Umeå University's logo

umu.sePublications
Change search
Link to record
Permanent link

Direct link
Publications (3 of 3) Show all publications
Hellström, M., Löfstedt, T. & Garpebring, A. (2023). Denoising and uncertainty estimation in parameter mapping with approximate Bayesian deep image priors. Magnetic Resonance in Medicine, 90(6), 2557-2571
Open this publication in new window or tab >>Denoising and uncertainty estimation in parameter mapping with approximate Bayesian deep image priors
2023 (English)In: Magnetic Resonance in Medicine, ISSN 0740-3194, E-ISSN 1522-2594, Vol. 90, no 6, p. 2557-2571Article in journal (Refereed) Published
Abstract [en]

Purpose: To mitigate the problem of noisy parameter maps with high uncertainties by casting parameter mapping as a denoising task based on Deep Image Priors.

Methods: We extend the concept of denoising with Deep Image Prior (DIP) into parameter mapping by treating the output of an image-generating network as a parametrization of tissue parameter maps. The method implicitly denoises the parameter mapping process by filtering low-level image features with an untrained convolutional neural network (CNN). Our implementation includes uncertainty estimation from Bernoulli approximate variational inference, implemented with MC dropout, which provides model uncertainty in each voxel of the denoised parameter maps. The method is modular, so the specifics of different applications (e.g., T1 mapping) separate into application-specific signal equation blocks. We evaluate the method on variable flip angle T1 mapping, multi-echo T2 mapping, and apparent diffusion coefficient mapping.

Results: We found that deep image prior adapts successfully to several applications in parameter mapping. In all evaluations, the method produces noise-reduced parameter maps with decreased uncertainty compared to conventional methods. The downsides of the proposed method are the long computational time and the introduction of some bias from the denoising prior.

Conclusion: DIP successfully denoise the parameter mapping process and applies to several applications with limited hyperparameter tuning. Further, it is easy to implement since DIP methods do not use network training data. Although time-consuming, uncertainty information from MC dropout makes the method more robust and provides useful information when properly calibrated.

Place, publisher, year, edition, pages
John Wiley & Sons, 2023
Keywords
deep image prior, denoising, parameter mapping, quantitative MRI, uncertainty estimation
National Category
Medical Image Processing Computer Vision and Robotics (Autonomous Systems)
Identifiers
urn:nbn:se:umu:diva-213711 (URN)10.1002/mrm.29823 (DOI)001049833500001 ()37582257 (PubMedID)2-s2.0-85168117341 (Scopus ID)
Funder
Swedish Research Council, 2019‐0432Region Västerbotten, RV‐970119Cancerforskningsfonden i Norrland, AMP 18‐912
Available from: 2023-09-14 Created: 2023-09-14 Last updated: 2023-12-20Bibliographically approved
Andersson, J., Bednarek, D. R., Bolch, W., Boltz, T., Bosmans, H., Gislason-Lee, A. J., . . . Zamora, D. (2021). Estimation of patient skin dose in fluoroscopy: summary of a joint report by AAPM TG357 and EFOMP. Medical physics (Lancaster), 48(7), e671-e696
Open this publication in new window or tab >>Estimation of patient skin dose in fluoroscopy: summary of a joint report by AAPM TG357 and EFOMP
Show others...
2021 (English)In: Medical physics (Lancaster), ISSN 0094-2405, Vol. 48, no 7, p. e671-e696Article in journal (Refereed) Published
Abstract [en]

Background: Physicians use fixed C-arm fluoroscopy equipment with many interventional radiological and cardiological procedures. The associated effective dose to a patient is generally considered low risk, as the benefit-risk ratio is almost certainly highly favorable. However, X-ray-induced skin injuries may occur due to high absorbed patient skin doses from complex fluoroscopically guided interventions (FGI). Suitable action levels for patient-specific follow-up could improve the clinical practice. There is a need for a refined metric regarding follow-up of X-ray-induced patient injuries and the knowledge gap regarding skin dose-related patient information from fluoroscopy devices must be filled. The most useful metric to indicate a risk of erythema, epilation or greater skin injury that also includes actionable information is the peak skin dose, that is, the largest dose to a region of skin.

