Enhancing recurrence detection in stage III colorectal cancer patients through molecular residual disease test-guided surveillance: A modeling study.

e15600 Background: Colorectal cancer (CRC) is a prevalent solid tumor cancer. Patients remain at risk of recurrence despite curative-intent treatment. Early detection of recurrence allows timely intervention and may improve outcomes, particularly in high-risk patients. Standard-of-care (SoC) guidelines recommend routine surveillance following post-definitive treatment to monitor recurrences. Molecular residual disease (MRD) testing via circulating tumor DNA (ctDNA) biomarker is an emerging approach for early recurrence detection. To date, there is no consensus on managing patients with detected ctDNA. This study evaluated the potential benefits of integrating MRD testing into routine surveillance for Stage III CRC patients. Methods: We developed a microsimulation model to replicate the natural history of 65-year-old Stage III CRC patients over a lifetime. Two surveillance strategies were evaluated: 1) SoC following current guidelines with routine tests including carcinoembryonic antigen (CEA), computed tomography (CT), and colonoscopy, and 2) integrating MRD testing into SoC, performed every three months for two years then every six months for three years. In the model, positive ctDNA led to intensified surveillance involving more frequent CEA testing (every three months) and earlier CT scans (semi-annually). Parameters for disease progression and test performances were derived from clinical trial data and published literature. The microsimulation model was validated against data from the National Cancer Database. Results: With the inclusion of MRD testing, the recurrence detection rates increased from 70.5% to 72.4% in year 1 and 83.3% to 87.8% in year 5. Similarly, the 5-year detection rates for locoregional recurrences increased from 89.3% to 92.7%, and for distant recurrences increased from 81.2% to 86.0%. The time to recurrence detection, defined as the duration between clinical recurrence and its detection, was reduced from 4.5 months under SoC to 3.8 months with the addition of MRD testing, leading to a 15.6% improvement in earlier detection. Further, earlier detection reduced the likelihood of recurrence progressing to unresectable stages. Specifically, over 5 years, per 10,000 patients, the number of unresectable locoregional recurrences decreased from 140 under the SoC to 121 when MRD testing was included (13.5% improvement), and unresectable distant recurrences decreased from 171 to 146 (14.6% improvement). Conclusions: Our clinically validated microsimulation model demonstrated that incorporating MRD testing into routine surveillance enhances recurrence detection in Stage III CRC patients by reducing the time to recurrence detection and, hence, identifying recurrences earlier. Further investigation using alternative surveillance intensification approaches is warranted to improve patient outcomes.