REACT-KD: Region-Aware Cross-modal Topological Knowledge Distillation for Interpretable Medical Image Classification
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909855748456448 |
|---|---|
| author | Chen, Hongzhao Ding, Hexiao Jiang, Yufeng Lan, Jing Li, Ka Chun Cheng, Gerald W. Y. Ng, Nga-Chun Pu, Yao Cai, Jing Lin, Liang-ting Yoo, Jung Sun |
| author_facet | Chen, Hongzhao Ding, Hexiao Jiang, Yufeng Lan, Jing Li, Ka Chun Cheng, Gerald W. Y. Ng, Nga-Chun Pu, Yao Cai, Jing Lin, Liang-ting Yoo, Jung Sun |
| contents | Reliable and interpretable tumor classification from clinical imaging remains a core challenge. The main difficulties arise from heterogeneous modality quality, limited annotations, and the absence of structured anatomical guidance. We present REACT-KD, a Region-Aware Cross-modal Topological Knowledge Distillation framework that transfers supervision from high-fidelity multi-modal sources into a lightweight CT-based student model. The framework employs a dual teacher design. One branch captures structure-function relationships through dual-tracer PET/CT, while the other models dose-aware features using synthetically degraded low-dose CT. These branches jointly guide the student model through two complementary objectives. The first achieves semantic alignment through logits distillation, and the second models anatomical topology through region graph distillation. A shared CBAM3D module ensures consistent attention across modalities. To improve reliability in deployment, REACT-KD introduces modality dropout during training, which enables robust inference under partial or noisy inputs. As a case study, we applied REACT-KD to hepatocellular carcinoma staging. The framework achieved an average AUC of 93.5\% on an internal PET/CT cohort and maintained 76.6\% to 81.5\% AUC across varying levels of dose degradation in external CT testing. Decision curve analysis further shows that REACT-KD consistently provides the highest net clinical benefit across all thresholds, confirming its value in real-world diagnostic practice. Code is available at: https://github.com/Kinetics-JOJO/REACT-KD |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_02104 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | REACT-KD: Region-Aware Cross-modal Topological Knowledge Distillation for Interpretable Medical Image Classification Chen, Hongzhao Ding, Hexiao Jiang, Yufeng Lan, Jing Li, Ka Chun Cheng, Gerald W. Y. Ng, Nga-Chun Pu, Yao Cai, Jing Lin, Liang-ting Yoo, Jung Sun Image and Video Processing Computer Vision and Pattern Recognition Reliable and interpretable tumor classification from clinical imaging remains a core challenge. The main difficulties arise from heterogeneous modality quality, limited annotations, and the absence of structured anatomical guidance. We present REACT-KD, a Region-Aware Cross-modal Topological Knowledge Distillation framework that transfers supervision from high-fidelity multi-modal sources into a lightweight CT-based student model. The framework employs a dual teacher design. One branch captures structure-function relationships through dual-tracer PET/CT, while the other models dose-aware features using synthetically degraded low-dose CT. These branches jointly guide the student model through two complementary objectives. The first achieves semantic alignment through logits distillation, and the second models anatomical topology through region graph distillation. A shared CBAM3D module ensures consistent attention across modalities. To improve reliability in deployment, REACT-KD introduces modality dropout during training, which enables robust inference under partial or noisy inputs. As a case study, we applied REACT-KD to hepatocellular carcinoma staging. The framework achieved an average AUC of 93.5\% on an internal PET/CT cohort and maintained 76.6\% to 81.5\% AUC across varying levels of dose degradation in external CT testing. Decision curve analysis further shows that REACT-KD consistently provides the highest net clinical benefit across all thresholds, confirming its value in real-world diagnostic practice. Code is available at: https://github.com/Kinetics-JOJO/REACT-KD |
| title | REACT-KD: Region-Aware Cross-modal Topological Knowledge Distillation for Interpretable Medical Image Classification |
| topic | Image and Video Processing Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2508.02104 |