In this article, we provide an extensive review of patient selection criteria and surgical approaches, as well as of controversies regarding resection for metastatic sarcoma.
Sarcoma metastasizes to the lungs in 20% to 40% of patients and in most cases does not involve any other organ. Systemic chemotherapy is of unproven benefit for stage IV sarcoma. Retrospective studies have shown 5-year survival rates of 21% to 38% with wedge resection of metastatic pulmonary nodules, and up to 30% to 40% of patients survive an additional 5 years with repeated metastasectomy. In this article, we provide an extensive review of patient selection criteria and surgical approaches, as well as of controversies regarding resection for metastatic sarcoma.
Sarcomas may arise from nearly any embryonic mesodermal tissue and constitute approximately 1% of all newly diagnosed malignancies annually. Through hematogenous dissemination, rates of metastatic disease to the lungs have been reported to be as high as 20% for osteogenic sarcoma and 40% for soft-tissue sarcoma, and pulmonary metastases are present in most patients who die from sarcomatous malignancies. Furthermore, the pulmonary bed often is the only site of distant disease.[1-3] Sarcomas are the second most common cause of pulmonary metastatic disease, following colorectal cancer. The youthful demographic most often affected by sarcoma warrants an aggressive approach to metastases.
Tumor progression from the primary site to clinically detectable pulmonary metastases involves multiple complex steps, including vascular invasion, survival without an extracellular matrix (anoikis), immune system evasion, extravasation into the target organ, dormancy followed by proliferation, and angiogenesis. This pathway applies to all solid tumors, yet certain solid tumors preferentially metastasize to specific organs. Attempts to characterize this phenomenon date back to Paget in the late 19th century, with his proposal of a “seed” and “soil” theory. Paget theorized that certain elements within the tumor cell (seed) have an affinity for elements in the target organ (soil). This was challenged by Ewing in the early 20th century, who maintained that target organ anatomic variables were more critical. Ewing postulated that the frequency of lung metastases was related to the vast capillary system that could snare tumor cells, and the high oxygen tension that could support their growth. Although both theories are likely relevant to the lung as a primary site of metastases, more recent data evaluating pulmonary and tumor chemokine production and receptors have suggested that the “seed” and “soil” theory may have a more significant impact on the development of pulmonary metastases. Specifically, a primary chemokine produced by the lungs is CXCL12, and its cognate receptor is CXCR4. In both breast cancer and melanoma, the overexpression of the receptor CXCR4 has been associated with a dramatic increase in the rate of pulmonary metastases.[5,6] CXCR4 overexpression has also been demonstrated in soft-tissue sarcoma and seems to correlate with the targeting of metastases to lung and bone. Blocking this chemokine-receptor interaction represents a potentially clinically useful intervention for preventing metastases, but it likely would have little to no impact on existing lung lesions.
Overall Actuarial Survival After Lung Metastasectomy: Complete Resection Vs Incomplete ResectionFIGURE 2
Survival of Four Prognostic Groups of Patients With Lung Metastases
Traditionally, metastatic solid tumor treatment consists of systemic chemotherapy. In soft-tissue sarcoma, anthracycline-based regimens are considered standard of care, yet there has never been a prospective randomized controlled trial comparing this to best supportive care. Pulmonary resection of secondary lung tumors was first performed over 125 years ago. Numerous retrospective studies have shown better than expected outcomes with surgical resection when compared to no treatment or chemotherapy only, but these studies are subject to substantial selection bias. To date, there have been no randomized controlled trials evaluating the effect of pulmonary metastasectomy on survival. The most convincing evidence of the benefit of this approach came from an analysis of the International Registry of Lung Metastases. Pastorino and colleagues retrospectively reviewed 5,206 cases of pulmonary metastasectomy accrued to the registry from 1991 to 1995 from 9 different countries for a variety of cancers. Of the cases in this series, 42% were soft-tissue sarcoma. A stepwise decrease in 5-year survival was seen with the number of metastases: 43% with a single metastasis, 34% with two or three metastases, and 27% with four or more metastases. Additionally, the disease-free interval between resection of the primary tumor and detection of pulmonary metastasis was directly correlated with survival. Resectability, however, was the most important determinant of survival: overall median survival was 35 months in patients who had undergone complete resection and only 15 months in those patients who had an incomplete resection (Figures 1 and 2). These data can be compared to historical outcomes of best supportive care, with a median survival of 8 to 13 months,[9,10] and of chemotherapy only, with a median survival of approximately 19 to 25 months.
Similar findings specific to soft-tissue sarcoma were seen at Memorial Sloan-Kettering Cancer Center, where 719 patients with lung metastases from soft-tissue sarcoma were evaluated. The median survival for patients who had undergone complete resection was 33 months, compared with 11 months for those patients treated with nonoperative therapy. The investigators also showed that the disease-free interval and, most importantly, resectability had significant roles in survival. A group at MD Anderson Cancer Center had similar findings: a median survival of 25 months was seen after complete resection vs 6 months in unresectable disease. The MD Anderson investigators also found that the number of nodules detected by preoperative computed tomography (CT) had significant prognostic value. Although Girard et al also found the number of pulmonary metastases to be a significant prognostic factor, survival rates after metastasectomy were still acceptable with multiple pulmonary metastases. They concluded that, if metastases were fully resectable, the number of lesions should have limited influence on the decision to proceed with resection. A short period of observation to ensure stability was suggested for situations in which numerous lesions are found. Blackmon et al evaluated 234 patients with pulmonary metastases from sarcoma and found a similar median survival-35.5 months. Interestingly, the authors also found that the presence of extrapulmonary metastases resected either previously or synchronously did not appear to decrease survival (median survival, 37.5 months). The current National Comprehensive Cancer Network (NCCN) guidelines for soft-tissue sarcoma recommend a broad range of choices for patients with single-organ metastatic disease: metastasectomy ± chemotherapy ± radiation; chemotherapy alone; stereotactic radiation or external beam radiation alone; or observation.
Osteogenic sarcoma often responds poorly to systemic therapies and rarely spreads to organs other than the lungs. Thus, metastasectomy plays an important role in this particular histology. Outcomes are significantly improved with resection of all sites of disease compared with any other combination of treatments. Bricolli et al have shown that even repeat metastasectomy for recurrent pulmonary lesions yields excellent outcomes, with 3-year and 5-year event-free survivals of 33% and 32%, respectively.
Patient Selection Checklist for Pulmonary Metastasectomy
Most patients are asymptomatic at presentation, with pulmonary metastatic disease found on routine surveillance imaging. In the fewer than 10% of patients who do present with symptoms, the most frequent complaints are hemoptysis, spontaneous pneumothorax, postobstructive pneumonia, and cough. In these patients, strong consideration should be given to surgical intervention even if it is of more palliative than curative intent.
Not all patients with sarcomatous pulmonary metastases will benefit from resection. Identification of negative prognostic variables through an appropriate history taking and adjunctive testing is essential. Generally, to be considered a candidate for resection, the primary tumor must be controlled, there must be no other distant unresectable disease, and it must be possible to achieve complete resection of the metastatic burden. There are, therefore, several key questions to ask when evaluating a patient for metastasectomy (Table 1):
An unresectable, or an incomplete (R2) resection of the primary tumor is generally considered a contraindication to pulmonary metastasectomy. Similarly, local recurrence should be addressed successfully before embarking on resection of metastatic lesions.
Although there are no large studies evaluating this question, Blackmon et al showed that if the distant disease can be sequentially or synchronously resected, outcomes similar to those seen with only pulmonary metastases can be achieved. This remains an area of controversy. However, given the generally young, fit patient population who present with extremity sarcomas, an aggressive approach is often warranted.
The disease-free interval (DFI) refers to the time from complete resection of the primary tumor to the diagnosis of metastases. DFI has been shown repeatedly to be an important prognostic factor. In the International Lung Metastases Registry, the cutoff point for a favorable DFI was 36 months. However, there is no consensus regarding a minimal DFI prior to pulmonary metastasectomy. In fact, synchronous presentation of both a primary sarcoma and lung lesions is not a contraindication to metastasectomy, presuming all gross disease can be resected in a staged fashion; however, such a presentation portends more aggressive tumor biology, and this should be taken into consideration in a patient who may be a marginal candidate for surgery. Additionally, for patients with synchronous disease or a short DFI, the use of systemic therapy prior to metastasectomy should be discussed in a multidisciplinary setting. This builds in a period of observation to determine whether the tumor is progressing rapidly and will be widely metastatic in a period of weeks to months, in which case the patient would not benefit from pulmonary resection.
