This study aims to automatically diagnose thoracic diseases depicted on the chest x-ray (CXR) images using deep convolutional neural networks. The existing methods generally used the entire CXR images for training purposes, but this strategy may suffer from two drawbacks. First, potential misalignment or the existence of irrelevant objects in the entire CXR images may cause unnecessary noise and thus limit the network performance. Second, the relatively low image resolution caused by the resizing operation, which is a common preprocessing procedure for training neural networks, may lead to the loss of image details, making it difficult to detect pathologies with small lesion regions. To address these issues, we present a novel method termed as segmentation-based deep fusion network (SDFN), which leverages the higher-resolution information of local lung regions. Specifically, the local lung regions were identified and cropped by the Lung Region Generator (LRG). Two CNN-based classification models were then used as feature extractors to obtain the discriminative features of the entire CXR images and the cropped lung region images. Lastly, the obtained features were fused by the feature fusion module for disease classification. Evaluated by the NIH benchmark split on the Chest X-ray 14 Dataset, our experimental result demonstrated that the developed method achieved more accurate disease classification compared with the available approaches via the receiver operating characteristic (ROC) analyses. It was also found that the SDFN could localize the lesion regions more precisely as compared to the traditional method.

Background The novel coronavirus disease 2019 (COVID-19) is an emerging worldwide threat to public health. While chest computed tomography (CT) plays an indispensable role in its diagnosis, the quantification and localization of lesions cannot be accurately assessed manually. We employed deep learning-based software to aid in detection, localization and quantification of COVID-19 pneumonia. Methods A total of 2460 RT-PCR tested SARS-CoV-2-positive patients (1250 men and 1210 women; mean age, 57.7 ± 14.0 years (age range, 11-93 years) were retrospectively identified from Huoshenshan Hospital in Wuhan from February 11 to March 16, 2020. Basic clinical characteristics were reviewed. The uAI Intelligent Assistant Analysis System was used to assess the CT scans. Results CT scans of 2215 patients (90%) showed multiple lesions of which 36 (1%) and 50 patients (2%) had left and right lung infections, respectively (> 50% of each affected lung's volume), while 27 (1%) had total lung infection (> 50% of the total volume of both lungs). Overall, 298 (12%), 778 (32%) and 1300 (53%) patients exhibited pure ground glass opacities (GGOs), GGOs with sub-solid lesions and GGOs with both sub-solid and solid lesions, respectively. Moreover, 2305 (94%) and 71 (3%) patients presented primarily with GGOs and sub-solid lesions, respectively. Elderly patients (≥ 60 years) were more likely to exhibit sub-solid lesions. The generalized linear mixed model showed that the dorsal segment of the right lower lobe was the favoured site of COVID-19 pneumonia. Conclusion Chest CT combined with analysis by the uAI Intelligent Assistant Analysis System can accurately evaluate pneumonia in COVID-19 patients. Keywords 2019 novel coronavirus. Viral pneumonia. Artificial intelligence (AI). Computed tomography (CT). Ground glass opacity (GGO) Hai-tao Zhang, Jin-song Zhang and Hai-hua Zhang contributed equally to this work.

Smoke inhalation induced acute respiratory distress syndrome (ARDS) has become more and more common throughout the world and it is hard to improve the outcome. The present research was to investigate possible roles of angiotensin-converting enzyme (ACE) and ACE2 in lung injury resulted from smoke exposure. Rats were exposed to dense smoke to induce ARDS. Histological changes, blood gases, bronchoalveolar lavage fluids (BALF) and wet-to-dry weight were analyzed to evaluate lung injury after smoke inhalation; beside, we also measured the expression of ACE and ACE2 at different time points to explore the possible mechanism of those changes. The results showed that pH of arterial blood, partial blood oxygen (PaO₂) and blood oxygen saturation (SO₂) decreased after smoke inhalation at different time points (P<0.01); while, partial blood carbon dioxide (PaCO₂), wet-to-dry weight ratio, leukocytes count, protein concentration and inflammatory cytokines in BALF increased after smoke exposure (P<0.01). More importantly, both immunohistochemical staining and Western blot results showed that ACE and ACE2 expression in lungs from the experimental groups significantly increased compared with that of the control group (P<0.05). This study indicated that inflammation pulmonary edema and histological changes resulted from smoke inhalation induced lung injury were possibly attributed to abnormal expression of ACE and ACE2 related pathway.

No abstract

Resveratrol is a plant-derived natural compound which possesses potential anticancer properties. However, there are scarce reports on its anticancer effects in non-small cell lung cancer and its auxiliary function on the anticancer effects of cisplatin. In the present study, we investigated the effects of resveratrol on the cell viability and apoptosis in human non-small cell lung cancer H838 and H520 cell lines. It has been found that resveratrol inhibited the proliferation of H838 and H520 cells in a dose- and time-dependent manner, and apoptosis was increased in cells treated with resveratrol which was associated with the depolarization of mitochondrial membrane potential, release of cytochrome c from mitochondria to cytosol, and abnormal expression of Bcl-2 and Bax proteins. Above all, resveratrol enhanced the effects of cisplatin on inhibition of cancer cell proliferation, induction of cell apoptosis, depolarization of mitochondrial membrane potential, release of cytochrome c and regulation on expression of Bcl-2 and Bax. Results from the present study demonstrated that resveratrol exhibited its anticancer effects on non-small cell lung cancer H838 and H520 cell lines, and enhanced the antitumor effects of cisplatin by regulating the mitochondrial apoptotic pathway. These results have put forward the rationale for further basic research and preclinical investigation on the anticancer effects of resveratrol against human non-small cell lung cancer.