AUTHOR=Yu Ze , Ye Xuan , Liu Hongyue , Li Huan , Hao Xin , Zhang Jinyuan , Kou Fang , Wang Zeyuan , Wei Hai , Gao Fei , Zhai Qing 

TITLE=Predicting Lapatinib Dose Regimen Using Machine Learning and Deep Learning Techniques Based on a Real-World Study

JOURNAL=Frontiers in Oncology

VOLUME=Volume 12 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.893966

DOI=10.3389/fonc.2022.893966

ISSN=2234-943X

ABSTRACT=Lapatinib is used for the treatment of metastatic HER2(+) breast cancer. We aim to establish a prediction model for lapatinib dose using machine learning and deep learning techniques based on real-world study. There were 149 breast cancer patients enrolled from July 2016 to June 2017 at Fudan University Shanghai Cancer Center. Sequential Forward Selection algorithm based on Random Forest was applied for variable selection. Ten machine learning and deep learning algorithms were compared their predictive abilities (Logistic Regression, SVM, Random Forest, Adaboost, XGBoost, GBDT, LightGBM, CatBoost, TabNet and ANN). As a result, TabNet was chosen to construct the prediction model with the best performance (accuracy=0.83 and AUC=0.83). Afterward, 4 variables that strongly correlated with lapatinib dose were ranked via importance score as follows: treatment protocols, weight, number of chemotherapy treatments, and number of metastases. Finally, confusion matrix was used to validate the model for dose regimen of 1250 mg lapatinib (precision=82% and recall=95%), and for dose regimen of 1000 mg lapatinib (precision=88% and recall=64%). To conclude, we established a deep learning model to predict lapatinib dose based on important influencing variables selected from real-world evidence, to achieve an optimal individualized dose regimen with good predictive performance.