English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Degenerative spinal pathologies are highly prevalent among the elderlypopulation. Timely diagnosis of osteoporotic fractures and other degenerativedeformities facilitates proactive measures to mitigate the risk of severe backpain and disability. In this study, we specifically explore the use of shapeauto-encoders for vertebrae, taking advantage of advancements in automatedmulti-label segmentation and the availability of large datasets forunsupervised learning. Our shape auto-encoders are trained on a large set ofvertebrae surface patches, leveraging the vast amount of available data forvertebra segmentation. This addresses the label scarcity problem faced whenlearning shape information of vertebrae from image intensities. Based on thelearned shape features we train an MLP to detect vertebral body fractures.Using segmentation masks that were automatically generated using theTotalSegmentator, our proposed method achieves an AUC of 0.901 on the VerSe19testset. This outperforms image-based and surface-based end-to-end trainedmodels. Additionally, our results demonstrate that pre-training the models inan unsupervised manner enhances geometric methods like PointNet and DGCNN. Ourfindings emphasise the advantages of explicitly learning shape features fordiagnosing osteoporotic vertebrae fractures. This approach improves thereliability of classification results and reduces the need for annotatedlabels. This study provides novel insights into the effectiveness of variousencoder-decoder models for shape analysis of vertebrae and proposes a newdecoder architecture: the point-based shape decoder.

Shape Matters: Detecting Vertebral Fractures Using Differentiable Point-Based Shape Decoding