1
|
Rytky SJO, Tiulpin A, Finnilä MAJ, Karhula SS, Sipola A, Kurttila V, Valkealahti M, Lehenkari P, Joukainen A, Kröger H, Korhonen RK, Saarakkala S, Niinimäki J. Clinical Super-Resolution Computed Tomography of Bone Microstructure: Application in Musculoskeletal and Dental Imaging. Ann Biomed Eng 2024; 52:1255-1269. [PMID: 38361137 PMCID: PMC10995025 DOI: 10.1007/s10439-024-03450-y] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/17/2023] [Accepted: 01/09/2024] [Indexed: 02/17/2024]
Abstract
PURPOSE Clinical cone-beam computed tomography (CBCT) devices are limited to imaging features of half a millimeter in size and cannot quantify the tissue microstructure. We demonstrate a robust deep-learning method for enhancing clinical CT images, only requiring a limited set of easy-to-acquire training data. METHODS Knee tissue from five cadavers and six total knee replacement patients, and 14 teeth from eight patients were scanned using laboratory CT as training data for the developed super-resolution (SR) technique. The method was benchmarked against ex vivo test set, 52 osteochondral samples are imaged with clinical and laboratory CT. A quality assurance phantom was imaged with clinical CT to quantify the technical image quality. To visually assess the clinical image quality, musculoskeletal and maxillofacial CBCT studies were enhanced with SR and contrasted to interpolated images. A dental radiologist and surgeon reviewed the maxillofacial images. RESULTS The SR models predicted the bone morphological parameters on the ex vivo test set more accurately than conventional image processing. The phantom analysis confirmed higher spatial resolution on the SR images than interpolation, but image grayscales were modified. Musculoskeletal and maxillofacial CBCT images showed more details on SR than interpolation; however, artifacts were observed near the crown of the teeth. The readers assessed mediocre overall scores for both SR and interpolation. The source code and pretrained networks are publicly available. CONCLUSION Model training with laboratory modalities could push the resolution limit beyond state-of-the-art clinical musculoskeletal and dental CBCT. A larger maxillofacial training dataset is recommended for dental applications.
Collapse
Affiliation(s)
- Santeri J O Rytky
- Research Unit of Health Sciences and Technology, University of Oulu, POB 5000, 90014, Oulu, Finland.
| | - Aleksei Tiulpin
- Research Unit of Health Sciences and Technology, University of Oulu, POB 5000, 90014, Oulu, Finland
- Neurocenter Oulu, Oulu University Hospital, Oulu, Finland
| | - Mikko A J Finnilä
- Research Unit of Health Sciences and Technology, University of Oulu, POB 5000, 90014, Oulu, Finland
- Medical Research Center, University of Oulu, Oulu, Finland
| | - Sakari S Karhula
- Research Unit of Health Sciences and Technology, University of Oulu, POB 5000, 90014, Oulu, Finland
- Department of Radiotherapy, Oulu University Hospital, Oulu, Finland
| | - Annina Sipola
- Department of Diagnostic Radiology, Oulu University Hospital, Oulu, Finland
| | - Väinö Kurttila
- Department of Oral and Maxillofacial Surgery, Oulu University Hospital, Oulu, Finland
| | - Maarit Valkealahti
- Department of Surgery and Intensive Care, Oulu University Hospital, Oulu, Finland
| | - Petri Lehenkari
- Department of Surgery and Intensive Care, Oulu University Hospital, Oulu, Finland
- Cancer and Translational Medical Research Unit, Faculty of Medicine, University of Oulu, Oulu, Finland
| | - Antti Joukainen
- Department of Orthopaedics, Traumatology and Hand Surgery, Kuopio University Hospital, Kuopio, Finland
| | - Heikki Kröger
- Department of Orthopaedics, Traumatology and Hand Surgery, Kuopio University Hospital, Kuopio, Finland
| | - Rami K Korhonen
- Department of Applied Physics, University of Eastern Finland, Kuopio, Finland
| | - Simo Saarakkala
- Research Unit of Health Sciences and Technology, University of Oulu, POB 5000, 90014, Oulu, Finland
- Department of Diagnostic Radiology, Oulu University Hospital, Oulu, Finland
| | - Jaakko Niinimäki
- Research Unit of Health Sciences and Technology, University of Oulu, POB 5000, 90014, Oulu, Finland
- Department of Diagnostic Radiology, Oulu University Hospital, Oulu, Finland
| |
Collapse
|
2
|
