Survival is higher after repeat lung metastasectomy than after a first metastasectomy: Too good to be true?
2015; Elsevier BV; Volume: 149; Issue: 5 Linguagem: Inglês
10.1016/j.jtcvs.2015.01.067
ISSN1097-685X
AutoresTom Treasure, Tommaso Claudio Mineo, Vincenzo Ambrogi, Francesca Fiorentino,
Tópico(s)Lung Cancer Diagnosis and Treatment
ResumoThe authors of the International Registry of Lung Metastases reported that survival was higher for a second compared with a first metastasectomy operation after 5 and 10 years (44% vs 39% and 29% vs 25%, respectively).1Pastorino U. Buyse M. Friedel G. Ginsberg R.J. Girard P. Goldstraw P. et al.Long-term results of lung metastasectomy: prognostic analyses based on 5206 cases.J Thorac Cardiovasc Surg. 1997; 113: 37-49Abstract Full Text Full Text PDF PubMed Scopus (1307) Google Scholar Pastorino and colleagues1Pastorino U. Buyse M. Friedel G. Ginsberg R.J. Girard P. Goldstraw P. et al.Long-term results of lung metastasectomy: prognostic analyses based on 5206 cases.J Thorac Cardiovasc Surg. 1997; 113: 37-49Abstract Full Text Full Text PDF PubMed Scopus (1307) Google Scholar wrote "The long-term outcome of patients who were treated by a second metastasectomy was remarkably good." The word "remarkably" provides the clue. The finding seemed to be too good to be true. Similar observations have been made repeatedly for sarcoma, which is the most established clinical indication for lung metastasectomy. In a systematic review of 14 follow-up studies composed of 1357 patients, 579 received a repeat metastasectomy (43%; range, 21%-79%).2Treasure T. Fiorentino F. Scarci M. Moller H. Utley M. Pulmonary metastasectomy for sarcoma: a systematic review of reported outcomes in the context of Thames Cancer Registry data.BMJ Open. 2012; 2: e001736Crossref Scopus (89) Google Scholar The observation that repeat metastasectomy was followed by higher survival has been made repeatedly.•"Prognostic factors for increased survival included 3 or greater redo pulmonary operations…"3Blackmon S.H. Shah N. Roth J.A. Correa A.M. Vaporciyan A.A. Rice D.C. et al.Resection of pulmonary and extrapulmonary sarcomatous metastases is associated with long-term survival.Ann Thorac Surg. 2009; 88: 877-884Abstract Full Text Full Text PDF PubMed Scopus (119) Google Scholar•"…patients with complete resection for recurrent pulmonary metastasis show a significantly better prognosis after repeat pulmonary metastasectomy."4Chen F. Miyahara R. Bando T. Okubo K. Watanabe K. Nakayama T. et al.Repeat resection of pulmonary metastasis is beneficial for patients with osteosarcoma of the extremities.Interact Cardiovasc Thorac Surg. 2009; 9: 649-653Crossref PubMed Scopus (36) Google Scholar•"…repeat metastasectomy for recurrent pulmonary metastasis also provided a favorable overall survival (P = .041)."5Chen F. Fujinaga T. Sato K. Sonobe M. Shoji T. Sakai H. et al.Significance of tumor recurrence before pulmonary metastasis in pulmonary metastasectomy for soft tissue sarcoma.Eur J Surg Oncol. 2009; 35: 660-665Abstract Full Text Full Text PDF PubMed Scopus (31) Google Scholar•"Repeated and aggressive pulmonary resections for leiomyosarcoma metastases extend survival."6Burt B.M. Ocejo S. Mery C.M. Dasilva M. Bueno R. Sugarbaker D.J. et al.Repeated and aggressive pulmonary resections for leiomyosarcoma metastases extends survival.Ann Thorac Surg. 2011; 92: 1202-1207Abstract Full Text Full Text PDF PubMed Scopus (60) Google Scholar The practice of repeated metastasectomy has become part of the mantra of "treating cancer as a chronic illness."7Treating cancer as a chronic illness. Available at: http://www.cancer.org/treatment/survivorshipduringandaftertreatment/understandingrecurrence/whenyourcancercomesback/when-cancer-comes-back-treating-cancer-as-chronic-illness. Accessed March 3, 2015.Google Scholar "…patients persistently free of the primary osteosarcoma who developed recurrent resectable metastatic disease of the lung should be considered for reoperation a second, third, or fourth time."8Briccoli A. Rocca M. Salone M. Bacci G. Ferrari S. Balladelli A. et al.Resection of recurrent pulmonary metastases in patients with osteosarcoma.Cancer. 