The design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs

Simple item page
dc.contributor.advisorCalvey, Tanya
dc.contributor.advisorRobertson, Frances
dc.contributor.authorLieberman, Isabel
dc.date.accessioned2026-09-08T07:41:29Z
dc.date.available2026-09-08T07:41:29Z
dc.date.issued2026
dc.date.updated2026-09-08T07:39:02Z
dc.description.abstractIntroduction Stroke is a leading cause of death globally, ranking second worldwide and third in South Africa. The burden on sub-Saharan African health systems is exacerbated by limited access to medical care, particularly radiological services. The application of artificial intelligence (AI) to medical diagnostics presents an exciting opportunity to address these challenges. This study aims to expand knowledge and infrastructure development for AI in medical diagnostics. Methods A dataset of publicly-available head computed tomography (CT) scans, including haemorrhagic and ischaemic stroke cases and controls, was used. Initially, SynthSeg, a deep learning model for automated segmentation and parcellation of Magnetic Resonance Imaging (MRI), was applied to estimate regional volumes, complemented by manual evaluation of segmentations. Regional volumes were compared between stroke types and to controls. Following this, two deep-learning models were developed: a modified OzNet for stroke classification and a modified Deeplabv3 for segmenting ischaemia and haematomas. Results Haemorrhagic stroke (HS) often involved multiple vascular territories (particularly middle cerebral artery/ posterior cerebral artery (MCA/ PCA)), whereas ischaemic stroke (IS) generally caused localized MCA infarcts. IS volumes were significantly larger in MCA-supplied regions, while HS volumes resembled controls except in the precentral gyrus, pallidum, and pericalcarine cortex. Both stroke types showed reduced CSF but enlarged temporal horn and third ventricle volumes. SynthSeg's performance declined for severe lesions, highlighting the challenges in volumetric analysis for advanced pathologies. The classification model achieved 0.49 AUC (indicating overfitting), while the segmentation model attained a 0.4 Dice Score, underscoring the need for refined data and further honing. Conclusion These findings highlight distinct stroke-related volume changes and segmentation challenges, emphasizing the need for pathology-adapted solutions. Enhanced datasets and computational resources could further improve AI-driven stroke diagnostics, offering a scalable option in settings with limited radiological expertise.
dc.identifier.apacitationLieberman, I. (2026). <i>The design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs</i>. (). University of Cape Town ,Faculty of Health Sciences ,Division of Clinical Anatomy and Biological Anthropology. Retrieved from http://hdl.handle.net/11427/43764en_ZA
dc.identifier.chicagocitationLieberman, Isabel. <i>"The design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs."</i> ., University of Cape Town ,Faculty of Health Sciences ,Division of Clinical Anatomy and Biological Anthropology, 2026. http://hdl.handle.net/11427/43764en_ZA
dc.identifier.citationLieberman, I. 2026. The design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs. . University of Cape Town ,Faculty of Health Sciences ,Division of Clinical Anatomy and Biological Anthropology. http://hdl.handle.net/11427/43764en_ZA
dc.identifier.ris TY - Thesis / Dissertation AU - Lieberman, Isabel AB - Introduction Stroke is a leading cause of death globally, ranking second worldwide and third in South Africa. The burden on sub-Saharan African health systems is exacerbated by limited access to medical care, particularly radiological services. The application of artificial intelligence (AI) to medical diagnostics presents an exciting opportunity to address these challenges. This study aims to expand knowledge and infrastructure development for AI in medical diagnostics. Methods A dataset of publicly-available head computed tomography (CT) scans, including haemorrhagic and ischaemic stroke cases and controls, was used. Initially, SynthSeg, a deep learning model for automated segmentation and parcellation of Magnetic Resonance Imaging (MRI), was applied to estimate regional volumes, complemented by manual evaluation of segmentations. Regional volumes were compared between stroke types and to controls. Following this, two deep-learning models were developed: a modified OzNet for stroke classification and a modified Deeplabv3 for segmenting ischaemia and haematomas. Results Haemorrhagic stroke (HS) often involved multiple vascular territories (particularly middle cerebral artery/ posterior cerebral artery (MCA/ PCA)), whereas ischaemic stroke (IS) generally caused localized MCA infarcts. IS volumes were significantly larger in MCA-supplied regions, while HS volumes resembled controls except in the precentral gyrus, pallidum, and pericalcarine cortex. Both stroke types showed reduced CSF but enlarged temporal horn and third ventricle volumes. SynthSeg's performance declined for severe lesions, highlighting the challenges in volumetric analysis for advanced pathologies. The classification model achieved 0.49 AUC (indicating overfitting), while the segmentation model attained a 0.4 Dice Score, underscoring the need for refined data and further honing. Conclusion These findings highlight distinct stroke-related volume changes and segmentation challenges, emphasizing the need for pathology-adapted solutions. Enhanced datasets and computational resources could further improve AI-driven stroke diagnostics, offering a scalable option in settings with limited radiological expertise. DA - 2026 DB - OpenUCT DP - University of Cape Town KW - brain CT's KW - analysis LK - https://open.uct.ac.za PB - University of Cape Town PY - 2026 T1 - The design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs TI - The design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs UR - http://hdl.handle.net/11427/43764 ER - en_ZA
dc.identifier.urihttp://hdl.handle.net/11427/43764
dc.identifier.vancouvercitationLieberman I. The design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs. []. University of Cape Town ,Faculty of Health Sciences ,Division of Clinical Anatomy and Biological Anthropology, 2026 [cited yyyy month dd]. Available from: http://hdl.handle.net/11427/43764en_ZA
dc.language.isoen
dc.language.rfc3066eng
dc.publisher.departmentDivision of Clinical Anatomy and Biological Anthropology
dc.publisher.facultyFaculty of Health Sciences
dc.publisher.institutionUniversity of Cape Town
dc.subjectbrain CT's
dc.subjectanalysis
dc.titleThe design and implementation of deep learning methods for the diagnosis and anatomical analysis of stroke in brain CTs
dc.typeThesis / Dissertation
dc.type.qualificationlevelMasters
dc.type.qualificationlevelMSc
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
thesis_hsf_2026_lieberman isabel.pdf
Size:
7.79 MB
Format:
Adobe Portable Document Format
Description:
Download
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.72 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Masters