Watch and Wait Versus Surgery for Rectal Cancer: Markov Modeling Derived Thresholds for Salvage, Regrowth, and Near-Complete Responders

Ann Surg Oncol. 2026 Mar 28. doi: 10.1245/s10434-026-19552-x. Online ahead of print.

Abstract

Introduction: Watch-and-Wait (WW) strategy is increasingly offered to rectal cancer patients achieving a clinical complete response to avoid major surgery. However, the safety of this approach relies on variables beyond response. The present study explores variables that determine the safety of a WW program using mathematical modelling.

Methods: This study utilized a Markov state-transition model to simulate 10,000 patients, aiming to quantify the specific tipping points where WW becomes oncologically inferior to immediate Total Mesorectal Excision (TME). Probabilistic and deterministic sensitivity analyses were performed.

Results: The simulation compared outcomes over a 5-year horizon for Disease-Free Survival (DFS) and 10 years for Overall Survival (OS). The base-case analysis showed that while 10-year overall survival was equivalent between the two strategies (83.0% for WW versus 81.5% for TME), the salvage rate, the success of surgery following tumor regrowth, emerged as the dominant driver of patient safety. The model identified that WW becomes inferior to immediate surgery if the salvage rate drops below 75% for DFS or below 61% for OS. Survival is also compromised if the proportion of near-complete responders exceeds 68%, or if the early regrowth rate surpasses 10% per six-month cycle.

Conclusions: Watch-and-Wait yields outcomes comparable to radical surgery only when specific quality benchmarks are met. Programs must maintain salvage rates above 75% and strictly monitor early regrowth velocity. Clinical teams should use these thresholds to guide patient counseling, suggesting that immediate TME is the safer option for patients falling outside these safety parameters.

Keywords: Chemoradiation; Local Regrowth; Markov model; Monte carlo; Rectal cancer; Salvage; Watch and Wait.