Skip to main content

Advertisement

SpringerLink
Log in

Application of deep learning algorithms in classification and localization of implant cutout for the postoperative hip

  • Scientific Article
  • Published:
Skeletal Radiology Aims and scope Submit manuscript

Abstract

Objective

This study aims to explore the feasibility of employing convolutional neural networks for detecting and localizing implant cutouts on anteroposterior pelvic radiographs.

Materials and methods

The research involves the development of two Deep Learning models. Initially, a model was created for image-level classification of implant cutouts using 40191 pelvic radiographs obtained from a single institution. The radiographs were partitioned into training, validation, and hold-out test datasets in a 6/2/2 ratio. Performance metrics including the area under the receiver operator characteristics curve (AUROC), sensitivity, and specificity were calculated using the test dataset. Additionally, a second object detection model was trained to localize implant cutouts within the same dataset. Bounding box visualizations were generated on images predicted as cutout-positive by the classification model in the test dataset, serving as an adjunct for assessing algorithm validity.

Results

The classification model had an accuracy of 99.7%, sensitivity of 84.6%, specificity of 99.8%, AUROC of 0.998 (95% CI: 0.996, 0.999) and AUPRC of 0.774 (95% CI: 0.646, 0.880). From the pelvic radiographs predicted as cutout-positive, the object detection model could achieve 95.5% localization accuracy on true positive images, but falsely generated 14 results from the 15 false-positive predictions.

Conclusion

The classification model showed fair accuracy for detection of implant cutouts, while the object detection model effectively localized cutout. This serves as proof of concept of using a deep learning-based approach for classification and localization of implant cutouts from pelvic radiographs.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Data availability

Data that supports the findings of this study is available from the corresponding author upon reasonable request. Data is located in controlled access data storage at Changi General Hospital.

References

  1. Veronese N, Maggi S. Epidemiology and social costs of hip fracture. Injury. 2018;49(8):1458–60.

    Article  PubMed  Google Scholar 

  2. Lu Y, Uppal HS. Hip fractures: relevant anatomy, classification, and biomechanics of fracture and fixation. Geriatr Orthop Surg Rehabil. 2019;10.

  3. O’Neill F, Condon F, McGloughlin T, Lenehan B, Coffey JC, Walsh M. Dynamic hip screw versus DHS blade: a biomechanical comparison of the fixation achieved by each implant in bone Journal of Bone and Joint Surgery - Series B. 2011; 93 B(5):616–21.

  4. Baumgaertner MR, Curtin SL, Lindskog DM, Keggi JM. The value of the tip-apex distance in predicting failure of fixation of peritrochanteric fractures of the hip. J Bone Joint Surg Am. 1995;77(7):1058–64.

    Article  CAS  PubMed  Google Scholar 

  5. Frei HC, Hotz T, Cadosch D, Rudin M, Käch K. Central head perforation, or “cut through”, caused by the helical blade of the proximal femoral nail antirotation. J Orthop Trauma. 2012;26(8):e102-7.

    Article  PubMed  Google Scholar 

  6. Nordin S, Zulkifli O, Faisham WI. Mechanical failure of dynamic hip screw (DHS) fixation in intertrochanteric fracture of the femur. Med J Malaysia. 2001;(Suppl D):12–7.

  7. Kukla C, Heinz T, Gaebler C, Heinze G, Vécsei V. The standard gamma nail: a critical analysis of 1,000 cases. J Trauma. 2001;51(1):77–83.

    Article  CAS  PubMed  Google Scholar 

  8. Lizaur Utrilla A, Reig JS, Miralles Muñoz F, Bendala Tufanisco C. Trochanteric Gamma Nail and Compression Hip Screw for Trochanteric Fractures.

  9. Griffiths EJ, Cash DJW, Kalra S, Hopgood PJ. Time to surgery and 30-day morbidity and mortality of periprosthetic hip fractures. Injury. 2013;44(12):1949–52.

    Article  CAS  PubMed  Google Scholar 

  10. Mulcahy H, Chew FS. Current concepts of hip arthroplasty for radiologists: Part 1, features and radiographic assessment AJR Am J Roentgenol. 2012; 199:559–69.

  11. Pietrzak JR, Donaldson MJ, Kayani B, Haddad FS. Painful total hip arthroplasty. Orthop Trauma. 2018;32(1):38–44.

    Article  Google Scholar 

  12. Berlin L. Accuracy of diagnostic procedures: has it improved over the past five decades? AJR Am J Roentgenol. 2007;188(5):1173–8.

    Article  PubMed  Google Scholar 

  13. Sabih DE, Sabih A, Sabih Q, Khan AN. Image perception and interpretation of abnormalities; can we believe our eyes? Can we do something about it? Insights Imaging. 2011;2(1):47–55.

    Article  PubMed  Google Scholar 

  14. Macri F, Niu BT, Erdelyi S, Mayo JR, Khosa F, Nicolaou S, et al. Impact of 24/7 onsite emergency radiology staff coverage on emergency department workflow. Can Assoc Radiol J. 2021; 73(1):249–58.

  15. Oakden-Rayner L, Gale W, Bonham TA, Lungren MP, Carneiro G, Bradley AP, et al. Validation and algorithmic audit of a deep learning system for the detection of proximal femoral fractures in patients in the emergency department: a diagnostic accuracy study. Lancet Digit Health. 2022;4(5):e351-8.

    Article  CAS  PubMed  Google Scholar 

  16. Gurung B, Liu P, Harris PDR, Sagi A, Field RE, Sochart DH, et al. Artificial intelligence for image analysis in total hip and total knee arthroplasty. Bone Joint J. 2022;104(8):929–37.

    Article  PubMed  Google Scholar 

  17. Shah RF, Bini SA, Martinez AM, Pedoia V, Vail TP. Incremental inputs improve the automated detection of implant loosening using machine-learning algorithms. Bone Joint J. 2020;102-B(6):101–6.

  18. Loppini M, Gambaro FM, Chiappetta K, Grappiolo G, Bianchi AM, Corino VDA. Automatic identification of failure in hip replacement: an artificial intelligence approach. Bioengineering. 2022;9(7).

  19. Alzaid A, Wignall A, Dogramadzi S, Pandit H, Xie SQ. Automatic detection and classification of peri-prosthetic femur fracture. Int J Comput Assist Radiol Surg. 2022;17(4):649–60.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Muscato F, Corti A, Manlio Gambaro F, Chiappetta K, Loppini M, Corino VDA. Combining deep learning and machine learning for the automatic identification of hip prosthesis failure: development, validation and explainability analysis. Int J Med Inform. 2023;176:105095.

    Article  PubMed  Google Scholar 

  21. Gao Y, Soh NYT, Liu N, Lim G, Ting D, Cheng LTE, et al. Application of a deep learning algorithm in the detection of hip fractures. iScience. 2023;26(8).

  22. Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. Densely connected convolutional networks. In: Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2017. Institute of Electrical and Electronics Engineers Inc. 2016; 2261–9

  23. Mehr G. Automating abnormality detection in musculoskeletal radiographs through deep learning. arXiv preprint. 2010;12030

  24. Kingma DP, Ba JL. Adam: a method for stochastic optimization. In: 3rd International Conference on Learning Representations, ICLR 2015. International Conference on Learning Representations, ICLR.

  25. Abadi M, Barham P, Chen J, Chen Z, Davis A, Dean J, et al. TensorFlow: a system for large-scale machine learning. In: 12th USENIX Symposium on Operating Systems Design and Implementation. 2016;265–83.

  26. models/research/object_detection/g3doc/tf2_detection_zoo.md at master tensorflow/models GitHub. https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/tf2_detection_zoo.md.  Accessed 20 Mar 2024.

  27. Thrall JH, Li X, Li Q, Cruz C, Do S, Dreyer K, et al. Artificial intelligence and machine learning in radiology: opportunities, challenges, pitfalls, and criteria for success. J Am Coll Radiol. 2018; 15(3 Pt B): 504–8.

  28. Kohli M, Prevedello LM, Filice RW, Geis JR. Implementing machine learning in radiology practice and research. AJR Am J Roentgenol. 2017;208(4):754–60.

    Article  PubMed  Google Scholar 

  29. Cohen JF, McInnes MDF. Deep learning algorithms to detect fractures: systematic review shows promising results but many limitations. Radiology. 2022;304(1):63–4.

    Article  PubMed  Google Scholar 

  30. Pei Y, Huang Y, Zou Q, Zhang X, Wang S. Effects of image degradation and degradation removal to CNN-based image classification. IEEE Trans Pattern Anal Mach Intell. 2021;43(4):1239–53. https://pubmed.ncbi.nlm.nih.gov/31689183/. Accessed 20 Mar 2024.

  31. Baumgaertner MR, Solberg BD. Awareness of tip-apex distance reduces failure of fixation of trochanteric fractures of the hip. J Bone Joint Surg Br. 1997; 79-B(6):969–71.

  32. Raghuraman R, Kam JW, Chua DTC. Predictors of failure following fixation of intertrochanteric fractures with proximal femoral nail antirotation. Singapore Med J. 2019;60(9):463–7.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Stern LC, Gorczyca JT, Kates S, Ketz J, Soles G, Humphrey CA. Radiographic review of helical blade versus lag screw fixation for cephalomedullary nailing of low-energy peritrochanteric femur fractures: there is a difference in cutout. J Orthop Trauma. 2017;31(6):305–10.

    Article  PubMed  Google Scholar 

  34. Esper GW, Fisher ND, Anil U, Ganta A, Konda SR, Egol KA. Cut-through versus cut-out: no easy way to predict how single lag screw design cephalomedullary nails used for intertrochanteric hip fractures will fail? Hip Pelvis. 2023;35(3):175.

    Article  PubMed  PubMed Central  Google Scholar 

  35. Caruso G, Bonomo M, Valpiani G, Salvatori G, Gildone A, Lorusso V, et al. A six-year retrospective analysis of cut-out risk predictors in cephalomedullary nailing for pertrochanteric fractures: can the tip-apex distance (TAD) still be considered the best parameter? Bone Joint Res. 2017;6(8):481.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. Caruso G, Corradi N, Caldaria A, Bottin D, Lo Re D, Lorusso V, et al. New tip-apex distance and calcar-referenced tip-apex distance cut-offs may be the best predictors for cut-out risk after intramedullary fixation of proximal femur fractures. Sci Rep. 2022;12(1):357.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Funding

This research was partially funded by the Early Career Grant awarded by the International Skeletal Society in 2023.

Author information

Authors and Affiliations

  1. Department of Diagnostic Radiology, Singapore General Hospital, Singapore General Hospital, Block 2, Level 1 Outram Road, Singapore, 169608, Singapore

    Jin Rong Tan & Lionel Tim-Ee Cheng

  2. Radiological Sciences ACP, Duke-NUS Medical School, Singapore, Singapore

    Jin Rong Tan, Kang Min Wong, Lionel Tim-Ee Cheng & Yet Yen Yan

  3. Health Services Research, Changi General Hospital, Singapore Health Services, Singapore, Singapore

    Yan Gao & Hong Choon Oh

  4. Department of Orthopaedic Surgery, Changi General Hospital, Singapore, Singapore

    Raghavan Raghuraman

  5. Duke-NUS Medical School, Singapore Health Service (SingHealth), Singapore, Singapore

    Daniel Ting

  6. Department of Radiology, Changi General Hospital, Singapore, Singapore

    Kang Min Wong & Yet Yen Yan

  7. Department of Emergency Medicine, Changi General Hospital, Singapore, Singapore

    Siang Hiong Goh

Authors
  1. Jin Rong Tan
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Yan Gao
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Raghavan Raghuraman
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Daniel Ting
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Kang Min Wong
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Lionel Tim-Ee Cheng
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Hong Choon Oh
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Siang Hiong Goh
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Yet Yen Yan
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Jin Rong Tan.

Ethics declarations

Competing interests

Dr Jin Rong Tan has received a speaker honorarium from Boston Scientific. The rest of the authors declare no competing interest.

Disclaimer

The funder played no role in study design, data collection, analysis and interpretation of data or the writing of this manuscript.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Tan, J.R., Gao, Y., Raghuraman, R. et al. Application of deep learning algorithms in classification and localization of implant cutout for the postoperative hip. Skeletal Radiol (2024). https://doi.org/10.1007/s00256-024-04692-6

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s00256-024-04692-6

Keywords

Search

Navigation