Address reprint requests to: W.M., Department of Radiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, Guangdong, China. E-mail: [email protected], or R.H., Department of Radiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, Guangdong, China. E-mail: [email protected]

Search for more papers by this author

Ruibin Huang MS,

Corresponding Author

Ruibin Huang MS

[email protected]

Department of Radiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, China

Search for more papers by this author

Lei Liu PhD,

Lei Liu PhD

Medical College, Shantou University, Shantou, China

Search for more papers by this author

Ruotao Zhong MS,

Ruotao Zhong MS

orcid.org/0009-0007-1413-2759

College of Engineering, Shantou University, Shantou, China

Search for more papers by this author

Yuzhen Zhang MS,

Yuzhen Zhang MS

College of Engineering, Shantou University, Shantou, China

Search for more papers by this author

Haoyang Wan BS,

Haoyang Wan BS

Medical College, Shantou University, Shantou, China

Search for more papers by this author

Shuju Chen BS,

Shuju Chen BS

Medical College, Shantou University, Shantou, China

Search for more papers by this author

Nanfeng Zhang BS,

Nanfeng Zhang BS

Guangdong Hangyu Satellite Technology Co, Shantou, China

Search for more papers by this author

JingJing Liu BS,

JingJing Liu BS

Guangdong Hangyu Satellite Technology Co, Shantou, China

Search for more papers by this author

Wei Mei MD,

Corresponding Author

Wei Mei MD

[email protected]

Department of Radiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, China

Search for more papers by this author

Ruibin Huang MS,

Corresponding Author

Ruibin Huang MS

[email protected]

Department of Radiology, The First Affiliated Hospital of Shantou University Medical College, Shantou, China

Search for more papers by this author

First published: 06 February 2025

https://doi.org/10.1002/jmri.29731

Citations: 1

Share a link

Email
Wechat
Bluesky

Abstract

Background

Sacroiliitis is a hallmark of ankylosing spondylitis (AS), and early detection plays an important role in managing the condition effectively. MRI is commonly used for diagnosing sacroiliitis, traditional methods often depend on subjective interpretation or limited automation which can introduce variability in diagnoses. The integration of semi-supervised segmentation and radiomics features may reduce reliance on expert interpretation and the need for large annotated datasets, potentially enhancing diagnostic workflows.

Purpose

To develop a diagnostic model for sacroiliitis and bone marrow edema (BME) using semi-supervised segmentation and radiomics analysis of MRI images.

Study Type

Retrospective cohort study.

Population

A total of 257 patients (161 males, 93 females; age 11–74 years), including 155 sacroiliitis and 175 BME patients. A total of 514 sacroiliac joint (SIJ) MRI images are analyzed, with 359 used for training and 155 for testing.

Field Strength/Sequence

3.0 T, spin echo T1-weighted imaging (T1WI) and short-tau inversion recovery (STIR).

Assessment

SIJ segmentation is automated using the semi-supervised segmentation-based Unimatch framework. Manual delineation of SIJ regions of interest (ROIs) on T1WI images by an experienced radiologist (W.M., 10-year experience) served as the reference standard for segmentation performance evaluation. Radiomics features from T1WI and STIR are used to train machine learning models, including support vector machine (SVM), logistic regression (LR), and light gradient boosting machine (LightGBM), for sacroiliitis and BME detection. Performance is assessed using area under the curve (AUC), sensitivity, specificity, and accuracy. The Dice coefficient is used to assess the performance of the semi-supervised segmentation model on SIJ segmentation.

Statistical Tests

Performance is evaluated using receiver operating characteristic (ROC) curves and decision curve analysis (DCA).

Result

The Unimatch model achieves an average Dice coefficient of 0.859 for SIJ segmentation. AUCs for sacroiliitis detection are 0.84 (LR), 0.86 (SVM), and 0.78 (LightGBM), while for BME detection, AUCs are 0.73 (LR), 0.76 (SVM), and 0.70 (LightGBM).

Data Conclusion

This study demonstrates that semi-supervised segmentation combined with radiomics features and machine learning models provides a promising approach for diagnosis of sacroiliitis and BME.

Plain Language Summary

This study aimed to improve the diagnosis of sacroiliitis and bone marrow edema in patients with ankylosing spondylitis. The researchers used a method that automatically segments MRI images and analyzes features from those images. By applying machine learning, they created models to help detect sacroiliitis and bone marrow edema more accurately. The results show that this approach can effectively assist in identifying these conditions, with the best accuracy for sacroiliitis and bone marrow edema reaching 81.2% and 74.2%, respectively. This method could help doctors make better decisions, offering a promising tool for improving diagnosis in clinical settings.

Level of Evidence

Technical Efficacy

Stage 2

References

1Feldtkeller E, Khan M, van der Heijde D, van der Linden S, Braun J. Age at disease onset and diagnosis delay in hla-b27 negative vs. positive patients with ankylosing spondylitis. Rheumatol Int 2003; 23: 61-66.
10.1007/s00296-002-0237-4
PubMed Web of Science® Google Scholar
2Mandl P, Navarro-Compán V, Terslev L, et al. Eular recommendations for the use of imaging in the diagnosis and management of spondyloarthritis in clinical practice. Ann Rheum Dis 2015; 74(7): 1327-1339.
10.1136/annrheumdis-2014-206971
CAS PubMed Web of Science® Google Scholar
3Bakker PA, van den Berg R, Lenczner G, et al. Can we use structural lesions seen on MRI of the sacroiliac joints reliably for the classification of patients according to the ASAS axial spondyloarthritis criteria? Data from the DESIR cohort. Ann Rheum Dis 2017; 76(2): 392-398.
10.1136/annrheumdis-2016-209405
PubMed Web of Science® Google Scholar
4Diekhoff T, Hermann K-GA, Greese J, et al. Comparison of mri with radiography for detecting structural lesions of the sacroiliac joint using ct as standard of reference: Results from the simact study. Ann Rheum Dis 2017; 76(9): 1502-1508.
10.1136/annrheumdis-2016-210640
PubMed Web of Science® Google Scholar
5Ye L, Liu Y, Xiao Q, et al. Mri compared with low-dose ct scanning in the diagnosis of axial spondyloarthritis. Clin Rheumatol 2020; 39: 1295-1303.
10.1007/s10067-019-04824-7
PubMed Web of Science® Google Scholar
6Maksymowych WP, Lambert RG, Østergaard M, et al. MRI lesions in the sacroiliac joints of patients with spondyloarthritis: An update of definitions and validation by the asas mri working group. Ann Rheum Dis 2019; 78(11): 1550-1558.
10.1136/annrheumdis-2019-215589
PubMed Web of Science® Google Scholar
7Semelka RC, Armao DM, Elias J, Huda W. Imaging strategies to reduce the risk of radiation in CT studies, including selective substitution with MRI. J Magn Reson Imaging 2007; 25(5): 900-909.
10.1002/jmri.20895
PubMed Web of Science® Google Scholar
8Renson T, Depicker A, De Craemer A-S, et al. High prevalence of spondyloarthritis-like mri lesions in postpartum women: A prospective analysis in relation to maternal, child and birth character istics. Ann Rheum Dis 2020; 79(7): 929-934.
10.1136/annrheumdis-2020-217095
PubMed Web of Science® Google Scholar
9de Winter J, de Hooge M, van de Sande M, et al. Magnetic resonance imaging of the sacroiliac joints indicating sacroiliitis according to the assessment of spondyloarthritis inter national society definition in healthy individuals, runners, and women with postpartum back pain. Arthritis Rheumatol 2018; 70(7): 1042-1048.
10.1002/art.40475
PubMed Web of Science® Google Scholar
10Weber U, Jurik AG, Zejden A, et al. Frequency and anatomic distribution of magnetic resonance imaging features in the sacroiliac joints of young athletes: Exploring “background noise” toward a data-driven definition of sacroiliitis in early spondyloarthritis. Arthritis Rheumatol 2018; 70(5): 736-745.
10.1002/art.40429
PubMed Web of Science® Google Scholar
11Varkas G, de Hooge M, Renson T, et al. Effect of mechanical stress on magnetic resonance imaging of the sacroiliac joints: Assessment of military recruits by magnetic resonance imaging study. Rheumatology 2018; 57(3): 508-513.
10.1093/rheumatology/kex491
PubMed Web of Science® Google Scholar
12Seven S, Østergaard M, Morsel-Carlsen L, et al. Anatomic distribution of sacroiliac joint lesions on magnetic resonance imaging in patients with axial spondyloarthritis and control subjects: A prospective cross-sectional study, including postpartum women, patients with disc herniation, cleaning staff, runners, and healthy individuals. Arthritis Care Res 2021; 73(5): 742-754.
10.1002/acr.24473
CAS Web of Science® Google Scholar
13Lambert RG, Bakker PA, van der Heijde D, et al. Defining active sacroiliitis on MRI for classification of axial spondyloarthritis: Update by the ASAS MRI working group. Ann Rheum Dis 2016; 75(11): 1958-1963.
10.1136/annrheumdis-2015-208642
PubMed Web of Science® Google Scholar
14Sudoł-Szopińska I, Jurik AG, Eshed I, et al. Recommendations of the ESSR arthritis subcommittee for the use of magnetic resonance imaging in musculoskeletal rheumatic diseases. Semin Musculoskelet Radiol 2015; 19(4): 396-411. https://doi.org/10.1055/s-0035-1564696.
10.1055/s-0035-1564696
PubMed Web of Science® Google Scholar
15Salih AM, Pujadas ER, Campello VM, et al. Image-based biological heart age estimation reveals differential aging patterns across cardiac chambers. J Magn Reson Imaging 2023; 58(6): 1797-1812.
10.1002/jmri.28675
PubMed Web of Science® Google Scholar
16Chen W, Gao C, Hu C, et al. Risk stratification and overall survival prediction in advanced gastric cancer patients based on whole-volume MRI radiomics. J Magn Reson Imaging 2023; 58(4): 1161-1174.
10.1002/jmri.28621
PubMed Web of Science® Google Scholar
17Souly N, Spampinato C, Shah M. Semi supervised semantic segmentation using generative adversarial network. In: Proceedings of the IEEE International Conference on Computer Vision; 2017, p 5688–5696.
Google Scholar
18Zhang K, Luo G, Li W, et al. Automatic image segmentation and grading diagnosis of sacroiliitis associated with as using a deep convolutional neural network on ct images. J Digit Imaging 2023; 36(5): 2025-2034.
10.1007/s10278-023-00858-1
PubMed Web of Science® Google Scholar
19Liu L, Zhang H, Zhang W, Mei W, Huang R. Sacroiliitis diagnosis based on interpretable features and multi-task learning. Phys Med Biol 2024; 69(4):045034.
10.1088/1361-6560/ad2010
CAS Web of Science® Google Scholar
20Rudwaleit M, Jurik A-G, Hermann KA, et al. Defining active sacroiliitis on magnetic resonance imaging (MRI) for classification of axial spondyloarthritis: A consensual approach by the ASAS/OMERACT MRI group. Ann Rheum Dis 2009; 68(10): 1520-1527.
10.1136/ard.2009.110767
CAS PubMed Web of Science® Google Scholar
21Yang L, Qi L, Feng L, Zhang W, Shi Y. Revisiting weak-to-strong consistency in semi-supervised semantic segmentation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition; 2023, p 7236–7246.
Google Scholar
22Haralick RM, Shanmugam K, Dinstein IH. Textural features for image classification. IEEE Trans Syst Man Cybern 1973; 6: 610-621.
10.1109/TSMC.1973.4309314
Google Scholar
23Joanes DN, Gill CA. Comparing measures of sample skewness and kurtosis. J R Stat Soc Ser D Statistician 1998; 47(1): 183-189.
10.1111/1467-9884.00122
Web of Science® Google Scholar
24Pal NR, Pal SK. Entropy: A new definition and its applications. IEEE Trans Syst Man Cybern 1991; 21(5): 1260-1270.
10.1109/21.120079
Web of Science® Google Scholar
25Thibault G, Angulo J, Meyer F. Advanced statistical matrices for texture characterization: Application to cell classification. IEEE Trans Biomed Eng 2013; 61(3): 630-637.
10.1109/TBME.2013.2284600
PubMed Google Scholar
26Vickers AJ, Cronin AM, Elkin EB, Gonen M. Extensions to decision curve analysis, a novel method for evaluating diagnostic tests, prediction models and molecular markers. BMC Med Inform Decis Mak 2008; 8: 1-17.
10.1186/1472-6947-8-53
PubMed Web of Science® Google Scholar
27He Z, Fang K, Lin X, et al. Enhancing preoperative diagnosis of subscapular muscle injuries with shoulder mri-based multimodal radiomics. Acad Radiol 2024.
10.1016/j.acra.2024.09.049
Google Scholar
28Bressem KK, Adams LC, Proft F, et al. Deep learning detects changes indicative of axial spondyloarthritis at mri of sacroiliac joints. Radiology 2022; 305(3): 655-665.
10.1148/radiol.212526
PubMed Web of Science® Google Scholar
29Liu W-X, Wu H, Cai C, Lai Q-Q, Wang Y, Li Y-Z. Research on automatic recognition radiomics algorithm for early sacroiliac arthritis based on sacroiliac MRI imaging. J Orthop Surg Res 2024; 19(1): 96.
10.1186/s13018-024-04569-3
PubMed Web of Science® Google Scholar

Citing Literature

Volume62, Issue2

August 2025

Pages 563-572

Diagnosis of Sacroiliitis Through Semi-Supervised Segmentation and Radiomics Feature Analysis of MRI Images

Abstract

Background

Purpose

Study Type

Population

Field Strength/Sequence

Assessment

Statistical Tests

Result

Data Conclusion

Plain Language Summary

Level of Evidence

Technical Efficacy

References

Citing Literature

References

Information

About Wiley Online Library

Help & Support

Opportunities

Connect with Wiley

Diagnosis of Sacroiliitis Through Semi-Supervised Segmentation and Radiomics Feature Analysis of MRI Images

Abstract

Background

Purpose

Study Type

Population

Field Strength/Sequence

Assessment

Statistical Tests

Result

Data Conclusion

Plain Language Summary

Level of Evidence

Technical Efficacy

References

Citing Literature

References

Related

Information