*Result*: Quantifying 3D foot and ankle alignment using an AI-driven framework: a pilot study.
Original Publication: Berlin, New York, Springer International.
McPoil TG, Hunt GC. Evaluation and management of foot and ankle disorders: present problems and future directions. J Orthop Sports Phys Ther. 1995;21(6):381–8. (PMID: 10.2519/jospt.1995.21.6.3817655482)
Lamm BM, Stasko PA, Gesheff MG, Bhave A. Normal foot and ankle radiographic angles, measurements, and reference points. J Foot Ankle Surg. 2016;55(5):991–8. (PMID: 10.1053/j.jfas.2016.05.00527320694)
Younger AS, Sawatzky B, Dryden P. Radiographic assessment of adult flatfoot. Foot Ankle Int. 2005O 1;26(10):820–5. (PMID: 10.1177/10711007050260100616221454)
AQTochigi Y, Suh JS, Amendola A, Saltzman CL. Ankle alignment on lateral radiographs. Part 2: reliability and validity of measures. Foot Ankle Int. 2006;27(2):88–92. (PMID: 10.1177/107110070602700203)
Li J, Fang M, Van Oevelen A, Peiffer M, Audenaert E, Burssens A. Diagnostic applications and benefits of weightbearing CT in the foot and ankle: a systematic review of clinical studies. Foot Ankle Surg. 2024;30:7–20. (PMID: 10.1016/j.fas.2023.09.00137704542)
Chun DI, Cho J, Won SH, Nomkhondorj O, Kim J, An CY, et al. Weight-bearing CT: advancing the diagnosis and treatment of hallux valgus, midfoot pathology, and progressive collapsing foot deformity. Diagnostics. 2025J 31;15(3):343. (PMID: 10.3390/diagnostics150303433994127311817285)
Burssens A, Peeters J, Buedts K, Victor J, Vandeputte G. Measuring hindfoot alignment in weight-bearing CT: a novel clinically relevant measurement method. Foot Ankle Surg. 2016D 1;22(4):233–8. (PMID: 10.1016/j.fas.2015.10.00227810020)
Netto CDC, Schon LC, Thawait GK, Da Fonseca LF, Chinanuvathana A, Zbijewski WB, et al. Flexible adult acquired flatfoot deformity: comparison between weight-bearing and non-weight-bearing measurements using cone-beam computed tomography. J Bone Joint Surg Am. 2017;99(18).
Carvalho KAM de, Mallavarapu V, Ehret A, Dibbern K, Lee HY, Barbachan Mansur NS, et al. The use of advanced semiautomated bone segmentation in hallux rigidus. Foot Ankle Orthop. 2022;7(4).
Peterson AC, Kruger KM, Lenz AL. Automatic anatomical foot and ankle coordinate toolbox. Front Bioeng Biotechnol. 2023;11. (PMID: 10.3389/fbioe.2023.12554643802687510644787)
Richter M, Schilke R, Duerr F, Zech S, Meissner SA, Naef I. Automatic software-based 3D-angular measurement for weight-bearing CT (WBCT) provides different angles than measurement by hand. Foot Ankle Surg. 2022;28(7):863–71. (PMID: 10.1016/j.fas.2021.11.01034876354)
Van den Borre I, Peiffer M, Huysentruyt R, et al. Development and validation of a fully automated tool to quantify 3D foot and ankle alignment using weight-bearing CT. Gait Posture. 2024;1(113):67–74. (PMID: 10.1016/j.gaitpost.2024.05.029)
Kuiper RJA, Seevinck PR, Viergever MA, Weinans H, Sakkers RJB. Automatic assessment of lower-limb alignment from computed tomography. J Bone Joint Surg. 2023;105(9):700–12. (PMID: 10.2106/JBJS.22.0089036947661)
Peiffer M, Van Den Borre I, Segers T, Ashkani-Esfahani S, Guss D, De Cesar NC, et al. Implementing automated 3D measurements to quantify reference values and side-to-side differences in the ankle syndesmosis. Sci Rep. 2023;13(1):1–10. (PMID: 10.1038/s41598-023-40599-3)
Choi M, Jang JS. Heatmap-based active shape model for landmark detection in lumbar X-ray images. Journal of imaging informatics in medicine. 2025; 38(1).
