Automated Deep Learning-based Bone Mineral Density Assessment for Opportunistic Osteoporosis Screening Using Various CT Protocols with Multi-vendor Scanners

Overview

Journal Sci Rep

Specialty Science

Date 2024 Oct 24

PMID 39443535

Authors

Heejun Park

Woo Young Kang

Ok Hee Woo

Jemyoung Lee

Zepa Yang

Sangseok Oh

Affiliations

Soon will be listed here.

Abstract

This retrospective study examined the diagnostic efficacy of automated deep learning-based bone mineral density (DL-BMD) measurements for osteoporosis screening using 422 CT datasets from four vendors in two medical centers, encompassing 159 chest, 156 abdominal, and 107 lumbar spine datasets. DL-BMD values on L1 and L2 vertebral bodies were compared with manual BMD (m-BMD) measurements using Pearson's correlation and intraclass correlation coefficients. Strong agreement was found between m-BMD and DL-BMD in total CT scans (r = 0.953, p < 0.001). The diagnostic performance of DL-BMD was assessed using receiver operating characteristic analysis for osteoporosis and low BMD by dual-energy x-ray absorptiometry (DXA) and m-BMD. Compared to DXA, DL-BMD demonstrated an AUC of 0.790 (95% CI 0.733-0.839) for low BMD and 0.769 (95% CI 0.710-0.820) for osteoporosis, with sensitivity, specificity, and accuracy of 80.8% (95% CI 74.2-86.3%), 56.3% (95% CI 43.4-68.6%), and 74.3% (95% CI 68.3-79.7%) for low BMD and 65.4% (95% CI 50.9-78.0%), 70.9% (95% CI 63.8-77.3%), and 69.7% (95% CI 63.5-75.4%) for osteoporosis, respectively. Compared to m-BMD, DL-BMD showed an AUC of 0.983 (95% CI 0.973-0.993) for low BMD and 0.972 (95% CI 0.958-0.987) for osteoporosis, with sensitivity, specificity, and accuracy of 97.3% (95% CI 94.5-98.9%), 85.2% (95% CI 78.8-90.3%), and 92.7% (95% CI 89.7-95.0%) for low BMD and 94.4% (95% CI 88.3-97.9%), 89.5% (95% CI 85.6-92.7%), and 90.8% (95% CI 87.6-93.4%) for osteoporosis, respectively. The DL-based method can provide accurate and reliable BMD assessments across diverse CT protocols and scanners.

References

Li Y, Jiang Y, Yu X, Ren B, Wang C, Chen S . Deep-learning image reconstruction for image quality evaluation and accurate bone mineral density measurement on quantitative CT: A phantom-patient study. Front Endocrinol (Lausanne). 2022; 13:884306. PMC: 9403270. DOI: 10.3389/fendo.2022.884306. View

Oh S, Kang W, Park H, Yang Z, Lee J, Kim C . Evaluation of deep learning-based quantitative computed tomography for opportunistic osteoporosis screening. Sci Rep. 2024; 14(1):363. PMC: 10770031. DOI: 10.1038/s41598-023-45824-7. View

Sirinukunwattana K, Raza S, Tsang Y, Snead D, Cree I, Rajpoot N . Locality Sensitive Deep Learning for Detection and Classification of Nuclei in Routine Colon Cancer Histology Images. IEEE Trans Med Imaging. 2016; 35(5):1196-1206. DOI: 10.1109/TMI.2016.2525803. View

Jain R, Vokes T . Dual-energy X-ray Absorptiometry. J Clin Densitom. 2017; 20(3):291-303. DOI: 10.1016/j.jocd.2017.06.014. View

Toussaint N, Lau K, Strauss B, Polkinghorne K, Kerr P . Determination and validation of aortic calcification measurement from lateral bone densitometry in dialysis patients. Clin J Am Soc Nephrol. 2008; 4(1):119-27. PMC: 2615701. DOI: 10.2215/CJN.03410708. View

