Motion grading of high-resolution quantitative computed tomography supported by deep convolutional neural networks

Simple item page
dc.contributor.authorWalle, Matthiasen
dc.contributor.authorEggemann, Dominicen
dc.contributor.authorAtkins, Penny R.en
dc.contributor.authorKendall, Jack J.en
dc.contributor.authorStock, Kerstinen
dc.contributor.authorMuller, Ralphen
dc.contributor.authorCollins, Caitlyn J.en
dc.date.accessioned2023-04-10T17:46:36Zen
dc.date.available2023-04-10T17:46:36Zen
dc.date.issued2023-01en
dc.description.abstractImage quality degradation due to subject motion confounds the precision and reproducibility of measurements of bone density, morphology and mechanical properties from high-resolution peripheral quantitative computed tomography (HR-pQCT). Time-consuming operator-based scoring of motion artefacts remains the gold standard to determine the degree of acceptable motion. However, due to the subjectiveness of manual grading, HR-pQCT scans of poor quality, which cannot be used for analysis, may be accepted upon initial review, leaving patients with incomplete or inaccurate imaging results. Convolutional Neural Networks (CNNs) enable fast image analysis with relatively few pre-processing requirements in an operator-independent and fully automated way for image classification tasks. This study aimed to develop a CNN that can predict motion scores from HR-pQCT images, while also being aware of uncertain predictions that require manual confirmation. The CNN calculated motion scores within seconds and achieved a high F1-score (86.8 +/- 2.8 %), with good precision (87.5 +/- 2.7 %), recall (86.7 +/- 2.9 %) and a substantial agreement with the ground truth measured by Cohen's kappa (kappa = 68.6 +/- 6.2 %); motion scores of the test dataset were predicted by the algorithm with comparable accuracy, precision, sensitivity and agreement as by the operators (p > 0.05). This post-processing approach may be used to assess the effect of motion scores on microstructural analysis and can be immediately implemented into clinical protocols, significantly reducing the time for quality assessment and control of HR-pQCT scans.en
dc.description.notesThis project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement Nos. 860898 and 841316.en
dc.description.sponsorshipEuropean Union [860898, 841316]en
dc.description.versionPublished versionen
dc.format.mimetypeapplication/pdfen
dc.identifier.doihttps://doi.org/10.1016/j.bone.2022.116607en
dc.identifier.eissn1873-2763en
dc.identifier.other116607en
dc.identifier.pmid36368464en
dc.identifier.urihttp://hdl.handle.net/10919/114458en
dc.identifier.volume166en
dc.language.isoenen
dc.publisherElsevieren
dc.rightsCreative Commons Attribution 4.0 Internationalen
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/en
dc.subjectHigh -resolution peripheral quantitativeen
dc.subjectcomputed tomographyen
dc.subjectMotion -gradingen
dc.subjectConvolutional neural networksen
dc.subjectMachine learningen
dc.subjectDeep learningen
dc.subjectArtificial intelligenceen
dc.titleMotion grading of high-resolution quantitative computed tomography supported by deep convolutional neural networksen
dc.title.serialBoneen
dc.typeArticle - Refereeden
dc.type.dcmitypeTexten

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
1-s2.0-S8756328222002848-main.pdf
Size:
1.76 MB
Format:
Adobe Portable Document Format
Description:
Published version
Download

Collections

Scholarly Works, Biomedical Engineering and Mechanics