Segmentation of the Proximal Femur from MR Images Using Deep Convolutional Neural Networks

Overview

Journal Sci Rep

Specialty Science

Date 2018 Nov 9

PMID 30405145

Citations 58

Authors

Cem M Deniz

Siyuan Xiang

R Spencer Hallyburton

Arakua Welbeck

James S Babb

Stephen Honig

Kyunghyun Cho

Gregory Chang

Affiliations

Soon will be listed here.

Abstract

Magnetic resonance imaging (MRI) has been proposed as a complimentary method to measure bone quality and assess fracture risk. However, manual segmentation of MR images of bone is time-consuming, limiting the use of MRI measurements in the clinical practice. The purpose of this paper is to present an automatic proximal femur segmentation method that is based on deep convolutional neural networks (CNNs). This study had institutional review board approval and written informed consent was obtained from all subjects. A dataset of volumetric structural MR images of the proximal femur from 86 subjects were manually-segmented by an expert. We performed experiments by training two different CNN architectures with multiple number of initial feature maps, layers and dilation rates, and tested their segmentation performance against the gold standard of manual segmentations using four-fold cross-validation. Automatic segmentation of the proximal femur using CNNs achieved a high dice similarity score of 0.95 ± 0.02 with precision = 0.95 ± 0.02, and recall = 0.95 ± 0.03. The high segmentation accuracy provided by CNNs has the potential to help bring the use of structural MRI measurements of bone quality into clinical practice for management of osteoporosis.

Citing Articles

The impact of data augmentation and transfer learning on the performance of deep learning models for the segmentation of the hip on 3D magnetic resonance images.

Montin E, Deniz C, Kijowski R, Youm T, Lattanzi R Inform Med Unlocked. 2024; 45.

PMID: 39119151 PMC: 11308385. DOI: 10.1016/j.imu.2023.101444.

Ultra-low dose hip CT-based automated measurement of volumetric bone mineral density at proximal femoral subregions.

Zhang X, Nadeem S, DiCamillo P, Shibli-Rahhal A, Regan E, Barr R Med Phys. 2024; 51(11):8213-8231.

PMID: 39042053 PMC: 11661458. DOI: 10.1002/mp.17319.

Towards Automatic Cartilage Quantification in Clinical Trials - Continuing from the 2019 IWOAI Knee Segmentation Challenge.

Dam E, Desai A, Deniz C, Rajamohan H, Regatte R, Iriondo C Osteoarthr Imaging. 2024; 3(1).

PMID: 39036792 PMC: 11258861. DOI: 10.1016/j.ostima.2023.100087.

Evaluation of fragility fracture risk using deep learning based on ultrasound radio frequency signal.

Luo W, Wu J, Chen Z, Guo P, Zhang Q, Lei B Endocrine. 2024; 86(2):800-812.

PMID: 38982023 DOI: 10.1007/s12020-024-03931-z.

Application of convolutional neural networks in medical images: a bibliometric analysis.

Jia H, Zhang J, Ma K, Qiao X, Ren L, Shi X Quant Imaging Med Surg. 2024; 14(5):3501-3518.

PMID: 38720828 PMC: 11074758. DOI: 10.21037/qims-23-1600.

References

Chang G, Hotca-Cho A, Rusinek H, Honig S, Mikheev A, Egol K . Measurement reproducibility of magnetic resonance imaging-based finite element analysis of proximal femur microarchitecture for in vivo assessment of bone strength. MAGMA. 2014; 28(4):407-12. PMC: 4605426. DOI: 10.1007/s10334-014-0475-y. View

Kazakia G, Hyun B, Burghardt A, Krug R, Newitt D, de Papp A . In vivo determination of bone structure in postmenopausal women: a comparison of HR-pQCT and high-field MR imaging. J Bone Miner Res. 2007; 23(4):463-74. DOI: 10.1359/jbmr.071116. View

Zoroofi R, Nishii T, Sato Y, Sugano N, Yoshikawa H, Tamura S . Segmentation of avascular necrosis of the femoral head using 3-D MR images. Comput Med Imaging Graph. 2001; 25(6):511-21. DOI: 10.1016/s0895-6111(01)00013-1. View

Arezoomand S, Lee W, Rakhra K, Beaule P . A 3D active model framework for segmentation of proximal femur in MR images. Int J Comput Assist Radiol Surg. 2014; 10(1):55-66. DOI: 10.1007/s11548-014-1125-6. View

