Barbano Carlo Alberto, Berton Luca, Renzulli Riccardo, Tricarico Davide, Rampado Osvaldo, Basile Domenico, Busso Marco, Grosso Marco, Grangetto Marco
Computer Science Dept., University of Turin, Italy.
Medical Physics Department, A.O.U. Città della Salute e della Scienza di Torino, Turin, Italy.
Comput Struct Biotechnol J. 2024 Dec 5;24:754-761. doi: 10.1016/j.csbj.2024.11.045. eCollection 2024 Dec.
In this paper, we present the significant results from the Covid Radiographic imaging System based on AI (Co.R.S.A.) project, which took place in Italy. This project aims to develop a state-of-the-art AI-based system for diagnosing Covid-19 pneumonia from Chest X-ray (CXR) images. The contributions of this work are manifold: the release of the public CORDA dataset, a deep learning pipeline for Covid-19 detection and prioritization, the clinical validation of the developed solution by expert radiologists, and an in-depth analysis of possible biases embedded in the data and in the models, in order to build more trust in our AI-based pipeline. The proposed detection model is based on a two-step approach that provides reliable results based on objective radiological findings. Our prioritization scheme ensures the ordering of the patients so that severe cases are presented first. We showcase the impact of our pipeline on radiologists' workflow with a clinical study, allowing us to assess the real benefits in terms of accuracy and time efficiency. Project homepage: https://corsa.di.unito.it/.
在本文中，我们展示了基于人工智能的新冠放射成像系统（Co.R.S.A.）项目在意大利开展后取得的重要成果。该项目旨在开发一种基于人工智能的先进系统，用于从胸部X光（CXR）图像诊断新冠病毒肺炎。这项工作的贡献是多方面的：发布公开的CORDA数据集、用于新冠病毒检测和优先级排序的深度学习管道、由放射科专家对所开发解决方案进行临床验证，以及对数据和模型中可能存在的偏差进行深入分析，以便在我们基于人工智能的管道中建立更多信任。所提出的检测模型基于两步法，该方法基于客观的放射学发现提供可靠结果。我们的优先级排序方案确保对患者进行排序，以便首先呈现重症病例。我们通过一项临床研究展示了我们的管道对放射科医生工作流程的影响，从而使我们能够在准确性和时间效率方面评估实际益处。项目主页：https://corsa.di.unito.it/ 。
