Machine learning to predict the individual Risk of Treatment-relevant Toxicity for Breast Cancer Patients undergoing Neoadjuvant Systemic Treatment
Senologie - Zeitschrift für Mammadiagnostik und -therapie · 01.06.2024 · 21 (02) · e5
DOI: 10.1055/s-0044-1786077 ↗Kurz gefasst
Fachartikel in der Zeitschrift Senologie - Zeitschrift für Mammadiagnostik und -therapie, erschienen am 01.06.2024. Verfasst von L. Cai und 11 weiteren Autor:innen.
Themen: Künstliche Intelligenz in der Pflege, Onkologische Pflege
Abstract (Originalsprache)
Background
Toxicity to systemic cancer treatment represents a major anxiety for patients and a challenge to treatment plans. We aimed to develop machine learning algorithms to accurately predict an individual’s risk of experiencing treatment-relevant toxicity prior to the initiation of treatment.
Methods
Clinical records were retrieved from a consecutive cohort of patients who underwent neoadjuvant treatment for early breast cancer. We developed and validated machine learning algorithms to predict Grade III or IV toxicity (anemia, neutropenia, deviation of liver enzymes, nephrotoxicity, thrombopenia, electrolyte disturbance, or neuropathy). We used 10-fold cross-validation to develop two algorithms (Logistic Regression with elastic net penalty (GLM) and Support Vector Machines (SVM)). Algorithm predictions were compared to documented toxicity events and diagnostic performance was evaluated via area under the curve (AUROC).
Results
A total of 590 patients were identified, 432 in the development set and 158 in the validation set. Median age was 51 years and 55.8% (329 of 590) experienced Grade III or IV toxicity. Performance significantly improved when adding referenced treatment information (referenced regimen, referenced Summation Dose Intensity Product) in addition to patient and tumor variables: GLM AUROC 0.59 (95% CI, 0.54 to 0.70) vs. 0.75 (95% CI, 0.68 to 0.83), P =0.02; SVM AUROC 0.64 (95% CI, 0.55 to 0.72) vs. 0.75 (95% 0.77 to 0.82), P =0.01.
Conclusion
The individual risk of treatment-relevant toxicity can be accurately predicted using machine learning algorithms. We demonstrate a promising way to improve efficacy and minimize toxicity of systemic cancer treatment. Publication History Article published online: 03 May 2024 © 2024. Thieme. All rights reserved. Georg Thieme Verlag KG Rüdigerstraße 14, 70469 Stuttgart, Germany
Zitieren
APA
Cai, L., Duetsch, T. M., Sidey-Gibbons, C., Kobel, M., Riedel, F., Smetanay, K., Fremd, C., Michel, L., Golatta, M., Heil, J., Schneeweiss, A., & Pfob, A. (2024). Machine learning to predict the individual Risk of Treatment-relevant Toxicity for Breast Cancer Patients undergoing Neoadjuvant Systemic Treatment. Senologie - Zeitschrift für Mammadiagnostik und -therapie, 21(02), e5. https://doi.org/10.1055/s-0044-1786077
Vancouver
Cai L, Duetsch TM, Sidey-Gibbons C, Kobel M, Riedel F, Smetanay K, et al. Machine learning to predict the individual Risk of Treatment-relevant Toxicity for Breast Cancer Patients undergoing Neoadjuvant Systemic Treatment. Senologie - Zeitschrift für Mammadiagnostik und -therapie. 2024;21(02):e5. doi:10.1055/s-0044-1786077
BibTeX
@article{Cai2024,
author = {Cai, L. and Duetsch, T. M. and Sidey-Gibbons, C. and Kobel, M. and Riedel, F. and Smetanay, K. and Fremd, C. and Michel, L. and Golatta, M. and Heil, J. and Schneeweiss, A. and Pfob, A.},
title = {Machine learning to predict the individual Risk of Treatment-relevant Toxicity for Breast Cancer Patients undergoing Neoadjuvant Systemic Treatment},
journal = {Senologie - Zeitschrift für Mammadiagnostik und -therapie},
year = {2024},
volume = {21},
number = {02},
pages = {e5},
doi = {10.1055/s-0044-1786077}
}
EUVIMED