall survival rates were 12.5% at 3 years and 8.3% at 5 years after brain surgery. The longest survival periods were 11.5 years after brain surgery and 15.4 years after lung surgery. The interval between lung and brain surgery (5 360 days vs. \ 360 days), differentiation of primary cancer (poor vs. moderate), size of primary site ( 5 5.0 cm vs. \ 5.0 cm), and operation of primary site (lobectomy vs. pneumonectomy) significantly affected survival as shown by univariate analysis (P B 0.05). Other clinical factors (age, gender, histology, T- and N-status, ‘resectability with curative intent’ of the primary site, location of the brain metastasis and postoperative radiation therapy) did not affect survival. Multivariate analysis using Cox’s proportional hazards model indicated that an interval of more than 360 days between the two surgical procedures (hazard ratio =0.2351, P =0.0136) and lobectomy (hazard ratio= 0.5274, P =0.0416) were independent prognostic factors. In conclusion, patients with solitary brain metastasis from non-small-cell lung cancer without other organ metastasis, in whom relapse in the brain occurred more than 1 year after resection of the primary site and in whom lobectomy was performed, should be treated surgically to maximize the chance of prolonged survival. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Surgery; Brain metastasis; Non-small-cell lung cancer
* Corresponding author. Tel.: +81-43-222-7171; fax: + 81-43-226-2172. 0169-5002/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 9 - 5 0 0 2 ( 9 9 ) 0 0 0 3 4 - 3
100
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
1. Introduction
2. Materials and methods
The brain is a common target of metastasis from non-small-cell lung cancer (NSCLC), and its relapse after lung surgery is correlated with poor prognosis [1]. In pathological examinations of autopsy cases, Knights [2] reported that 53% of pulmonary carcinoma cases had brain metastases. In addition, five of 21 cases showed only intracranial metastases without metastasis in other organs, and three of these 21 cases showed solitary brain lesions. These findings showed that there may be indications for surgical resection in selected cases with brain metastasis from NSCLC. Clinically, brain metastases occur frequently as the first relapse site during the follow-up period after resection of NSCLC [3,4]. Other investigators have reported good results in symptom-free conditions and quality of life after resection of brain metastases, and a small number of these cases have shown good long-term survival [5 – 7]. Recently, two randomized trials have indicated that surgical resection of solitary brain metastasis followed by radiation therapy results in better survival rate after treatment than radiation therapy alone [8,9]. On the other hand, Mintz et al. [10] detected no difference in survival rate between radiation therapy alone and surgery plus radiation therapy. Median survival periods of groups with surgery in these studies were very short, i.e. 40 weeks [8], 12 months [9] and 5.6 months [10], respectively, and the majority of patients had early death, suggesting that not all cases with solitary brain metastasis should be treated with brain surgery. One reason for these unsatisfactory survival results was that the above three studies included many cases with extracranial metastasis at the time of brain surgery. Therefore, in addition to a lack of extracranial metastasis, other clinical factors influencing survival should be investigated prior to scheduling brain surgery. To elucidate prognostic factors for long-term survival and clarify the indications of surgical treatment for solitary brain metastasis, we reviewed the clinical records of cases treated at our institute.
2.1. Patients From August 1977, through June 1993, 1049 patients underwent resection of NSCLC at our hospital. Of these, 30 underwent combined lung surgery for the primary site and brain surgery for brain metastases. Four cases had brain surgery prior to lung surgery because of their neurological symptoms and two cases had multiple brain metastases. Twenty-four patients (2.3%), excluding the above six patients, in whom the solitary brain metastases were surgically treated during the follow-up period after lung surgery, were analyzed in this study. The resected brain tumors were confirmed pathologically to be metastases from NSCLC. These patients included 20 men and four women ranging in age from 40 to 75 years old (average, 57.8 years old).
2.2. Histological examination Histology and T, N classification were confirmed according to the criteria reported by Mountain [11] and the general rule for clinical and pathological recording for lung carcinoma, edited by the Japan Lung Cancer Society [12]. Cellular differentiation was graded as well, moderately or poorly differentiated for those patients with adenocarcinoma and squamous cell carcinoma. Large cell carcinomas were labeled as undifferentiated [13].
2.3. Statistical analysis Clinical data obtained from medical records were classified into two groups and used as variables (Table 1). Relations between clinical variables were tested by Fisher’s exact test or t-test. Survival time was measured from the date of brain surgery until death or the most-recent date of follow-up for surviving patients. Survival rate was calculated by the Kaplan–Meier method and comparisons of survival were estimated by generalized Wilcoxon test [14] for univariate analysis. Cox’s proportional hazards model (forward stepwise procedure) [15] was used for multivariate
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
analysis. Significance was defined as P B 0.05. All these analyses were performed with the SPSS software package (SPSS version 6.1 for Windows, SPSS, Chicago, IL).
