Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning Analysis
Detection and segmentation of brain abnormalities using Magnetic Resonance Imaging (MRI) is an important task that, nowadays, the role of AI algorithms as supporting tools is well established both at the research and clinical-production level. While the performance of the state-of-the-art models is...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-12-01
|
| Series: | Journal of Imaging |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2313-433X/11/1/6 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832588305241210880 |
|---|---|
| author | Ioannis Stathopoulos Roman Stoklasa Maria Anthi Kouri Georgios Velonakis Efstratios Karavasilis Efstathios Efstathopoulos Luigi Serio |
| author_facet | Ioannis Stathopoulos Roman Stoklasa Maria Anthi Kouri Georgios Velonakis Efstratios Karavasilis Efstathios Efstathopoulos Luigi Serio |
| author_sort | Ioannis Stathopoulos |
| collection | DOAJ |
| description | Detection and segmentation of brain abnormalities using Magnetic Resonance Imaging (MRI) is an important task that, nowadays, the role of AI algorithms as supporting tools is well established both at the research and clinical-production level. While the performance of the state-of-the-art models is increasing, reaching radiologists and other experts’ accuracy levels in many cases, there is still a lot of research needed on the direction of in-depth and transparent evaluation of the correct results and failures, especially in relation to important aspects of the radiological practice: abnormality position, intensity level, and volume. In this work, we focus on the analysis of the segmentation results of a pre-trained U-net model trained and validated on brain MRI examinations containing four different pathologies: Tumors, Strokes, Multiple Sclerosis (MS), and White Matter Hyperintensities (WMH). We present the segmentation results for both the whole abnormal volume and for each abnormal component inside the examinations of the validation set. In the first case, a dice score coefficient (DSC), sensitivity, and precision of 0.76, 0.78, and 0.82, respectively, were found, while in the second case the model detected and segmented correct (True positives) the 48.8% (DSC ≥ 0.5) of abnormal components, partially correct the 27.1% (0.05 > DSC > 0.5), and missed (False Negatives) the 24.1%, while it produced 25.1% False Positives. Finally, we present an extended analysis between the True positives, False Negatives, and False positives versus their position inside the brain, their intensity at three MRI modalities (FLAIR, T2, and T1ce) and their volume. |
| format | Article |
| id | doaj-art-c1584582d481462da69db135e4a63de7 |
| institution | Kabale University |
| issn | 2313-433X |
| language | English |
| publishDate | 2024-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Journal of Imaging |
| spelling | doaj-art-c1584582d481462da69db135e4a63de72025-01-24T13:36:14ZengMDPI AGJournal of Imaging2313-433X2024-12-01111610.3390/jimaging11010006Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning AnalysisIoannis Stathopoulos0Roman Stoklasa1Maria Anthi Kouri2Georgios Velonakis3Efstratios Karavasilis4Efstathios Efstathopoulos5Luigi Serio62nd Department of Radiology, Medical School, Attikon University Hospital, National and Kapodistrian University of Athens, 11527 Athens, GreeceTechnology Department, CERN, 1211 Geneva, Switzerland2nd Department of Radiology, Medical School, Attikon University Hospital, National and Kapodistrian University of Athens, 11527 Athens, Greece2nd Department of Radiology, Medical School, Attikon University Hospital, National and Kapodistrian University of Athens, 11527 Athens, GreeceMedical Physics Laboratory, School of Medicine, Democritus University of Thrace, 68100 Alexandroupolis, Greece2nd Department of Radiology, Medical School, Attikon University Hospital, National and Kapodistrian University of Athens, 11527 Athens, GreeceTechnology Department, CERN, 1211 Geneva, SwitzerlandDetection and segmentation of brain abnormalities using Magnetic Resonance Imaging (MRI) is an important task that, nowadays, the role of AI algorithms as supporting tools is well established both at the research and clinical-production level. While the performance of the state-of-the-art models is increasing, reaching radiologists and other experts’ accuracy levels in many cases, there is still a lot of research needed on the direction of in-depth and transparent evaluation of the correct results and failures, especially in relation to important aspects of the radiological practice: abnormality position, intensity level, and volume. In this work, we focus on the analysis of the segmentation results of a pre-trained U-net model trained and validated on brain MRI examinations containing four different pathologies: Tumors, Strokes, Multiple Sclerosis (MS), and White Matter Hyperintensities (WMH). We present the segmentation results for both the whole abnormal volume and for each abnormal component inside the examinations of the validation set. In the first case, a dice score coefficient (DSC), sensitivity, and precision of 0.76, 0.78, and 0.82, respectively, were found, while in the second case the model detected and segmented correct (True positives) the 48.8% (DSC ≥ 0.5) of abnormal components, partially correct the 27.1% (0.05 > DSC > 0.5), and missed (False Negatives) the 24.1%, while it produced 25.1% False Positives. Finally, we present an extended analysis between the True positives, False Negatives, and False positives versus their position inside the brain, their intensity at three MRI modalities (FLAIR, T2, and T1ce) and their volume.https://www.mdpi.com/2313-433X/11/1/6deep learningmagnetic resonance imaging (MRI)AI algorithmstumorsstrokesmultiple sclerosis (MS) |
| spellingShingle | Ioannis Stathopoulos Roman Stoklasa Maria Anthi Kouri Georgios Velonakis Efstratios Karavasilis Efstathios Efstathopoulos Luigi Serio Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning Analysis Journal of Imaging deep learning magnetic resonance imaging (MRI) AI algorithms tumors strokes multiple sclerosis (MS) |
| title | Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning Analysis |
| title_full | Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning Analysis |
| title_fullStr | Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning Analysis |
| title_full_unstemmed | Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning Analysis |
| title_short | Exploring Multi-Pathology Brain Segmentation: From Volume-Based to Component-Based Deep Learning Analysis |
| title_sort | exploring multi pathology brain segmentation from volume based to component based deep learning analysis |
| topic | deep learning magnetic resonance imaging (MRI) AI algorithms tumors strokes multiple sclerosis (MS) |
| url | https://www.mdpi.com/2313-433X/11/1/6 |
| work_keys_str_mv | AT ioannisstathopoulos exploringmultipathologybrainsegmentationfromvolumebasedtocomponentbaseddeeplearninganalysis AT romanstoklasa exploringmultipathologybrainsegmentationfromvolumebasedtocomponentbaseddeeplearninganalysis AT mariaanthikouri exploringmultipathologybrainsegmentationfromvolumebasedtocomponentbaseddeeplearninganalysis AT georgiosvelonakis exploringmultipathologybrainsegmentationfromvolumebasedtocomponentbaseddeeplearninganalysis AT efstratioskaravasilis exploringmultipathologybrainsegmentationfromvolumebasedtocomponentbaseddeeplearninganalysis AT efstathiosefstathopoulos exploringmultipathologybrainsegmentationfromvolumebasedtocomponentbaseddeeplearninganalysis AT luigiserio exploringmultipathologybrainsegmentationfromvolumebasedtocomponentbaseddeeplearninganalysis |