Department of Diagnostic Radiology, Fondazione IRCCS San Gerardo dei Tintori, Via Pergolesi 33, 20900 Monza, MB, Italy.
Department of Diagnostic Radiology, Fondazione IRCCS San Gerardo dei Tintori, Via Pergolesi 33, 20900 Monza, MB, Italy.
Eur J Radiol. 2024 Sep;178:111637. doi: 10.1016/j.ejrad.2024.111637. Epub 2024 Jul 17.
To evaluate the diagnostic performance of an Artificial Intelligence (AI) algorithm, previously trained using both adult and pediatric patients, for the detection of acute appendicular fractures in the pediatric population on conventional X-ray radiography (CXR).
In this retrospective study, anonymized extremities CXRs of pediatric patients (age <17 years), with or without fractures, were included. Six hundred CXRs (maintaining the positive-for-fracture and negative-for-fracture balance) were included, grouping them per body part (shoulder/clavicle, elbow/upper arm, hand/wrist, leg/knee, foot/ankle). Follow-up CXRs and/or second-level imaging were considered as reference standard. A deep learning algorithm interpreted CXRs for fracture detection on a per-patient, per-radiograph, and per-location level, and its diagnostic performance values were compared with the reference standard. AI diagnostic performance was computed by using cross-tables, and 95 % confidence intervals [CIs] were obtained by bootstrapping.
The final cohort included 312 male and 288 female with a mean age of 8.9±4.5 years. Three undred CXRs (50 %) were positive for fractures, according to the reference standard. For all fractures, the AI tool showed a per-patient 91.3% (95%CIs = 87.6-94.3) sensitivity, 76.7% (71.5-81.3) specificity, and 84% (82.1-86.0) accuracy. In the per-radiograph analysis the AI tool showed 85% (81.9-87.8) sensitivity, 88.5% (86.3-90.4) specificity, and 87.2% (85.7-89.6) accuracy. In the per-location analysis, the AI tool identified 606 bounding boxes: 472 (77.9 %) were correct, 110 (18.1 %) incorrect, and 24 (4.0 %) were not-overlapping.
The AI algorithm provides good overall diagnostic performance for detecting appendicular fractures in pediatric patients.
评估一种人工智能（AI）算法在常规 X 射线摄影（CXR）中检测儿科人群急性阑尾骨折的诊断性能，该算法先前使用成人和儿科患者进行了训练。
在这项回顾性研究中，纳入了有或无骨折的儿科患者（年龄 <17 岁）的四肢 CXR。共纳入 600 张 CXR，按身体部位（肩部/锁骨、肘部/上臂、手部/腕部、腿部/膝盖、脚部/脚踝）分组。随访 CXR 和/或二级影像学被认为是参考标准。深度学习算法在患者、每张 X 光片和每个部位水平上对骨折进行解读，并将其诊断性能值与参考标准进行比较。AI 诊断性能通过交叉表进行计算，95%置信区间[CI]通过自举法获得。
最终队列包括 312 名男性和 288 名女性，平均年龄为 8.9±4.5 岁。根据参考标准，300 张 CXR（50%）为阳性骨折。对于所有骨折，AI 工具的患者总体敏感性为 91.3%（95%CI=87.6-94.3），特异性为 76.7%（71.5-81.3），准确性为 84%（82.1-86.0）。在每张 X 光片的分析中，AI 工具的敏感性为 85%（81.9-87.8），特异性为 88.5%（86.3-90.4），准确性为 87.2%（85.7-89.6）。在每个部位的分析中，AI 工具共识别了 606 个边界框：472 个（77.9%）是正确的，110 个（18.1%）是错误的，24 个（4.0%）是不重叠的。
AI 算法为检测儿科患者的四肢骨折提供了良好的整体诊断性能。
