MICROSIMULATION OF THE NATURAL HISTORY OF STAGE II/III COLORECTAL CANCER AND EVALUATION OF FOLLOW-UP TESTING STRATEGIES (MONET-MODEL)
Author(s)
Tara C. Boute, MSc.
PhD student, Amsterdam UMC, Amsterdam, Netherlands.
PhD student, Amsterdam UMC, Amsterdam, Netherlands.
OBJECTIVES: The effectiveness of intensive follow-up strategies for colorectal cancer (CRC) remains debated. We developed a decision model to synthesize available evidence and evaluate follow-up strategies for stage II/III CRC patients.
METHODS: Our microsimulation model simulates patient trajectories from CRC-diagnosis until death using seven health states: primary resected CRC (RES), 90-day mortality (90DM), detectable recurrence (REC), detected recurrence with non-curative (NCT) and curative-intent treatment (CIT), death of disease (DOD), and death from other causes (DOC). Model parameters were derived from two Netherlands Cancer Registry datasets. Transition probabilities were estimated using cure and parametric survival models incorporating relevant covariates. Three follow-up strategies were compared with no follow-up: (A) pre-2021 Dutch guideline with realistic adherence (semi-annual imaging for 2 years, annual thereafter; CEA every 3-6 months for 3 years, then every 6 months; colonoscopy at 1 and 4 years), (B) Strategy A with full adherence, and (C) post-2021 guideline with full adherence (single CT at 1 year; CEA every 3-6 months for 2 years, then every 6 months; colonoscopy at 1 and 4 years).
RESULTS: Localization (colon vs rectum) and high-risk features (pT4 and/or pN2) were included as covariates in the transition RES-REC. Mode of detection (symptomatic vs follow-up) and recurrence pattern (locoregional, one distant site, or ≥2 sites) were covariates in transition REC-NCT/CIT. The model showed good internal validity. In a simulated cohort, 15.5% of patients with recurrent disease were eligible for curative-intent treatment without follow-up versus 39.5%, 41.2%, and 39.3% in strategies A, B, and C, respectively. Disease-specific mortality among patients with recurrence was 22.8% without follow-up and 18.7%, 18.5%, and 18.7% for strategies A, B, and C, respectively.
CONCLUSIONS: The MONET model enables evaluation of CRC follow-up strategies. No survival advantage of intensive follow-up was observed, supporting recent de-intensification of Dutch guidelines. Future applications will assess long-term outcomes and cost-effectiveness of risk-based strategies.
METHODS: Our microsimulation model simulates patient trajectories from CRC-diagnosis until death using seven health states: primary resected CRC (RES), 90-day mortality (90DM), detectable recurrence (REC), detected recurrence with non-curative (NCT) and curative-intent treatment (CIT), death of disease (DOD), and death from other causes (DOC). Model parameters were derived from two Netherlands Cancer Registry datasets. Transition probabilities were estimated using cure and parametric survival models incorporating relevant covariates. Three follow-up strategies were compared with no follow-up: (A) pre-2021 Dutch guideline with realistic adherence (semi-annual imaging for 2 years, annual thereafter; CEA every 3-6 months for 3 years, then every 6 months; colonoscopy at 1 and 4 years), (B) Strategy A with full adherence, and (C) post-2021 guideline with full adherence (single CT at 1 year; CEA every 3-6 months for 2 years, then every 6 months; colonoscopy at 1 and 4 years).
RESULTS: Localization (colon vs rectum) and high-risk features (pT4 and/or pN2) were included as covariates in the transition RES-REC. Mode of detection (symptomatic vs follow-up) and recurrence pattern (locoregional, one distant site, or ≥2 sites) were covariates in transition REC-NCT/CIT. The model showed good internal validity. In a simulated cohort, 15.5% of patients with recurrent disease were eligible for curative-intent treatment without follow-up versus 39.5%, 41.2%, and 39.3% in strategies A, B, and C, respectively. Disease-specific mortality among patients with recurrence was 22.8% without follow-up and 18.7%, 18.5%, and 18.7% for strategies A, B, and C, respectively.
CONCLUSIONS: The MONET model enables evaluation of CRC follow-up strategies. No survival advantage of intensive follow-up was observed, supporting recent de-intensification of Dutch guidelines. Future applications will assess long-term outcomes and cost-effectiveness of risk-based strategies.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
EE97
Topic
Economic Evaluation, Health Technology Assessment, Methodological & Statistical Research
Disease
Oncology