Combining computed tomography and biologically effective dose in radiomics and deep learning improves prediction of tumor response to robotic lung stereotactic body radiation therapy

Abstract

Purpose The aim of this study is to improve the performance of machine learning (ML) models in predicting response of non‐small cell lung cancer (NSCLC) to stereotactic body radiation therapy (SBRT) by integrating image features from pre‐treatment computed tomography (CT) with features from the biologically effective dose (BED) distribution. Materials and methods Image features, consisting of crafted radiomic features or machine‐learned features extracted using a convolutional neural network, were calculated from pre‐treatment CT data and from dose distributions converted into BED for 80 NSCLC lesions over 76 patients treated with robotic guided SBRT. ML models using different combinations of features were trained to predict complete or partial response according to response criteria in solid tumors, including radiomics CT (RadCT), radiomics CT and BED (RadCT,BED), deep learning (DL) CT (DLCT), and DL CT and BED (DLCT,BED). Training of ML included feature selection by neighborhood component analysis followed by ensemble ML using robust boosting. A model was considered as acceptable when the sum of average sensitivity and specificity on test data in repeated cross validations was at least 1.5. Results Complete or partial response occurred in 58 out of 80 lesions. The best models to predict the tumor response were those using BED variables, achieving significantly better area under curve (AUC) and accuracy than those using only features from CT, including a RadCT,BED model using three radiomic features from BED, which scored an accuracy of 0.799 (95% confidence intervals (0.75–0.85)) and AUC of 0.773 (0.688–0.846), and a DLCT,BED model also using three variables with an accuracy of 0.798 (0.649–0.829) and AUC of 0.812 (0.755–0.867). Conclusion According to our results, the inclusion of BED features improves the response prediction of ML models for lung cancer patients undergoing SBRT, regardless of the use of radiomic or DL features.

A Combined Model to Improve the Prediction of Local Control for Lung Cancer Patients Undergoing Stereotactic Body Radiotherapy Based on Radiomic Signature Plus Clinical and Dosimetric Parameters

“…It failed to achieve statistical significance due to the small sample size, subtle differences in the data set or the mixed effect of other parameters. Similar results were obtained by Avanzo et al who have demonstrated that combining BED features and image features in radiomics and deep learning improves the tumor response prediction of machine learning models for lung SBRT (42). This trend is in agreement with past studies, showing it is highly valuable to predict tumor local control in lung SBRT using multivariate factors (43).…”

Based on the Development and Verification of a Risk Stratification Nomogram: Predicting the Risk of Lung Cancer-Specific Mortality in Stage IIIA-N2 Unresectable Large Cell Lung Neuroendocrine Cancer Compared With Lung Squamous Cell Cancer and Lung Adenocarcinoma

“…Furthermore, a 2021 study by Avanzo et al. found that the application of radiomics is of great value in improving lung stereotactic body radiation therapy ( 30 ). In clinical practice, radiomics is an emerging field of research, and it is used as a predictive tool for responses and treatment outcomes.…”

“…Therefore, in addition to chemotherapy and radiotherapy, targeted therapy, immunotherapy and combination therapy may provide more personalized and specialized options to improve patient survival. Furthermore, a 2021 study by Avanzo et al found that the application of radiomics is of great value in improving lung stereotactic body radiation therapy(30). In clinical practice, radiomics is an emerging field of research, and it is used as a predictive tool for responses and treatment outcomes.…”

Imbalanced Data Correction Based PET/CT Radiomics Model for Predicting Lymph Node Metastasis in Clinical Stage T1 Lung Adenocarcinoma

“…To alleviate this negative effect, we use several re-sampling techniques to balance the distribution in training cohort for model construction. The re-sampling techniques should be applied only to training data to avoid data leakage, while the validation cohort was untouched as it represents the real situation in clinical practice (24).Ten re-sampling techniques worked in our study: random over-sampling (ROS), Adaptive Synthetic (ADASYN), Synthetic Minority Oversampling Technique (SMOTE), Borderline SMOTE (bSMOTE), Random under-sampling (RUS), NearMiss (NM), Tomek links (TL), Edited Nearest Neighbours (ENN), Over-sampling using SMOTE and cleaning using Tomek links (SMOTE-TL) and Over-sampling using SMOTE and cleaning using ENN (SMOTE-ENN) (25)(26)(27). The re-sampling methods are described in the Supplementary Method S7.…”