An Evolutionary Algorithm to Personalize Stool-Based Colorectal Cancer Screening.

Area of Science:

• Oncology
• Preventive Medicine
• Computational Biology

Background:

• Fecal immunochemical testing (FIT) is a standard colorectal cancer (CRC) screening method.
• FIT concentrations correlate with CRC risk, enabling personalized screening strategies.
• Evaluating personalized screening is computationally complex.

Purpose of the Study:

• To develop a broadly applicable algorithm for optimizing personalized CRC screening policies.
• The algorithm aims to efficiently determine screening intervals and FIT cutoffs based on patient age and FIT history.

Main Methods:

• A mathematical framework and bi-objective evolutionary algorithm were developed to optimize screening policies.
• The algorithm identifies policies balancing minimal costs and maximal health benefits.
• A microsimulation model (MISCAN-Colon) was integrated to accurately estimate policy costs and benefits.

Main Results:

• The algorithm achieved near-optimal results on a benchmark problem (0.007% difference).
• Personalized screening strategies increased health benefits by up to 14% (Experiment 2) and 4.3% (Experiment 3) without additional costs.
• Generated policies align with current screening recommendations.

Conclusions:

• The presented method effectively optimizes personalized screening policies using complex simulation models.
• This approach can guide screening strategies for CRC and other diseases.
• Further discussion is required to establish practical implementation criteria for CRC screening policies.