2005; 104: 1721-1725Crossref PubMed Scopus (123) Google Scholar For colorectal cancer, there is a similar acceptance of repeated lung metastasectomy that increased from a rate of 15% to more than 20% from the 1960s to 2000s in an analysis of 51 studies including 3504 patients.9Fiorentino F. Hunt I. Teoh K. Treasure T. Utley M. Pulmonary metastasectomy in colorectal cancer: a systematic review and quantitative synthesis.J R Soc Med. 2010; 103: 60-66Crossref PubMed Scopus (105) Google Scholar Repeated metastasectomy was noted as practice in the review of lung metastasectomy in 1539 patients10Pfannschmidt J. Hoffmann H. Dienemann H. Reported outcome factors for pulmonary resection in metastatic colorectal cancer.J Thorac Oncol. 2010; 5: S172-S178Crossref PubMed Scopus (113) Google Scholar that formed part of the European Society of Thoracic Surgeons Lung Metastasectomy Project.11Van Raemdonck D. Friedel G. The European Society of Thoracic Surgeons lung metastasectomy project.J Thorac Oncol. 2010; 5: S127-S129Crossref PubMed Scopus (47) Google Scholar It was also recorded in a later comprehensive analysis of colorectal cancer lung metastasectomy in 2925 patients.12Pfannschmidt J. Dienemann H. Surgical treatment of oligometastatic non-small cell lung cancer.Lung Cancer. 2010; 69: 251-258Abstract Full Text Full Text PDF PubMed Scopus (123) Google Scholar In this statistical analysis, we take a closer look at how survival data are presented and suggest that a more skeptical view of this conclusion is warranted. The Roman surgeons Tommaso Mineo and Vincenzo Ambrogi had for 25 years followed a policy of repeat metastasectomy. A Kaplan–Meier analysis undertaken for them showed that of patients who underwent multiple metastasectomy operations, 65% were alive at 5 years compared with 42% who underwent only 1 metastasectomy operation (Figure 1). In common with the authors of the landmark International Registry of Lung Metastases, they found this somewhat counterintuitive and invited the co-authors (Francesca Fiorentino and Tom Treasure) to review the analysis. Figure 1 shows a familiar generic approach to displaying survival data that would be entirely appropriate if there were 2 treatment strategies randomly assigned. To permit comparison, a specified starting point is taken as time zero on the horizontal axis and the vertical axis shows the loss of patients due to death, accounting for censoring. The explicit purpose is to allow visual comparison. If the patients had been randomly assigned to either a policy of once only metastasectomy or a policy of repeated metastasectomy, then the inference might be drawn that a difference in survival was due to the different treatment plans. The method fails as a means of fair comparison if the patients themselves differ in ways that influence survival. Then the difference may be due to the patients' characteristics rather than the treatment they received. We will call the summary of these patient characteristics "survivability." Clinicians are aware of these patient characteristics, but they cannot all be retrieved from even a comprehensive research database and certainly not from a registry.13Williams W.G. Uses and limitations of registry and academic databases.Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu. 2010; 13: 66-70Abstract Full Text Full Text PDF PubMed Scopus (16) Google Scholar Estimates from various sources indicate that of all patients who are found to have metastases, less than 1 in 20 will have a lung metastasectomy.14Wade T.P. Virgo K.S. Li M.J. Callander P.W. Longo W.E. Johnson F.E. Outcomes after detection of metastatic carcinoma of the colon and rectum in a national hospital system.J Am Coll Surg. 1996; 182: 353-361PubMed Google Scholar, 15Embun R. Fiorentino F. Treasure T. Rivas J.J. Molins L. Pulmonary metastasectomy in colorectal cancer: a prospective study of demography and clinical characteristics of 543 patients in the Spanish colorectal metastasectomy registry (GECMP-CCR).BMJ Open. 2013; 3: e002787Crossref Scopus (50) Google Scholar, 16Robinson B.J. Rice T.W. Strong S.A. Rybicki L.A. Blackstone E.H. Is resection of pulmonary and hepatic metastases warranted in patients with colorectal cancer?.J Thorac Cardiovasc Surg. 1999; 117: 66-75Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar, 17Watanabe K. Saito N. Sugito M. Ito M. Kobayashi A. Nishizawa Y. Incidence and predictive factors for pulmonary metastases after curative resection of colon cancer.Ann Surg Oncol. 