Kratter Tom, Shapira Noam, Lev Yarden, Mauda Or, Moshkovitz Yehonatan, Shitrit Roni, Konyo Shani, Ukashi Offir, Dar Lior, Shlomi Oranit, Albshesh Ahmad, Soffer Shelly, Klang Eyal, Ben Horin Shomron, Eliakim Rami, Kopylov Uri, Margalit Yehuda Reuma
Penta-AI, Tel Aviv 6701101, Israel.
Faculty of Medicine, Ben-Gurion University of the Negev, Be'er Sheva 8410501, Israel.
Diagnostics (Basel). 2022 Oct 14;12(10):2490. doi: 10.3390/diagnostics12102490.
The aim of our study was to create an accurate patient-level combined algorithm for the identification of ulcers on CE images from two different capsules.
We retrospectively collected CE images from PillCam-SB3's capsule and PillCam-Crohn's capsule. ML algorithms were trained to classify small bowel CE images into either normal or ulcerated mucosa: a separate model for each capsule type, a cross-domain model (training the model on one capsule type and testing on the other), and a combined model.
The dataset included 33,100 CE images: 20,621 PillCam-SB3 images and 12,479 PillCam-Crohn's images, of which 3582 were colonic images. There were 15,684 normal mucosa images and 17,416 ulcerated mucosa images. While the separate model for each capsule type achieved excellent accuracy (average AUC 0.95 and 0.98, respectively), the cross-domain model achieved a wide range of accuracies (0.569-0.88) with an AUC of 0.93. The combined model achieved the best results with an average AUC of 0.99 and average mean patient accuracy of 0.974.
A combined model for two different capsules provided high and consistent diagnostic accuracy. Creating a holistic AI model for automated capsule reading is an essential part of the refinement required in ML models on the way to adapting them to clinical practice.
我们研究的目的是创建一种准确的患者水平组合算法,用于识别来自两种不同胶囊的胶囊内镜(CE)图像上的溃疡。
我们回顾性收集了来自PillCam - SB3胶囊和PillCam - 克罗恩病胶囊的CE图像。使用机器学习(ML)算法将小肠CE图像分类为正常或溃疡黏膜:每种胶囊类型分别建立一个模型、一个跨域模型(在一种胶囊类型上训练模型并在另一种上测试)以及一个组合模型。
数据集包括33,100张CE图像:20,621张PillCam - SB3图像和12,479张PillCam - 克罗恩病图像,其中3582张为结肠图像。有15,684张正常黏膜图像和17,416张溃疡黏膜图像。虽然每种胶囊类型的单独模型都取得了优异的准确率(平均曲线下面积分别为0.95和0.98),但跨域模型的准确率范围较广(0.569 - 0.88),曲线下面积为0.93。组合模型取得了最佳结果,平均曲线下面积为0.99,平均患者准确率为0.974。
针对两种不同胶囊的组合模型提供了高且一致的诊断准确率。创建一个用于自动胶囊解读的整体人工智能模型是ML模型在适应临床实践过程中所需改进的重要组成部分。