Attention-guided and MIL-constrained CycleGAN for high-fidelity virtual p16 and Ki-67 staining
Xinli Lei, Yuefeng Xie, Andi Duan, Yuhang Wu, Linhong Liao, Xi Wang · 2026
Read Attention-guided and MIL-constrained CycleGAN for high-fidelity virtual p16 and Ki-67 staining by Xinli Lei, Yuefeng Xie, Andi Duan, Yuhang Wu, Linhong Liao, Xi 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/pmc-13270425.txt.
Description
Xinli Lei and co-authors build in 2026 an image-generating system that paints two diagnostic protein stains onto cervical tissue images without extra lab work. Grading early cervical cell changes normally needs chemical stains for p16 and Ki-67 on separate slices, costing time, money, and tissue. Their model learns from roughly matched slice pairs, using attention guidance and a learning scheme that tolerates the slices' imperfect alignment.
Two pathologists graded cases with and without the virtual stains, and their performance improved when the generated stains were available. Checks on an outside dataset suggested the model transfers beyond its home data. The authors also guard against the system inventing false stain signals, adding consistency constraints and expert review. Virtual staining, they argue, can support real diagnosis.
Catalog record
| CREATOR | Xinli Lei, Yuefeng Xie, Andi Duan, Yuhang Wu, Linhong Liao, Xi Wang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |