|The following article features coverage from the 2019 San Antonio Breast Cancer Symposium. Click here to read more of Cancer Therapy Advisor‘s conference coverage.|
Prediction models may one day help forecast how patients with hormone receptor (HR)-positive, HER2-negative advanced or metastatic breast cancer respond to cyclin dependent kinase 4/6 inhibitor (CDKI) therapy, according to the preliminary findings of a pooled analysis by members of the US and Drug Administration (FDA).
The findings were presented by the FDA-affiliated authors during a poster discussion at the 2019 San Antonio Breast Cancer Symposium (SABCS) in Texas.
The study researchers pooled data from 4580 breast cancer patients from 8 randomized controlled pivotal trials that were submitted to the FDA. Patients were divided into 2 groups on the basis of treatment received: hormonal therapy (1778 patients) or CDKI therapy (2802 patients).
Using patient characteristics at baseline, models were developed to predict progression-free survival (PFS) and overall survival (OS) risk for a given therapy, as well as the probability of developing adverse events.
According to the study authors, the machine learning models “can predict treatment efficacy and adverse event development.”
The model to predict outcomes from CDKI therapy had a prediction accuracy of 69.2%, and the model to predict outcomes from hormonal therapy had a prediction accuracy of 70.6%, according to the authors of the poster.
Although the study has strengths, such as prediction models being “simple, straightforward, and interpretable,” the study researchers pointed out important study limitations. Namely, the prediction models only apply to HR-positive, HER2-negative advanced breast cancer and the data used in the models do not capture all patient variables, “thus reducing the accuracy and leaving room for error.”
Read more of Cancer Therapy Advisor‘s coverage of SABCS by visiting the conference page.
Mason J, Gong Y, Amiri-Kordestani L, et al. Prediction of CDK inhibitor efficacy in ER+/HER2- breast cancer using machine learning algorithms. Poster presentation at: 2019 San Antonio Breast Cancer Symposium; December 10-14, 2019; San Antonio, TX. Abstract PD2-07.