Materials and Methods: The report is based on a comprehensive review of best practices and methods to estimate peak skin dose found in the scientific literature and situates the importance of the Digital Imaging and Communication in Medicine (DICOM) standard detailing pertinent information contained in the Radiation Dose Structured Report (RDSR) and DICOM image headers for FGI devices. Furthermore, the expertise of the task group members and consultants have been used to bridge and discuss different methods and associated available DICOM information for peak skin dose estimation.

Results: The report contributes an extensive summary and discussion of the current state of the art in estimating peak skin dose with FGI procedures with regard to methodology and DICOM information. Improvements in skin dose estimation efforts with more refined DICOM information are suggested and discussed.

Conclusions: The endeavor of skin dose estimation is greatly aided by the continuing efforts of the scientific medical physics community, the numerous technology enhancements, the dose-controlling features provided by the FGI device manufacturers, and the emergence and greater availability of the DICOM RDSR. Refined and new dosimetry systems continue to evolve and form the infrastructure for further improvements in accuracy. Dose-related content and information systems capable of handling big data are emerging for patient dose monitoring and quality assurance tools for large-scale multihospital enterprises.

Place, publisher, year, edition, pages
John Wiley & Sons, 2021
Keywords
fluoroscopically guided interventions, peak skin dose, x-ray fluoroscopy
National Category
Radiology, Nuclear Medicine and Medical Imaging
Identifiers
urn:nbn:se:umu:diva-183708 (URN)10.1002/mp.14910 (DOI)000652210100001 ()33930183 (PubMedID)2-s2.0-85106011476 (Scopus ID)
Available from: 2021-05-31 Created: 2021-05-31 Last updated: 2023-03-24Bibliographically approved
Löfstedt, T., Hellström, M., Bylund, M. & Garpebring, A. (2020). Bayesian non-linear regression with spatial priors for noise reduction and error estimation in quantitative MRI with an application in T1 estimation. Physics in Medicine and Biology, 65(22), Article ID 225036.
Open this publication in new window or tab >>Bayesian non-linear regression with spatial priors for noise reduction and error estimation in quantitative MRI with an application in T1 estimation
2020 (English)In: Physics in Medicine and Biology, ISSN 0031-9155, E-ISSN 1361-6560, Vol. 65, no 22, article id 225036Article in journal (Refereed) Published
Abstract [en]

Purpose. To develop a method that can reduce and estimate uncertainty in quantitative MR parameter maps without the need for hand-tuning of any hyperparameters.

Methods. We present an estimation method where uncertainties are reduced by incorporating information on spatial correlations between neighbouring voxels. The method is based on a Bayesian hierarchical non-linear regression model, where the parameters of interest are sampled, using Markov chain Monte Carlo (MCMC), from a high-dimensional posterior distribution with a spatial prior. The degree to which the prior affects the model is determined by an automatic hyperparameter search using an information criterion and is, therefore, free from manual user-dependent tuning. The samples obtained further provide a convenient means to obtain uncertainties in both voxels and regions. The developed method was evaluated on T1 estimations based on the variable flip angle method.

Results. The proposed method delivers noise-reduced T1 parameter maps with associated error estimates by combining MCMC sampling, the widely applicable information criterion, and total variation-based denoising. The proposed method results in an overall decrease in estimation error when compared to conventional voxel-wise maximum likelihood estimation. However, this comes with an increased bias in some regions, predominately at tissue interfaces, as well as an increase in computational time.

Conclusions. This study provides a method that generates more precise estimates compared to the conventional method, without incorporating user subjectivity, and with the added benefit of uncertainty estimation.

Place, publisher, year, edition, pages
Institute of Physics (IOP), 2020
Keywords
Bayesian statistics, quantitative MRI, noise reduction, tissue parameter estimation, WAIC
National Category
Radiology, Nuclear Medicine and Medical Imaging Computer Vision and Robotics (Autonomous Systems) Probability Theory and Statistics
Identifiers
urn:nbn:se:umu:diva-177258 (URN)10.1088/1361-6560/abb9f5 (DOI)000591796100001 ()32947277 (PubMedID)2-s2.0-85097228803 (Scopus ID)
Funder
Swedish Research Council, 2019-0432Region Västerbotten, RV-738491
Available from: 2020-12-03 Created: 2020-12-03 Last updated: 2023-10-13Bibliographically approved
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0002-0200-6567

Search in DiVA

Show all publications