AI-MDT: an automatic and intelligent multidisciplinary team consultations platform for lung cancer diagnosis

AI-MDT:一个用于肺癌诊断的自动化智能多学科团队会诊平台

阅读:1

Abstract

PURPOSE: Multidisciplinary team (MDT) consultations are crucial for managing pulmonary nodules, yet face challenges in efficiency, evidence-based decision support, and data utilization within the MDT process. We present an integrated artificial intelligence (AI)-MDT platform that serves as an assistive tool for lung cancer MDT workflows by incorporating AI across various processes.The aim of this study is to evaluate the clinical utility and preliminary efficacy of the AI-MDT platform. METHODS: The platform comprises three core modules: process automation, intelligent decision support, and diagnostic assistance. It integrates a real-time, evidence-based knowledge base powered by large language models and deep learning, with computer vision for automatic lesion detection and feature analysis. A web-based interface allows users to interact seamlessly with the AI-MDT platform. RESULTS: Since its implementation in November 2023 at a tertiary Grade A hospital in China, the platform has been involved in 879 consultations, including 811 patients. AI-generated diagnostic recommendations were utilized 852 times, and decision-making support was used in 744 cases. The platform significantly increased consultation volume, reduced expert time, and enhanced data utilization compared to traditional MDT. CONCLUSIONS: It offers clinicians tools to improve diagnostic quality and work efficiency, highlighting its significant clinical application value. These findings suggest that the proposed platform contributes to the emerging research on advances precision lung cancer management by integrating a continually updated evidence base and intelligent imaging methodologies, having potential implications for MDT processes across various medical specialties.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。