In their article, Chao and Goldberg provide a concise overview of the literature on pulmonary metastasectomy for sarcoma, including a brief history of the procedure, guidelines for preoperative evaluation, conduct of the operation, and probable outcomes achieved. Several points that they review deserve further discussion.
Thoracoscopy vs Thoracotomy
First, as the authors mention, it is unclear which operation should be performed in the patient with radiologically resectable disease. The standard approach of the Thoracic Surgery Service at Memorial Sloan-Kettering Cancer Center (MSKCC) is only to explore a hemithorax if radiographs convincingly suggest that at least one metastasis will be found. Thus, if only unilateral disease is evident, a unilateral thoracotomy is performed, and if bilateral disease is evident, either sequential posterolateral thoracotomies or a bilateral concurrent approach is chosen.
We have not considered thoracoscopic resection of all radiographic disease appropriate, because the available data suggest that manual palpation of the lung will reveal further disease in approximately 30% of patients. On the other hand, if it is so important to detect and resect all disease, why not explore the contralateral side as well? It has been shown that contralateral exploration will also find disease not evident on x-ray in approximately 30% of patients, and if we are willing to leave contralateral disease until it becomes evident radiographically, why is it inappropriate to leave it on the ipsilateral side in one-third of patients? Cancer and Leukemia Group B (CALGB) trial 39804, a randomized prospective comparison of thoracoscopy and thoracotomy, is underway and may help answer these questions.
Retrospective Data Reliable in Guiding Surgical Decisions?
There are few data to support which approach is best and all of the data supporting surgical resection of pulmonary metastases by any approach are retrospective and, therefore, subject to bias. The retrospective nature of the data is clouded further by the fact that reports often mix tumors of varied histologies, often fail to provide information on adjuvant therapies, and/or do not provide information on the denominator of all patients with pulmonary recurrence from which the treated patients are drawn.
It is entirely possible that patients with resectable (and re-resectable) lesions are a subset with favorable (ie, less aggressive) disease who may have enjoyed prolonged survival without surgery. The most coherent argument of this viewpoint is made in an editorial by Aberg.
Reliance on retrospective data to support claims of efficacy is not limited to metastasectomy for sarcoma, but also guides metastasectomy for any histology, including such common diseases as colorectal cancer. Proof that metastasectomy improves survival for any histology would probably require a prospective randomized trial, which is unlikely to be conducted given the belief that the retrospective data are strong enough to make investigators reluctant to randomize patients to an arm without surgery.
Even if such a trial were to be initiated for sarcoma, the limited number of patients with resectable pulmonary metastases (approximately 500 per year in the United States) would make accrual difficult. In addition, the varied sarcomatous histologies (eg, liposarcoma, malignant fibrous histiocytoma, angiosarcoma) are often lumped together because of the limited number of patients available. As a result, it is extremely difficult to identify subgroups within the total group of patients with pulmonary sarcomatous metastases who are most or least likely to benefit from resection.
Benefit Seen in a Small Number
Even if it were demonstrated that lung resection for sarcomatous metastases definitely improved the likelihood of survival, the number of patients truly afforded benefit is small. Data from the prospectively maintained MSKCC Sarcoma Database suggest that of 3,149 patients seen over approximately 15 years, 719 (23%) will either present with or develop lung metastasis. Of these, 248 (8% of all patients, and 34% of patients with metastases) have undergone complete resection and of these 248, approximately 60% will suffer pulmonary re-recurrence. Half will undergo a second resection, and another half of this half will be resected further for re-recurrent disease. Overall, therefore, from the time the diagnosis of pulmonary metastases is first made, only 10% of all patients with pulmonary metastases will benefit from this tratment.
Patients are most likely to suffer rerecurrent disease in the lung, presumably developing from occult micrometastatic disease present at the time of first resection. There is not a great deal of information available on adjuvant or neoadjuvant therapy to eradicate such residual disease, and certainly no treatment of proven efficacy. The European Organization for Research and Treatment of Cancer (EORTC) protocol 62933 is currently randomizing patients to either metastasectomy alone, or induction chemotherapy followed by metastasectomy. The trial has been open since 1996 but to date has accrued only a few patients.
Current Efforts and Future Directions
Current efforts at Memorial Sloan-Kettering are focusing on intensifiying chemotherapy within the lung by minimally invasive techniques (regional lung infusion), after a phase I trial of an intraoperative perfusion circuit to deliver and retrieve chemotherapeutic agents has led to an unacceptably high loss of pulmonary function. If phase I studies of regional lung infusion prove safe, we plan to perform phase II and III studies of regional lung infusion as adjuvant therapy after complete resection.
In the meantime, given the low long-term survival after surgery alone for pulmonary metastases, any efforts investigating more conventional adjuvant or neoadjuvant therapies should be encouraged and supported. Such trials are well worth the effort, because even if sarcoma is an orphan disease affecting relatively few patients each year, the techniques developed are likely to be applicable to patients with metastases of more common histologies, such as colon cancer.