Abstract
We developed and externally validated a hierarchical deep learning pipeline that automates cervical sagittal measurements, explicitly addressing C7 obscuration on lateral radiographs. The model combines a global keypoint detector with C2/C7 specialists localized via a multilayer perceptron to refine landmarks on high‑resolution patches. Trained on 5604 images and tested internally and on a challenging external cohort enriched for C7 obscuration (82%), it achieved excellent agreement with ground truth. Externally, intraclass correlation coefficients (ICCs) were 0.97 for lordosis (mean absolute error [MAE] 2.6°), >0.99 for C2 slope (MAE 0.8°), and 0.93 for C7 slope (MAE 2.3°), with minimal bias and narrower limits of agreement than a single‑stage baseline. The model showed near-perfect repeatability (ICC > 0.99) and higher artificial intelligence-expert agreement (ICC 0.81–0.84) for C7 slope than inter-expert reliability (ICC 0.67). In failure cases, the pipeline corrected large global model errors (e.g., 10.22°– 0.22°). This robust, coarse‑to‑fine approach advances reliable, generalizable cervical alignment assessment in real‑world conditions.
Data availability
The institutional datasets generated and/or analyzed during the current study are not publicly available due to patient privacy regulations and restrictions imposed by the IRB of Samsung Medical Center. However, anonymized data may be made available from the corresponding author upon reasonable request and pending approval of a formal data sharing agreement.
Code availability
The code used in this study is available from the corresponding author upon reasonable request.
References
Ames, C. P. et al. Cervical radiographical alignment: comprehensive assessment techniques and potential importance in cervical myelopathy. Spine 38, S149–160 (2013).
Kim, S. W. Latest knowledge on a comprehensive understanding of cervical deformity and selection of effective treatment methods using recent classification systems: a narrative review. Asian Spine J. 18, 608–620 (2024).
Lee, D.-H., Lee, H. R. & Riew, K. D. An algorithmic roadmap for the surgical management of degenerative cervical myelopathy: a narrative review. Asian Spine J. 18, 274–286 (2024).
Kim, M. W., Park, Y. S., Kang, C. N. & Choi, S. H. Cervical spondylotic myelopathy and radiculopathy: a stepwise approach and comparative analysis of surgical outcomes: a narrative review of recent literature. Asian Spine J. 19, 121–132 (2025).
Park, M. S. et al. Sagittal alignment as a predictor of clinical adjacent segment pathology requiring surgery after anterior cervical arthrodesis. Spine J. 14, 1228–1234 (2014).
Ling, F. P. et al. Which parameters are relevant in sagittal balance analysis of the cervical spine? A literature review. Eur. Spine J. 27, 8–15 (2018).
Lee, S. H., Hyun, S. J. & Jain, A. Cervical Sagittal Alignment: Literature Review and Future Directions. Neurospine 17, 478–496 (2020).
Wang, Z. et al. Can C7 slope accurately substitute for an invisible T1 slope according to age and cervical morphology in cervical lateral radiographs? Glob. Spine J. 15, 406–416 (2025).
Kang, D. H. et al. Automated measurement of pelvic parameters using convolutional neural network in complex spinal deformities: overcoming challenges in coronal deformity cases. Spine J. 25, 1688–1697 (2025).
Fujimori, T. et al. Development of artificial intelligence for automated measurement of cervical lordosis on lateral radiographs. Sci. Rep. 12, 15732 (2022).
Grover, P. et al. Can artificial intelligence support or even replace physicians in measuring sagittal balance? A validation study on preoperative and postoperative full spine images of 170 patients. Eur. Spine J. 31, 1943–1951 (2022).
Song, S. Y. et al. AI-driven segmentation and automated analysis of the whole sagittal spine from X-ray images for spinopelvic parameter evaluation. Bioengineering 10, 1229 (2023).
Löchel, J. et al. Deep learning algorithm for fully automated measurement of sagittal balance in adult spinal deformity. Eur. Spine J. 6, 4119–4124 (2024).
Nakarai, H. et al. Automatic calculation of cervical spine parameters using deep learning: development and validation on an external dataset. Glob. Spine J. 15, 710–721 (2025).
Vogt, S. et al. Novel AI-based algorithm for the automated measurement of cervical sagittal balance parameters. A validation study on pre- and postoperative radiographs of 129 patients. Glob. Spine J. 15, 1155–1165 (2025).
Liawrungrueang, W., Cho, S. T., Sarasombath, P., Kim, I. & Kim, J. H. Current trends in artificial intelligence-assisted spine surgery: a systematic review. Asian Spine J. 18, 146–157 (2024).
Cina, A. et al. 2-step deep learning model for landmarks localization in spine radiographs. Sci. Rep. 11, 9482 (2021).
Arslanoglu, M. C., Albayrak, A. & Acar, H. Vision transformers versus convolutional neural networks: comparing robustness by exploiting varying local features. IEEE Access 13, 65232–65245 (2025).
Ran, Y. et al. A high-quality dataset featuring classified and annotated cervical spine X-ray atlas. Sci. Data 11, 625 (2024).
Zhang, M. CLX-34: Cervical lateral X-ray 34-point dataset. Harvard Dataverse (2023).
He, K., Gkioxari, G., Dollar, P. & Girshick, R. Mask R-CNN. IEEE Trans. Pattern Anal. Mach. Intell. 42, 386–397 (2020).
Goodfellow, I. Deep Learning (MIT Press, 2017).
Acknowledgements
This study utilized the open-access Cervical Spine X-ray Atlas (CSXA) dataset, which was developed and released by Ran et al. (Scientific Data 2024;11:625) under the Creative Commons Attribution 4.0 International License. We gratefully acknowledge the authors and contributors of the CSXA dataset for making this valuable resource available to the public. In addition, this study used the open-source CLX-34 dataset, a cervical lateral X-ray benchmark with 34 anatomical keypoints designed for landmark detection and parameter computation, which has been publicly released for research purposes. The authors received no financial support for the research, authorship, or publication of this article.
Author information
Authors and Affiliations
Contributions
Conceptualization: D.H.K. and S.J.P.; Methodology: D.H.K. and S.J.P.; Software: D.H.K.; Validation: D.H.K., S.J.P., J.S.P., and H.K.; Formal analysis: D.H.K. and S.J.P.; Investigation: D.H.K. and S.J.P.; Resources: D.H.K., J.S.P., and C.S.L.; Data curation: D.H.K., S.J.P., J.S.P., and H.K.; Writing—original draft: D.H.K.; Writing—review and editing: D.H.K., S.J.P., J.S.P., and C.S.L.; Visualization: D.H.K.; Supervision: C.S.L.; Project administration: D.H.K.; Funding acquisition: None declared.
Corresponding authors
Ethics declarations
Competing interests
The authors declare the following competing interests: A patent application entitled “Method and apparatus for calculating spinal alignment parameters” has been filed by Samsung Medical Center (Samsung Life Public Welfare Foundation) (Application No. 10-2025-0164287, Inventors: Dong-Ho Kang, Status: Pending). This patent covers the hierarchical deep learning architecture for spinal parameter measurement described in this manuscript. The other authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Kang, DH., Park, SJ., Park, JS. et al. Hierarchical deep learning pipeline for robust cervical parameter measurement in radiographs with C7 obscuration. npj Digit. Med. (2026). https://doi.org/10.1038/s41746-026-02455-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41746-026-02455-2