3. Results
3.1. O6erall sur6i6al Of the 24 cases examined, two patients survived for 11.5 and 9.7 years following brain surgery,
101
respectively. Of the remaining 22 patients, two died due to operation-related complications within 30 days after brain surgery (8%) and 20 died due to the cancer. The causes of operationrelated death were pulmonary infarction on the seventh day and cerebral bleeding on the twentyfirst day after brain surgery. Among the cases of cancer-related death, 16 patients died within a year after brain surgery because of systemic recurrence (15 cases) and multiple brain metastases (one case), while four survived for over 1 year. The overall survival rates were 25% at 1 year,
Table 1 Clinical characteristics of 24 patients underwent surgery for brain metastasis and results of univariate analysis for prognostic factors
Gender Age (year; 57.8 9 10.3)b Interval between two operations (days; 396 9 327) Primary lung tumor Histology Differentiation
Location
T-factor Size (cm; 5.4 9 2.7) N-factor Resectability Operation Brain metastasis Location Resectability Radiation
a b
Generalized Wilcoxon test. Mean9 SD.
n
Median survival time (days)
P valuea
Male Female ]60 B60 5360 \360
20 4 12 12 12 12
203 173 216 117 117 292
0.5328
Adenocarcinoma Others Moderate Poor, undifferentiated Right Left Upper lobe Others T1 or T2 T3 55.0 \5.0 N0 N1 or N2 Complete resection Incomplete resection Lobectomy Pneumonectomy
15 9 11 13
224 203 292 134
9 15 18 6 15 9 15 9 11 13 20 4 18 6
117 254 203 173 224 203 239 102 267 189 203 134 239 75
Supratentorial Infratentorial Complete resection Incomplete resection With Without
16 8 20 4 10 14
224 173 216 189 203 173
0.1923 0.0069
0.7894 0.0152
0.063 0.7203 0.9057 0.023 0.4178 0.6386 0.0057
0.4206 0.9399 0.4374
102
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
3.3. Brain metastasis
Fig. 1. Actuarial survival according to the types of pulmonary resection. Cases undergoing lobectomy had a significantly prolonged survival.
12.5% at 3 years and 8.3% at 5 years. The median survival period was 6.8 months.
3.2. Primary site Histological types of primary lung tumors were adenocarcinoma in 15 cases (63%), squamous cell carcinoma in eight (33%), and large cell carcinoma in one case (4%). Furthermore, adenocarcinoma and squamous cell carcinoma were graded as moderately differentiated in 11 cases (46%) and poorly differentiated in 12 cases (50%). The types of pulmonary resection consisted of lobectomy in 18 cases and pneumonectomy in six. Twenty patients underwent complete resection of lung tumor and four underwent incomplete resection (one in the bronchial stump and three in the chest wall). Three cases with chest wall resection received radiation therapy of the chest wall following surgery. On pathological examination, the lymph node status was classified as N0 in 11 cases, N1 in three and N2 in ten, and the primary tumor status was classified as T1 in four cases, T2 in 11 and T3 in nine. Before lung surgery, the results of Gallium whole-body scan and bone scan indicated no extrathoracic spread of the tumor, and no clinical symptoms of brain metastasis were detected in any of the cases. Use of brain CT was introduced for routine preoperative evaluation in 1982 (19 cases). At the time of lung surgery, six cases had stage I disease, two stage II and 16 stage IIIA.
In all cases, neurological symptoms were detected and brain CT or brain MRI (ten cases since 1990) showed solitary mass lesions. Before brain surgery, chest CT showed no intrathoracic recurrence and clinically there were no signs of metastasis in other organs. The interval from lung surgery to brain surgery in 24 cases ranged from 49 to 1339 days with a mean of 396 days. Location of the brain tumor included the supratentorial region in 16 and infratentorial region in eight. All of the former and four of the latter had complete removal of the brain tumor. Four cases with infratentorial lesions underwent incomplete resection. After brain surgery, five cases received localized radiation therapy (50 Gy) and five were treated with both whole-brain radiation (30 Gy) and localized radiation (ranging from 10 to 20 Gy) therapy. No cases had systemic chemotherapy after brain surgery. Two cases showed intracranial relapse without systemic metastasis. One underwent reoperation for a supratentorial lesion 2 months after the initial treatment, and another received radiosurgery using a gamma knife for a cerebellar lesion 3 months after the first brain surgery.
3.4. Uni6ariate analysis We analyzed 15 variables to evaluate their influence on survival: i.e. age, gender, interval between lung and brain surgery, nine variables related to primary lung tumor and three variables related to brain metastasis (Table 1). Univariate analysis indicated that lobectomy (P= 0.0057, Fig. 1), an interval between the two operations longer than 360 days (P= 0.0069, Fig. 2), moderate differentiation (P=0.0152), and primary lesion less than 5 cm in maximum diameter (P=0.023) were significantly good prognostic factors for survival.
3.5. Multi6ariate analysis Six variables showing differences with PB 0.2 on univariate analysis were selected for multivariate analysis. Multivariate analysis using Cox’s proportional hazards model revealed two signifi-
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
103
4. Discussion
Fig. 2. Actuarial survival according to interval between lung and brain surgery. Cases with interval more than 360 days had a significantly prolonged survival.
cant variables (Table 2). The interval between the two operations was an independent factor with a hazard ratio of 0.2351 (P= 0.0136). The type of pulmonary resection had a hazard ratio of 0.5274 (P= 0.0416), indicating that this was another independent factor.
3.6. Clinical factors associated with inter6al between the two operations and the type of pulmonary resection The interval between the two operations was significantly associated with histological type (P= 0.018) and differentiation of the lung tumor (P= 0.003). Adenocarcinomas and moderately differentiated tumors had longer intervals between the two operations (Table 3). The type of pulmonary resection was significantly associated with differentiation of lung tumor (P= 0.0168) and tumor size (lobectomy; 4.691.9 cm, pneumonectomy; 7.99 3.2 cm, P =0.049). All of the pneumonectomy cases had poorly differentiated carcinomas.
There have been many reports concerning the factors influencing survival after combined resection of NSCLC and brain metastasis, but the results of these studies have been controversial. Cases with these combined resections are grouped into those with synchronous and metachronous presentation, or simultaneous and sequential presentation [16], based on the time of diagnosis of lung tumor and brain metastasis. In general, the former cases had resection of brain metastasis prior to resection of lung tumor, because of their neurological symptoms. The above two groups should be analyzed independently to accurately elucidate the prognostic factors. In this study of patients undergoing lung and brain surgery, those in whom brain metastases were not detected at the time of lung surgery were selected and analyzed. The results of this study provided useful information for deciding whether brain surgery should be performed when the first recurrent site of NSCLC is the brain during follow-up after lung surgery. The 3-year survival rate of 12.5%, 5-year survival rate of 8.3%, and the median survival of 6.8 months after brain surgery reported in this study are unsatisfactory from the viewpoint of prolonged survival. Two of the 24 cases examined showed long-term survival (11.5 and 9.7 years) with no recurrence and little neurological deficits following brain surgery. On the other hand, ten of the 24 patients (42%) died within 6 months and 18 (75%) died within the first year. From these observations, it is obvious that the elucidation of prognostic factors and the selection of candidates for brain surgery are of major clinical importance. Previous studies showed that age and gender [17,18], histological type of tumor [19], N-status
Table 2 Results of multivariate analysis for six prognostic variables shown to be significant (PB0.2) by univariate analysis Variables
Interval between two operations (5360 vs. \360) Type of operation (lobectomy vs. pneumonectomy)
P value
0.0136 0.0416
Hazard ratio
0.2351 0.5274
95% C.I. Lower
Upper
0.0744 0.2851
0.7427 0.9759
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
104
Table 3 Clinical factors associated with interval between the two operations Interval between two operations (days)
P valuea
Gender Male Female
385 9349b 451 9 199
N.S.
Age (year) ]60 B60
3319442 4429385
N.S.
514 9350
0.018
Histology Adenocarcinoma Others Differentiation Moderate Poor, undifferentiated
199 9148 599 9326 224 9214
0.003
Location Right Left
385 9277 402 9362
N.S.
Upper lobe Others
342 9336 558 9255
N.S.
T-factor T1 or T2 T3
425 9249 347 9 440
N.S.
Size (cm) 55.0 \5.0
481 9 363 2539 196
N.S.
N-factor N0 N1 or N2
380 9 224 409 9 402
N.S.
Resectability Complete resec- 428 9 344 tion Incomplete re234 9 161 section Operation Lobectomy Pneumonectomy a b
451 9 355 230 9 136
N.S.
N.S.
t-test. Mean9 SD.
