Brain atrophy measurement over a MRI scanner change in multiple sclerosis

Sinnecker, Tim; Schädelin, Sabine; Benkert, Pascal; Ruberte, Esther; Amann, Michael; Lieb, Johanna; Naegelin, Yvonne; Müller, Jannis; Kühle, Jens; Kappos, Ludwig; Wuerfel, Jens; Granziera, Cristina; Yaldizli, Özgür

doi:10.1016/j.nicl.2022.103148

Cited by 8 publications

(8 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tracking brain volume changes over different scanners is prone to an error related to the scanner change (Lee et al, 2019). (Sinnecker et al, 2022) showed that given a paired data set, a corrective term for the volume computation can be calculated. To achieve a higher level of generalizability, images should be mapped to the target contrast.…”

Section: Related Workmentioning

confidence: 99%

“…We used a data set exclusively created to track a scanner change from a 1.5 T Siemens Magnetom Avanto to a 3 T Siemens Magnetom Skyra f it whole-body scanner in 2016 (Sinnecker et al, 2022). Scanner details can be found in Appendix A.…”

Section: Data Set and Training Detailsmentioning

confidence: 99%

“…with automated tools such as SIENA (Smith et al, 2002(Smith et al, , 2004) is highly difficult. Consequently, a scanner change in the middle of a long-term study affects the automatic evaluation and longitudinal comparability of scans (Sinnecker et al, 2022). Longitudinal studies are important to monitor progressive diseases such as Multiple Sclerosis (MS), a demyelinating central nervous system disease (Mahad et al, 2015).…”

Section: Introductionmentioning

confidence: 99%

“…The goal of this work is to restore the comparability of the images by mapping all images to the same target contrast. A data set was acquired by (Sinnecker et al, 2022), for which healthy subjects as well as MS patients were scanned in the 1.5 T and the 3 T scanner within approximately 3.5 months. Due to the short time span between the acquisitions, we assume that the differences in the images of the same participant are only related to the different scanner types.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Diffusion Models for Contrast Harmonization of Magnetic Resonance Images

Durrer¹,

Wolleb²,

Bieder³

et al. 2023

Preprint

View full text Add to dashboard Cite

Magnetic resonance (MR) images from multiple sources often show differences in image contrast related to acquisition settings or the used scanner type. For long-term studies, longitudinal comparability is essential but can be impaired by these contrast differences, leading to biased results when using automated evaluation tools. This study presents a diffusion model-based approach for contrast harmonization. We use a data set consisting of scans of 18 Multiple Sclerosis patients and 22 healthy controls. Each subject was scanned in two MR scanners of different magnetic field strengths (1.5 T and 3 T), resulting in a paired data set that shows scanner-inherent differences. We map images from the source contrast to the target contrast for both directions, from 3 T to 1.5 T and from 1.5 T to 3 T. As we only want to change the contrast, not the anatomical information, our method uses the original image to guide the image-to-image translation process by adding structural information. The aim is that the mapped scans display increased comparability with scans of the target contrast for downstream tasks. We evaluate this method for the task of segmentation of cerebrospinal fluid, grey matter and white matter. Our method achieves good and consistent results for both directions of the mapping.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Data Set and Training Detailsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Diffusion Models for Contrast Harmonization of Magnetic Resonance Images

Durrer¹,

Wolleb²,

Bieder³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Lee et al (2019) estimates the fixed effect of scanner changes with a linear model and accounts for this factor during measurement of BVC. Sinnecker et al (2022) estimates an additive fixed corrective term for scanner change by comparing BVC rates during scanner change and no scanner change for healthy control subjects. These methods are subject to scanner-specific variation and require group level test results to estimate the correction factor.…”

Section: Introductionmentioning

confidence: 99%

Learning from pseudo-labels: deep networks improve consistency in longitudinal brain volume estimation

Zhan¹,

Wang²,

Cabezas³

et al. 2023

Preprint

View full text Add to dashboard Cite

Our deep learning model is trained with pseudo-labels from SIENA and the impact of label noise is ameliorated by a regularisation method.• Data augmentation is introduced to cope with inconsistent imaging acquisitions that are largely unavoidable in clinical settings.• Experiments in two datasets demonstrate that DeepBVC has better accuracy, robustness, reliability and consistency compared with SIENA, even though the intermediate outputs from SIENA were used for training.

show abstract