Editorial Comment: Forty years on: pulmonary metastasectomy for sarcoma Free

We searched the literature for evidence concerning pulmonary metastasectomy for sarcoma. A systematic review includes data on 1357 patients from 18 surgical follow-up studies [1]. The report from Nagoya University in this issue [2] is characteristic of the genre: a 15-year experience is described of a mixed series of 52 consecutive patients with bone or soft-tissue sarcoma. The majority (28 of 52) had a solitary metastasis, and most of the rest had two or three metastases (17 of 52). Overall survival was 51% five years after the first pulmonary metastasectomy.

From the 1970s, rather than being an occasional individualized decision, as it was for other cancers, pulmonary metastasectomy has become an established practice in the care of sarcoma [3]. Of patients with sarcoma who have metastatic disease, the lungs are affected in 80% and are the only site in 60% [4]. For patients in whom the primary site is rendered free of cancer, pulmonary metastasectomy is performed with curative intent and, when there is recurrence in the lung, it is hoped that further operations will complete cancer clearance. Requests for removal of sarcomatous metastases are therefore familiar to thoracic surgeons. Forty years on, we know that the anticipation of cures has not been borne out. In an admirable example of assiduous long-term follow-up, the outcomes for 22 patients first reported in 1971 [3] were again reported 20 years later [5]. At the time of the first metastasectomy, the patients were 9–19 years old [3]. The great majority (at least 19 of 22) went on to eventually die of sarcoma after up to nine thoracotomies. There were six survivors at 10 years, and of four alive at 20 years, only two had no evident disease. We cannot know for certain whether the lives of any of these 22 patients would have been shorter or longer without pulmonary metastasectomy.

This insight—that the majority of patients who have pulmonary metastasectomy eventually die of this cancer—is not usually emphasized in follow-up studies. Five-year survival rates are provided as if they represent a cure rate; there is often an implicit accompanying assumption that no patient would have been alive at 5 years were it not for metastasectomy. Neither assumption is correct. Of patients, in the Thames Cancer Registry 1995–2004 who had metastatic bone cancer at the time of registration, 25% were alive at 5 years. It follows that if a surgeon were able to correctly identify and not to operate on the less-favourable half of a cohort, by a simple calculation, a 50% five-year survival might be seen in the half of the cohort selected for metastasectomy. Over 30 years ago, it was postulated by Åberg that rather than metastasectomy being curative, it was the selection of patients with a better prognosis that provided a higher survival rate at 5 years than would generally be expected [6]. Mathematical modelling and simple reasoning support Åberg's postulate [7].

Apart from the possibility of occasional cure, does pulmonary metastasectomy for sarcoma prolong life? There is no way of knowing from the reported series because no studies include a control group [1]. Is there amelioration of symptoms? There are no data reported and, in practice, these patients are not symptomatic at the outset so there is no breathlessness to palliate at the time of metastasectomy. It has not been shown that metastasectomy helps maintain lung function. Reading through surgical accounts, it appears that it is the loss of parenchyma by successive resections that eventually calls a halt to further surgery, so while it is clear that breathlessness may preclude metastasectomy, it is not evident that metastasectomy benefits breathing.

Another frequent misconception is that because similar survival rates are seen after a second and subsequent metastasectomy, repeated operations are effectively resetting the clock. For example, the assertion that repeated and aggressive pulmonary resections for leiomyosarcoma metastases extends survival [8] cannot be claimed without some evidence as to what the survival would have been without surgery, whether aggressive or repeated. These conclusions cannot be derived from uncontrolled follow-up studies of selected patients [9]. A diminishing number of patients have a subsequent metastasectomy and only the fittest and most favourable have multiple successive operations. So the relationship might be reversed: rather than being alive because of the subsequent metastasectomy, an essential prerequisite to having a further metastasectomy is being alive and tolerably well. A statistical association between length of life and the number of operations may thus be misinterpreted.

We offer a disheartening view of the role of surgery in advanced cancer, but the inescapable reality is that pulmonary metastases are incontrovertible evidence of blood-borne systemic disease. Surgery is powerful and effective, but only in those diseases that are amenable to it. In the case of cancer, that generally means disease localized to the primary site. Considerable physical suffering, time away from home and everyday life, and expense for families and health care funders are entailed in unavailing surgery for advanced cancer. Metastasectomy should be confined to those groups of patients for whom benefit has been proven.

Follow-up studies without a record of the features and outcome for unoperated patients are common and are familiar to surgeons, but they have repeatedly been misleading [8]. We need better evidence than they provide. Mathematical modelling studies are one place to start; they narrow and help frame the research question [7]. A clinical trial with a contemporaneous and unbiased control group, against which any positive or negative effect can be quantified, is the best way to answer the question. The commonest disease for which pulmonary metastasectomy is performed is colorectal cancer. The PulMiCC trial is currently recruiting patients. It incorporates several innovative trial-design features that respect patients' preferences and the cancer team's judgment. Only where there is uncertainty, and the decision could go either way, is random allocation offered [9]. Entering patients into trials requires a culture change among surgeons. The reasons given for not doing randomized trials are familiar: the patients will not accept randomization, the oncologists demand metastasectomy, and if we do not do it, ‘someone else will’. However, these are insufficient reasons for continuing to do surgery of doubtful benefit [10]. If surgeons do not lead research in surgery, ‘someone else will’.

REFERENCES