2013; 20: 1374-1380Crossref PubMed Scopus (26) Google Scholar, 18Tampellini M. Ottone A. Bellini E. Alabiso I. Baratelli C. Bitossi R. et al.The role of lung metastasis resection in improving outcome of colorectal cancer patients: results from a large retrospective study.Oncologist. 2012; 17: 1430-1438Crossref PubMed Scopus (56) Google Scholar With each successive metastasectomy, the degree of selection from the original denominator is greater; the apparent denominator diminishes as seen in the Roman surgeons' data (Table 1). We can deduce that approximately 1 in 100 of the original denominator received a second metastasectomy and 1 in 1000 received a fifth metastasectomy. These were not randomly selected from the original denominator; at the very least, they had to still be alive. Difference in survival cannot be ascribed to the number of metastasectomy operations they had undergone.Table 1Intraoperative intervalsMetastasectomy operationsPatientsInterval beforeMedian monthsMinimumMaximum25%75%First14070524≥2113Second226961434≥354Third2310641739≥431Fourth226441227≥58Fifth13496111364Sixth877079Unpublished data (T. Mineo, V. Ambrogi, June 2014). The available data can be obtained by request to the authors. Open table in a new tab Unpublished data (T. Mineo, V. Ambrogi, June 2014). The available data can be obtained by request to the authors. In a study of liver and lung metastasectomy from the Cleveland Clinic, denominators are provided.16Robinson B.J. Rice T.W. Strong S.A. Rybicki L.A. Blackstone E.H. Is resection of pulmonary and hepatic metastases warranted in patients with colorectal cancer?.J Thorac Cardiovasc Surg. 1999; 117: 66-75Abstract Full Text Full Text PDF PubMed Scopus (110) Google Scholar The upper line on the graph in Figure 2 relates to 25 patients who underwent both a liver and a lung metastasectomy. Seventeen patients who underwent a first liver metastasectomy had a subsequent lung metastasectomy, and 6 patients underwent a lung metastasectomy first. The lower line relates to 23 patients who had both liver and lung metastases but did not undergo any liver or lung metastasectomy. During the same time frame, 5787 patients had resections of primary colorectal carcinomas, 466 patients had resections of isolated colorectal hepatic metastases, and 72 patients had resections of isolated lung metastases. The selection of patients for metastasectomy was by individual clinical evaluation, and proceeding with the second metastasectomy operation was contingent on further individual clinical evaluation of the patient. The outcome was not reported on intention to treat but on completed treatment. However sophisticated the analysis, it is impossible to know how much of the survival difference shown was due to inadvertent selection of patients inherently most likely to live longer (survivability) rather than an effect of sequential metastasectomy operations. Although the analysis is carefully justified in the text, it might be argued that it was misleading to put the lines next to each other in the depiction, implicitly inviting a comparison between 25 and 23 patients with lung and liver metastases who in the event (but not by intention to treat) had both or neither resected. If a picture is worth a thousand words, it is worth studying the survival graph to see what the picture tells us. We are familiar with cancer survival curves that generally follow the shape of the lower lines in Figure 1, Figure 2, Figure 3. The upper line in each has an early more horizontal component not seen in natural cancer survival graphs. Lung metastases are rarely a cause of death, so it is unlikely that the early horizontal component (Figure 1, Figure 2) is due to survival benefit from lung metastasectomy. It is more likely to be due to conditional entry of patients judged clinically to be likely to survive for a reasonable time after operation.19Glasziou P. Chalmers I. Rawlins M. McCulloch P. When are randomised trials unnecessary? Picking