Yang J, Ren P, Xin P, Wang Y, Ma Y, Liu W, et al. Automatic measurement of lower limb alignment in portable devices based on deep learning for knee osteoarthritis. J Orthop Surg Res. 2024D 1;19(1):1–8.
Takahashi K, Shimamura Y, Tachiki C, Nishii Y, Hagiwara M. Cephalometric landmark detection without X-rays combining coordinate regression and heatmap regression. Sci Rep. 2023;13(1):20011. (PMID: 10.1038/s41598-023-46919-x3797401810654665)
Wang X, Rigall E, Chen Q, Zhang S, Dong J. Efficient and stable cephalometric landmark localization using two-stage heatmaps’ regression. IEEE Trans Instrum Meas. 2022. https://doi.org/10.1109/TIM.2022.3206762 . (PMID: 10.1109/TIM.2022.3206762)
Xue H, Artico J, Fontana M, Moon JC, Davies RH, Kellman P. Landmark detection in cardiac MRI by using a convolutional neural network. Radiology: Artificial Intelligence. 2021. https://doi.org/10.1148/ryai.2021200197 . (PMID: 10.1148/ryai.2021200197346170228489464)
Tan Z, Feng J, Zhou J. Multi-task learning network for landmark detection in anatomical tree structures. Proceedings - International Symposium on Biomedical Imaging. 2021;1975–9.
Barbosa RM, Serrador L, da Silva MV, Macedo CS, Santos CP. Knee landmarks detection via deep learning for automatic imaging evaluation of trochlear dysplasia and patellar height. Eur Radiol. 2024S 1;34(9):5736–47. (PMID: 10.1007/s00330-024-10596-93833707211364617)
Baltus SC, Geitenbeek RTJ, Frieben M, et al. Deep learning-based pelvimetry in pelvic MRI volumes for pre-operative difficulty assessment of total mesorectal excision. Surg Endosc. 2025. https://doi.org/10.1007/s00464-024-11485-4 . (PMID: 10.1007/s00464-024-11485-44082589612408677)
Jonkers J, Duchateau L, van Wallendael G, van Hoecke S. landmarker: a toolkit for anatomical landmark localization in 2D/3D images. SoftwareX. 2025;2(30).
Otsu N. Threshold selection method from gray-level histograms. IEEE Trans Syst Man Cybern. 1979;SMC-9(1):62–6.
Audenaert EA, Pattyn C, Steenackers G, De Roeck J, Vandermeulen D, Claes P. Statistical shape modeling of skeletal anatomy for sex discrimination: their training size, sexual dimorphism, and asymmetry. Front Bioeng Biotechnol. 2019;1:7.
Audenaert EA, Van Houcke J, Almeida DF, et al. Cascaded statistical shape model based segmentation of the full lower limb in CT. Comput Methods Biomech Biomed Engin. 2019A 26;22(6):644–57. (PMID: 10.1080/10255842.2019.157782830822149)
Li, S., Xiang, X. Lightweight human pose estimation using loss weighted by target heatmap. Pattern Recognition, Computer Vision, and Image Processing. ICPR 2022.
*Further Information*
*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.
(© 2025. The Author(s), under exclusive licence to International Skeletal Society (ISS).)*
*Declarations. Ethical approval and consent to participate: All procedures involving human participants were approved by the Institutional Review Board of the University Hospital of Ghent (B6702022000639). The study was conducted in accordance with the Declaration of Helsinki and the Guidelines for Good Clinical Practice. Informed consents were obtained from the participants. Competing interests: The authors declare no conflict of interests.*