![[2602.19005] GUIDE-US: Grade-Informed Unpaired Distillation of Encoder Knowledge from Histopathology to Micro-UltraSound](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.19005] GUIDE-US: Grade-Informed Unpaired Distillation of Encoder Knowledge from Histopathology to Micro-UltraSound
Summary
The paper presents a novel method for non-invasive grading of prostate cancer using micro-ultrasound, leveraging knowledge distillation from histopathology without requiring patient-level pairing.
Why It Matters
This research is significant as it enhances the accuracy of prostate cancer detection through imaging, potentially improving patient outcomes by facilitating earlier diagnosis and treatment planning. The method's reliance on unpaired data makes it applicable in clinical settings where patient data is limited.
Key Takeaways
- Introduces a grade-informed unpaired distillation strategy for micro-ultrasound imaging.
- Improves sensitivity to clinically significant prostate cancer by 3.5% at 60% specificity.
- Eliminates the need for patient-level pairing or image registration in training.
- Enhances overall sensitivity of prostate cancer detection from imaging.
- Source code will be publicly available, promoting further research and application.
Computer Science > Computer Vision and Pattern Recognition arXiv:2602.19005 (cs) [Submitted on 22 Feb 2026] Title:GUIDE-US: Grade-Informed Unpaired Distillation of Encoder Knowledge from Histopathology to Micro-UltraSound Authors:Emma Willis, Tarek Elghareb, Paul F. R. Wilson, Minh Nguyen Nhat To, Mohammad Mahdi Abootorabi, Amoon Jamzad, Brian Wodlinger, Parvin Mousavi, Purang Abolmaesumi View a PDF of the paper titled GUIDE-US: Grade-Informed Unpaired Distillation of Encoder Knowledge from Histopathology to Micro-UltraSound, by Emma Willis and 8 other authors View PDF HTML (experimental) Abstract:Purpose: Non-invasive grading of prostate cancer (PCa) from micro-ultrasound (micro-US) could expedite triage and guide biopsies toward the most aggressive regions, yet current models struggle to infer tissue micro-structure at coarse imaging resolutions. Methods: We introduce an unpaired histopathology knowledge-distillation strategy that trains a micro-US encoder to emulate the embedding distribution of a pretrained histopathology foundation model, conditioned on International Society of Urological Pathology (ISUP) grades. Training requires no patient-level pairing or image registration, and histopathology inputs are not used at inference. Results: Compared to the current state of the art, our approach increases sensitivity to clinically significant PCa (csPCa) at 60% specificity by 3.5% and improves overall sensitivity at 60% specificity by 1.2%. Conclusion: By enabling earlie...
