Figure 1 from GPSai: A Clinically Validated AI Tool for Tissue of Origin Prediction during Routine Tumor Profiling
<p>Specimen selection for GPSai model training and validation. A total of 201,612 non-CUP specimens were used to train the model, 21,106 non-CUP specimens and 443 CUP specimens were used for retrospective validation, 6,820 TCGA cases were used for external validation, and 72,421 non-CUP and 3,850 CUP specimens were used for prospective validation, respectively.</p>
Authors
- Jennifer R. Ribeiro (ORCID: https://orcid.org/0000-0002-1685-8568)
- Patricia Pittman
- Anthony Helmstetter
- Hassan Ghani
- Jaclyn F. Hechtman (ORCID: https://orcid.org/0000-0003-2645-0985)
- Rebecca Feldman (ORCID: https://orcid.org/0000-0003-4502-1473)
- Ari M. Vanderwalde (ORCID: https://orcid.org/0000-0002-6842-2563)
- Elena Florento
- Milan M. Radovich (ORCID: https://orcid.org/0000-0001-9667-156X)
- Jeffrey Swensen (ORCID: https://orcid.org/0000-0002-8204-4220)
- Matthew James Oberley (ORCID: https://orcid.org/0000-0001-6419-2513)
- Jim Abraham (ORCID: https://orcid.org/0000-0001-5038-1735)
- Casey Bales (ORCID: https://orcid.org/0000-0002-1016-6042)
- George W. Sledge (ORCID: https://orcid.org/0000-0003-0297-0775)
- Todd Maney
- Farah Abdulla (ORCID: https://orcid.org/0000-0002-3455-7566)
- David B. Spetzler (ORCID: https://orcid.org/0000-0002-9531-0970)
- Myra M. George
- David Bryant
- Stephanie Rock (ORCID: https://orcid.org/0000-0001-5700-6954)
Publication Details
- Published
- 2025-09-01
- DOI
- https://doi.org/10.1158/2767-9764.30022019
- Primary Topic
- Radiomics and Machine Learning in Medical Imaging
- Type
- article
- Field-Weighted Citation Impact
- 0.00