Harnessing Advanced Machine Learning Techniques for Microscopic Vessel Segmentation in Pulmonary Fibrosis Using Novel Hierarchical Phase-Contrast Tomography Images

Abstract

Background Fibrotic lung disease is a progressive illness that causes scarring and ultimately respiratory failure, with irreversible damage by the time it is diagnosed on computed tomography imaging. Recent research postulates the role of the lung vasculature on the pathogenesis of the disease. With the recent development of high-resolution hierarchical phase-contrast tomography (HiP-CT), we have the potential to understand and detect changes in the lungs long before conventional imaging. However, to gain quantitative insight into vascular changes you first need to be able to segment the vessels before further downstream analysis can be conducted. Aside from this, HiP-CT generates large-volume, high-resolution data which is time-consuming and expensive to label.

Objectives This project aims to qualitatively assess the latest machine learning methods for vessel segmentation in HiP-CT data to enable label propagation as the first step for imaging biomarker discovery, with the goal to identify early-stage interstitial lung disease amenable to treatment, before fibrosis begins.

Methods Semisupervised learning (SSL) has become a growing method to tackle sparsely labeled datasets due to its leveraging of unlabeled data. In this study, we will compare two SSL methods; Seg PL, based on pseudo-labeling, and MisMatch, using consistency regularization against state-of-the-art supervised learning method, nnU-Net, on vessel segmentation in sparsely labeled lung HiP-CT data.

Results On initial experimentation, both MisMatch and SegPL showed promising performance on qualitative review. In comparison with supervised learning, both MisMatch and SegPL showed better out-of-distribution performance within the same sample (different vessel morphology and texture vessels), though supervised learning provided more consistent segmentations for well-represented labels in the limited annotations.

Conclusion Further quantitative research is required to better assess the generalizability of these findings, though they show promising first steps toward leveraging this novel data to tackle fibrotic lung disease.

Keywords

vessel segmentation - hierarchical phase-contrast tomography - semi-supervised learning - pulmonary fibrosis - interstitial lung disease

Ethical Approval Statement

For the use of novel HiP-CT in this study, original ethics approval of the data was obtained at Hannover Medical School, Germany, for the use of Human tissue culture as ex vivo models for the analysis of end-stage lung disease and lung tumors on February 4, 2022, under the following ethics approval number: 10194_BO_K_2022 (Ethics Review Board Chair Prof. Dr. Bernhard Schmidt). Approval for this retrospective study was obtained from the local research ethics committees and Leeds East Research Ethics Committee: 20/YH/0120.

