MRI-Based Artificial Intelligence Study for Detection, Classification, and Grading of Renal Tumors Using Preoperative Imaging

What this Trial Is About

Researchers are conducting a retrospective and prospective observational study to develop and evaluate artificial intelligence AI methods for detecting, classifying, and grading renal tumors using preoperative magnetic resonance imaging MRI. The study involves about 900 adult patients who have undergone preoperative renal MRI and have corresponding postoperative pathological results. The pathological findings serve as the reference standard for AI model training and validation. Additional demographic, clinical, and laboratory data may be included to enhance model performance. This research aims to create an MRI-based AI diagnostic tool to aid clinicians in assessing and managing renal tumors before surgery. The study uses existing MRI data, including various sequences like T1-weighted, T2-weighted, diffusion-weighted, fat-suppressed, apparent diffusion coefficient, and contrast-enhanced imaging when available. AI models are developed to detect, localize, segment renal tumors, classify common and rare pathological subtypes, and predict histological grades for malignant tumors. The dataset is divided into development and validation groups, with additional cases used for independent testing. No new imaging or diagnostic procedures, treatments, or follow-up visits are required, as the study is non-interventional and observational. Participants involvement consists of using previously collected de-identified MRI, pathological, clinical, and laboratory data. No changes to diagnosis, treatment, or clinical care will occur. Researchers will compare AI predictions with pathological diagnoses to assess accuracy in tumor detection, subtype classification, and grading. The study will measure AI model accuracy at the completion of evaluation using data collected from January 2021 through December 2026. This approach ensures participant safety and confidentiality while aiming to improve future preoperative decision-making for renal tumor patients.

An MRI-Based Study of Intelligent Pathological Subtyping and Grading of Renal Tumors