Discrimination, calibration, and variable importance in statistical and machine learning models for predicting overall survival in advanced non–small cell lung cancer patients treated with immune checkpoint inhibitors
There was no clear best-performing model for either discrimination or calibration, although XGBoost models showed possible superior calibration numerically. Performance of a given model varied across evaluation cohorts, highlighting the importance of model assessment using multiple independent datasets. All models identified pretreatment NLR and ECOGPS as the key prognostic factors.
Authors
- Richard John Woodman (ORCID: https://orcid.org/0000-0002-4094-1222)
- Andrew Rowland (ORCID: https://orcid.org/0000-0002-8946-3954)
- Lee X. Li (ORCID: https://orcid.org/0000-0001-5259-5198)
- Ashley M. Hopkins (ORCID: https://orcid.org/0000-0001-7652-4378)
- Michael J. Sorich (ORCID: https://orcid.org/0000-0003-1999-866X)
- Ahmad Y. Abuhelwa (ORCID: https://orcid.org/0000-0002-4182-065X)
- Yuan Gao (ORCID: https://orcid.org/0000-0001-6647-2156)
- Natalie Parent
Institutions
- Flinders University (AU)
- University of Sharjah (AE)
Publication Details
- Journal
- Journal of Clinical Epidemiology
- Published
- 2025-11-22
- DOI
- https://doi.org/10.1016/j.jclinepi.2025.112082
- Citations
- 1
- Primary Topic
- Cancer Immunotherapy and Biomarkers
- Type
- article
- Field-Weighted Citation Impact
- 1.05