Dual energy CT and deep learning for an automated volumetric segmentation of the major intracranial tissues : Feasibility and initial findings
(2026) In Medical Physics 53(1).- Abstract
Background: Magnetic resonance imaging (MRI) has traditionally been preferred over computed tomography (CT) for segmentation of intracranial structures due to its superior low contrast resolution. However, a reliable CT-based segmentation could improve patient management when MRI is not practical. Despite advancements in CT imaging, such as enhanced tissue differentiation using virtual monoenergetic imaging (VMI) from dual energy CT, volumetric analysis remains underexplored. Purpose: The aim was to evaluate the feasibility of using deep learning (DL) models for segmentation of gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF)—using virtual monoenergetic images (VMI). Methods: The study included 26 patients... (More)
Background: Magnetic resonance imaging (MRI) has traditionally been preferred over computed tomography (CT) for segmentation of intracranial structures due to its superior low contrast resolution. However, a reliable CT-based segmentation could improve patient management when MRI is not practical. Despite advancements in CT imaging, such as enhanced tissue differentiation using virtual monoenergetic imaging (VMI) from dual energy CT, volumetric analysis remains underexplored. Purpose: The aim was to evaluate the feasibility of using deep learning (DL) models for segmentation of gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF)—using virtual monoenergetic images (VMI). Methods: The study included 26 patients (training/validation: 21, test: 5) who underwent brain imaging on a dual-layer CT and a T1-weighted MR scan. MR-based segmentation of GM, WM, and CSF served as the ground truth for training and testing of the DL models. Models included a baseline U-Net++ trained on 70 keV VMIs and several U-Net and U-Net++ extensions designed to leverage spectral information from multiple VMIs (50, 70, and 120 keV). Model performance was evaluated using Dice Similarity Coefficient (DSC) and volumetric accuracy. Results: The U-Net++ (Aug) model, using VMIs as augmentations of the input data, outperformed the baseline and other models with DSC 0.84, 0.77, and 0.88 for WM, GM, and CSF, respectively. The superiority was significant compared to several of the other models, and most notably compared to the baseline model with DSC of 0.81 for WM (p = 0.002) and 0.75 for GM (p = 0.002). U-Net++ (Aug) had an average volumetric error of 12%, while U-Net (Gated) had the lowest error at 10%. Conclusions: This study demonstrates the feasibility of CT-based segmentation of intracranial tissue using DL and VMI. The improved accuracy of the U-Net++ (Aug) compared to the baseline model suggests that spectral information may enhance segmentation performance.
(Less)
- author
- Fransson, Veronica
LU
; Winzell, Filip
LU
; Ramgren, Birgitta
LU
; Christensen, Søren
LU
; Ydström, Kristina
LU
; Arvidsson, Ida
LU
; Overgaard, Niels Christian
LU
; Åström, Kalle
LU
; Heyden, Anders
LU
and Wassélius, Johan
LU
- organization
-
- Medical Radiation Physics, Malmö (research group)
- LTH Profile Area: AI and Digitalization
- LTH Profile Area: Engineering Health
- Computer Vision and Machine Learning (research group)
- LU Profile Area: Natural and Artificial Cognition
- ELLIIT: the Linköping-Lund initiative on IT and mobile communication
- Neurological injury in acute type A aortic dissection (research group)
- Stroke Imaging Research group (research group)
- Neuroradiology (research group)
- Diagnostic Radiology, (Lund)
- Department of Clinical Sciences, Lund
- ID-A workshop
- Medical Radiation Physics, Lund
- LU Profile Area: Proactive Ageing
- eSSENCE: The e-Science Collaboration
- Lund Laser Centre, LLC
- LTH Profile Area: Photon Science and Technology
- Mathematics (Faculty of Engineering)
- LU Profile Area: Nature-based future solutions
- LU Profile Area: Light and Materials
- publishing date
- 2026-01
- type
- Contribution to journal
- publication status
- published
- subject
- keywords
- Brain, Deep Learning, Tissue segmentation, X-Ray Computed Tomography Scanner
- in
- Medical Physics
- volume
- 53
- issue
- 1
- article number
- e70217
- publisher
- John Wiley & Sons Inc.
- external identifiers
-
- scopus:105025378296
- pmid:41423435
- ISSN
- 0094-2405
- DOI
- 10.1002/mp.70217
- language
- English
- LU publication?
- yes
- additional info
- Publisher Copyright: © 2025 The Author(s). Medical Physics published by Wiley Periodicals LLC on behalf of American Association of Physicists in Medicine.
- id
- c4c0bfd1-f0d8-43c7-94c0-476dae4babfb
- date added to LUP
- 2026-03-06 15:28:55
- date last changed
- 2026-05-30 04:23:28
@article{c4c0bfd1-f0d8-43c7-94c0-476dae4babfb,
abstract = {{<p>Background: Magnetic resonance imaging (MRI) has traditionally been preferred over computed tomography (CT) for segmentation of intracranial structures due to its superior low contrast resolution. However, a reliable CT-based segmentation could improve patient management when MRI is not practical. Despite advancements in CT imaging, such as enhanced tissue differentiation using virtual monoenergetic imaging (VMI) from dual energy CT, volumetric analysis remains underexplored. Purpose: The aim was to evaluate the feasibility of using deep learning (DL) models for segmentation of gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF)—using virtual monoenergetic images (VMI). Methods: The study included 26 patients (training/validation: 21, test: 5) who underwent brain imaging on a dual-layer CT and a T1-weighted MR scan. MR-based segmentation of GM, WM, and CSF served as the ground truth for training and testing of the DL models. Models included a baseline U-Net++ trained on 70 keV VMIs and several U-Net and U-Net++ extensions designed to leverage spectral information from multiple VMIs (50, 70, and 120 keV). Model performance was evaluated using Dice Similarity Coefficient (DSC) and volumetric accuracy. Results: The U-Net++ (Aug) model, using VMIs as augmentations of the input data, outperformed the baseline and other models with DSC 0.84, 0.77, and 0.88 for WM, GM, and CSF, respectively. The superiority was significant compared to several of the other models, and most notably compared to the baseline model with DSC of 0.81 for WM (p = 0.002) and 0.75 for GM (p = 0.002). U-Net++ (Aug) had an average volumetric error of 12%, while U-Net (Gated) had the lowest error at 10%. Conclusions: This study demonstrates the feasibility of CT-based segmentation of intracranial tissue using DL and VMI. The improved accuracy of the U-Net++ (Aug) compared to the baseline model suggests that spectral information may enhance segmentation performance.</p>}},
author = {{Fransson, Veronica and Winzell, Filip and Ramgren, Birgitta and Christensen, Søren and Ydström, Kristina and Arvidsson, Ida and Overgaard, Niels Christian and Åström, Kalle and Heyden, Anders and Wassélius, Johan}},
issn = {{0094-2405}},
keywords = {{Brain; Deep Learning; Tissue segmentation; X-Ray Computed Tomography Scanner}},
language = {{eng}},
number = {{1}},
publisher = {{John Wiley & Sons Inc.}},
series = {{Medical Physics}},
title = {{Dual energy CT and deep learning for an automated volumetric segmentation of the major intracranial tissues : Feasibility and initial findings}},
url = {{http://dx.doi.org/10.1002/mp.70217}},
doi = {{10.1002/mp.70217}},
volume = {{53}},
year = {{2026}},
}