signal from noise.BMJ. 2007; 334: 349-351Crossref PubMed Google Scholar Figure 3 come from an analysis of 114,155 patients who underwent resection for a colorectal tumor, among whom 47,430 patients had liver metastases and 3116 (6.6%) had them resected. We again see the horizontal component at the beginning of the survival curve of patients who had liver metastases resected.20Morris E.J. Forman D. Thomas J.D. Quirke P. Taylor E.F. Fairley L. et al.Surgical management and outcomes of colorectal cancer liver metastases.Br J Surg. 2010; 97: 1110-1118Crossref PubMed Scopus (270) Google Scholar These plateau curves in Figure 1, Figure 2, Figure 3 are not natural cancer survival patterns but manifestations of immortal time bias.21Suissa S. Immortal time bias in observational studies of drug effects.Pharmacoepidemiol Drug Saf. 2007; 16: 241-249Crossref PubMed Scopus (361) Google Scholar, 22Suissa S. Immortal time bias in pharmaco-epidemiology.Am J Epidemiol. 2008; 167: 492-499Crossref PubMed Scopus (1045) Google Scholar, 23Levesque L.E. Hanley J.A. Kezouh A. Suissa S. Problem of immortal time bias in cohort studies: example using statins for preventing progression of diabetes.BMJ. 2010; 340: b5087Crossref PubMed Scopus (707) Google Scholar No surgeon would disagree that it is important to be selective in offering patients lung metastasectomy. Less readily accepted is that survival attributed to metastasectomy may be due to the selection of patients who are inherently destined to survive longer.24Aberg T. Malmberg K.A. Nilsson B. Nou E. The effect of metastasectomy: fact or fiction?.Ann Thorac Surg. 1980; 30: 378-384Abstract Full Text PDF PubMed Scopus (108) Google Scholar, 25Aberg T. Selection mechanisms as major determinants of survival after pulmonary metastasectomy.Ann Thorac Surg. 1997; 63: 611-612Abstract Full Text PDF PubMed Scopus (48) Google Scholar Selected patients have fewer metastases, longer intervals, and, in colorectal cancer, lower levels of carcinoembryonic antigen.26Gonzalez M. Poncet A. Combescure C. Robert J. Ris H.B. Gervaz P. Risk factors for survival after lung metastasectomy in colorectal cancer patients: a systematic review and meta-analysis.Ann Surg Oncol. 2013; 20: 572-579Crossref PubMed Scopus (291) Google Scholar These are not predictive of effectiveness of metastasectomy; they are prognostic factors for survival.27Simms L. Barraclough H. Govindan R. Biostatistics primer: what a clinician ought to know-prognostic and predictive factors.J Thorac Oncol. 2013; 8: 808-813Crossref PubMed Scopus (32) Google Scholar Patients with slow progression and thus longer interval before the (re)appearance of metastases are more likely to have them surgically removed.28Fiorentino F. Treasure T. Pulmonary metastasectomy for colorectal cancer: making the case for a randomized controlled trial in the zone of uncertainty.J Thorac Cardiovasc Surg. 2013; 146: 748-752Abstract Full Text Full Text PDF PubMed Scopus (24) Google Scholar The intraoperative intervals are in reality determined by conscious surgical decision rather than being governed by the behavior of the disease, as implied by the misnomer "disease-free interval" (Table 1). The process of clinical assessment over time allows survivability to be identified in part, but not only because of slow progression of measureable lesions. This is deliberately or inadvertently included in the decision to offer metastasectomy. Selection on survivability is likely to be associated with subsequent survival. If there is a sequence of treatments, a patient must necessarily have survived the previous treatment and have been judged a likely future survivor to proceed to the next treatment. Clinical judgment of survivability is thus a confounding factor for the outcome of interest, which is survival. Symptomatic or life-threatening metastases may be removed on clinical grounds. This is uncommon because lung metastases from carcinoma are rarely symptomatic and infrequently lethal.29Treasure T. Milosevic M. Fiorentino F. Pfannschmidt J. History and present status of pulmonary metastasectomy in colorectal cancer.World J Gastroenterol. 2014; 20: 14517-14526Crossref PubMed Scopus (14) Google Scholar
Referência(s)