A workflow utilizing general-purpose large language models for efficient structuring and data mining of bone scintigraphy narratives
WenXin Tang, Jun Zhou, Wenxin Chen, Yunze Xie, Lichao Zhang, Shuting Chen, Xiaoguang Hou, Shuo Wang, Hongcheng Shi · 2026
Read A workflow utilizing general-purpose large language models for efficient structuring and data mining of bone scintigraphy narratives by WenXin Tang, Jun Zhou, Wenxin Chen, Yunze Xie, Lichao Zhang, Shuting Chen, Xiaoguang Hou, Shuo Wang, Hongcheng Shi (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-13376209.txt.
Description
Tang and co-authors build a workflow that uses general purpose language models to turn bone scan reports into structured data. Whole body bone scans guide cancer monitoring, but their free text reports block easy data use across centers. Their 2026 study tests whether expert designed prompts can steer models without special fine tuning.
Four models were benchmarked on a multicenter truth set, and one model, DeepSeek-R1, proved most accurate and stable. With a human in the loop, model help cut manual work time by roughly three quarters and raised accuracy. The team then processed records from more than eighteen thousand patients to build a bone spread atlas for eight common cancers.
Catalog record
| CREATOR | WenXin Tang, Jun Zhou, Wenxin Chen, Yunze Xie, Lichao Zhang, Shuting Chen, Xiaoguang Hou, Shuo Wang, Hongcheng Shi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 39 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |