Abstract:e12567 Background: Radiation-induced lymphopenia accompanied with radiation therapy is associated with inferior clinical outcomes in a wide variety of solid malignancies. This study aimed to examine the potential determines of radiation-induced lymphocyte decrease and radiation-induced lymphopenia in breast cancer patients who underwent radiotherapy. Methods: Patients with breast cancer treated who underwent radiotherapy were enrolled in University of Hong Kong-Shenzhen Hospital (our cohort). Circulating lymphocyte levels were evaluated within 7 days prior to and end of radiation therapy. Feature groups including clinical data, tumor characteristics, radiotherapy dosimetrics, treatment regiments were also collected. We applied machine learning algorithms (Extreme Gradient Boosting, XGboost) to predict the ratio of lymphocyte level after radiotherapy to baseline lymphocyte level and the event of lymphopenia and compared with Lasso regression approaches. Next, we used Shapley additive explanation (SHAP) to explore the directional contribution of each feature for lymphocyte decrease and lymphopenia. For the purpose of model validation and proof-of-concept validation, an independent cohort of patients enrolled in prospective trial was eligible (IP cohort). Results: A total of 589 patients were enrolled in our cohort and 203 patients in IP cohort. XGboost models which trained in our cohort with performances of a mean RMSE: 0.157 and R2: 53.9% for the ratio of lymphocyte levels; a mean accuracy: 0.757 and ROC-AUC: 0.733 for the lymphopenia events, separately. These models can predict the ratio of lymphocyte levels with a mean RMSE: 0.175 and R2: 47%; predict the lymphopenia events with a mean accuracy: 0.739 and ROC-AUC: 0.737 in the totally independent IP cohort. The feature group of dosimetrics had the largest predictive power with RMSE: 0.192, R2: 29.8%, accuracy: 0.678 and ROC-AUC: 0.667; followed by the group of baseline blood cells with predictive power as RMSE: 0.207, R2: 18.9%, accuracy: 0.669 and ROC-AUC: 0.645. Next, by SHAP value analysis, we investigated that integral dose of the total body, V5 dose, mean lung dose and V20 dose of ipsilateral lung/bilateral lungs were in consequence important promote factors for lymphocyte decrease and for the event of lymphopenia, while the features of baseline monocyte, mean heart dose and tumor size played a role of protection at some extend. Conclusions: In this study, we constructed robust XGboost models for predicting the lymphocyte decrease and the event of lymphopenia in breast cancer patients who underwent radiation therapy. We also applied SHAP analysis for revealing the directional contribution of features. These results are important either for the understanding the contributions of dosimetrics on immune response or for the refine of radiation dosimetrics before treatment in future clinical usages.

A Light Gradient Boosting Machine-Enabled Early Prediction of Cardiotoxicity for Breast Cancer Patients

A Machine Learning Framework for Early Prediction of Radiation Dermatitis in Patients with Breast Cancer Receiving Radiation Treatment: A Multicenter Retrospective Analysis Study

Identifying Cancer Patients at Risk for Heart Failure Using Machine Learning Methods

LightGBM: A Leading Force in Breast Cancer Diagnosis Through Machine Learning and Image Processing

Potential Determinants for Radiation-Induced Lymphopenia in Patients With Breast Cancer Using Interpretable Machine Learning Approach

Prediction of radiation-induced acute skin toxicity in breast cancer patients using data encapsulation screening and dose-gradient-based multi-region radiomics technique: A multicenter study

LightGBM: An Effective and Scalable Algorithm for Prediction of Chemical Toxicity-Application to the Tox21 and Mutagenicity Data Sets

Machine Learning in Cardio-Oncology: New Insights from an Emerging Discipline

Potential determinants of radiation-induced lymphocyte decrease and lymphopenia in breast cancer patients by machine learning approaches.

Early detection of chemotherapy-induced cardiotoxicity in breast cancer patients: a comprehensive analysis using speckle tracking echocardiography

Biology-guided deep learning predicts prognosis and cancer immunotherapy response

XML‐LightGBMDroid: A self‐driven interactive mobile application utilizing explainable machine learning for breast cancer diagnosis

Prediction of Extubation Failure for Intensive Care Unit Patients Using Light Gradient Boosting Machine

Advancing patient care: Machine learning models for predicting grade 3+ toxicities in gynecologic cancer patients treated with HDR brachytherapy

Prognostic prediction of breast cancer patients using machine learning models: a retrospective analysis

Cancer immunotherapy response prediction from multi-modal clinical and image data using semi-supervised deep learning

Deep learning-based multi-modal data integration enhancing breast cancer disease-free survival prediction

Risk prediction of heart diseases in patients with breast cancer: A deep learning approach with longitudinal electronic health records data

On Weather Data-Based Prediction of Gamma Exposure Rates Using Gradient Boosting Learning for Environmental Radiation Monitoring

234 Predictive Radiomics in Glioblastoma: A Machine Learning Approach for Personalized Survival Outcomes and a Novel Web Application

Blood pressure stratification using photoplethysmography and light gradient boosting machine