No abstract

BackgroundCervical cancer is the fifth most common cancer among women, which is the third leading cause of cancer death in women worldwide. Brachytherapy is the most effective treatment for cervical cancer. For brachytherapy, computed tomography (CT) imaging is necessary since it conveys tissue density information which can be used for dose planning. However, the metal artifacts caused by brachytherapy applicators remain a challenge for the automatic processing of image data for image-guided procedures or accurate dose calculations. Therefore, developing an effective metal artifact reduction (MAR) algorithm in cervical CT images is of high demand.MethodsA novel residual learning method based on convolutional neural network (RL-ARCNN) is proposed to reduce metal artifacts in cervical CT images. For MAR, a dataset is generated by simulating various metal artifacts in the first step, which will be applied to train the CNN. This dataset includes artifact-insert, artifact-free, and artifact-residual images. Numerous image patches are extracted from the dataset for training on deep residual learning artifact reduction based on CNN (RL-ARCNN). Afterwards, the trained model can be used for MAR on cervical CT images.ResultsThe proposed method provides a good MAR result with a PSNR of 38.09 on the test set of simulated artifact images. The PSNR of residual learning (38.09) is higher than that of ordinary learning (37.79) which shows that CNN-based residual images achieve favorable artifact reduction. Moreover, for a 512 × 512 image, the average removal artifact time is less than 1 s.ConclusionsThe RL-ARCNN indicates that residual learning of CNN remarkably reduces metal artifacts and improves critical structure visualization and confidence of radiation oncologists in target delineation. Metal artifacts are eliminated efficiently free of sinogram data and complicated post-processing procedure.

Purpose Accurately segmenting organs‐at‐risk (OARs) is a key step in the effective planning of radiation therapy for nasopharyngeal carcinoma (NPC) treatment. In OAR segmentation of the head and neck computed tomography (CT), the low‐contrast and surrounding adhesion tissues of the parotids, thyroids, and optic nerves result in the difficulty in segmentation and lower accuracy of automatic segmentation for these organs than the other organs. In this paper, we propose a cascaded network structure to delineate these three OARs for NPC radiotherapy by combining deep learning and Boosting algorithm. Materials and methods The CT images of 140 NPC patients treated with radiotherapy were collected, and each of the three OAR annotations was respectively delineated by an experienced rater and reviewed by a professional radiologist (with 10 yr of experience). The datasets (140 patients) were divided into a training set (100 patients), a validation set (20 patients), and a test set (20 patients). From the Boosting method for combining multiple classifiers, three cascaded CNNs for segmentation were combined. The first network was trained with the traditional approach. The second one was trained on patterns (pixels) filtered by the first net. That is, the second machine recognized a mix of patterns (pixels), 50% of which was accurately identified by the first net. Finally, the third net was trained on the new patterns (pixels) screened jointly by the first and second networks. During the test, the outputs of the three nets were considered to obtain the final output. Dice similarity coefficient (DSC), 95th percentile of the Hausdorff distance (95% HD), and volume overlap error (VOE) were used to assess the method performance. Results The mean DSC (%) values were above 92.26 for the parotids, above 92.29 for the thyroids, and above 89.37 for the optic nerves. The mean 95% HDs (mm) were approximately 3.08 for the parotids, 2.64 for the thyroids, and 2.03 for the optic nerves. The mean VOE (%) values were approximately 14.16 for the parotids, 14.94 for the thyroids, and 19.07 for the optic nerves. Conclusions The proposed cascaded deep learning structure could achieve high performance compared with existing single‐network or other segmentation algorithms.

Although persistent hepatitis B virus (HBV) infection is associated with natural killer (NK) cell dysfunction, it remains obscure whether HBV viral antigens are responsible for NK cell dysfunction in patients with chronic hepatitis B (CHB) infection. In this study, we found that the percentage of NK cells expressing the inhibitory receptor, NKG2A, was increased in CHB patients, and NKG2A blockade restored NK cell function. Furthermore, in CHB patients, the frequency of NK cells expressing NKG2A positively correlated with the number of regulatory T cells (Tregs) and production of interleukin-10 (IL-10) in these Tregs. Moreover, exposure of peripheral blood mononuclear cells (PBMCs) isolated from healthy controls to sera from CHB patients resulted in increased proportion of NKG2A + NK cells; IL-10 blockade reduced the frequency of NKG2A + NK cells while increasing the percentage of IFN-γ + NK cells. In addition, stimulation of NK cells and Tregs from healthy controls with CHB sera together with anti-IL-10 antibody increased IFN-γ production in the culture supernatant. The frequencies of NKG2A + NK cells and IL-10 + Tregs, along with serum levels of alanine transferase and HBV DNA, were significantly increased in CHB patients positive for the Hepatitis B e antigen (HBeAg, a marker of viral replication) when compared to HBeAg-negative CHB patients. Importantly, exposure of PBMCs from healthy controls to HBeAg resulted in increased IL-10 production but reduced levels of TNF and IFN-γ, and IL-10 blockade rescued the generation of TNF and IFN-γ in this assay. The reduced production of TNF and IFN-γ was also observed in NK cells and Tregs from healthy controls that were stimulated with HBeAg, while IL-10 blockade increased the secretion of these two cytokines. We conclude that HBeAg induces IL-10 production in Tregs, thereby leading to increased expression of NKG2A on NK cells, which contributes to NK cell dysfunction during CHB infection. These data suggest that HBeAg is associated with NK cell dysfunction in CHB.

