RESUMEN
Radiation encephalopathy (RE) refers to radiation-induced brain necrosis and is a life-threatening complication in patients with nasopharyngeal carcinoma (NPC) after radiotherapy (RT), and radiation-induced pre-symptomatic glymphatic alterations have not yet been investigated. We used diffusion tensor image analysis along the perivascular space (DTI-ALPS) index to examine the pre-symptomatic glymphatic alterations in NPC patients following RT. A total of 109 patients with NPC consisted of Pre-RT (n = 35) and Post-RT (n = 74) cohorts were included. The post-RT NPC patients, with normal-appearing brain structure at the time of MRI, were further divided into Post-RT-RE- (n = 58) and Post-RT-RE+ (n = 16) subgroups based on the detection of RE in follow-up. We observed lower DTI-ALPS left index, DTI-ALPS right index and DTI-ALPS whole brain index in post-RT patients than that in pre-RT patients (p < 0.05). We further found that post-RT-RE+ patients demonstrated significantly lower DTI-ALPS right (p = 0.013), DTI-ALPS whole brain (p = 0.011) and marginally lower DTI-ALPS left (p = 0.07) than Post-RT non-RE patients. Significant negative correlations were observed between the maximum dosage of radiation-treatment (MDRT) and DTI-ALPS left index (p = 0.003) as well as DTI-ALPS whole brain index (p = 0.004). Receiver operating characteristic (ROC) curve analysis showed that DTI-ALPS whole brain index exhibited good performance (AUC = 0.706) in identifying patients more likely developing RE. We concluded that glympathic function was impaired in NPC patients following RT and DTI-ALPS index may serve as a novel imaging biomarker for diagnosis of RE.
RESUMEN
Depth estimation, as a necessary clue to convert 2D images into the 3D space, has been applied in many machine vision areas. However, to achieve an entire surrounding 360° geometric sensing, traditional stereo matching algorithms for depth estimation are limited due to large noise, low accuracy, and strict requirements for multi-camera calibration. In this work, for a unified surrounding perception, we introduce panoramic images to obtain a larger field of view. We extend PADENet [IEEE 23rd International Conference on Intelligent Transportation Systems, (2020), pp. 1-610.1109/ITSC45102.2020.9294206], which first appeared in our previous conference work for outdoor scene understanding, to perform panoramic monocular depth estimation with a focus for indoor scenes. At the same time, we improve the training process of the neural network adapted to the characteristics of panoramic images. In addition, we fuse the traditional stereo matching algorithm with deep learning methods and further improve the accuracy of depth predictions. With a comprehensive variety of experiments, this research demonstrates the effectiveness of our schemes aiming for indoor scene perception.