Two Examples of Pulmonary Metastases From Sarcoma
Pulmonary function tests are mandatory prior to lung surgery, and the usual guidelines for operability should be followed. Depending on age and comorbidities, cardiac testing may be appropriate as well. Removing two or three small peripheral nodules by wedge resection will have little impact on respiratory function status in most patients. A parenchyma-sparing approach is always preferred, especially given the frequency of recurrent disease in the lungs, often warranting multiple lung resections over several years. Anatomic resections, especially pneumonectomy, should be reserved for cases in which no other option for complete resection exists. Surgery is contraindicated when there is invasion of unresectable structures, or if all gross disease cannot be removed (eg, miliary disease). Ideal characteristics that allow maximal lung-sparing surgery include: unilateral and peripheral location, small tumor size, and a limited number of metastases. In contrast, large tumors, central masses, or miliary disease would be less likely to result in adequate postoperative pulmonary function (Figure 3).
There are no randomized studies comparing the effectiveness of chemotherapy to metastasectomy in sarcoma. Because all of these patients are by definition stage IV, consideration of systemic chemotherapy is recommended. This is particularly important in patients with high-grade histology and aggressive tumors. Doxorubicin and ifosfamide have become the chemotherapeutics of choice for adjuvant therapy in metatstatic soft-tissue sarcoma, which may explain the improved outcomes seen with metastasectomy over the past decade. However, controversy still remains; several studies have shown no benefit for chemotherapy, and one study has indicated that it is not cost-effective compared with metastasectomy. For patients with osteogenic sarcoma or alveolar soft parts sarcoma, in which systemic therapies are known to be less effective-and in patients who may not tolerate chemotherapy toxicities-metastasectomy may be a particularly useful alternative.
Preoperative pulmonary function testing and assessment of functional status are vital to determining how much lung can safely be resected. If there is concern, a quantitative pulmonary perfusion scan may be of benefit and often is necessary with repeat resections. Patients treated with bleomycin may be at higher risk for postoperative acute respiratory distress syndrome (ARDS) and pulmonary fibrosis and thus warrant more careful evaluation. Cardiac testing should be considered based on age and comorbidities; in addition, adriamycin cardiotoxicity can be of concern regardless of age. A CT scan of the chest with fine cuts and IV contrast should be obtained on all patients to assess the number, size, and location of the pulmonary metastases. Positron emission tomography (PET)/CT scans are frequently performed in cancer patients, but their routine use in the evaluation of sarcoma is controversial. A retrospective review of patients evaluated at the Mayo Clinic found that PET scans were only positive for 44% of malignant nodules in patients with sarcoma, making the utility of such scans in this group questionable. However, another study indicated that PET/CT scans altered management in 14% of all oncologic patients under consideration for lung resection for metastatic disease, perhaps by identifying other, unsuspected sites of disease. In our experience, many patients are referred for surgery after they have already had a PET/CT scan. If a PET/CT scan has not been done, it is up to the discretion of the surgeon whether or not to obtain one. In any patient with neurologic symptoms, a brain magnetic resonance imaging (MRI) scan should be obtained. Finally, flexible bronchoscopy is an important step before lung resection commences, since occult endobronchial disease has been reported in up to 3% of patients.
Surgical Approaches for Pulmonary Metastasectomy
Many options exist for the resection of pulmonary metastases. Unilateral disease can be approached via a thoracotomy or video-assisted thoracoscopic surgery (VATS); bilateral disease can be approached via a clamshell incision, median sternotomy, or sequential thoracotomies (Figure 4).
The advantage of open resection includes excellent visualization, greater ability to preserve lung parenchyma, and perhaps most importantly, the ability to palpate for nodules. Parsons and others have shown that in up to 46% of patients, preoperative CT scans failed to predict the number of nodules found at thoracotomy.This was also seen by the University of Alabama group, who found that 1 in 5 patients had ipsilateral non-imaged malignant metastases found at thoracotomy. In the most elegant study to date evaluating the importance of open lung palpation, a prospective trial from Memorial Sloan-Kettering Cancer Center, patients were preoperatively evaluated with a CT scan and subsequently underwent VATS pulmonary metastasectomy, but they then were immediately converted to a thoracotomy to evaluate for missed nodules. The investigators found that 10 of the first 18 patients had lesions missed by CT scan and VATS alone. Because of the dramatic results, the study was closed early.
