Skip to main content

Advertisement

SpringerLink
Log in

Opportunistic osteoporosis screening in multi-detector CT images using deep convolutional neural networks

  • Imaging Informatics and Artificial Intelligence
  • Published:
European Radiology Aims and scope Submit manuscript

Abstract

Objective

To explore the application of deep learning in patients with primary osteoporosis, and to develop a fully automatic method based on deep convolutional neural network (DCNN) for vertebral body segmentation and bone mineral density (BMD) calculation in CT images.

Materials and methods

A total of 1449 patients were used for experiments and analysis in this retrospective study, who underwent spinal or abdominal CT scans for other indications between March 2018 and May 2020. All data was gathered from three different CT vendors. Among them, 586 cases were used for training, and other 863 cases were used for testing. A fully convolutional neural network, called U-Net, was employed for automated vertebral body segmentation. The manually sketched region of vertebral body was used as the ground truth for comparison. A convolutional neural network, called DenseNet-121, was applied for BMD calculation. The values post-processed by quantitative computed tomography (QCT) were identified as the standards for analysis.

Results

Based on the diversity of CT vendors, all testing cases were split into three testing cohorts: Test set 1 (n = 463), test set 2 (n = 200), and test set 3 (n = 200). Automated segmentation correlated well with manual segmentation regarding four lumbar vertebral bodies (L1–L4): the minimum average dice coefficients for three testing sets were 0.823, 0.786, and 0.782, respectively. For testing sets from different vendors, the average BMDs calculated by automated regression showed high correlation (r > 0.98) and agreement with those derived from QCT.

Conclusions

A deep learning–based method could achieve fully automatic identification of osteoporosis, osteopenia, and normal bone mineral density in CT images.

Key Points

• Deep learning can perform accurate fully automated segmentation of lumbar vertebral body in CT images.

• The average BMDs obtained by deep learning highly correlates with ones derived from QCT.

• The deep learning–based method could be helpful for clinicians in opportunistic osteoporosis screening in spinal or abdominal CT scans.

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

Abbreviations

ACR:

American College of Radiology

BMD:

Bone mineral density

DCNN:

Deep convolutional neural network

DSC:

Dice similarity coefficient

DXA:

Dual-energy X-ray absorptiometry

ESP:

European spine phantom

HU :

Hounsfield units

ISCD:

International Society for Clinical Densitometry

PET:

Positron emission tomography

PPV:

Positive predictive value

QCT:

Quantitative computed tomography

QUS:

Quantitative ultrasound

WL:

Window level

WW:

Window width

References

  1. Hofbauer LC, Rachner TD (2015) More DATA to guide sequential osteoporosis therapy. Lancet 386:1116–1118. https://doi.org/10.1016/S0140-6736(15)61175-8

    Article  PubMed  Google Scholar 

  2. Pisani P, Renna MD, Conversano F et al (2016) Major osteoporotic fragility fractures: risk factor updates and societal impact. World J Orthop 7:171–181. https://doi.org/10.5312/wjo.v7.i3.171

    Article  PubMed  PubMed Central  Google Scholar 

  3. Yang J, Pham SM, Crabbe DL (2003) Effects of oestrogen deficiency on rat mandibular and tibial microarchitecture. Dentomaxillofac Radiol 32:247–251. https://doi.org/10.1259/dmfr/12560890

    Article  CAS  PubMed  Google Scholar 

  4. Watanabe M, Sakai D, Yamamoto Y, Sato M, Mochida J (2010) Upper cervical spine injuries: age-specific clinical features. J Orthop Sci 15:485–492. https://doi.org/10.1007/s00776-010-1493-x

  5. Devlin HB, Goldman M (1966) Backache due to osteoporosis in an industrial population. A survey of 481 patients. Ir J Med Sci 6:141–148. https://doi.org/10.1007/bf02943677

    Article  CAS  PubMed  Google Scholar 

  6. Ensrud KE, Blackwell TL, Cawthon PM et al (2016) Degree of trauma differs for major osteoporotic fracture events in older men versus older women. J Bone Miner Res 31:204–207. https://doi.org/10.1002/jbmr.2589

    Article  PubMed  Google Scholar 

  7. Fechtenbaum J, Etcheto A, Kolta S, Feydy A, Roux C, Briot K (2016) Sagittal balance of the spine in patients with osteoporotic vertebral fractures. Osteoporos Int 27:559–567. https://doi.org/10.1007/s00198-015-3283-y

  8. Yang Z, Griffith JF, Leung PC, Lee R (2009) Effect of osteoporosis on morphology and mobility of the lumbar spine. Spine (Phila Pa 1976) 34:E115–E121. https://doi.org/10.1097/brs.0b013e3181895aca

  9. Lee JJ, Aghdassi E, Cheung AM et al (2012) Ten-year absolute fracture risk and hip bone strength in Canadian women with systemic lupus erythematosus. J Rheumatol 39(7):1378–1384. https://doi.org/10.3899/jrheum.111589