Pan Y, Shi D, Wang H, Chen T, Cui D, Cheng X . Automatic opportunistic osteoporosis screening using low-dose chest computed tomography scans obtained for lung cancer screening. Eur Radiol. 2020; 30(7):4107-4116. PMC: 7305250. DOI: 10.1007/s00330-020-06679-y. View

Li N, Li X, Xu L, Sun W, Cheng X, Tian W . Comparison of QCT and DXA: Osteoporosis Detection Rates in Postmenopausal Women. Int J Endocrinol. 2013; 2013:895474. PMC: 3623474. DOI: 10.1155/2013/895474. View

Wolterink J, Leiner T, de Vos B, van Hamersvelt R, Viergever M, Isgum I . Automatic coronary artery calcium scoring in cardiac CT angiography using paired convolutional neural networks. Med Image Anal. 2016; 34:123-136. DOI: 10.1016/j.media.2016.04.004. View

Loffler M, Sekuboyina A, Jacob A, Grau A, Scharr A, El Husseini M . A Vertebral Segmentation Dataset with Fracture Grading. Radiol Artif Intell. 2021; 2(4):e190138. PMC: 8082364. DOI: 10.1148/ryai.2020190138. View

10.

Gulshan V, Peng L, Coram M, Stumpe M, Wu D, Narayanaswamy A . Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA. 2016; 316(22):2402-2410. DOI: 10.1001/jama.2016.17216. View

11.

Fang Y, Li W, Chen X, Chen K, Kang H, Yu P . Opportunistic osteoporosis screening in multi-detector CT images using deep convolutional neural networks. Eur Radiol. 2020; 31(4):1831-1842. DOI: 10.1007/s00330-020-07312-8. View

12.

Krishnaraj A, Barrett S, Bregman-Amitai O, Cohen-Sfady M, Bar A, Chettrit D . Simulating Dual-Energy X-Ray Absorptiometry in CT Using Deep-Learning Segmentation Cascade. J Am Coll Radiol. 2019; 16(10):1473-1479. DOI: 10.1016/j.jacr.2019.02.033. View

13.

Ito M, Hayashi K, Yamada M, Uetani M, Nakamura T . Relationship of osteophytes to bone mineral density and spinal fracture in men. Radiology. 1993; 189(2):497-502. DOI: 10.1148/radiology.189.2.8210380. View

14.

Lee S, Choe E, Kang H, Yoon J, Kim H . The exploration of feature extraction and machine learning for predicting bone density from simple spine X-ray images in a Korean population. Skeletal Radiol. 2019; 49(4):613-618. DOI: 10.1007/s00256-019-03342-6. View

15.

Gonzalez G, Ash S, Vegas-Sanchez-Ferrero G, Onieva J, Rahaghi F, Ross J . Disease Staging and Prognosis in Smokers Using Deep Learning in Chest Computed Tomography. Am J Respir Crit Care Med. 2017; 197(2):193-203. PMC: 5768902. DOI: 10.1164/rccm.201705-0860OC. View

16.

Ong W, Liu R, Makmur A, Low X, Sng W, Tan J . Artificial Intelligence Applications for Osteoporosis Classification Using Computed Tomography. Bioengineering (Basel). 2023; 10(12). PMC: 10741220. DOI: 10.3390/bioengineering10121364. View

17.

Gausden E, Nwachukwu B, Schreiber J, Lorich D, Lane J . Opportunistic Use of CT Imaging for Osteoporosis Screening and Bone Density Assessment: A Qualitative Systematic Review. J Bone Joint Surg Am. 2017; 99(18):1580-1590. DOI: 10.2106/JBJS.16.00749. View

18.

Peng T, Zeng X, Li Y, Li M, Pu B, Zhi B . A study on whether deep learning models based on CT images for bone density classification and prediction can be used for opportunistic osteoporosis screening. Osteoporos Int. 2023; 35(1):117-128. PMC: 10786975. DOI: 10.1007/s00198-023-06900-w. View

19.

Yu W, Gluer C, Fuerst T, Grampp S, Li J, Lu Y . Influence of degenerative joint disease on spinal bone mineral measurements in postmenopausal women. Calcif Tissue Int. 1995; 57(3):169-74. DOI: 10.1007/BF00310253. View

20.

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