From routine full-spine radiographs to decision-oriented Risser stratification: an interpretable deep-learning approach
Zexi Wang, Yuan Zhang, Yixi Wang, Feng Xue, Yi Yang, Wen Zhao · 2026
Read From routine full-spine radiographs to decision-oriented Risser stratification: an interpretable deep-learning approach by Zexi Wang, Yuan Zhang, Yixi Wang, Feng Xue, Yi Yang, Wen Zhao (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/PMC13318898.txt.
Description
Zexi Wang and colleagues present a 2026 deep learning study for judging remaining growth in teenagers with sideways spine curvature. Risser staging guides treatment, but readers often disagree, especially on full-spine X-rays. The team built an interpretable model that sorts patients into younger or older maturity bands. Data came from one centre.
The sample held 875 standing full-spine radiographs from patients aged 10 to 18. Three surgeons set the reference standard by consensus. The chosen ResNet-18 model reached an AUC of 0.938 on held-out test data. Accuracy was 0.875 for the two-band split in that testing. Heat maps showed the model looking at the hip growth plate, as humans do. Surgeons using it read each case about ten seconds faster and agreed more.
Catalog record
| CREATOR | Zexi Wang, Yuan Zhang, Yixi Wang, Feng Xue, Yi Yang, Wen Zhao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47798 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |