2024 World Conference on Lung Cancer (WCLC) - ePosters-EP.08F.11 Development of Clinical Prediction Model Using Digital Device for ILD in the Stage III NSCLC Patients Treated with Durvalumab

EP.08F.11 Development of Clinical Prediction Model Using Digital Device for ILD in the Stage III NSCLC Patients Treated with Durvalumab

Back to course

Pdf Summary

The study aimed to develop a machine learning-based clinical prediction model to anticipate grade 2 or higher interstitial lung diseases/radiation pneumonitis (ILD/RP) in patients with stage III non-small cell lung cancer (NSCLC) receiving durvalumab after chemoradiation therapy (CRT). Prompt prediction of ILD/RP is crucial as severe cases can necessitate cessation of treatment. For this, physiological data from wearable devices and background patient information were utilized. <br /><br />The research incorporated a multicenter, prospective, non-interventional design and involved 145 patients from Japanese sites, with 123 included in the analysis. Data collection focused on features like heart rate and oxygen saturation (SpO2) recorded via wearable devices. Several machine learning models were examined, including XGBoost, random forest, support vector machine, and recurrent neural networks.<br /><br />The XGBoost model demonstrated the best predictive performance with a ROC-AUC of 0.789, indicating rational potential for clinical application. The study highlighted the importance of heart rate and SpO2 data while noting that cough frequency data was less significant to the model's performance. <br /><br />Results suggested that integrating such a prediction model could enhance early detection of ILD/RP, potentially informing timely medical interventions. However, implementation in clinical settings requires further validation through larger sample sizes and external verification to establish effectiveness.<br /><br />In summary, the development of a prediction model using physiological and background data shows promise in improving patient outcomes in NSCLC treatment by enabling healthcare professionals to better anticipate complications and tailor interventions effectively. The research supports future exploration into the integration of wearable technology in healthcare for proactive disease management.

Asset Subtitle

Hirotsugu Kenmotsu

Meta Tag

Speaker Hirotsugu Kenmotsu

Topic Local-Regional Non-small Cell Lung Cancer

Keywords

machine learning

clinical prediction model

interstitial lung diseases

radiation pneumonitis

non-small cell lung cancer

durvalumab

wearable devices

XGBoost

heart rate

oxygen saturation