Computational analysis of variability and uncertainty in the clinical reference on magnetic resonance imaging radiomics: modelling and performance
Cindy Xue, Jing Yuan, Gladys G Lo, Darren MC Poon, Winnie CW Chu · 2024
Read Computational analysis of variability and uncertainty in the clinical reference on magnetic resonance imaging radiomics: modelling and performance by Cindy Xue, Jing Yuan, Gladys G Lo, Darren MC Poon, Winnie CW Chu (2024) 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-11573982.txt.
Description
Cindy Xue and co-authors study, in 2024, how errors in the clinical labels used for training affect MRI radiomics models of prostate cancer. Radiomics turns scan texture into numbers that may predict tumor grade, but the grade labels themselves can be wrong. The team used two public datasets and deliberately shuffled parts of the training labels.
Feature selection grew unstable as label noise rose. Once a fifth of labels were shuffled, model accuracy on held-out data dropped sharply from very high to middling levels, and results varied more run to run. The lesson is blunt: label quality caps model quality, and uncertainty in the reference standard must be reported, not ignored.
Catalog record
| CREATOR | Cindy Xue, Jing Yuan, Gladys G Lo, Darren MC Poon, Winnie CW Chu |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |