Exploiting Longitudinal Context in Clinician-Verified Interactive Lesion Tracking
A new approach to tracking tumor lesions in CT scans has been proposed, focusing on clinician verification to enhance accuracy. This method combines automated tracking with user input to resolve ambiguities in lesion segmentation. The researchers also introduced a new benchmark dataset for pancreatic cancer to improve generalization in tracking models.
- ▪The proposed Verified Tracking paradigm allows clinicians to verify prompts for improved lesion tracking accuracy.
- ▪The new framework combines early spatial prompt fusion with latent temporal difference weighting.
- ▪The approach achieved first place in the MICCAI autoPET IV challenge and outperformed previous methods in both automatic and verified tracking settings.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Computer Vision and Pattern Recognition arXiv:2605.23118 (cs) [Submitted on 22 May 2026] Title:Exploiting Longitudinal Context in Clinician-Verified Interactive Lesion Tracking Authors:Yannick Kirchhoff, Maximilian Rokuss, Daniel Philipp Mertens, David Füller, Benjamin Hamm, Andreas Schreyer, Oliver Ritter, Klaus Maier-Hein View a PDF of the paper titled Exploiting Longitudinal Context in Clinician-Verified Interactive Lesion Tracking, by Yannick Kirchhoff and 7 other authors View PDF HTML (experimental) Abstract:Tracking tumor lesions across serial CT scans is essential for oncological response assessment.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
