Lily S. Lau MD, I. Benjamin Paz MD, Frederic W. Grannis Jr. MD, Joshua D.I. Ellenhorn MD, Chul Ahn Ph.D., Lawrence D. Wagman MD
Division of General and Oncologic Surgery, Thoracic Surgery and Department of Biostatistics
City of Hope National Medical Center, Duarte, California 91010
Please address all correspondence to I. Benjamin Paz MD at City of Hope National Medical Center 1500 E. Duarte Road Duarte, CA 91010 (818) 359-8111 ext. 2656 Fax (818( 301-8260
The median follow-up was 28 months, with a range of one to 149 months. The median overall survival for this group of patients was 50 months with a median disease-free interval of 39 months.
By multivariate analysis, only tissue invasion was predictive of decrease in overall survival. The other variables studied (age at time of lung metastasis, margins of resection, tumor differentiation, lymph node involvement, number of metastases, size of metastases, tumor histology, prior history of metastases, evidence of metastases at another site, bilateral metastases, and adjuvant therapy) were not predictive of overall survival or of disease-free interval.
On the basis of the survival data presented, we advocate resection of pulmonary metastases in selected patients.
The most difficult issue in these patients is that of selection (i.e. taking those patients with a potential for cure to operation, while excluding those without hope of cure). Multiple studies have demonstrated the varying importance of different prognostic factors in predicting survival. Putnam et al evaluated 487 patients and found that survival was significantly correlated with tumor doubling time, number of metastases and disease free survival . A subsequent study from Stanford University demonstrated that the size of the largest metastases as well as the disease free interval predicted survival post-op, but the number of metastases did not. At NCI, survival curves were affected most by the ability to render the patient disease free at the time of surgery.
We have reviewed our experience at City of Hope in an attempt to evaluate the potential for long term survival and to investigate possible parameters predictive of success or failure.
General criteria for patients undergoing pulmonary metastasis resection were:
Metastasectomy was performed either through a posterolateral thoracotomy or a median sternotomy. Lesions were removed whenever possible with wedge resection. If necessary, lobectomy or pneumonectomy was performed. At the time of surgery all clinically evident disease was resected if possible. There was one post-operative death following pneumonectomy.
In addition to the clinical factors that may affect survival, histologic considerations were also addressed. Pathology reports verified the tumor histology, tumor differentiation, margins of resection, and tumor invasion of the tissues submitted.
The median follow-up was 28 months (range 1-149 months). Survival curves for time to lung recurrence, time to any recurrence and time to death were estimated using Kaplan-Meier product limit estimator. Cox regression analyses were used to see if time to death, any recurrence, or lung recurrence was significantly associated with prognostic variables.
A lateral thoracotomy approach was utilized in 90% of cases, sternotomy in 10%. Wedge resections alone were done in 72% of cases. 32 patients had lobectomy. Ninety six patients had unilateral metastases, while 29 had bilateral lesions. The majority of the lesions were solitary (64% n=81)); 18 patients had four or more lesions at the time of diagnosis. Over 95% of the margins of resection were negative. Metastases in other areas (liver, nodes) were noted in 15 patients. Lymph node biopsy or hilar and mediastinal node dissection were performed in 61 patients. Forty-three (34%) had negative nodes and 18 (14%) had positive nodes. Tumor invasion was noted in 12 patients (10%) of cases. Tumor histology and differentiation were obtained from the pathological reports. Surgical resection was followed by adjuvant therapy in 41% (n=51) of the patients. This included chemotherapy and hormonal therapy or both. Chemotherapy was given as adjuvant therapy to patients in each of the primary cancer groups. Hormonal therapy with or without chemotherapy was given only to breast cancer patients.
Clinical and pathological variables were evaluated as prognostic indicators of survival. (Table 2) The only variable that was predictive of overall survival was tumor invasion as noted in the pathologic specimen. (p-0.014) (Figure 1)
The five year disease-free survival (DFS) and the overall survival (OS) are 41% and 48% respectively. (Figure 2, Table 1) In breast cancer, the OS was 49%; while the DFS was 35%. The DFS and OS for colorectal cancer in this study was 41% and 48% . Soft tissue sarcomas had a DFS of 9% and an OS of 29%. The DFS and OS for osteosarcoma was 33% and 51%. In renal cell carcinoma, the DSF and OS were 50% and 45%. The DFS and OS for testicular carcinoma are 66% and 75%. The median DFS of renal and testicular cancer as well as the median OS of bone tumors and testicular cancer had not been reached yet because of their better survival.
Other malignancies made up 19% of cases. Each primary tumor site including uterus (n=6), ovary (n=2), melanoma (n=6), head and neck (n=4) and bladder (n=3) had too small a number of cases to derive accurate statistical information.
We prefer posterolateral thoracotomy for resection of pulmonary metastases unless bilateral metastases on CT scan are all anteriorly located, especially on the left side, where access to posterior portions of the lower lobe can be quite difficult. In Roth's series, five year survival was lower by approximately 10% in the median sternotomy group than in the sternotomy group. The difference was not statistically significant.
The DSF and OS in our patients is comparable to that of other studies previously published. General OS in these studies ranged between 35 and 48.2%.5, 9 Resection of metastases from primary breast cancer produced OS of 36 to 49.5%.12 In the case of colorectal cancer, OS ranged from 38-47%. Cerfolio et al had a five year OS of 35.9% in patients with metastatic renal cell carcinoma. Survival is lower (OS 22--29%) in patients with soft tissue sarcomas. The highest cure rates have been obtained in patients with testicular cancer. Two studies have described 5 year OS of 59%. The primary reason for good survival is effective chemotherapy. Surgical resection is an adjuvant treatment which clears residual teratomas and helps to guide salvage chemotherapy.
Clinical and pathological factors were evaluated in this study to determine if any predicted improved OS and DSF after resection of lung metastases. Previous studies pointed to increased number of metastases, increased size of metastases, short disease free interval and unresectable disease as unfavorable prognostic factors. In this study, tumor invasion, found in the surgical specimen, was the only adverse predictor of OS in the multivariate statistical analysis. Failure to prove other adverse variables may be due to small sample size, limiting the ability to detect significant differences. In addition, our study consisted of a heterogeneous group of cancers whose biological behaviors may differ from one another. Each tumor type may have different variables that could affect OS. Complete resection of disease, number of metastases and disease free interval have been found important in soft tissue sarcoma. The number of metastases was an important predictor in testicular cancer patients. Liver metastases may be the most important variable in colon cancer patients in determining survival with or without lung metastases. The DFS and OS attained after lung resection, regardless of primary cancer, in a carefully selected group of patients (with controlled primary tumors, resectable metastases and a low risk of surgery) demonstrates improved survival over historical controls. None of the various prognostic factors we studied provides the surgeon with an absolute indicator of when to offer or withhold surgical treatment.
Another major challenge that we face is in the education of primary physicians of the appropriate indications for surgery in patients with pulmonary metastases.
Finally, information regarding treatment of the various primary tumors is in a constant state of evolution. The thoracic surgeon must work in close cooperation with general and oncologic, orthopedic, urologic and gynecologic surgeons as well as pediatric and medical oncologists and radiation oncologists in the interests of optimizing care of patients with pulmonary metastases. REFERENCES:
