Portal de Periódicos da CAPES

Mesothelioma: Benefit from Surgical Resection is Questionable

2007; Elsevier BV; Volume: 2; Issue: 10 Linguagem: Inglês

10.1097/jto.0b013e318156af3b

1556-1380

Tom Treasure, Martin Utley,

Interstitial Lung Diseases and Idiopathic Pulmonary Fibrosis

Deaths from malignant pleural mesothelioma are increasing in number in Europe and are expected to peak between 2011 and 2015 in Britain. The burden of this disease in the developing world is likely to be great. Is there anything we can do as surgeons to influence a patient's survival? Does radical surgery have a useful role? The report on 945 patients, from Memorial Sloan-Kettering Cancer Centre is by far the largest series of patients reviewed to date.1Flores R et al.Prognostic factors in the treatment of malignant pleural mesothelioma (MPM) at a large tertiary referral centre.J Thorac Oncol. 2007; 2: 957-965Abstract Full Text Full Text PDF PubMed Scopus (172) Google Scholar The data will inevitably be used by those planning and designing randomized trials of treatment of MPM so the paper merits close attention and very careful reading. It is useful to first take a general look at the data to get a sense of their reliability. The sample in this report was obtained from the pathology department database which provided a list of patients in whom a diagnosis of mesothelioma was recorded. Clinical information on these patients was then retrieved from the clinical data base, and the case records, for analysis. In about half of the 945 patients, there was no record of the histological classification and there was no pathological material available for review so mesothelioma was entered as unclassified into the Cox proportional hazards model that was constructed. There was also a high level of missing data; in half of the patients, the stage was missing. This rings an alarm bell, particularly as there is likely to be more complete histological and staging data in those operated upon within the institution than unoperated patients, creating an imbalance that could distort comparisons between treatment groups. No details are provided concerning data completeness by treatment group. In this practice the objective of surgery is implicitly extension of life; the report makes no mention of symptom relief or quality of life despite extrapleural pneumonectomy being a severe intervention with a high complication rate.2Sugarbaker DJ Jaklitsch MT Bueno R Richards W Lukanich J Mentzer SJ et al.Prevention, early detection, and management of complications after 328 consecutive extrapleural pneumonectomies.J Thorac Cardiovasc Surg. 2004 Jul; 128: 138-146Abstract Full Text Full Text PDF PubMed Scopus (314) Google Scholar The abstract concludes with the sentence “Surgical resection in a multimodality setting was associated with improved survival.― Having been operated upon may be associated with longer survival but the word “improved― implies a belief that there is cause and effect. Is that belief justified on these data? Is longer survival conferred by the act of surgery or merely associated with the fact of surgery? Surgical resection emerged as being significantly associated with longer survival from a Cox proportional hazards model. The implication is that other factors including tumor stage, smoking history, asbestos exposure, gender, pain, histology and laterality (but apparently not age) are controlled for. The authors dichotomized each factor before entering the data into the model; whether the dichotomics used were devised prior to analysis or on inspection of the data is not specified. It is worth pointing out that putting confounding variables into such a Cox model is not the same as “controlling― for the effect of those confounding variables. For instance, consider the association between survival and laterality, which we focus on because data are present for all 945 patients. The authors combine right sided disease and bilateral disease despite bilateral disease being associated with much poorer survival than right sided disease. None of those patients with bilateral disease will have had resection. Resection could therefore emerge as a favorable factor independent of dichotomized laterality solely as it would exclude those patients with bilateral disease and hence very poor prognosis. This type of effect undermines any interpretation of the association of resection and longer survival as a causal relationship. There is considerable scope for such effects, particularly given the fact that staging and histology data were missing in half of the cases. There is another worrisome aspect to the analysis, that being the exclusion of those patients who had a thoracotomy, but in whom the intended resection was not achieved. These 174 patients represent about a third (31%) of those subjected to thoracotomy and would of course remain within an “operated― group in an intention to treat analysis. Not only are they excluded from the surgical resection group, but they are included in the group with which surgical results are being compared. To spell it out, those patients entering the model as having had “surgical resection― are twice selected: once clinically in the hands of an expert team, and then again on the basis of favorable features discovered at surgery. The implications for the analysis aside, we consider 31% having major exploratory surgery to represent a concerning burden in a population of patients nearing the end of their lives. Furthermore, would it not be as reasonable to attribute survival differences between treatment groups to the adjuvant therapies as to the surgery? The largest survival difference was observed between those patients that had multi-modality therapy compared to those that had resection alone. Indeed, comparing figure 9 and 10 there is little difference between the survival among those that had resection alone and those that had no operation. That said, in order to have received multiple treatments you have to be a survivor in reasonable clinical state for a sufficient length of time. This is a further reminder of the importance of intention to treat analysis. Our reservations about the possible interpretation of this paper as providing evidence of benefit from surgical resection does not stem from an obstinate refusal to accept any evidence other than a randomized trial. There are many operations for which we rely on other forms of evidence3Treasure T The evidence on which to base practice: Different tools for different times.Eur J Cardiothorac Surg. 2006 Dec; 30: 819-824Crossref PubMed Scopus (11) Google Scholar but with surgery as severe as this, within the context of multimodality treatment and relatively small differences, analyses of further case series are unlikely to produce an answer we can trust.4Treasure T Utley M Ten traps for the unwary in surgical series: a case study in mesothelioma reports.J Thorac Cardiovasc Surg. 2007 Jun; 133: 1414-1418Abstract Full Text Full Text PDF PubMed Scopus (59) Google Scholar It is for exactly these reasons that we embarked on the MARS (Mesothelioma and Radical Surgery) trial which is in its pilot phase recruiting in Great Britain.5Treasure T Tan C Lang-Lazdunski L Waller D The MARS trial: mesothelioma and radical surgery.Interact CardioVasc Thorac Surg. 2006; 5: 58-59Crossref PubMed Scopus (38) Google Scholar This trial aims to evaluate EPP within tri-modality therapy with a control arm of any other treatment. Trials are not easy in surgery but it is up to surgeons to deliver them. We must be willing to both take a lead in developing them and to put our patients into them.6Rusch VW Thoracic surgical clinical trials: Y2K and beyond.Ann Thorac Surg. 1999 Jul; 68: 2-3Abstract Full Text Full Text PDF PubMed Scopus (210) Google Scholar The study of Flores et al is the largest of its kind and contains valuable data but the question of whether radical surgery confers a net benefit on patients is still open and will not be resolved by further case series.