Anatomy-aware lymphoma lesion detection in whole-body PET/CT - Takeaways - MDSpire

Anatomy-aware lymphoma lesion detection in whole-body PET/CT

  • By

  • Simone Bendazzoli

  • Antonios Tzortzakakis

  • Andréas Abrahamsson

  • Björn Engelbrekt Wahlin

  • Örjan Smedby

  • Maria Holstensson

  • Rodrigo Moreno

  • May 22, 2026

  • 0 min

Share

  • 1

    Incorporating anatomical prior information improves deep learning-based lesion detection performance in PET/CT imaging.

  • 2

    The study utilized organ segmentation masks as auxiliary input channels in nnDetection and Swin UNETR frameworks.

  • 3

    nnDetection with anatomical masks showed improved performance, with mAP@0.1–0.5 increasing from 0.288 to 0.335.

  • 4

    Anatomical priors had minimal impact on the Swin Transformer, indicating challenges in integrating this information.

  • 5

    The findings highlight the importance of anatomical context for enhancing lesion detection in CNN-based models.

Original Source(s)

Anatomy-aware lymphoma lesion detection in whole-body PET/CT

  • by Simone Bendazzoli, Antonios Tzortzakakis, Andréas Abrahamsson, Björn Engelbrekt Wahlin, Örjan Smedby, Maria Holstensson, Rodrigo Moreno

  • May 22, 2026

Related Content

Is full adherence mandatory? Real-world outcomes of completing perioperative chemoimmunotherapy in resectable non-small cell lung cancer

Validation of a cervical CDO1/CELF4 methylation test for endometrial cancer: a prospective paired-sample comparison with intrauterine specimen

Management of toxicities from antibody – drug conjugates in breast cancer