Search

We aimed to build radiomic models for classifying non-small cell lung cancer (NSCLC) histopathological subtypes through a dual-centre dataset and comprehensively evaluate the effect of ComBat harmonisation on the performance of single- and multimodality radiomic models. A public dataset of NSCLC patients from two independent centres was used. Two image fusion methods, namely guided filtering-based fusion and image fusion based on visual saliency map and weighted least square optimisation, were used. Radiomic features were extracted from each scan, including first-order, texture and moment-invariant features. Subsequently, ComBat harmonisation was applied to the extracted features from computed tomography (CT), positron emission tomography (PET) and fused images to correct the centre effect. For feature selection, least absolute shrinkage and selection operator (Lasso) and recursive feature elimination (RFE) were investigated. For machine learning, logistic regression (LR), support vector machine (SVM) and AdaBoost were evaluated for classifying NSCLC subtypes. Training and evaluation of the models were carried out in a robust framework to offset plausible errors and performance was reported using area under the curve, balanced accuracy, sensitivity and specificity before and after harmonisation. N-way ANOVA was used to assess the effect of different factors on the performance of the models. Support vector machine fed with selected features by recursive feature elimination from a harmonised PET feature set achieved the highest performance (area under the curve = 0.82) in classifying NSCLC histopathological subtypes. Although the performance of the models did not significantly improve for CT images after harmonisation, the performance of PET and guided filtering-based fusion feature signatures significantly improved for almost all models. Although the selection of the image modality and feature selection methods was effective on the performance of the model (ANOVA P -values <0.001), machine learning and harmonisation did not change the performance significantly (ANOVA P -values = 0.839 and 0.292, respectively). This study confirmed the potential of radiomic analysis on PET, CT and hybrid images for histopathological classification of NSCLC subtypes. • Radiomic models built for the classification of NSCLC histopathological subtypes on PET, CT and hybrid images using a dual-centre dataset. • The effect of image harmonisation on the performance of single- and multimodality radiomic models was evaluated. • The effect of harmonisation varied for the different classifiers and feature selection algorithms. [ABSTRACT FROM AUTHOR]

Breast cancer is the leading cause of cancer death in worldwide. In order to enhance the chances of a favorable outcome for the survivors, it is essential to carefully balance the potential advantages of treatment methods with the risks of harm to healthy tissues, including the heart. There is currently a lack of comprehensive, data-driven evidence on effective risk stratification strategies. The aim of this study is to investigate whether prediction of cardiotoxicity could improve with machine learning methods using radiomics imaging and clinical features along with dosimetric parameters. Fifty left side breast cancer patients without history of cardiac disease were participated in this study. 2D and 3D echocardiography before and after 6 month of treatment completion was performed to evaluate cardiac toxicity. Cardiac dose volume histograms (DVH), demographic data, echocardiographic parameters and echocardiograph's radiomics features were collected for all patients. Toxicity modelling performed with feature selection methods and five classifiers (K- Nearest Neighbor, Decision Tree, Neural Network, Random Forrest, and Support Vector Machine (SVM)) in 4 separated groups (DVH, DVH + Demographic, DVH + Demographic+ Clinical, DVH + Demographic + Clinical + ultrasound imaging). Prediction performance of models was validated by 5-fold cross validation and evaluated by AUCs. Fifty-eight percent of papulation showed cardiotoxicity after 6 months of treatment. Mean left ventricular Ejection Fraction and Global Longitudinal Strain were decreased significantly after treatment (P-Value < 0.001). After feature selection and prediction modelling using SVM, the DVH, DVH + Demographic, DVH + Demographic + Clinical, DVH + Demographic + Clinical + Radiomics models showed prediction performance (AUC) up to 70%, 75%, 85% and 95% respectively. Incorporation of patient's data, cardiac and dose descriptors combined with machine learning algorithms is beneficial for cardiac toxicity prediction after chemo-radiotherapy. [ABSTRACT FROM AUTHOR]