A Cross sectional study of the utility of Artificial Intelligence in Direct Immunofluorescence microscopy: An image based approach
试验速览
- 阶段
- 不适用
- 状态
- 尚未招募
- 发起方
- 入组人数
- 1,600
- 试验地点
- 1
- 主要终点
- 1) To assess the utility of AI in diagnostic algorithm of DIF microscopy.
研究概览
简要总结
This study is a cross-sectional study which aims to assess the utility of artificial intelligence (AI) in interpreting direct immunofluorescence (DIF) images for diagnosing autoimmune skin diseases and vasculitis. DIF microscopy is the gold standard for diagnosis of Autoimmune bullous diseases and useful in diagnosis of vasculitis but depends heavily on the expertise of dermatopathologists and is time-consuming. The study will compare the diagnostic accuracy and time taken by AI models with that of human experts using 1600 DIF images from skin biopsy slides. Various AI approaches, including machine learning, deep learning, and vision transformers, will be evaluated to identify which provides the best performance. The goal is to determine if AI can assist dermatopathologists by offering faster and reliable diagnoses, potentially improving diagnostic accuracy and efficiency in clinical practice.
研究设计
- 研究类型
- Observational
入排标准
- 年龄范围
- 1.00 Day(s) 至 99.00 Year(s)（—）
- 性别
- All
入选标准
- •All DIF images (obtained from slides from skin biopsy samples) from patients with clinically suspected AIBDs and Vasculitis – that are received in DIF lab 2) All DIF Images in DIF lab with confirmed and labelled diagnosis (AIBDs and Vasculitis) 3)Slides diagnosed as negative (For training of AI).
排除标准
- •Poor quality images obtained from slides which were obtained from biopsies with insufficient dermis 2)Formalin stained samples.
结局指标
主要结局
1) To assess the utility of AI in diagnostic algorithm of DIF microscopy.
时间窗: 24 months
2) To compare human interpretation of DIF slides with that of AI diagnosis.
时间窗: 24 months
次要结局
- To compare different AI algorithms for their accuracy in diagnosis of DIF slides
- To establish potential advantages of AI in improving time required in diagnosing and accuracy of interpretation of diseases : a) Autoimmune bullous diseases b) Vasculitis(24 months)
研究者
Aishwarya Dhanuka
Kasturba Medical College, Manipal