Gomberg B, Saha P, Song H, Hwang S, Wehrli F . Topological analysis of trabecular bone MR images. IEEE Trans Med Imaging. 2000; 19(3):166-74. DOI: 10.1109/42.845175. View

Majumdar S . Magnetic resonance imaging of trabecular bone structure. Top Magn Reson Imaging. 2002; 13(5):323-34. DOI: 10.1097/00002142-200210000-00004. View

Chang G, Honig S, Brown R, Deniz C, Egol K, Babb J . Finite element analysis applied to 3-T MR imaging of proximal femur microarchitecture: lower bone strength in patients with fragility fractures compared with control subjects. Radiology. 2014; 272(2):464-74. PMC: 4263634. DOI: 10.1148/radiol.14131926. View

Liu F, Zhou Z, Jang H, Samsonov A, Zhao G, Kijowski R . Deep convolutional neural network and 3D deformable approach for tissue segmentation in musculoskeletal magnetic resonance imaging. Magn Reson Med. 2017; 79(4):2379-2391. PMC: 6271435. DOI: 10.1002/mrm.26841. View

Schmid J, Magnenat-Thalmann N . MRI bone segmentation using deformable models and shape priors. Med Image Comput Comput Assist Interv. 2008; 11(Pt 1):119-26. DOI: 10.1007/978-3-540-85988-8_15. View

10.

Cody D, Gross G, Hou F, Spencer H, Goldstein S, Fyhrie D . Femoral strength is better predicted by finite element models than QCT and DXA. J Biomech. 1999; 32(10):1013-20. DOI: 10.1016/s0021-9290(99)00099-8. View

11.

Brosch T, Tang L, Yoo Y, Li D, Traboulsee A, Tam R . Deep 3D Convolutional Encoder Networks With Shortcuts for Multiscale Feature Integration Applied to Multiple Sclerosis Lesion Segmentation. IEEE Trans Med Imaging. 2016; 35(5):1229-1239. DOI: 10.1109/TMI.2016.2528821. View

12.

Youden W . Index for rating diagnostic tests. Cancer. 1950; 3(1):32-5. DOI: 10.1002/1097-0142(1950)3:1<32::aid-cncr2820030106>3.0.co;2-3. View

13.

LeCun Y, Bengio Y, Hinton G . Deep learning. Nature. 2015; 521(7553):436-44. DOI: 10.1038/nature14539. View

14.

Chen H, Dou Q, Yu L, Qin J, Heng P . VoxResNet: Deep voxelwise residual networks for brain segmentation from 3D MR images. Neuroimage. 2017; 170:446-455. DOI: 10.1016/j.neuroimage.2017.04.041. View

15.

Krug R, Banerjee S, Han E, Newitt D, Link T, Majumdar S . Feasibility of in vivo structural analysis of high-resolution magnetic resonance images of the proximal femur. Osteoporos Int. 2005; 16(11):1307-14. DOI: 10.1007/s00198-005-1907-3. View

16.

Rajapakse C, Leonard M, Bhagat Y, Sun W, Magland J, Wehrli F . Micro-MR imaging-based computational biomechanics demonstrates reduction in cortical and trabecular bone strength after renal transplantation. Radiology. 2012; 262(3):912-20. PMC: 3285225. DOI: 10.1148/radiol.11111044. View

17.

Ladinsky G, Vasilic B, Popescu A, Wald M, Zemel B, Snyder P . Trabecular structure quantified with the MRI-based virtual bone biopsy in postmenopausal women contributes to vertebral deformity burden independent of areal vertebral BMD. J Bone Miner Res. 2007; 23(1):64-74. PMC: 2663589. DOI: 10.1359/jbmr.070815. View

18.

Kamnitsas K, Ledig C, Newcombe V, Simpson J, Kane A, Menon D . Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation. Med Image Anal. 2016; 36:61-78. DOI: 10.1016/j.media.2016.10.004. View

19.

Carballido-Gamio J, Harnish R, Saeed I, Streeper T, Sigurdsson S, Amin S . Structural patterns of the proximal femur in relation to age and hip fracture risk in women. Bone. 2013; 57(1):290-9. PMC: 3809121. DOI: 10.1016/j.bone.2013.08.017. View

20.

Macneil J, Boyd S . Bone strength at the distal radius can be estimated from high-resolution peripheral quantitative computed tomography and the finite element method. Bone. 2008; 42(6):1203-13. DOI: 10.1016/j.bone.2008.01.017. View