    Article  PubMed  Google Scholar 

  10. Roski F, Hammel J, Mei K et al (2019) Bone mineral density measurements derived from dual-layer spectral CT enable opportunistic screening for osteoporosis. Eur Radiol 29(11):6355–6363. https://doi.org/10.1007/s00330-019-06263-z

    Article  PubMed  PubMed Central  Google Scholar 

  11. Li GW, Tang GY, Liu Y, Tang RB, Peng YF, Li W (2012) MR spectroscopy and Micro-CT in evaluation of osteoporosis model in rabbits: comparison with histopathology. Eur Radiol 22:923–929. https://doi.org/10.1007/s00330-011-2325-x

  12. Dyke JP, Aaron RK (2010) Noninvasive methods of measuring bone blood perfusion. Ann N Y Acad Sci 1192:95–102. https://doi.org/10.1111/j.1749-6632.2009.05376.x

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Ott SM (1991) Methods of determining bone mass. J Bone Miner Res 6:S71–S76. https://doi.org/10.1002/jbmr.5650061416

    Article  PubMed  Google Scholar 

  14. Rand T, Seidl G, Kainberger F et al (1997) Impact of spinal degenerative changes on the evaluation of bone mineral density with dual energy X-ray absorptiometry (DXA). Calcif Tissue Int 60:430e3. https://doi.org/10.1007/s002239900258

    Article  Google Scholar 

  15. Ito M, Hayashi K, Yamada M, Uetani M, Nakamura T (1993) Relationship of osteophytes to bone mineral density and spinal fracture in men. Radiology 189:497–502. https://doi.org/10.1148/radiology.189.2.8210380

  16. Engelke K (2017) Quantitative computed tomography—current status and new developments. J Clin Densitom 20(3):309–321. https://doi.org/10.1016/j.jocd.2017.06.017

    Article  PubMed  Google Scholar 

  17. Engelke K, Libanati C, Liu Y et al (2009) Quantitative computed tomography (QCT) of the forearm using general purpose spiral whole-body CT scanners: accuracy, precision and comparison with dual-energy X-ray absorptiometry (DXA). Bone 45(1):110e8. https://doi.org/10.1016/j.bone.2009.03.669

    Article  Google Scholar 

  18. Li N, Li XM, Xu L, Sun WJ, Cheng XG, Tian W (2013) Comparison of QCT and DXA: osteoporosis detection rates in postmenopausal women. Int J Endocrinol 2013:1–5. https://doi.org/10.1155/2013/895474

  19. Xiao-ming X, Na L, Li K et al (2019) Discordance in diagnosis of osteoporosis by quantitative computed tomography and dual-energy X-ray absorptiometry in Chinese elderly men. J Orthop Translat 18:59–64. https://doi.org/10.1016/j.jot.2018.11.003

    Article  Google Scholar 

  20. Löffler MT, Jacob A, Valentinitsch A et al (2019) Improved prediction of incident vertebral fractures using opportunistic QCT compared to DXA. Eur Radiol 29:4980–4989. https://doi.org/10.1007/s00330-019-06018-w

    Article  PubMed  PubMed Central  Google Scholar 

  21. Valentinitsch A, Trebeschi S, Kaesmacher J et al (2019) Opportunistic osteoporosis screening in multi-detector CT images via local classification of textures. Osteoporos Int 30:1275–1285. https://doi.org/10.1007/s00198-019-04910-1

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Gausden EB, Nwachukwu BU, Schreiber JJ, Lorich DG, Lane JM (2017) Opportunistic use of CT imaging for osteoporosis screening and bone density assessment. J Bone Joint Surg 99:1580–1590. https://doi.org/10.2106/JBJS.16.00749

  23. Feng-tan L, Dong L, Zhang Y-t (2013) Influence of tube voltage on CT attenuation, radiation dose, and image quality: phantom study. Chin J Radiol 47:458–461. https://doi.org/10.3760/cma.j.issn.1005-1201.2013.05.016

    Article  Google Scholar 

  24. Wolterink JM, Leiner T, de Vos BD, van Hamersvelt RW, Viergever MA, Išgum I (2017) Automatic coronary artery calcium scoring in cardiac CT angiography using paired convolutional neural networks. Med Image Anal 34:123–136. https://doi.org/10.1016/j.media.2016.04.004

  25. Sirinukunwattana K, Ahmed Raza SE, Yee-Wah T, Snead DRJ, Cree IA, Rajpoot NM (2016) Locality sensitive deep learning for detection and classification of nuclei in routine colon cancer histology images. IEEE Trans Med Imaging 35:1196–1206. https://doi.org/10.1109/TMI.2016.2525803

  26. Gulshan V, Peng L, Coram M et al (2016) Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs. JAMA 316:2402–2410. https://doi.org/10.1001/jama.2016.17216

    Article  PubMed  Google Scholar 

  27. Esteva A, Kuprel B, Novoa RA et al (2017) Dermatologist-level classification of skin cancer with deep neural networks. Nature 542:115–118. https://doi.org/10.1038/nature21056

