Enhancing the Etiological Diagnosis of Uveitis with Artificial Intelligence: A Predictive Approach

Artificial Intelligence (AI) has showcased its potential across various domains, particularly within the healthcare industry. This research explores the intersection of AI and healthcare, with a focus on developing a predictive method discerning the underlying cause of uveitis, an inflammation condition affecting the eye's uvea. With multiple origins contributing to uveitis, its identification is difficult due to the overlap of symptoms. The prevention of severe vision impairment through the effective prediction and management of uveitis is of utmost urgency. In this paper, we present a Random Forest-based predictive model adept at handling imbalanced datasets and identifying critical features. The study aims to comprehensively address primary etiologies while selecting meaningful features. Results show the model performs well with a global sensitivity of 78% for the most probable cause, validating AI's potential to assist clinicians with difficult diagnoses. Subsequent work will integrate the model into a web application facilitating practical evaluation by ophthalmologists. This evaluation will leverage expert contributed patient data to improve diagnostic accuracy and efficiency.