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/

Zendy Plus

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

Zendy Tools

Zendy Open

This study addresses critical gaps in automated lymphoma segmentation fromPET/CT images, focusing on issues often overlooked in existing literature.While deep learning has been applied for lymphoma lesion segmentation, fewstudies incorporate out-of-distribution testing, raising concerns about modelgeneralizability across diverse imaging conditions and patient populations. Wehighlight the need to compare model performance with expert human annotators,including intra- and inter-observer variability, to understand task difficultybetter. Most approaches focus on overall segmentation accuracy but overlooklesion-specific measures important for precise lesion detection and diseasequantification. To address these gaps, we propose a clinically relevantframework for evaluating deep segmentation networks. Using this lesionmeasure-specific evaluation, we assess the performance of four deep networks(ResUNet, SegResNet, DynUNet, and SwinUNETR) across 611 cases frommulti-institutional datasets, covering various lymphoma subtypes and lesioncharacteristics. Beyond standard metrics like the Dice similarity coefficient,we evaluate clinical lesion measures and their prediction errors. We alsointroduce detection criteria for lesion localization and propose a newdetection Criterion 3 based on metabolic characteristics. We show that networksperform better on large, intense lesions with higher metabolic activity.Finally, we compare network performance to physicians via intra- andinter-observer variability analyses, demonstrating that network errors closelyresemble those made by experts, i.e., the small and faint lesions remainchallenging for both humans and networks. This study aims to improve automatedlesion segmentation's clinical relevance, supporting better treatment decisionsfor lymphoma patients. The code is available at:https://github.com/microsoft/lymphoma-segmentation-dnn.

Comprehensive framework for evaluation of deep neural networks in  detection and quantification of lymphoma from PET/CT images: clinical  insights, pitfalls, and observer agreement analyses