Bao Mingbo, Liu Wenjia, Shi Haifeng, Meng Mingzhu, Cao Jian
Department of Radiology, Chinese Medical Hospital of Wujin, Changzhou, China.
Department of Gastroenterology, The Second People's Hospital of Changzhou, The Third Affiliated Hospital of Nanjing Medical University, Changzhou Medical Center, Nanjing Medical University, Changzhou, China.
Gastroenterol Res Pract. 2025 Jul 23;2025:1506567. doi: 10.1155/grp/1506567. eCollection 2025.
Inflammatory bowel disease (IBD) is an immune-mediated disorder characterized by intestinal inflammation and includes two subtypes: Crohn's disease (CD) and ulcerative colitis (UC). The computed tomography manifestations of colonic CD (cCD) and UC are similar, and differential diagnosis is challenging. Our study aimed to investigate the feasibility of using a modified YOLOv5 algorithm for differentiating between cCD and UC on computed tomography enterography (CTE) images. This multicenter retrospective study analyzed data from a total of 29 cCD patients and 29 UC patients. Five submodels (YOLOv5n, YOLOv5s, YOLOv5m, YOLOv5l, and YOLOv5x) of YOLOv5 were trained and evaluated on the datasets. The CTE images of the cCD group and UC group were divided into a training set, validation set, and test set at a ratio of 8:1:1. Finally, the precision (Pr), recall rate (Rc), and mean average precision (mAP and mAP) of the models were compared. The YOLOv5x model showed the best performance among the five submodels, with mAP of 0.97 and mAP of 0.97 and 0.84 in the validation set and mAP and mAP of 0.97 and 0.83 in the test set, respectively. These results demonstrated similar diagnostic accuracy to the two radiologists (84.5%). The modified YOLOv5 algorithm is a feasible approach to distinguish between cCD and UC on CTE images. These findings may facilitate the early detection and differential diagnosis of IBD.
炎症性肠病（IBD）是一种以肠道炎症为特征的免疫介导性疾病，包括两种亚型：克罗恩病（CD）和溃疡性结肠炎（UC）。结肠型CD（cCD）和UC的计算机断层扫描表现相似，鉴别诊断具有挑战性。我们的研究旨在探讨使用改进的YOLOv5算法在计算机断层扫描小肠造影（CTE）图像上区分cCD和UC的可行性。这项多中心回顾性研究分析了总共29例cCD患者和29例UC患者的数据。在数据集上对YOLOv5的五个子模型（YOLOv5n、YOLOv5s、YOLOv5m、YOLOv5l和YOLOv5x）进行了训练和评估。cCD组和UC组的CTE图像按8:1:1的比例分为训练集、验证集和测试集。最后，比较了模型的精度（Pr）、召回率（Rc）和平均精度均值（mAP和mAP）。YOLOv5x模型在五个子模型中表现最佳，在验证集中mAP为0.97，mAP为0.97和0.84，在测试集中mAP和mAP分别为0.97和0.83。这些结果显示出与两位放射科医生相似的诊断准确性（84.5%）。改进的YOLOv5算法是在CTE图像上区分cCD和UC的一种可行方法。这些发现可能有助于IBD的早期检测和鉴别诊断。