Despite these findings, VATS still offers perceived advantages over open techniques, with less pain and shorter hospital stays. In lung cancer patients, compliance with adjuvant therapy regimens is greatly improved, and this may be relevant in sarcoma patients as well. Reoperation after VATS is considered less morbid and generally “easier” on patient and surgeon, again a relevant concern given the frequency of new metastases years after initial resection. For all these reasons, VATS is appealing to both patients and surgeons. Proponents of open metastasectomy argue that missed nodules will guarantee the need for reoperation in the future, yet there is no evidence to support the contention that the lesions missed by VATS (generally < 5 mm) are of any substantial clinical significance. In several studies, survival outcomes do not seem to be negatively impacted by the use of VATS for metastasectomy.[24,25] Perhaps a compromise exists between these two opposing viewpoints: use of a subxiphoid port that can allow the benefits of VATS with the possibility of manual lung palpation.[26,27]
We generally reserve thoracoscopy for patients with solitary, peripheral lesions that are large enough to localize with instrument palpation. Patients are counseled extensively about the risk of missed additional lesions with a VATS approach compared to thoracotomy. Careful postoperative surveillance is recommended in either case to check for local recurrence, new lesions, or missed lesions that have enlarged.
Although bilateral metastases can be approach via either sequential VATS or thoracotomies, they often require a recovery period of 4 to 8 weeks between operations. Median sternotomy and clamshell incisions have the advantage of a single trip to the operating room. The benefit of a median sternotomy incision over bilateral thoracotomy or a clamshell incision is less postoperative pain and perhaps fewer postoperative pulmonary complications. Unfortunately, with sternotomy, exposure of posterior or hilar lesions and left lower lung fields is often suboptimal. In comparison, a clamshell incision offers excellent access to all lung fields, but it results in significant postoperative pain and functional impairment. Additionally, a clamshell incision requires the sacrifice of both internal mammary arteries. It is important to know the advantages and shortcomings of each incision in order to individualize recommendations to patients.
Pulmonary metastases from carcinomas are occasionally accompanied by, or perhaps lead to, mediastinal lymph node metastases. While consideration of this possibility can be clinically important in carcinomas, it is rarely seen with sarcoma.[28,29] Therefore, lymphadenectomy is not routinely recommended during resection of pulmonary metastases from sarcoma.
Five-year survival for both soft-tissue and osteogenic sarcoma after complete resection of lung metastases ranges from 21% to 38%.[11-13,16] Pastorino et al found that approximately 60% of patients had recurrent metastatic disease, and about two-thirds of these recurrences were intrathoracic. More than 50% of patients underwent a second metastasectomy, with 44% surviving 5 additional years. Similar results were reported by Briccoli and et al, who showed a 5-year survival of 32% from repeat metastasectomy. Given these outcomes with repeat resection, it is logical that close radiologic follow-up is useful to identify new lung nodules before they become inoperable. NCCN guidelines suggest chest CT or plain chest radiography every 3 to 6 months for 2 to 3 years, then every 6 months for 2 years, then annually.
Several new alternatives and adjuncts to surgery have been proposed, including stereotactic radiosurgery, radiofrequency ablation (RFA), and isolated lung chemotherapeutic perfusion. Stereotactic radiosurgery has previously been used for patients with inoperable primary lung tumors. A study evaluating its use in 35 patients with pulmonary metastases showed that 77% were alive at a median of 18 months. At this time, it is not considered a substitute for surgery, but rather an alternative when patients are unfit for surgery or have had multiple reoperations and do not have adequate lung function for further surgery. RFA has also been used in situations in which surgery was felt to be too great a risk. One report evaluated 47 metastases in 29 nonsurgical patients. It reported 1- and 3-year survival rates of 92% and 65%. Despite these promising early results, there are significant limitations to RFA. Centrally located nodules are not generally amenable to RFA, especially when lesions are in close proximity to large vessels. Additionally, because margins are not attainable and there is usually a residual mass (scar tissue, necrotic tumor, or perhaps residual tumor), it is difficult to determine whether local control has been achieved.
Isolated lung perfusion utilizes chemotherapeutic agents infused directly into the pulmonary vascular bed, allowing local drug levels that would be toxic systemically. Promising results have previously been seen in animal studies. A phase I clinical trial in humans illustrated the safety and feasibility of this approach and its potential to at least stabilize disease. Multiple drugs alone or in combination can be used with this modality, including doxorubicin, cisplatin, tumor necrosis factor alpha, interferon gamma, and melphalan. Pending phase II trials will help determine whether this modality has a role in the treatment of sarcomatous pulmonary metastases.
Mentioned in This Article
Tumor necrosis factor alpha
Brand names are listed in parentheses only if a drug is not available generically and is marketed as no more than two trademarked or registered products. More familiar alternative generic designations may also be included parenthetically.
While there is a lack of phase III studies comparing pulmonary metastasectomy to other options in metastatic sarcoma, substantial retrospective data support a survival advantage for aggressive resection in these patients. Careful preoperative assessment both physiologically and oncologically is key to identifying those patients most likely to benefit from resection of sarcomatous lung metastases. Controversy persists regarding the optimal surgical approach; however, VATS is frequently preferred by patients and surgeons and is probably acceptable in cases with a low number of nodules that are periphally located. For medically inoperable patients, RFA or stereotactic radiosurgery provide some degree of local control and represent a reasonable alternative to surgery. The use of chemotherapy remains unproven in terms of survival benefit; optimal timing in relation to metastasectomy is also unknown. Review of complex cases in a multidisciplinary setting is highly recommended.
Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.
1. Kager L, Zoubek A, Potschger U, et al. Primary metastatic osteosarcoma: presentation and outcome of patients treated on neoadjuvant Cooperative Osteosarcoma Study Group protocols. J Clin Oncol. 2003;21:2011-8.
2. Coindre JM, Terrier P, Guillou L, et al. Predictive value of grade for metastasis development in the main histologic types of adult soft tissue sarcomas: a study of 1240 patients from the French Federation of Cancer Centers Sarcoma Group. Cancer. 2001;91:1914-26.
3. Raney RB, Jr., Tefft M, Maurer HM, et al. Disease patterns and survival rate in children with metastatic soft-tissue sarcoma. A report from the Intergroup Rhabdomyosarcoma Study (IRS)-I. Cancer. 1988;
4. Krishnan K, Khanna C, Helman LJ. The molecular biology of pulmonary metastasis. Thorac Surg Clin. 2006;16:115-24.
5. Muller A, Homey B, Soto H, et al. Involvement of chemokine receptors in breast cancer metastasis. Nature. 2001;410:50-6.
6. Murakami T, Maki W, Cardones AR, et al. Expression of CXC chemokine receptor-4 enhances the pulmonary metastatic potential of murine B16 melanoma cells. Cancer Res. 2002;62:7328-34.
7. Kim RH, Li BD, Chu QD. The role of chemokine receptor CXCR4 in the biologic behavior of human soft tissue sarcoma. Sarcoma. 2011;2011:593708.
8. Long-term results of lung metastasectomy: prognostic analyses based on 5206 cases. The International Registry of Lung Metastases. J Thorac Cardiovasc Surg. 1997;113:37-49.
9. Casson AG, Putnam JB, Natarajan G, et al. Five-year survival after pulmonary metastasectomy for adult soft tissue sarcoma. Cancer. 1992;69:662-8.
10. Porter GA, Cantor SB, Walsh GL, et al. Cost-effectiveness of pulmonary resection and systemic chemotherapy in the management of metastatic soft tissue sarcoma: a combined analysis from the University of Texas M. D. Anderson and Memorial Sloan-Kettering Cancer Centers. J Thorac Cardiovasc Surg. 2004;127:1366-72.
11. Billingsley KG, Burt ME, Jara E, et al. Pulmonary metastases from soft tissue sarcoma: analysis of patterns of diseases and postmetastasis survival. Ann Surg. 1999;229:602-10.
12. Girard P, Spaggiari L, Baldeyrou P, et al. Should the number of pulmonary metastases influence the surgical decision? Eur J Cardiothorac Surg. 1997;12:385-91.
13. Blackmon SH, Shah N, Roth JA, et al. Resection of pulmonary and extrapulmonary sarcomatous metastases is associated with long-term survival. Ann Thorac Surg. 2009;88:877-84.
