2022 World Conference on Lung Cancer (ePosters)-EP05.01-033. Stimulation CT-Based Radiomics Predict Radiation Pneumonitis after Chemoradiotherapy in Locally Advanced NSCLC

EP05.01-033. Stimulation CT-Based Radiomics Predict Radiation Pneumonitis after Chemoradiotherapy in Locally Advanced NSCLC

Back to course

Pdf Summary

This study aimed to develop a machine learning model using radiomics features from CT images to predict radiation pneumonitis (RP) in patients with non-small cell lung cancer (NSCLC). The study included 134 NSCLC patients, of which 24 had RP grade 2. The radiomics model, which included 20 radiomics features, achieved an area under the curve (AUC) of 0.94 in the training cohort and an AUC of 0.90 in the validation cohort. In comparison, the dosimetric model had limited predictive power with an AUC of 0.70. Calibration curves and decision curve analysis indicated favorable consistency and satisfactory clinical utility of the radiomics model. Patient characteristics, including age, gender, smoking status, pathology, chemotherapy, and dose metrics, did not show significant differences between RP2 and non-RP2 groups. The authors concluded that the radiomics model outperformed traditional dosimetry models in predicting RP2 and suggested that external validation is needed for future research.

Asset Subtitle

Haihua Yang

Meta Tag

Speaker Haihua Yang

Topic Locally Advanced Non-small Cell Lung Cancer - Chemoradiotherapy and Radiotherapy

Keywords

machine learning model

radiomics features

CT images

radiation pneumonitis

non-small cell lung cancer

NSCLC

area under the curve

dosimetric model

calibration curves

decision curve analysis