Capturing Data on Colostomy Formation in Anal Cancer

Abstract

TO THE EDITOR: The authors of the recent article by Peiffert et al are to be congratulated because randomized trials in anal cancer are difficult toperform.Theirprimaryendpointwascolostomy-freesurvival, which was defined as a definitive colostomy for progression, relapse, or complicationatthetimeofanalysis,andtheydidnotincludepatientswho had undergone a colostomy that was eventually reversed. Therefore, we are surprised that, in Figure 3, which depicts actuarial colostomy-free survival, the y-axis starts at 100%. Were there no patients who had an initial colostomy for symptoms in whom this was never reversed? Current conventional descriptions of colostomy in anal cancer trials fail to capture the data accurately. Colostomy-free survival or cumulative colostomy rates are frequently reported, or the rate of colostomies at specific time-points is reported. However, several series have noted that, among patients who receive a pretreatment colostomy, despite subsequent local disease control, failure to reverse the colostomy is almost invariable. However, although this could initially be considered cause specific, the definition would seem inaccurate in patients who achieve long-term control. This difficult issue was highlighted in a recent correspondence. A baseline pretreatment colostomy that was that was fashioned because of symptoms before radiotherapy was considered by Sunesen et al as a tumor-related colostomy. It is vital to capture this information if we are to compare different treatments. The RTOG (Radiation Therapy Oncology Group) 98-11 study described persistent or recurrent anal carcinoma as being responsible for 62 of 80 patients having a colostomy (78%), but 16 patients (20%) were coded as treatment-related in patients with no evidence of disease (treatment complications or sphincter dysfunction as a result of a previous tumor invasion or destruction), and two patients were coded as having both reasons. The recent classification of Sunesen et al should be extended to include separately a third group of patients (5% to 12% of all patients, depending on the stage) who require a baseline colostomy pretreatment in anal cancer for incontinence, fistula formation, and/or pain on defecation. We recommend the following definitions: (1) a colostomy is classified as a tumor-related colostomy if the patient had evidence of anal cancer at the time of the surgery either for a defunctioning stoma to allay symptoms or a salvage APER (which clearly will not be reversed); (2) a colostomy is classified as a treatment-related colostomy if it was performed either during chemoradiotherapy or after the completion of chemoradiotherapy, and the patient had no histologic evidence of cancer; or (3) a colostomy is classified as a baseline pretreatment-colostomy if the colostomy was created before chemoradiotherapy because of a fistula or intolerable symptoms from the tumor. This may be reversed if the patient achieves local control, although in practice this is unlikely. If the tumor never goes into remission and the patient never achieves a clinical CR, this colostomy should be considered tumor-related at the 6-month time point (Fig 1). A common language according to these definitions would give greater clarity in the future.