使用人工智能驱动框架量化足踝三维对线:一项先导研究
本研究开发并评估了一种基于热图预测的3D U-Net深度学习框架,可直接从负重CT图像自动检测22个解剖标志点,无需分割和迭代网格配准。在74例骨科患者(含高弓足和扁平外翻足)中,五折交叉验证显示标志点平均绝对距离误差为1.00–1.88 mm,自动测量的临床角度平均绝对误差为0.91°–2.90°。作者认为该人工智能方法可实现足踝对线的自动评估,精度与既往报道的人工测量者间变异相当,但作为探索性研究仍需更大数据集进一步评估其真实临床适用性。
要点
- 3D U-Net热图预测可直接从负重CT检测22个足踝解剖标志点,无需分割和迭代网格配准。
- 标志点平均绝对距离误差1.00–1.88 mm,临床角度平均绝对误差0.91°–2.90°。
- 自动测量精度与既往报道的人工测量者间变异相当,但本研究为探索性先导研究。
核心结论
- 基于热图的人工智能方法能够从负重CT图像自动评估足踝对线,精度与既往研究中报道的人工测量者间变异相当,该人工智能驱动方法可能是评估足踝形态的有价值方法,但这项探索性研究需要更大数据集进一步评估其真实临床适用性。
研究设计与样本
- 先导研究,使用3D U-Net模型通过热图预测直接从负重CT图像预测22个解剖标志点,采用五折交叉验证评估每个标志点和每个角度的平均绝对误差;数据集为74例骨科患者,包括高弓足和扁平外翻足等足部畸形病例。
诊断标准与关键数值
- 标志点平均绝对距离误差从第一趾骨近端头中心的最小1.00 mm至跟骨最低点的最大1.88 mm;自动临床测量平均绝对误差从后足角的最小0.91°至Böhler角的最大2.90°。
局限性
- 这项探索性研究需要更大数据集进一步评估其真实临床适用性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVE: Accurate assessment of foot and ankle alignment through clinical measurements is essential for diagnosing deformities, treatment planning, and monitoring outcomes. The traditional 2D radiographs fail to fully represent the 3D complexity of the foot and ankle. In contrast, weight-bearing CT provides a 3D view of bone alignment under physiological loading. Nevertheless, manual landmark identification on WBCT remains time-intensive and prone to variability. This study presents a novel AI framework automating foot and ankle alignment assessment via deep learning landmark detection.
MATERIALS AND METHODS: By training 3D U-Net models to predict 22 anatomical landmarks directly from weight-bearing CT images, using heatmap predictions, our approach eliminates the need for segmentation and iterative mesh registration methods. A small dataset of 74 orthopedic patients, including foot deformity cases such as pes cavus and planovalgus, was used to develop and evaluate the model in a clinically relevant population. The mean absolute error was assessed for each landmark and each angle using a fivefold cross-validation.
RESULTS: Mean absolute distance errors ranged from 1.00 mm for the proximal head center of the first phalanx to a maximum of 1.88 mm for the lowest point of the calcaneus. Automated clinical measurements derived from these landmarks achieved mean absolute errors between 0.91° for the hindfoot angle and a maximum of 2.90° for the Böhler angle.
CONCLUSION: The heatmap-based AI approach enables automated foot and ankle alignment assessment from WBCT imaging, achieving accuracies comparable to the manual inter-rater variability reported in previous studies. This novel AI-driven method represents a potentially valuable approach for evaluating foot and ankle morphology. However, this exploratory study requires further evaluation with larger datasets to assess its real clinical applicability.
引用
- Huysentruyt R, Audenaert E, Van den Borre I, et al. Quantifying 3D foot and ankle alignment using an AI-driven framework: a pilot study. Skeletal Radiol. 2025. PMID: 41014330. doi:10.1007/s00256-025-05038-6
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。