While reporting a comparative analysis of 409 patients with pulmonary metastases because of colorectal cancer, Tampellini and colleagues write: "To solve the debate, a phase III, prospective, randomized clinical trial … has been advocated" [1]. The trial cited [2] is PulMiCC (Pulmonary Metastasectomy in Colorectal Cancer). At the time of writing, 123 U.K. patients have been recruited and 36 of them have been randomized to either pulmonary metastasectomy or not, with both groups being actively monitored. PulMiCC International is open for recruitment in Catania, Italy.We are running the trial because we agree with the statement of Tampellini and colleagues, "From a statistical point of view, our findings are insufficient to definitively solve the question of whether or not surgery is beneficial in resectable patients because this can be demonstrated only with a prospective, randomized, phase III trial" [1]. They nicely show the statistical limitations of observational data. They identified 1,411 patients with metastatic colorectal cancer between 1994 and 2010. Of these patients, 409 (29%) had pulmonary metastases and in 154 (11%), metastases were confined to the lung; 50 patients (3.5%) were selected to have pulmonary metastasectomy. This illustrates the problem of the shrinking denominator in selection of patients for pulmonary metastasectomy. As the authors state, this information is not generally captured in surgical follow-up studies where the starting point is a retrieved list of patients who had pulmonary metastasectomy. Exceptions include reports from the U.S. DepartmentofVeteransAffairs[3]andaClevelandClinicstudyof liver and lung versus lung only metastasectomy [4], which showed a similarly high degrees of selection.If we consider the 154 patients with lung-only metastases, 50 (32%) were selected for surgery (group 3) while the remaining 104 (group 2) were selected to not have surgery [1]. The groups are very different on critical baseline characteristics. Compared with the 104 patients in group 2, the 50 patients selected for pulmonary metastasectomy had far fewer metastases (median 1, range 1-5 vs. 5, 1-15) and a much longer interval before metastases were evident (median 24 months, range 0 -119 months vs. 5, 0 -76). These are well-established factors associated with survival variation in patients with pulmonary metastases, irrespectiveofwhethertheydoordonothavemetastasectomy [5].So,to ask the 30-year-old question [6], is it the selection rather than the surgery that explains the difference?We agree with the authors' final conclusion: "From a medical oncology point of view, however, our results add evidence to the debate on whether or not such a study is still necessary." Their data, which show the large differences in selection criteria between those who did and did not have pulmonary metastasectomy, emphasize the need for an even-handed comparison. PulMiCCisdesignedsothatpatientsandclinicalteamscandecide for or against pulmonary metastasectomy. If, after standard clinical assessment, the better course ...