On Unsupervised Methods for Medical Image Segmentation: Investigating Classic Approaches in Breast Cancer DCE-MRI
Source abstract
Abstract
Unsupervised segmentation techniques, which do not require labeled data for training and can be more easily integrated into the clinical routine, represent a valid solution especially from a clinical feasibility perspective. Indeed, large-scale annotated datasets are not always available, undermining their immediate implementation and use in the clinic. Breast cancer is the most common cause of cancer death in women worldwide. In this study, breast lesion delineation in Dynamic Contrast Enhanced MRI (DCE-MRI) series was addressed by means of four popular unsupervised segmentation approaches: Split-and-Merge combined with Region Growing (SMRG), k-means, Fuzzy C-Means (FCM), and spatial FCM (sFCM). They represent well-established pattern recognition techniques that are still widely used in clinical research. Starting from the basic versions of these segmentation approaches, during our analysis, we identified the shortcomings of each of them, proposing improved versions, as well as developing ad hoc pre- and post-processing steps. The obtained experimental results, in terms of area-based—namely, Dice Index (DI), Jaccard Index (JI), Sensitivity, Specificity, False Positive Ratio (FPR), False Negative Ratio (FNR)—and distance-based metrics—Mean Absolute Distance (MAD), Maximum Distance (MaxD), Hausdorff Distance (HD)—encourage the use of unsupervised machine learning techniques in medical image segmentation. In particular, fuzzy clustering approaches (namely, FCM and sFCM) achieved the best performance. In fact, for area-based metrics, they obtained DI = 78.23% ± 6.50 (sFCM), JI = 65.90% ± 8.14 (sFCM), sensitivity = 77.84% ± 8.72 (FCM), specificity = 87.10% ± 8.24 (sFCM), FPR = 0.14 ± 0.12 (sFCM), and FNR = 0.22 ± 0.09 (sFCM). Concerning distance-based metrics, they obtained MAD = 1.37 ± 0.90 (sFCM), MaxD = 4.04 ± 2.87 (sFCM), and HD = 2.21 ± 0.43 (FCM). These experimental findings suggest that further research would be useful for advanced fuzzy logic techniques specifically tailored to medical image segmentation.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: AI in cancer detection · Radiomics and Machine Learning in Medical Imaging · Medical Image Segmentation Techniques
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Institute of Molecular Bioimaging and Physiology, Italian National Research Council (IBFM-CNR), 90015 Cefalu, Italy
Department of Mechanical and Aerospace Engineering, Sapienza University of Rome, 00184 Roma, Italy
Department of Radiology, University of Cambridge, Cambridge CB2 0QQ, UK
Cancer Research UK Cambridge Centre, Cambridge CB2 0RE, UK
Department of Information and Electrical Engineering and Applied Mathematics (DIEM), University of Salerno, 84084 Fisciano, Italy
Department of Radiology, Fondazione Istituto “G. Giglio”, 90015 Cefalu, Italy
Breast Unit, Fondazione Istituto “G. Giglio”, 90015 Cefalu, Italy
Section of Radiology—Department of Biomedicine, Neuroscience and Advanced Diagnostics (BiND), University Hospital “Paolo Giaccone”, 90127 Palermo, Italy
Funding
Italian MISE · The Mark Foundation for Cancer Research and Cancer Research UK Cambridge Centre · Royal Society for the International Exchanges 2020 Cost Share with the Italian CNR · National Institute of Health Research (NIHR) Cambridge Biomedical Research Centre
References (61)
Show source-provided references
1. Images in Space and Time: Real Big Data in Healthcare
2. Medical image analysis: Progress over two decades and the challenges ahead10.1109/34.824822 ↗
