To the Editor:

I read with great interest the recent article by Ajani et al,¹ who reported prognostic factors for time to colostomy (TTC) in a group of 644 assessable patients treated on the US Gastrointestinal Intergroup Radiation Therapy Oncology group 98-11 trial with either mitomycin-based or cisplatin-based chemoradiotherapy. They found that a tumor diameter of more than 5 cm and treatment with the cisplatin-based protocol were independent predictors of the time to colostomy. In a previous publication, they had reported the cumulative colostomy rate at 5 years to be 10% after mitomycin-based and 19% after cisplatin-based chemoradiotherapy.²

Tumor control with a preserved and functional anal sphincter is the ultimate goal of chemoradiotherapy in anal carcinoma. Failing to control the primary tumor and failure to maintain the sphincter function are two alternative potential causes of colostomy after chemoradiotherapy in this disease. Therefore, freedom from colostomy or colostomy-free survival have frequently been reported in anal carcinoma trials as end points incorporating both the tumor and the normal tissue effects of radiotherapy or chemoradiotherapy for anal cancer. As the anal sphincter is both a tumor-bearing organ and an organ of which the function is to be preserved at the same time; the selection of treatment intensity (eg, radiotherapy dose) is particularly critical to obtain an optimal ratio between tumor control and function preservation. In contrast to other pelvic tumors (prostate cancer, gynecological cancer), treatment intensification via selective escalation of radiotherapy dose in the tumor-bearing organ with sparing of the anorectum by means of methods such as brachytherapy or intensity-modulated radiotherapy can not be achieved in anal cancer.

Therefore, Ajani et al¹ must be congratulated for their detailed reporting of the occurrence of cancer-related colostomy versus treatment-related colostomy. They found that after total doses of 55 to 59 Gy to the primary tumor (ie, the anal canal) of 80 colostomies observed, 50 were in the cisplatin arm and 30 in the mitomycin arm. Although the median follow-up in the current report was only 2.2 years, the plateau in the colostomy curves for both arms reached between 2 and 3 years suggests that the crude occurrence of colostomy stated already includes nearly all toxicity-related surgery to be expected. Therefore, the crude numbers could also be interpreted as follows: After chemoradiotherapy, the rates of colostomy for mitomycin-based versus cisplatin-based protocols were cancer related, 6% versus 13%; treatment related, 3% versus 2%; and related to both cancer and treatment; 0% versus 1%; respectively. This suggests that both protocols are equally toxic (regarding preservation of sphincter integrity and function), but that the cisplatin-containing protocol was less effective in achieving local control.

The difference between the two indications for colostomy is that colostomy for tumor, but not colostomy for treatment toxicity, is associated with uncontrolled primary tumor and an increased risk of tumor-related death. An imbalance of causes for colostomy toward treatment-related colostomy in anal cancer, as previously observed by our group for a cohort of patients with mostly favorable prognostic factors treated with chemoradiotherapy and intracavitary brachytherapy,³ would suggest that the treatment intensity can potentially be reduced. The rates of colostomy for the different indications in the mitomycin-based treatment in the present trial¹ appear favorable, and future trials of chemoradiotherapy in anal canal carcinoma should report cause-specific colostomy rates. It would be highly interesting to see a Kaplan-Meier plot of the time course of cause-specific colostomy incidence for the present data set.

AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.