Decision tree-based methodology to select a proper approach for wart treatment

Area of Science:

• Dermatology
• Medical Informatics
• Computational Biology

Background:

• Human papillomaviruses (HPVs) cause benign tumors known as warts, commonly affecting hands and feet.
• Current treatments like cryotherapy and immunotherapy require careful selection based on individual patient factors for optimal outcomes.

Purpose of the Study:

• To develop accurate predictive models using the Classification and Regression Tree (CART) algorithm.
• To analyze patient response to cryotherapy and immunotherapy for common and plantar warts.
• To provide a reliable tool for selecting the best healing method based on patient circumstances.

Main Methods:

• Utilized the CART algorithm to build predictive models for treatment response.
• Incorporated patient variables such as age, gender, wart type, number, surface area, and treatment duration.
• Included initial induration diameter for immunotherapy model parameters.

Main Results:

• The implemented CART models achieved high accuracy in predicting patient outcomes for wart treatments.
• The decision tree-based models demonstrated simplicity and reliability compared to existing fuzzy rule-based methods.
• Error analysis confirmed the high performance of the developed predictive models.

Conclusions:

• The developed CART models offer a reliable and efficient method for predicting patient response to wart therapies.
• These models can assist healthcare providers in optimizing treatment selection, saving time and resources.
• The study highlights the potential of machine learning in personalized dermatological treatment planning.