3. Automatic segmentation of prostate MRI using convolutional neural networks: Investigating the impact of network architecture on the accuracy of volume measurement and MRI-ultrasound registration10.1016/j.media.2019.101558 ↗
4. Lee, B., Yamanakkanavar, N., and Choi, J.Y. (2020). Automatic segmentation of brain MRI using a novel patch-wise U-net deep architecture. PLoS ONE, 15.10.1371/journal.pone.0236493 ↗
5. Pasero, E., and Castagneri, C. (2017, January 11–13). Application of an automatic ulcer segmentation algorithm. Proceedings of the International Forum on Research and Technologies for Society and Industry (RTSI), Modena, Italy.10.1109/RTSI.2017.8065954 ↗
6. Quantitative imaging in cancer clinical trials10.1158/1078-0432.CCR-14-3336 ↗
7. Decoding tumour phenotype by noninvasive imaging using a quantitative radiomics approach10.1038/ncomms5006 ↗
8. Radiomics: The bridge between medical imaging and personalized medicine10.1038/nrclinonc.2017.141 ↗
9. Current perspectives in medical image perception10.3758/APP.72.5.1205 ↗
10. Recent advances of HCI in decision-making tasks for optimized clinical workflows and precision medicine10.1016/j.jbi.2020.103479 ↗
11. Lee, N.Y., and Lu, J.J. (2013). Target Volume Delineation and Field Setup: A Practical Guide for Conformal and Intensity-Modulated Radiation Therapy, Springer. [1st ed.].10.1007/978-3-642-28860-9 ↗
12. Tumor-Cut: Segmentation of brain tumors on contrast enhanced MR images for radiosurgery applications10.1109/TMI.2011.2181857 ↗
13. GTVcut for neuro-radiosurgery treatment planning: An MRI brain cancer seeded image segmentation method based on a cellular automata model10.1007/s11047-017-9636-z ↗
14. A survey on deep learning in medical image analysis10.1016/j.media.2017.07.005 ↗
15. Embracing imperfect datasets: A review of deep learning solutions for medical image segmentation10.1016/j.media.2020.101693 ↗
16. Convolutional neural networks for medical image analysis: Full training or fine tuning?10.1109/TMI.2016.2535302 ↗
17. Deep learning for health informatics10.1109/JBHI.2016.2636665 ↗
18. Semi-automated and interactive segmentation of contrast-enhancing masses on breast DCE-MRI using spatial fuzzy clustering10.1016/j.bspc.2021.103113 ↗
19. Biomedical imaging modalities: A tutorial10.1016/0895-6111(94)00043-3 ↗
20. Effect Of Global Thresholding On Tumor-Bearing Brain MRI Images
21. A fully automatic 2D segmentation method for uterine fibroid in MRgFUS treatment evaluation10.1016/j.compbiomed.2015.04.030 ↗
22. Islam, M.R., Imteaz, M.R. (2018, January 8–9). Detection and analysis of brain tumor from MRI by Integrated Thresholding and Morphological Process with Histogram based method. Proceedings of the International Conference on Computer, Communication, Chemical, Material and Electronic Engineering (IC4ME2), Rajshahi, Bangladesh.10.1109/IC4ME2.2018.8465663 ↗
23. Combining Split-and-Merge and Multi-Seed Region Growing Algorithms for Uterine Fibroid Segmentation in MRgFUS Treatments10.1007/s11517-015-1404-6 ↗
24. Picture Segmentation by a Tree Traversal Algorithm10.1145/321941.321956 ↗
25. Split-and-Merge Segmentation of Magnetic Resonance Medical Images: Performance Evaluation and Extension to Three Dimensions10.1006/cbmr.1998.1489 ↗
26. Saad, N.M., Abu-Bakar, S.A.R., Muda, S., and Mokji, M. (December, January 30). Automated segmentation of brain lesion based on diffusion-weighted MRI using a split and merge approach. Proceedings of the IEEE EMBS Conference on Biomedical Engineering and Sciences (IECBES), Kuala Lumpur, Malaysia.
27. Saad, N.M., Abu-Bakar, S.A.R., Muda, S., Mokji, M., and Abdullah, A.R. (2012, January 14–16). Automated region growing for segmentation of brain lesion in diffusion-weighted MRI. Proceedings of the International MultiConference of Engineers and Computer Scientists, IMECS 2012, Hong Kong, China.
