With the onset of the COVID-19 pandemic, quantifying the condition of positively diagnosed patients is of paramount importance. Chest CT scans can be used to measure the severity of a lung infection and the isolate involvement sites in order to increase awareness of a patient's disease progression. In this work, we developed a deep learning framework for lung infection severity prediction. To this end, we collected a dataset of 232 chest CT scans and involved two public datasets with an additional 59 scans for our model's training and used two external test sets with 21 scans for evaluation. On an input chest Computer Tomography (CT) scan, our framework, in parallel, performs a lung lobe segmentation utilizing a pre-trained model and infection segmentation using three distinct trained SE-ResNet18 based U-Net models, one for each of the axial, coronal, and sagittal views. By having the lobe and infection segmentation masks, we calculate the infection severity percentage in each lobe and classify that percentage into 6 categories of infection severity score using a k-nearest neighbors (k-NN) model. The lobe segmentation model achieved a Dice Similarity Score (DSC) in the range of [0.918, 0.981] for different lung lobes and our infection segmentation models gained DSC scores of 0.7254 and 0.7105 on our two test sets, respectfully. Similarly, two resident radiologists were assigned the same infection segmentation tasks, for which they obtained a DSC score of 0.7281 and 0.6693 on the two test sets. At last, performance on infection severity score over the entire test datasets was calculated, for which the framework's resulted in a Mean Absolute Error (MAE) of 0.505 ± 0.029, while the resident radiologists' was 0.571 ± 0.039.

No abstract

Purpose The aim of this study was to design a predictive radiobiological model of normal brain tissue in low-grade glioma following radiotherapy based on imaging and molecular biomarkers. Methods Fifteen patients with primary brain tumors prospectively participated in this study and underwent radiation therapy. Magnetic resonance imaging (MRI) were obtained from the patients, including T1 and T2 weighted imaging and diffusion tensor imaging (DTI), and a generalized equivalent dose (gEUD) was calculated. The radiobiological model of the normal tissue complication probability (NTCP) was performed using the variables gEUD; axial diffusivity (AD) and radial diffusivity (RD) of the corpus callosum; and serum protein S100B by univariate and multivariate logistic regression. Results Changes in AD, RD, and S100B from baseline up to the six months after treatment had an increasing trend and were significant in some time points (P-Value < 0.05). The model resulting from RD changes in the six months after treatment was significantly more predictable of necrosis than other univariate models. The bivariate model combining RD changes in Gy40 dose-volume and gEUD, as well as the trivariate model obtained using gEUD, RD and S100B, had a higher predictive value among multivariate models at the sixth month of the treatment. Conclusion Changes in RD diffusion indices and in serum protein S100B value were used in the early delayed stage as reliable biomarkers for predicting late-delayed damage (necrosis) caused by radiation in the corpus callosum. Current findings could pave the way for intervention therapies to delay the severity of damage to white matter structures, minimize cognitive impairment, and improve the quality of life of patients with low-grade glioma.