14. von Mehren M, Benjamin RS, Bui MM, et al. NCCN Guidelines: Soft Tissue Sarcoma, Version 2.2011. 1-90. 8-4-0011. National Comprehensive Cancer Network, Inc.
15. Meyers PA, Heller G, Healey JH, et al. Osteogenic sarcoma with clinically detectable metastasis at initial presentation. J Clin Oncol. 1993;11:449-53.
16. Briccoli A, Rocca M, Salone M, et al. Resection of recurrent pulmonary metastases in patients with osteosarcoma. Cancer. 2005;104:1721-5.
17. Fortes DL, Allen MS, Lowe VJ, et al. The sensitivity of 18F-fluorodeoxyglucose positron emission tomography in the evaluation of metastatic pulmonary nodules. Eur J Cardiothorac Surg. 2008;34:1223-7.
18. Bar-Shalom R, Yefremov N, Guralnik L, et al. Clinical performance of PET/CT in evaluation of cancer: additional value for diagnostic imaging and patient management. J Nucl Med. 2003;44:1200-9.
19. Downey RJ. Surgical treatment of pulmonary metastases. Surg Oncol Clin N Am. 1999;8:341.
20. Parsons AM, Ennis EK, Yankaskas BC, et al. Helical computed tomography inaccuracy in the detection of pulmonary metastases: can it be improved? Ann Thorac Surg. 2007;84:1830-6.
21. Cerfolio RJ, Bryant AS, McCarty TP, Minnich DJ. A prospective study to determine the incidence of non-imaged malignant pulmonary nodules in patients who undergo metastasectomy by thoracotomy with lung palpation. Ann Thorac Surg. 2011;91:1696-700.
22. McCormack PM, Bains MS, Begg CB, et al. Role of video-assisted thoracic surgery in the treatment of pulmonary metastases: results of a prospective trial. Ann Thorac Surg. 1996;62:213-6.
23. Petersen RP, Pham D, Burfeind WR, et al. Thoracoscopic lobectomy facilitates the delivery of chemotherapy after resection for lung cancer. Ann Thorac Surg. 2007;83:1245-9.
24. Gossot D, Radu C, Girard P, et al. Resection of pulmonary metastases from sarcoma: can some patients benefit from a less invasive approach? Ann Thorac Surg. 2009;87:238-43.
25. Mutsaerts EL, Zoetmulder FA, Meijer S, et al. Long term survival of thoracoscopic metastasectomy vs metastasectomy by thoracotomy in patients with a solitary pulmonary lesion. Eur J Surg Oncol. 2002;28:864-8.
26. Detterbeck FC, Egan TM. Thoracoscopy using a substernal handport for palpation. Ann Thorac Surg. 2004;78:1031-6.
27. Mineo TC, Ambrogi V, Mineo D, Pompeo E. Transxiphoid hand-assisted videothoracoscopic surgery. Ann Thorac Surg. 2007;83:1978-84.
28. Martini N, McCormack PM. Evolution of the surgical management of pulmonary metastases. Chest Surg Clin N Am. 1998;8:13-27.
29. Putnam JB, Jr., Roth JA. Prognostic indicators in patients with pulmonary metastases. Semin Surg Oncol. 1990;6:291-6.
30. Brown WT, Wu X, Fowler JF, et al. Lung metastases treated by CyberKnife image-guided robotic stereotactic radiosurgery at 41 months. South Med J. 2008;101:376-82.
31. Palussiere J, Italiano A, Descat E, et al. Sarcoma lung metastases treated with percutaneous radiofrequency ablation: results from 29 patients. Ann Surg Oncol. 2011 Jun 3. [Epub ahead of print]
32. van Putte BP, Hendriks JM, Romijn S, et al. Combination chemotherapy with gemcitabine with isolated lung perfusion for the treatment of pulmonary metastases. J Thorac Cardiovasc Surg. 2005;130:125-30.
33. Burt ME, Liu D, Abolhoda A, et al. Isolated lung perfusion for patients with unresectable metastases from sarcoma: a phase I trial. Ann Thorac Surg. 2000;69:1542-9.
34. The International Registry of Lung Metastases Writing Committee: Pastorino U, Buyse M, FRiedel G, et al. Long-term results of lung metastasectomy: prognostic analyses based on 5206 cases. J Thor Cardiovasc Surg. 1997;113:37-49.