Search

Hepatic echinococcosis (HE) is a life-threatening liver disease caused by parasites that requires a precise diagnosis and proper treatments. To assess HE lesions accurately, we propose a novel automatic HE lesion segmentation and classification network that contains lesion region positioning (LRP) and lesion region segmenting (LRS) modules. First, we used the LRP module to obtain the probability map of the lesion distribution and the position of the lesion. Then, based on the result of the LRP module, we used the LRS module to precisely segment the HE lesions within the high-probability region. Finally, we classified the HE lesions and identified the lesion types by a convolutional neural network (CNN). The entire dataset was delineated by the hospital's senior radiologist. We collected CT slices of 160 patients from Qinghai Provincial People's Hospital. The Dice score of the final segmentation result reached 89.89%. The Dice scores, indicating the classification accuracy, for cystic vs. alveolar echinococcosis and calcified vs. noncalcified lesions were 80.32% and 82.45%, the sensitivities were 72.41% and 75.17%, the specificities were 83.72% and 86.04%, the NPVs were 80.01% and 86.96%, the PPVs were 80.45% and 81.74%, and the areas under the ROC curves were 0.8128 and 0.8205, respectively. Graphical abstract.

Knee arthroscopy suffers from a lack of depth information and easy occlusion of the visual field. To solve these limitations, we propose an arthroscopic navigation system based on self-positioning technology, with the guidance of virtual-vision views. This system can work without any external tracking devices or added markers, thus increasing the working range and improving the robustness of the rotating operation. The fly-through view and global positioning view for surgical guidance are rendered through virtual-vision rendering in real time. The fly-through view provides surgeons with navigating the arthroscope in the internal anatomical structures using a virtual camera perspective. The global positioning view shows the posture of the arthroscope relative to the preoperative model in a transparent manner. The posture of the arthroscope is estimated from the fusion of visual and inertial data based on the visual–inertial stereo slam. A flexible calibration method that transforms the posture of the arthroscope in the physical world into the virtual-vision rendering framework is proposed for the arthroscopic navigation system with self-positioning information. Quantitative experiments for evaluating self-positioning accuracy were performed. For translation, the acquired mean error was 0.41 ± 0.28 mm; for rotation, it was 0.11° ± 0.07°. The tracking range of the proposed system was approximately 1.4 times that of the traditional external optical tracking system for the rotating operation. Simulated surgical operations were performed on the phantom. The fly-through and global positing views were paired with original arthroscopic images for intuitive surgical guidance. The proposed system provides surgeons with both fly-through and global positioning views without a dependence on the traditional external tracking systems for surgical guidance. The feasibility and robustness of the system are evaluated, and it shows promise for medical applications.