[20 – 22] and complete or incomplete resection of the primary site [23], and supratentorial or infratentorial location of brain metastasis [17] are
significant prognostic factors. In our experience, cases with metachronous presentation after lung surgery, solitary brain lesion and no extracranial recurrence, the clinical factors described above as significant prognostic factors failed to show significant differences in survival after brain surgery. However, the number of cases examined in this study was too small to conclusively exclude the influence of the above factors on survival, especially the resectability of the primary site and the location of brain metastasis. In this study, multivariate analysis indicated that interval between brain and lung surgery was a significant prognostic factor. Mussi et al. [21] and Macchiarini et al. [24] reported similar results concerning the interval between the two operations using univariate analysis. It is important to clarify what the interval between the two operations actually reflects. In our study, this interval was associated with histological type and differentiation of the tumor, but showed no relation with local extent of primary lesion indicated by Tstatus, N-status, tumor size or ‘resectability with curative intent’. In reports of cases with long-term survival after brain surgery, the majority of tumors were histologically classified as adenocarcinomas [6]. Furthermore, in cases with recurrent brain metastasis after brain surgery [25] and in patients treated with radiation therapy alone [26], the histology of adenocarcinoma was also associated with longer survival than other types of tumor. Adenocarcinoma has a tendency to develop brain metastasis [27], but it is well known that this type of tumor shows biological heterogeneity. It seems that this interval reflects the biological characteristics of the tumor, and a longer interval between brain and lung surgery indicates low-grade malignancy of the tumor. Another significant prognostic factor selected in this study was the type of pulmonary resection. There have been a few previous reports concerning this factor. Mussi et al. [21] reported that patients who underwent lobectomy had longer survival than those who underwent pneumonectomy. Our results were consistent with these observations. Magilligan et al. [28] reported that wedge resection of small peripheral tumors showed better results than lobectomy or pneu-
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106
monectomy, but this study included cases of synchronous presentation with brain surgery performed before lung surgery. Usually, a major determinant of the type of pulmonary resection is local extent of the primary site. However, in this study this factor was associated with differentiation of lung tumor and primary tumor size. Based on these results, it seems that the biological characteristics of the tumor have the most important effect on long-term survival, and the two prognostic factors selected by multivariate analysis were considered to be good clinical indicators of the biological characteristics of the tumor. Kiss et al. [29] reported that DNA ploidy level in brain metastasis is related to patient survival. It is possible that better prognostic factors will be determined by biological investigation of carcinoma. In this study, ten patients received radiation therapy after brain surgery. There was no significant difference in survival between cases with and without radiation therapy. Since these cases were not randomized, the efficacy of postoperative radiation therapy could not be evaluated in this study. In conclusion, patients with solitary brain metastasis from non-small-cell lung cancer without metastasis to other organs, in whom relapse in the brain occurred more than 1 year after resection of the primary site and in whom the lung operation was lobectomy, should be treated surgically, with a good chance of prolonged survival.
References [1] Markesbery WR, Brooks WH, Gupta GD, Young AB. Treatment for patients with cerebral metastases. Arch Neurol 1978;35:754 –6. [2] Knights EM Jr. Metastatic tumors of the brain and their relation to primary and secondary pulmonary cancer. Cancer 1954;7:259 – 65. [3] Figlin RA, Piantadosi S, Feld R. Intracranial recurrence of carcinoma after complete surgical resection of stage,, and non-small-cell lung cancer. New Engl J Med 1988;318:1300– 5. [4] Yano T, Yokoyama H, Inoue T, Asoh H, Tayama K, Takai E, Ichinose Y. The first site of recurrence after complete resection in non-small-cell carcinoma of the lung. Comparison between pN0 disease and pN2 disease. J Thorac Cardiovasc Surg 1994;108:680–3.