28. Seeded region growing10.1109/34.295913 ↗
29. Adaptive image region-growing10.1109/83.336259 ↗
30. A threshold selection method from gray-level histograms
31. Brain tumor MRI image segmentation and detection in image processing
32. An efficient k-means clustering algorithm: Analysis and implementation10.1109/TPAMI.2002.1017616 ↗
33. Bezdek, J.C. (1981). Objective function clustering. Pattern Recognition with Fuzzy Objective Function Algorithms, Springer. [1st ed.].10.1007/978-1-4757-0450-1 ↗
34. An automatic fuzzy c-means algorithm for image segmentation10.1007/s00500-009-0442-0 ↗
35. Gamma Knife treatment planning: MR brain tumor segmentation and volume measurement based on unsupervised Fuzzy C-Means clustering10.1002/ima.22139 ↗
36. Rundo, L., Militello, C., Russo, G., Garufi, A., Vitabile, S., Gilardi, M.C., and Mauri, G. (2017). Automated prostate gland segmentation based on an unsupervised fuzzy c-means clustering technique using multispectral T1w and T2w MR imaging. Information, 8.10.3390/info8020049 ↗
37. Caponetti, L., Castellano, G., and Corsini, V. (2017). MR brain image segmentation: A framework to compare different clustering techniques. Information, 8.10.3390/info8040138 ↗
38. NeXt for neuro-radiosurgery: A fully automatic approach for necrosis extraction in brain tumor MRI using an unsupervised machine learning technique10.1002/ima.22253 ↗
39. Unusual breast lesions: Radiologic-pathologic correlation10.1148/radiographics.19.suppl_1.g99oc07s11 ↗
40. Fuzzy c-means clustering with spatial information for image segmentation10.1016/j.compmedimag.2005.10.001 ↗
41. Tissue-specific and interpretable sub-segmentation of whole tumour burden on CT images by unsupervised fuzzy clustering10.1016/j.compbiomed.2020.103751 ↗
42. Preprocessing To Be Considered For MR and CT Images Containing Tumors
43. Vasuki, P., Kanimozhi, J., and Devi, M.B. (2017, January 27–28). A survey on image preprocessing techniques for diverse fields of medical imagery. Proceedings of the 2017 IEEE International Conference on Electrical, Instrumentation and Communication Engineering (ICEICE), Tamilnadu, India.10.1109/ICEICE.2017.8192443 ↗
44. Image filtering techniques for medical image post-processing: An overview10.1259/bjr/17464219 ↗
45. Image Postprocessing in Digital Radiology—A Primer for Technologists10.1016/j.jmir.2008.01.004 ↗
46. AI applications to medical images: From machine learning to deep learning10.1016/j.ejmp.2021.02.006 ↗
47. Survey of image denoising techniques
48. Marrone, S., Piantadosi, G., Fusco, R., Petrillo, A., Sansone, M., and Sansone, C. (2016, January 4–8). Breast segmentation using Fuzzy C-Means and anatomical priors in DCE-MRI. Proceedings of the 23rd International Conference on Pattern Recognition (ICPR), Cancun, Mexico.10.1109/ICPR.2016.7899845 ↗
49. Piantadosi, G., Fusco, R., Petrillo, A., Sansone, M., and Sansone, C. (2015). LBP-TOP for volume lesion classification in breast DCE-MRI. International Conference on Image Analysis and Processing, Springer. Lecture Notes in Computer Science.10.1007/978-3-319-23231-7_58 ↗
50. Individual comparisons by ranking methods
51. A Simple Sequentially Rejective Multiple Test Procedure
52. Breast cancer statistics, 201910.3322/caac.21583 ↗
53. Cancer Statistics, 202110.3322/caac.21654 ↗
54. International Agency for Research on Cancer (2021, July 28). The Global Cancer Observatory. Available online: https://gco.iarc.fr/today/data/factsheets/cancers/20-Breast-fact-sheet.pdf.
55. Frackiewicz, M., Koper, Z., Palus, H., Borys, D., and Psiuk-Maksymowicz, K. (2018, January 26–29). Breast lesion segmentation in DCE-MRI Imaging. Proceedings of the 14th International Conference on Signal-Image Technology &Internet-Based Systems (SITIS), Las Palmas de Gran Canaria, Spain.10.1109/SITIS.2018.00054 ↗
56. MRI-based radiomics in breast cancer: Feature robustness with respect to inter-observer segmentation variability10.1038/s41598-020-70940-z ↗
57. Han, C., Rundo, L., Murao, K., Noguchi, T., Shimahara, Y., Milacski, Z.Á., Koshino, S., Sala, E., Nakayama, H., and Satoh, S. (2021). MADGAN: Unsupervised medical anomaly detection GAN using multiple adjacent brain MRI slice reconstruction. BMC Bioinf., 22.10.1186/s12859-020-03936-1 ↗
58. Wu, X., Bi, L., Fulham, M., and Kim, J. (2020, January 13–15). Unsupervised Positron Emission Tomography Tumor Segmentation via GAN based Adversarial Auto-Encoder. Proceedings of the 16th International Conference on Control, Automation, Robotics and Vision (ICARCV), Shenzhen, China.10.1109/ICARCV50220.2020.9305364 ↗
59. A fully automatic approach for multimodal PET and MR image segmentation in Gamma Knife treatment planning10.1016/j.cmpb.2017.03.011 ↗
60. Hyperpolarized 13C MRI of tumor metabolism demonstrates early metabolic response to neoadjuvant chemotherapy in breast cancer10.1148/rycan.2020200017 ↗
61. Hypoxia and perfusion in breast cancer: Simultaneous assessment using PET/MR imaging10.1007/s00330-020-07067-2 ↗