The Management of Patients with T1 Adenocarcinoma of the Low Rectum

BACKGROUND: Decision making for patients with T1 adenocarcinoma of the low rectum, when treatment options are limited to a transanal local excision or abdominoperineal resection, is challenging. OBJECTIVES: The aim of this study was to develop a contemporary decision analysis to assist patients and clinicians in balancing the goals of maximizing life expectancy and quality of life in this situation. DESIGN: We constructed a Markov-type microsimulation in open-source software. Recurrence rates and quality-of-life parameters were elicited by systematic literature reviews. Sensitivity analyses were performed on key model parameters. PATIENTS AND SETTING: Our base case for analysis was a 65-year-old man with low-lying T1N0 rectal cancer. We determined the sensitivity of our model for sex, age up to 80, and T stage. MAIN OUTCOME MEASURES: The main outcome measured was quality-adjusted life-years. RESULTS: In the base case, selecting transanal local excision over abdominoperineal resection resulted in a loss of 0.53 years of life expectancy but a gain of 0.97 quality-adjusted life-years. One-way sensitivity analysis demonstrated a health state utility value threshold for permanent colostomy of 0.93. This value ranged from 0.88 to 1.0 based on tumor recurrence risk. There were no other model sensitivities. LIMITATIONS: Some model parameter estimates were based on weak data. CONCLUSIONS: In our model, transanal local excision was found to be the preferable approach for most patients. An abdominoperineal resection has a 3.5% longer life expectancy, but this advantage is lost when the quality-of-life reduction reported by stoma patients is weighed in. The minority group in whom abdominoperineal resection is preferred are those who are unwilling to sacrifice 7% of their life expectancy to avoid a permanent stoma. This is estimated to be approximately 25% of all patients. The threshold increases to 12% of life expectancy in high-risk tumors. No other factors are found to be relevant to the decision.