105
[5] Sarma DP, Weilbaecher TG. Long-term survival after brain metastasis from lung cancer. Cancer 1986;58:1366 – 70. [6] Shahidi H, Kvale PA. Long-term survival following surgical treatment of solitary brain metastasis in non-small cell lung cancer. Chest 1996;109:271 – 6. [7] Kelly K, Bunn PA Jr. Is it time to reevaluate our approach to the treatment of brain metastases in patients with non-small cell lung cancer? Lung Cancer 1998;20:85 – 91. [8] Patchell RA, Tibbs PA, Walsh JW, Dempsey RJ, Maruyama Y, Kryscio RJ, Markesbery WR, Macdonald JS, Young B. A randomized trial of surgery in the treatment of single metastases to the brain. New Engl J Med 1990;322:494 – 500. [9] Vecht CJ, Haaxma-Reiche H, Noordijk EM, Padberg GW, Voormolen JH, Hoekstra FH, Tans JT, Lambooij N, Metsaars JA, Wattendorff AR, et al. Treatment of single brain metastasis: radiotherapy alone or combined with neurosurgery? Ann Neurol 1993;33:583 – 90. [10] Mintz AH, Kestle J, Rathbone MP, Gaspar L, Hugenholtz H, Fisher B, Duncan G, Skingley P, Foster G, Levine M. A randomized trial to assess the efficacy of surgery in addition to radiotherapy in patients with a single cerebral metastasis. Cancer 1996;78:1470 – 6. [11] Mountain CF. A new international staging system for lung cancer. Chest 1986;89:225S – 33. [12] The Japan Lung Cancer Society. General Rule for Clinical and Pathological Record for Lung Cancer, 4th ed. Tokyo: Kanehara Press, 1995:24 – 30. [13] Robbins SL, Cotran RS. Differentiation and anaplasia. In: Robbins SL, Cotran RS, editors. Pathologic Basis of Disease, 2nd ed. Philadelphia, PA: Saunders, 1979:146 – 50. [14] Gehan EA. A generalized two-sample Wilcoxon test for doubly censored data. Biometrika 1965;52:650 – 3. [15] Cox DR. Regression models and life tables. J R Stat Soc B 1972;34:187 – 220. [16] Vecht CJ. Clinical management of brain metastasis. J Neurol 1998;245:127 – 31. [17] Wronski M, Arbit E, Burt M, Galicich JH. Survival after surgical treatment of brain metastases from lung cancer: a follow-up study of 231 patients treated between 1976 and 1991. J Neurosurg 1995;83:605 – 16. [18] Noordijk EM, Vecht CJ, Haaxma-Reiche H, Padberg GW, Voormolen JH, Hoekstra FH, Tans JT, Lambooij N, Metsaars JA, Wattendorff AR, et al. The choice of treatment of single brain metastasis should be based on extracranial tumor activity and age. Int J Radiat Oncol Biol Phys 1994;29:711 – 7. [19] Nakagawa H, Miyawaki Y, Fujita T, Kubo S, Tokiyoshi K, Tsuruzono K, Kodama K, Higashiyama M, Doi O, Hayakawa T. Surgical treatment of brain metastases of lung cancer: retrospective analysis of 89 cases. J Neurol Neurosurg Psychiatry 1994;57:950 – 6. [20] Torre M, Quaini E, Chiesa G, Ravini M, Soresi E, Belloni PA. Synchronous brain metastasis from lung can-
106
[21]
[22]
[23]
[24]
Y. Saitoh et al. / Lung Cancer 24 (1999) 99–106 cer. Result of surgical treatment in combined resection. J Thorac Cardiovasc Surg 1988;95:994–7. Mussi A, Pistolesi M, Lucchi M, Janni A, Chella A, Parenti G, Rossi G, Angeletti CA. Resection of single brain metastasis in non-small-cell lung cancer: prognostic factors. J Thorac Cardiovasc Surg 1996;112:146–53. Harpole D, Amos A, Alexander E, Loeffler J, Swanson S, DeCamp M, Mentzer S, Skarin A, Richards W, Black P, Sugarbaker D. Stage of the primary is important when treating isolated brain metastases from lung cancer. Proc ASCO 1996;15:382. Burt M, Wronski M, Arbit E, Galicich JH. Resection of brain metastases from non-small-cell lung carcinoma. Results of therapy. J Thorac Cardiovasc Surg 1992;103:399– 411. Macchiarini P, Buonaguidi R, Hardin M, Mussi A, Angeletti CA. Results and prognostic factors of surgery in the management of non-small cell lung cancer with solitary brain metastasis. Cancer 1991;68:300–4.
.
[25] Arbit E, Wronski M, Burt M, Galicich JH. The treatment of patients with recurrent brain metastasis. Cancer 1995;76:765 – 73. [26] Bergqvist M, Brattstrom D, Bennmarker H, Wagenius G, Riska H, Brodin O. Irradiation of brain metastases from lung cancer: a retrospective study. Lung Cancer 1998;20:57 – 63. [27] Komaki R, Cox JD, Stark R. Frequency of brain metastasis in adenocarcinoma and large cell carcinoma of the lung: correlation with survival. Int J Radiat Oncol Biol Phys 1983;9:1467 – 70. [28] Magilligan DJ Jr., Duvernoy C, Malik G, Lewis JW Jr., Knighton R, Ausman JI. Surgical approach to lung cancer with solitary cerebral metastasis: twenty-five years’ experience. Ann Thorac Surg 1986;42:360 – 4. [29] Kiss R, Rorive S, Camby I, Pasteels JL, Brotchi J, Salmon I. DNA ploidy level assessments in 83 human brain metastases. Relationship to the survival of 35 patients. J Cancer Res Clin Oncol 1996;122:127 – 31.