Automated analysis of spinal radiographs with calculation of sagittal balance parameters: development and validation of a clinical measurement support system

Objective. To develop and evaluate a system for automated analysis of standing anteroposterior and lateral spinal radiographs with subsequent calculation of sagittal balance parameters and determination of personalized target spinopelvic parameters based on the pelvic incidence (PI).

Material and Methods. A retrospective two-center study was conducted. The analysis included anonymized digital radiographs of the spine and pelvis in anteroposterior and lateral projections obtained in a standing position from 1,076 patients. A total of 973 fully annotated and expert-verified studies were used for training, validation, and testing. Automatic segmentation was performed using a U-Net architecture, followed by calculation of sagittal balance parameters with the possibility of deriving coronal parameters. A separate analysis of agreement between expert and automated measurements was conducted for the PI, PT, SS, GLL, and L4–S1 parameters. Segmentation quality was assessed using the Dice coefficient and IoU (Jaccard index). Inter-expert variability of annotations was additionally analyzed, and results were compared using “model–expert” and “expert–expert” schemes.

Results. On the test dataset, the mean Dice coefficient between the model and the expert was 90.1%, falling within the range of inter-expert variability (88.5–93.4%). A comparison of automated and expert measurements of PI, PT, SS, GLL, and L4–S1 parameters in a subset of 46 patients demonstrated high agreement between the methods: ICC(A,1) values ranged from 0.989 to 0.996; the 95% confidence intervals for ICC(A,1) ranged from 0.977 to 0.998; the mean absolute error ranged from 0.84° to 1.14°, and the proportion of measurements with an absolute difference of no more than 2.5° was 93.5–97.8%. The complete analysis cycle for a single patient required less than 3 seconds on a graphics processing unit (GPU). The prototype components successfully underwent pilot testing.

Conclusion. The developed system enables the automatic recognition of anatomical structures on spinal radiographs and the calculation of sagittal balance parameters with a high level of agreement with expert measurements. The system may be used to standardize radiographic measurements; however, its use in preoperative planning requires further clinical validation.

1. Glassman SD, Bridwell K, Dimar JR, Horton W, Berven S, Schwab F. The impact of positive sagittal balance in adult spinal deformity. Spine (Phila Pa 1976). 2005;30:2024–2029. DOI: http://dx.doi.org/10.1097/01.BRS.0000179086.30449.96

2. Schwab F, Ungar B, Blondel B, Buchowski J, Coe J, Deinlein D, Dewald C, Mehdian H, Shaffrey C, Tribus C, Lafage V. Scoliosis Research Society-Schwab adult spinal deformity classification: a validation study. Spine (Phila Pa 1976). 2012;37:1077–1082. DOI: http://dx.doi.org/10.1097/BRS.0b013e31823e15e2

3. Lafage R, Schwab F, Glassman S, Bess S, Harris B, Sheer J, Hart R, Line B, Henry J, Burton D, Kim HJ, Klineberg E, Ames C, Lafage V. Age-adjusted alignment goals have the potential to reduce PJK. Spine (Phila Pa 1976). 2017;42:1275–1282. DOI: http://dx.doi.org/10.1097/BRS.0000000000002146

4. Le Huec JC, Charosky S, Barrey C, Rigal J, Aunoble S. Sagittal imbalance cascade for simple degenerative spine and consequences: algorithm of decision for appropriate treatment. Eur Spine J. 2011;20 Suppl 5:S699–S703. DOI: http://dx.doi.org/10.1007/s00586-011-1938-8

5. Макиров С.К., Юз А.А., Джахаф М.Т. Методика оценки параметров сагиттального позвоночно-тазового баланса. Хирургия позвоночника. 2015;12(3):55–63. [Makirov SK, Yuz AA, Jahaf MT. Method of assessing the parameters of the sagittal spinal pelvic balance. Russian Journal of Spine Surgery (Khirurgiya Pozvonochnika). 2015;12(3):55–63]. DOI: http://dx.doi.org/10.14531/ss2015.3.55-63 EDN: https://elibrary.ru/UMGWQX

6. Байков Е.С., Пелеганчук А.В., Сангинов А.Д., Леонова О.Н., Крутько А.В. Хирургическая коррекция сагиттального дисбаланса поясничного отдела позвоночника дегенеративного генеза. Хирургия позвоночника. 2020;17(2):49–57. [Baikov ES, Peleganchuk AV, Sanginov AD, Leonova ON, Krutko AV. Surgical correction of degenerative sagittal imbalance of the lumbar spine. Russian Journal of Spine Surgery (Khirurgiya Pozvonochnika). 2020;17(2):49–57]. DOI: http://dx.doi.org/10.14531/ss2020.2.49-57 EDN: https://elibrary.ru/PTCFZB

7. Прудникова О.Г., Аранович А.М., Муштаева Ю.А., Губин А.В. Биомеханические аспекты сагиттального баланса позвоночника у пациентов с ахондроплазией при удлинении конечностей методом Илизарова. Хирургия позвоночника. 2018;15(4):7–14. [Prudnikova OG, Aranovich AM, Mushtaeva YuA, Gubin AV. Biomechanical aspects of spinal sagittal balance in achondroplasia patients during Ilizarov limb lengthening. Russian Journal of Spine Surgery (Khirurgiya Pozvonochnika). 2018;15(4):7–14]. DOI: http://dx.doi.org/10.14531/2018.4.7-14 EDN: https://elibrary.ru/PPXBWX

8. Litjens G, Kooi T, Bejnordi BE, Setio AAA, Ciompi F, Ghafoorian M, van der Laak JAWM, van Ginneken B, Sánchez CI. A survey on deep learning in medical image analysis. Med Image Anal. 2017;42:60–88. DOI: http://dx.doi.org/10.1016/j.media.2017.07.005

9. Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation. In: Navab N, Hornegger J, Wells WM, Frangi AF, eds. Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015: 18th International Conference Munich, Germany, October 5-9, 2015 proceedings, part III. Cham: Springer, 2015;234–241. DOI: http://dx.doi.org/10.1007/978-3-319-24574-4_28

10. Chen Y, Mo Y, Readie A, Ligozio G, Mandal I, Jabbar F, Coroller T, Papież BW. VertXNet: an ensemble method for vertebral body segmentation and identification from cervical and lumbar spinal X-rays. Sci Rep. 2024;14:3341. DOI: http://dx.doi.org/10.1038/s41598-023-49923-3

11. Karimi D, Dou H, Warfield SK, Gholipour A. Deep learning with noisy labels: exploring techniques and remedies in medical image analysis. Med Image Anal. 2020;65:101759. DOI: http://dx.doi.org/10.1016/j.media.2020.101759

12. Galbusera F, Casaroli G, Bassani T. Artificial intelligence and machine learning in spine research. JOR Spine. 2019;2:e1044. DOI: http://dx.doi.org/10.1002/jsp2.1044

13. Korez R, Putzier M, Vrtovec T. A deep learning tool for fully automated measurements of sagittal spinopelvic balance from X-ray images: performance evaluation. Eur Spine J. 2020;29:2295–2305. DOI: http://dx.doi.org/10.1007/s00586-020-06406-7

14. Vrtovec T, Ibragimov B. Spinopelvic measurements of sagittal balance with deep learning: systematic review and critical evaluation. Eur Spine J. 2022;31:2031–2045. DOI: http://dx.doi.org/10.1007/s00586-022-07155-5

15. Hoppe BF, Rueckel J, Rudolph J, Fink N, Weidert S, Hohlbein W, Cavalcanti-Kussmaul A, Trappmann L, Munawwar B, Ricke J, Sabel BO. Automated spinopelvic measurements on radiographs with artificial intelligence: a multi-reader study. Radiol Med. 2025;130:359–367. DOI: http://dx.doi.org/10.1007/s11547-025-01957-5

16. Bishara A, Patel S, Warman A, Jo J, Hughes LP, Khalifeh JM, Azad TD. Artificial intelligence automated measurements of spinopelvic parameters in adult spinal deformity - a systematic review. Spine Deform. 2025;13:1289–1304. DOI: http://dx.doi.org/10.1007/s43390-025-01111-1

17. Vila-Casademunt A, Pellisé F, Acaroglu E, Pérez-Grueso FJ, Martín-Buitrago MP, Sanli T, Yakici S, de Frutos AG, Matamalas A, Sánchez-Márquez JM, Obeid I, Yam-an O, Bagó J. The reliability of sagittal pelvic parameters: the effect of lumbosacral instrumentation and measurement experience. Spine (Phila Pa 1976). 2015;40:E253–E258. DOI: http://dx.doi.org/10.1097/BRS.0000000000000720

18. Langella F, Villafañe JH, Damilano M, Cecchinato R, Pejrona M, Ismael M, Berjano P. Predictive accuracy of Surgimap surgical planning for sagittal imbalance: a cohort study. Spine (Phila Pa 1976). 2017;42:E1297–E1304. DOI: http://dx.doi.org/10.1097/BRS.0000000000002230

19. Захарин В.Р., Леонова О.Н., Байков Е.С., Кокорев А.И., Дарчия Л.Ю., Балычев Г.Е., Иванов Д.В., Доль А.В., Бессонов Л.В., Крутько А.В. Валидация мобильного приложения для расчета сагиттальных параметров позвоночника «SmartPlan Balance». Хирургия позвоночника. 2025;22(2):96–103. [Zakharin VR, Leonova ON, Baikov ES, Kokorev AI, Darchia LYu, Balychev GE, Ivanov DV, Dol AV, Bessonov LV, Krutko AV. Validation of the SmartPlan Balance mobile application to measure sagittal parameters of the spine. Russian Journal of Spine Surgery (Khirurgiya Pozvonochnika). 2025;22(2):96–103. DOI: http://dx.doi.org/10.14531/ss2025.2.96-103 EDN: https://elibrary.ru/TEUJWR