Background Treatment of pulmonary tuberculosis (TB) requires at least six months and is compromised by poor adherence. In the directly observed therapy (DOT) scheme recommended by the World Health Organization, the patient is directly observed taking their medications at a health post. An alternative to DOT is video-observed therapy (VOT), in which the patients take videos of themselves taking the medication and the video is uploaded into the app and reviewed by a health care worker. We developed a comprehensive TB management system by using VOT that is installed as an app on the smartphones of both patients and health care workers. It was implemented into the routine TB control program of the Nanshan District of Shenzhen, China. Objective The aim of this study was to compare the effectiveness of VOT with that of DOT in managing the treatment of patients with pulmonary TB and to evaluate the acceptance of VOT for TB management by patients and health care workers. Methods Patients beginning treatment between September 2017 and August 2018 were enrolled into the VOT group and their data were compared with the retrospective data of patients who began TB treatment and were managed with routine DOT between January 2016 and August 2017. Sociodemographic characteristics, clinical features, treatment adherence, positive findings of sputum smears, reporting of side effects, time and costs of transportation, and satisfaction were compared between the 2 treatment groups. The attitudes of the health care workers toward the VOT-based system were also analyzed. Results This study included 158 patients in the retrospective DOT group and 235 patients in the VOT group. The VOT group showed a significantly higher fraction of doses observed (P<.001), less missed observed doses (P<.001), and fewer treatment discontinuations (P<.05) than the DOT group. Over 79.1% (186/235) of the VOT patients had >85% of their doses observed, while only 16.4% (26/158) of the DOT patients had >85% of their doses observed. All patients were cured without recurrences. The VOT management required significantly (P<.001) less median patient time (300 minutes vs 1240 minutes, respectively) and transportation costs (¥53 [US $7.57] vs ¥276 [US $39.43], respectively; P<.001) than DOT. Significantly more patients (191/235, 81.3%) in the VOT group preferred their treatment method compared to those on DOT (37/131, 28.2%) (P<.001), and 92% (61/66) of the health care workers thought that the VOT method was more convenient than DOT for managing patients with TB. Conclusions Implementation of the VOT-based system into the routine program of TB management was simple and it significantly increased patient adherence to their drug regimens. Our study shows that a comprehensive VOT-based TB management represents a viable and improved evolution of DOT.

To better understand early brain growth patterns in health and disorder, it is critical to accurately segment infant brain magnetic resonance (MR) images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). Deep learning-based methods have achieved state-of-the-art performance; however, one of major limitations is that the learning-based methods may suffer from the multi-site issue, that is, the models trained on a dataset from one site may not be applicable to the datasets acquired from other sites with different imaging protocols/scanners. To promote methodological development in the community, iSeg-2019 challenge (http://iseg2019.web.unc.edu) provides a set of 6-month infant subjects from multiple sites with different protocols/scanners for the participating methods. Training/validation subjects are from UNC (MAP) and testing subjects are from UNC/UMN (BCP), Stanford University, and Emory University. By the time of writing, there are 30 automatic segmentation methods participating in iSeg-2019.We review the 8 top-ranked teams by detailing their pipelines/implementations, presenting experimental results and evaluating performance in terms of the whole brain, regions of interest, and gyral landmark curves. We also discuss their limitations and possible future directions for the multi-site issue. We hope that the multi-site dataset in iSeg-2019 and this review article will attract more researchers on the multi-site issue.

No abstract

In Traditional Chinese Medicine (TCM), most of the algorithms used to solve problems of syndrome diagnosis are superficial structure algorithms and not considering the cognitive perspective from the brain. However, in clinical practice, there is complex and nonlinear relationship between symptoms (signs) and syndrome. So we employed deep leaning and multilabel learning to construct the syndrome diagnostic model for chronic gastritis (CG) in TCM. The results showed that deep learning could improve the accuracy of syndrome recognition. Moreover, the studies will provide a reference for constructing syndrome diagnostic models and guide clinical practice.