Beirinckx Q, Bladt P, van der Plas MCE, van Osch MJP, Jeurissen B, den Dekker AJ, Sijbers J. Model-based super-resolution reconstruction for pseudo-continuous Arterial Spin Labeling. Neuroimage 2024; 286:120506. [PMID: 38185186 DOI: 10.1016/j.neuroimage.2024.120506] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/30/2023] [Revised: 12/28/2023] [Accepted: 01/02/2024] [Indexed: 01/09/2024] Open
Abstract
Arterial spin labeling (ASL) is a promising, non-invasive perfusion magnetic resonance imaging technique for quantifying cerebral blood flow (CBF). Unfortunately, ASL suffers from an inherently low signal-to-noise ratio (SNR) and spatial resolution, undermining its potential. Increasing spatial resolution without significantly sacrificing SNR or scan time represents a critical challenge towards routine clinical use. In this work, we propose a model-based super-resolution reconstruction (SRR) method with joint motion estimation that breaks the traditional SNR/resolution/scan-time trade-off. From a set of differently oriented 2D multi-slice pseudo-continuous ASL images with a low through-plane resolution, 3D-isotropic, high resolution, quantitative CBF maps are estimated using a Bayesian approach. Experiments on both synthetic whole brain phantom data, and on in vivo brain data, show that the proposed SRR Bayesian estimation framework outperforms state-of-the-art ASL quantification.
Collapse
Affiliation(s)
- Quinten Beirinckx
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium; μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| | - Piet Bladt
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium; μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| | - Merlijn C E van der Plas
- C.J. Gorter MRI Center, Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands
| | - Matthias J P van Osch
- C.J. Gorter MRI Center, Department of Radiology, Leiden University Medical Center, Leiden, The Netherlands
| | - Ben Jeurissen
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium; Lab for Equilibrium Investigations and Aerospace, Department of Physics, University of Antwerp, Antwerp, Belgium; μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| | - Arnold J den Dekker
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium; μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| | - Jan Sijbers
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium; μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium.
| |
Collapse
|
3
|
Nicastro M, Jeurissen B, Beirinckx Q, Smekens C, Poot DHJ, Sijbers J, den Dekker AJ. To shift or to rotate? Comparison of acquisition strategies for multi-slice super-resolution magnetic resonance imaging. Front Neurosci 2022; 16:1044510. [PMID: 36440272 PMCID: PMC9694825 DOI: 10.3389/fnins.2022.1044510] [Citation(s) in RCA: 2] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/14/2022] [Accepted: 10/18/2022] [Indexed: 07/27/2023] Open
Abstract
Multi-slice (MS) super-resolution reconstruction (SRR) methods have been proposed to improve the trade-off between resolution, signal-to-noise ratio and scan time in magnetic resonance imaging. MS-SRR consists in the estimation of an isotropic high-resolution image from a series of anisotropic MS images with a low through-plane resolution, where the anisotropic low-resolution images can be acquired according to different acquisition schemes. However, it is yet unclear how these schemes compare in terms of statistical performance criteria, especially for regularized MS-SRR. In this work, the estimation performance of two commonly adopted MS-SRR acquisition schemes based on shifted and rotated MS images respectively are evaluated in a Bayesian framework. The maximum a posteriori estimator, which introduces regularization by incorporating prior knowledge in a statistically well-defined way, is put forward as the estimator of choice and its accuracy, precision, and Bayesian mean squared error (BMSE) are used as performance criteria. Analytic calculations as well as Monte Carlo simulation experiments show that the rotated scheme outperforms the shifted scheme in terms of precision, accuracy, and BMSE. Furthermore, the superior performance of the rotated scheme is confirmed in real data experiments and in retrospective simulation experiments with and without inter-image motion. Results show that the rotated scheme allows regularized MS-SRR with a higher accuracy and precision than the shifted scheme, besides being more resilient to motion.
Collapse
Affiliation(s)
- Michele Nicastro
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium
- μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| | - Ben Jeurissen
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium
- μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
- Lab for Equilibrium Investigations and Aerospace, Department of Physics, University of Antwerp, Antwerp, Belgium
| | - Quinten Beirinckx
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium
- μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| | - Céline Smekens
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium
- μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
- Siemens Healthcare NV/SA, Groot-Bijgaarden, Belgium
| | - Dirk H. J. Poot
- Biomedical Imaging Group Rotterdam, Department of Radiology and Nuclear Medicine, Erasmus MC, Rotterdam, Netherlands
| | - Jan Sijbers
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium
- μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| | - Arnold J. den Dekker
- imec-Vision Lab, Department of Physics, University of Antwerp, Antwerp, Belgium
- μNEURO Research Centre of Excellence, University of Antwerp, Antwerp, Belgium
| |
Collapse
|
4
|
Koktzoglou I, Huang R, Ankenbrandt WJ, Walker MT, Edelman RR. Super-resolution head and neck MRA using deep machine learning. Magn Reson Med 2021; 86:335-345. [PMID: 33619802 DOI: 10.1002/mrm.28738] [Citation(s) in RCA: 12] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/21/2020] [Revised: 01/25/2021] [Accepted: 01/26/2021] [Indexed: 12/29/2022]
Abstract
PURPOSE To probe the feasibility of deep learning-based super-resolution (SR) reconstruction applied to nonenhanced MR angiography (MRA) of the head and neck. METHODS High-resolution 3D thin-slab stack-of-stars quiescent interval slice-selective (QISS) MRA of the head and neck was obtained in eight subjects (seven healthy volunteers, one patient) at 3T. The spatial resolution of high-resolution ground-truth MRA data in the slice-encoding direction was reduced by factors of 2 to 6. Four deep neural network (DNN) SR reconstructions were applied, with two based on U-Net architectures (2D and 3D) and two (2D and 3D) consisting of serial convolutions with a residual connection. SR images were compared to ground-truth high-resolution data using Dice similarity coefficient (DSC), structural similarity index measure (SSIM), arterial diameter, and arterial sharpness measurements. Image review of the optimal DNN SR reconstruction was done by two experienced neuroradiologists. RESULTS DNN SR of up to twofold and fourfold lower-resolution (LR) input volumes provided images that resembled those of the original high-resolution ground-truth volumes for intracranial and extracranial arterial segments, and improved DSC, SSIM, arterial diameters, and arterial sharpness relative to LR volumes (P < .001). The 3D DNN SR outperformed 2D DNN SR reconstruction. According to two neuroradiologists, 3D DNN SR reconstruction consistently improved image quality with respect to LR input volumes (P < .001). CONCLUSION DNN-based SR reconstruction of 3D head and neck QISS MRA offers the potential for up to fourfold reduction in acquisition time for neck vessels without the need to commensurately sacrifice spatial resolution.
Collapse
Affiliation(s)
- Ioannis Koktzoglou
- Department of Radiology, NorthShore University HealthSystem, Evanston, Illinois, USA.,Pritzker School of Medicine, University of Chicago, Chicago, Illinois, USA
| | - Rong Huang
- Department of Radiology, NorthShore University HealthSystem, Evanston, Illinois, USA
| | - William J Ankenbrandt
- Department of Radiology, NorthShore University HealthSystem, Evanston, Illinois, USA.,Pritzker School of Medicine, University of Chicago, Chicago, Illinois, USA
| | - Matthew T Walker
- Department of Radiology, NorthShore University HealthSystem, Evanston, Illinois, USA.,Pritzker School of Medicine, University of Chicago, Chicago, Illinois, USA
| | - Robert R Edelman
- Department of Radiology, NorthShore University HealthSystem, Evanston, Illinois, USA.,Northwestern University Feinberg School of Medicine, Chicago, Illinois, USA
| |
Collapse
|