Abstract
Category:
Hindfoot; Ankle
Introduction/Purpose:
Traditionally, foot and ankle alignment is quantified on 2D standing radiographs. This imaging method present limitations in capturing the 3D nature of our foot and ankle alignment and the obtained manual measurements are prone to observer error. This study aims to integrate weight-bearing CT (WBCT) imaging and advanced deep learning techniques to automate and enhance quantification our 3D foot and ankle alignment.
Methods:
Thirty-two patients who underwent a WBCT of the foot and ankle were retrospectively included. Training and validation of a 3D nnU-Net model on 45 cases to automate the segmentation into bony models was assessed by Hausdorff distance (measures how far two 3D bone model scans are from each other) and dice coefficient (measures the similarity between 3D bone model scans). Afterwards, 35 clinically relevant 3D measurements were automatically computed using a custom-made tool. Automated measurements were assessed for accuracy against manual measurements, while the latter were analyzed for inter-observer reliability.
Results:
Deep learning segmentation results showed a mean Hausdorff distance of 1.41mm and a mean dice coefficient of 0.95. A good to excellent reliability and mean prediction error of under 2 degrees was found for all angles except the talonavicular coverage angle and distal metatarsal articular angle, which were under 2.5 degrees.
Conclusion:
In summary, this study introduces a fully automated framework (Fig. 1) to quantify foot and ankle alignment and showcases reliability comparable to current clinical practice measurements. This operator-friendly and time-efficient tool holds promise for implementation in a clinical setting, benefitting both radiologists and surgeons. Future studies are encouraged to assess the tool's impact on streamlining image assessment workflows in clinical practice.
