Imaging Radiomic Biomarkers of Mandibular Osteoradionecrosis for Head and Neck Cancer

Purpose/Objective(s) This study aims to identify radiomic features extracted from contrast-enhanced CT scans that differentiate osteoradionecrosis (ORN) from normal mandibular bone in head and neck cancer patients treated with radiotherapy. Materials/Methods Contrast-enhanced CT images were collected for patients with confirmed ORN diagnosis at MD Anderson Cancer Center between 2008 and 2018. The ORN regions of interest (ROIs) were segmented manually in each image. The control ROIs of the contralateral health mandible were generated by a Python script then adjusted manually in each image. Commercial software was then used to extract the radiomic features from both ORN and control ROIs after the application of intrinsic filters. The pairwise correlation filter was used to remove radiomic features whose pairwise correlation was ≥0.99. Filter algorithms were then used to further reduce the number of radiomic features. After that, wrapper and embedded methods were applied on the resulting radiomic features. Finally, Gini importance and Recursive Feature Elimination (RFE) were used to select the final radiomic features for the predictive model. The support vector machine (SVM) with linear kernel was used for the binary classification of ORN and normal mandibular bone. The performance of the model was evaluated using the Area Under Curve (AUC). Results A total of 150 patients with radiologically established ORN were included in our study. The mean age was 62.3 years (range 27-82). The mean duration between the end of RT and ORN diagnosis was 32.6 months. The pairwise correlation omitted 432 features with a correlation ≥ 0.99. After that, the first step of the radiomic features engineering (using the filter algorithm) resulted in the selection of 33 radiomic features with statistically significant results in all the following three statistical methods: Pearson correlation, Chi-square test, and F-score. The RFE based on the Gini index selected 5 radiomics features. The final classifier used SVM with linear Kernel. The input for this classifier was the final set of radiomic features (N=5). We validated this binary classification model using 5-fold cross-validation. During this validation, the range of AUC was (0.84–0.95) & the average AUC was 0.90. Conclusion We successfully used imaging radiomic features to construct an accurate model (AUC= 0.90) to discriminate ORN versus normal mandibular bone in head and neck cancer patients. Future studies are needed to validate this model in prospective studies to early detect ORN in head and neck cancer patients after radiation treatment. This study aims to identify radiomic features extracted from contrast-enhanced CT scans that differentiate osteoradionecrosis (ORN) from normal mandibular bone in head and neck cancer patients treated with radiotherapy. Contrast-enhanced CT images were collected for patients with confirmed ORN diagnosis at MD Anderson Cancer Center between 2008 and 2018. The ORN regions of interest (ROIs) were segmented manually in each image. The control ROIs of the contralateral health mandible were generated by a Python script then adjusted manually in each image. Commercial software was then used to extract the radiomic features from both ORN and control ROIs after the application of intrinsic filters. The pairwise correlation filter was used to remove radiomic features whose pairwise correlation was ≥0.99. Filter algorithms were then used to further reduce the number of radiomic features. After that, wrapper and embedded methods were applied on the resulting radiomic features. Finally, Gini importance and Recursive Feature Elimination (RFE) were used to select the final radiomic features for the predictive model. The support vector machine (SVM) with linear kernel was used for the binary classification of ORN and normal mandibular bone. The performance of the model was evaluated using the Area Under Curve (AUC). A total of 150 patients with radiologically established ORN were included in our study. The mean age was 62.3 years (range 27-82). The mean duration between the end of RT and ORN diagnosis was 32.6 months. The pairwise correlation omitted 432 features with a correlation ≥ 0.99. After that, the first step of the radiomic features engineering (using the filter algorithm) resulted in the selection of 33 radiomic features with statistically significant results in all the following three statistical methods: Pearson correlation, Chi-square test, and F-score. The RFE based on the Gini index selected 5 radiomics features. The final classifier used SVM with linear Kernel. The input for this classifier was the final set of radiomic features (N=5). We validated this binary classification model using 5-fold cross-validation. During this validation, the range of AUC was (0.84–0.95) & the average AUC was 0.90. We successfully used imaging radiomic features to construct an accurate model (AUC= 0.90) to discriminate ORN versus normal mandibular bone in head and neck cancer patients. Future studies are needed to validate this model in prospective studies to early detect ORN in head and neck cancer patients after radiation treatment.