    Article  CAS  PubMed  Google Scholar 

  28. González G, Ash SY et al (2017) Disease staging and prognosis in smokers using deep learning in chest computed tomography. Am J Respir Crit Care Med 197:193–203. https://doi.org/10.1164/rccm.201705-0860OC

    Article  Google Scholar 

  29. Lee S, Choe EK, Kang HY, Yoon JW, Kim HS (2019) The exploration of feature extraction and machine learning for predicting bone density from simple spine X-ray images in a Korean population. Skeletal Radiology pp 1–6. https://doi.org/10.1007/s00256-019-03342-6

  30. Bergman Amitai O, Bar A, Toledano E et al (2017) Computing DEXA score from CT using Deep segmentation networks cascade. Available at: www.zebra-med.com/research-publications/computing-dexa-score-from-ctusing-deep-segmentation-networks-cascade/

  31. Pan Y, Shi D, Wang H et al (2020) Automatic opportunistic osteoporosis screening using low-dose chest computed tomography scans obtained for lung cancer screening. Eur Radiol 30(7):4107–4116. https://doi.org/10.1007/s00330-020-06679-y

    Article  PubMed  PubMed Central  Google Scholar 

  32. Yasaka K, Akai H, Kunimatsu A, Kiryu S, Abe O (2020) Prediction of bone mineral density from computed tomography: application of deep learning with a convolutional neural network. Eur Radiol 30(6):3549–3557. https://doi.org/10.1007/s00330-020-06677-0

  33. Wu Y, Guo Z, Fu X et al (2019) The study protocol for the China Health Big Data (China Biobank) project. Quant Imaging Med Surg 96:1095–1102. https://doi.org/10.21037/qims.2019.06.16

    Article  Google Scholar 

  34. Link TM, Lang TF (2014) Axial QCT: clinical applications and new developments. J Clin Densitom 17:438–448. https://doi.org/10.1016/j.jocd.2014.04.119

    Article  PubMed  Google Scholar 

  35. American College of Radiology (2018) ACR-SPR-SSR practice parameter for the performance of musculoskeletal quantitative computed tomography (QCT). American College of Radiology, Reston. Available at: https://www.acr.org/-/media/ACR/Files/Practice-Parameters/QCT.pdf

  36. Ronneberger O, Fischer P, Brox T (2015) U-Net: convolutional networks for biomedical image segmentation. In: International Conference on Medical image computing and computer-assisted intervention. Springer, Cham, pp 234–241. https://doi.org/10.1007/978-3-319-24574-4_28

    Chapter  Google Scholar 

  37. Iandola F, Moskewicz M, Karayev S, Girshick R, Darrell T, Keutzer K (2014) Densenet: implementing efficient convnet descriptor pyramids. arXiv preprint arXiv:1404.1869

  38. Russakovsky O, Deng J, Su H et al (2015) Imagenet large scale visual recognition challenge. Int J Comput Vision 115:211–252. https://doi.org/10.1007/s11263-015-0816-y

    Article  Google Scholar 

Download references

Funding

This study has received funding by Science and Technology Planning Project of Zhuhai (ZH2202200010HJL), by Investigator-Initiated Clinical Trial of The Fifth Affiliated Hospital of Sun Yat-sen University (YNZZ003).

Author information

Author notes

  1. Yijie Fang and Wei Li have equal contributions to the study.

Authors and Affiliations

  1. Department of Radiology, The Fifth Affiliated Hospital of Sun Yat-sen University, Zhuhai, 519000, Guangdong Province, China

    Yijie Fang, Wei Li, Xiaojun Chen, Keming Chen, Jianwei Liao, Guobin Hong & Shaolin Li

  2. Guangdong Provincial Key Laboratory of Biomedical Imaging, The Fifth Affiliated Hospital of Sun Yat-sen University, Zhuhai, 519000, Guangdong Province, China

    Yijie Fang, Wei Li, Xiaojun Chen, Keming Chen, Jianwei Liao, Guobin Hong & Shaolin Li

  3. Institute of Advanced Research, Infervision, Beijing, 100025, China

    Han Kang, Pengxin Yu & Rongguo Zhang

Authors
  1. Yijie Fang
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Wei Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Xiaojun Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Keming Chen
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Han Kang
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Pengxin Yu
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Rongguo Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Jianwei Liao
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Guobin Hong
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Shaolin Li
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Guobin Hong or Shaolin Li.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Shaolin Li.

Conflict of interest

The authors of this manuscript declare no relationships with any companies whose products or services may be related to the subject matter of the article.

Statistics and biometry

No complex statistical methods were necessary for this paper.

Informed consent

Written informed consent was waived by the Institutional Review Board.

Ethical approval

Institutional Review Board approval was obtained.

Methodology

• retrospective

• diagnostic or prognostic study

• performed at one institution

Additional information

Publisher’s note

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

Electronic supplementary material

ESM 1

(DOCX 176 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Fang, Y., Li, W., Chen, X. et al. Opportunistic osteoporosis screening in multi-detector CT images using deep convolutional neural networks. Eur Radiol 31, 1831–1842 (2021). https://doi.org/10.1007/s00330-020-07312-8

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-020-07312-8

Keywords

Search

Navigation