CLEAR-HPV: Interpretable concept discovery for human-papillomavirus-associated morphology in whole-slide histology
Weiyi Qin, Yingci Liu-Swetz, Shiwei Tan, Hao Wang · 2026
Read CLEAR-HPV: Interpretable concept discovery for human-papillomavirus-associated morphology in whole-slide histology by Weiyi Qin, Yingci Liu-Swetz, Shiwei Tan, Hao Wang (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/PMC13576518.txt.
Description
Weiyi Qin and three colleagues present CLEAR-HPV, a 2026 method that makes AI predictions of virus status in cancer slides both accurate and readable. HPV status strongly shapes outlook and treatment in head and neck and cervical cancers. Deep learning models can read it from stained tissue slides, yet they rarely explain what they see.
Their framework reshapes the model's inner space so it sorts tissue into plain concepts without being taught labels for them. It found keratinising, basaloid, and stromal patterns on its own, mapped where each appears, and summed each slide as shares of just ten concepts. Those shares kept the predictive power of the original 1,536-number fingerprint. The concept structure held steady across three large cancer atlas datasets. The authors offer the method as a general add-on for slide-reading models.
Catalog record
| CREATOR | Weiyi Qin, Yingci Liu-Swetz, Shiwei Tan, Hao Wang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 87215 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |