Nonoperative Management Versus Radical Surgery of Rectal Cancer After Neoadjuvant Therapy-Induced Clinical Complete Response: A Markov Decision Analysis

BACKGROUND: Nonoperative management of rectal cancer was introduced for patients with clinical complete response after neoadjuvant chemoradiotherapy to avoid short- and long-term surgical morbidity related to radical resection. OBJECTIVE: The purpose of this study was to determine the expected life-years and quality-adjusted life-years for nonoperative management and radical resection of locally advanced rectal cancer after clinical complete response following neoadjuvant chemoradiotherapy. DESIGN: Markov modeling was used to simulate nonoperative management and radical surgery for a base case scenario over a 10-year time horizon. Estimates for various clinical variables were obtained after extensive literature search. Outcome was expressed in both life-years and quality-adjusted life-years. Deterministic sensitivity analyses were completed to assess the impact of variation in key parameters. SETTING: A decision model using a Markov model was designed. PATIENTS: The base case was a 65-year-old man with a distal rectal tumor who had achieved clinical complete response after neoadjuvant chemoradiotherapy. MAIN OUTCOME MEASURES: Life-years and quality-adjusted life-years were measured. RESULTS: Quality-adjusted life-years (5.79 for nonoperative management vs 5.62 for radical surgery) and life-years (6.92 for nonoperative management vs 6.96 for radical surgery) were similar between nonoperative management and radical surgery. The preferred treatment strategy changed with variations in the probability of local regrowth in nonoperative management, the probability of salvage surgery for regrowth in nonoperative management, utilities associated with nonoperative management and low anterior resection, and the utility of low anterior resection syndrome. The model was not sensitive to (dis)utilities associated with stoma, chemotherapy, or postoperative morbidity and mortality. LIMITATIONS: The study was limited by assumptions inherent to modeling studies. CONCLUSIONS: Nonoperative management and radical surgery resulted in similar (quality-adjusted) life-years. Nonoperative management should therefore be considered as a reasonable treatment option.