| Sumario: | Background: Machine learning models learn feature connections from data to learn general behavior. The goal was to build a prediction model to identify the percentage of patients with colorectal cancer who are at increased risk of chemotherapy-induced toxicity and to determine the factors that affect treatment-related side effects. Methods: Ninety-five features of the health of 74 patients prior to the first round of chemotherapy were chosen for training data, using general toxicity as the predictor. Following data processing, Random Forest models were constructed to balance accuracy and interpretability. Results: We developed a machine learning predictor that ranks numerical and categorical features for toxicity. Conclusions: The use of artificial intelligence to predict and manage toxicities in the treatment of colorectal cancer is a major step forward in the direction of more individualized and precise medical care.
|
|---|