Publication History

Received: 19 June 2024

Accepted: 14 February 2025

Accepted Manuscript online:
18 February 2025

Article published online:
09 May 2025

• References

• 1 NHS. . Lung Health Checks. Accessed July 15, 2023 at: https://www.nhs.uk/conditions/lung-health-checks/
• 2 National Institute for Health and Care Excellence (NICE). . Lung Cancer: Diagnosis and Management. NICE Guideline [NG122]. Updated March 8, 2024 . Accessed May 30, 2024 at: https://www.nice.org.uk/guidance/ng122/chapter/Recommendations-for-research#diagnosis-and-staging
• 3 Gov.UK. . New lung cancer screening roll-out to detect cancer sooner. 2023 . Accessed July 17, 2023 at: https://www.gov.uk/government/news/new-lung-cancer-screening-roll-out-to-detect-cancer-sooner
• 4 Hewitt RJ, Bartlett EC, Ganatra R. et al. Lung cancer screening provides an opportunity for early diagnosis and treatment of interstitial lung disease. Thorax 2022; 77 (11) 1149-1151
• 5 Mayo Clinic. . Pulmonary Fibrosis: Symptoms and Causes. Updated 2018 . Accessed June 13, 2023 at: https://www.mayoclinic.org/diseases-conditions/pulmonary-fibrosis/symptoms-causes/syc-20353690
• 6 Chua F, Desai SR, Nicholson AG. et al. Pleuroparenchymal fibroelastosis. a review of clinical, radiological, and pathological characteristics. Ann Am Thorac Soc 2019; 16 (11) 1351-1359
• 7 Fujimoto H, Kobayashi T, Azuma A. Idiopathic pulmonary fibrosis: treatment and prognosis. Clin Med Insights Circ Respir Pulm Med 2016; 9 (Suppl. 01) 179-18
• 8 May J, Mitchell JA, Jenkins RG. Beyond epithelial damage: vascular and endothelial contributions to idiopathic pulmonary fibrosis. J Clin Invest 2023; 133 (18) e172058
• 9 Gudmundsson E, Zhao A, Mogulkoc N. et al. Delineating associations of progressive pleuroparenchymal fibroelastosis in patients with pulmonary fibrosis. ERJ Open Res 2023; 9 (02) 00637-2022
• 10 Gudmundsson E, Zhao A, Mogulkoc N. et al. Pleuroparenchymal fibroelastosis in idiopathic pulmonary fibrosis: Survival analysis using visual and computer-based computed tomography assessment. EClinicalMedicine 2021; 38: 101009
• 11 Shioya M, Otsuka M, Yamada G. et al. poorer prognosis of idiopathic pleuroparenchymal fibroelastosis compared with idiopathic pulmonary fibrosis in advanced stage. Can Respir J 2018; 2018: 6043053
• 12 Ishii H, Kinoshita Y, Kushima H, Nagata N, Watanabe K. The similarities and differences between pleuroparenchymal fibroelastosis and idiopathic pulmonary fibrosis. Chron Respir Dis 2019; 16: 1479973119867945
• 13 Maher TM, Strek ME. Antifibrotic therapy for idiopathic pulmonary fibrosis: time to treat. Respir Res 2019; 20 (01) 205
• 14 Gaikwad AV, Lu W, Dey S. et al. Vascular remodelling in idiopathic pulmonary fibrosis patients and its detrimental effect on lung physiology: potential role of endothelial-to-mesenchymal transition. ERJ Open Res 2022; 8 (01) 00571-2021
• 15 Barratt S, Millar A. Vascular remodelling in the pathogenesis of idiopathic pulmonary fibrosis. QJM 2014; 107 (07) 515-519
• 16 Walsh CL, Tafforeau P, Wagner WL. et al. Imaging intact human organs with local resolution of cellular structures using hierarchical phase-contrast tomography. Nat Methods 2021; 18 (12) 1532-1541
• 17 Viermetz M, Birnbacher L, Willner M, Achterhold K, Pfeiffer F, Herzen J. High resolution laboratory grating-based X-ray phase-contrast CT. Sci Rep 2018; 8 (01) 15884
• 18 Moccia S, De Momi E, El Hadji S, Mattos LS. Blood vessel segmentation algorithms - Review of methods, datasets and evaluation metrics. Comput Methods Programs Biomed 2018; 158: 71-91
• 19 Willemink MJ, Koszek WA, Hardell C. et al. Preparing medical imaging data for machine learning. Radiology 2020; 295 (01) 4-15
• 20 Tommasi T, Patricia N, Caputo B, Tuytelaars T. A deeper look at dataset bias. arXiv preprint; arXiv:1505.01257 2015 . doi: 10.48550/arXiv.1505.01257
• 21 Anderson M, Sadiq S, Nahaboo Solim M. et al. Biomedical data annotation: an OCT imaging case study. J Ophthalmol 2023; 2023: 5747010
• 22 Zhu X, Goldberg AB. Introduction to Semi-Supervised Learning. Springer; 2009
• 23 Chapelle O, Scholkopf B, Zien A. . Semi-Supervised Learning. MIT Press; 2010
• 24 Grandvalet Y, Bengio Y. . Semi-supervised Learning by Entropy Minimization. Paper presented at: Advances in Neural Information Processing Systems 17 (NIPS 2004). 2004 . Accessed February 21, 2025 at: https://proceedings.neurips.cc/paper_files/paper/2004/file/96f2b50b5d3613adf9c27049b2a888c7-Paper.pdf
• 25 Xu M-C, Zhou Y, Jin C. et al Expectation maximization pseudo labelling for segmentation with limited annotations. arXiv preprint; arXiv:2305.01747 2023 . doi: 10.48550/arXiv.2305.01747
• 26 Laine S, Aila T. Temporal ensembling for semi-supervised learning. arXiv:1610.02242 2016 . doi: 10.48550/arXiv.1610.02242
• 27 Sajjadi M, Javanmardi M, Tasdizen T. Regularization with stochastic transformations and perturbations for deep semi-supervised learning. arXiv:1606.04586 2016 . doi: 10.48550/arXiv.1606.04586
• 28 Bachman P, Alsharif O, Precup D. Learning with pseudo-ensembles. arXiv:1412.4864. 2014. . Doi: 10.48550/arXiv.1412.4864
• 29 Lee D-H. . Pseudo-Label: The Simple and Efficient Semi-Supervised Learning Method for Deep Neural Networks. Paper presented at: ICML 2013 Workshop: Challenges in Representation Learning (WREPL); 2013
• 30 Arazo E, Ortego D, Albert P, O’Connor NE, McGuinness K. . Pseudo-labeling and confirmation bias in deep semi-supervised learning. Paper presented at: 2020 International Joint Conference on Neural Networks (IJCNN). 2020 :1–8. doi: 10.1109/IJCNN48605.2020.9207304
• 31 Sohn K, Berthelot D, Carlini N. et al. Fixmatch: Simplifying semi-supervised learning with consistency and confidence. Adv Neural Inf Process Syst 2020; 33: 596-608
• 32 Xu M-C, Zhou Y, Jin C. et al MisMatch: Calibrated segmentation via consistency on differential morphological feature perturbations with limited labels. arXiv preprint arXiv:211012179. 2023 . doi: 10.48550/arXiv.2110.12179
• 33 Xu M-C, Zhou Y, Jin C. , et al. Bayesian Pseudo Labels: Expectation Maximization for Robust and Efficient Semi-supervised Segmentation. Paper presented at: Medical Image Computing and Computer Assisted Intervention – MICCAI 2022. 2022. :580–590. doi: 10.48550/arXiv.2208.04435
• 34 Ronneberger O, Fischer P, Brox T. U-Net: Convolutional networks for biomedical image segmentation. arXiv:1505.04597 2015 . doi: 10.48550/arXiv.1505.04597
• 35 Isensee F, Jaeger PF, Kohl SAA, Petersen J, Maier-Hein KH. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nat Methods 2021; 18 (02) 203-211
• 36 Oliver A, Odena A, Raffel CA, Cubuk ED, Goodfellow I. Realistic evaluation of deep semi-supervised learning algorithms. Adv Neural Inf Process Syst arXiv:1804.09170. 2018;
• 37 Wang R, Jia X, Wang Q, Wu Y, Meng D. Imbalanced semi-supervised learning with bias adaptive classifier. arXiv preprint arXiv:220713856 2022 . doi: 10.48550/arXiv.2207.13856
• 38 Heim E, Roß T, Seitel A. et al. Large-scale medical image annotation with crowd-powered algorithms. J Med Imaging (Bellingham) 2018; 5 (03) 034002
• 39 Lin F, Xia Y, Ravikumar N, Liu Q, MacRaild M, Frangi AF. . Adaptive semi-supervised segmentation of brain vessels with ambiguous labels. Paper presented at: International Conference on Medical Image Computing and Computer-Assisted Intervention. 2023 :106–116
• 40 Hou J, Ding X, Deng JD. . Semi-supervised semantic segmentation of vessel images using leaking perturbations. Paper presented at: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 2022 :2625–2634
• 41 Li C, Ma W, Sun L. et al Hierarchical deep network with uncertainty-aware semi-supervised learning for vessel segmentation. arXiv:2105.14732 2022 doi: 10.48550/arXiv.2105.14732
• 42 Merveille O, Talbot H, Najman L, Passat N. Curvilinear structure analysis by ranking the orientation responses of path operators. IEEE Trans Pattern Anal Mach Intell 2018; 40 (02) 304-317
• 43 Lebre M-A, Vacavant A, Grand-Brochier M. et al. Automatic segmentation methods for liver and hepatic vessels from CT and MRI volumes, applied to the Couinaud scheme. Comput Biol Med 2019; 110 (110) 42-51