Noninvasive molecular subtyping of breast cancer using multimodal ultrasound spatiotemporal transformer
Tao Chen, Qinghua Niu, Zichen An, Chenhui Wang, Zhihao Chen, Lianfang Du, Chao Jia, Chunxiao Li, Jiying Gu, Ting Sun, Yuxiang Wang, Fan Li, Hongming Shan · 2026
Read Noninvasive molecular subtyping of breast cancer using multimodal ultrasound spatiotemporal transformer by Tao Chen, Qinghua Niu, Zichen An, Chenhui Wang, Zhihao Chen, Lianfang Du, Chao Jia, Chunxiao Li, Jiying Gu, Ting Sun, Yuxiang Wang, Fan Li, Hongming Shan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13346553.txt.
Description
Chen and co-authors build a scan-based test that sorts breast tumors by type without a biopsy. Biopsy now sets the type that guides drug care, but it cuts tissue. Plain B-mode scans show shape well yet miss blood flow cues. Their tool, MUST-Sub, joins B-mode shape traits with flow traits from dye-based scans. A time-aware Transformer net reads both streams at once. It sorts tumors as Luminal, HER2-rich, or triple-negative.
Trained on past cases, MUST-Sub was checked on held-out, new, and multi-site sets. Its mean scores were 0.94, 0.90, and 0.92 on those sets. It beat nets that used B-mode scans alone. Two score marks from its focus maps also tracked tumor size and a growth index.
Catalog record
| CREATOR | Tao Chen, Qinghua Niu, Zichen An, Chenhui Wang, Zhihao Chen, Lianfang Du, Chao Jia, Chunxiao Li, Jiying Gu, Ting Sun, Yuxiang Wang, Fan Li, Hongming Shan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |