This paper performs a comprehensive study on the deep-learning-based computer-aided diagnosis (CADx) for the differential diagnosis of benign and malignant nodules/lesions by avoiding the potential errors caused by inaccurate image processing results (e.g., boundary segmentation), as well as the classification bias resulting from a less robust feature set, as involved in most conventional CADx algorithms. Specifically, the stacked denoising auto-encoder (SDAE) is exploited on the two CADx applications for the differentiation of breast ultrasound lesions and lung CT nodules. The SDAE architecture is well equipped with the automatic feature exploration mechanism and noise tolerance advantage, and hence may be suitable to deal with the intrinsically noisy property of medical image data from various imaging modalities. To show the outperformance of SDAE-based CADx over the conventional scheme, two latest conventional CADx algorithms are implemented for comparison. 10 times of 10-fold cross-validations are conducted to illustrate the efficacy of the SDAE-based CADx algorithm. The experimental results show the significant performance boost by the SDAE-based CADx algorithm over the two conventional methods, suggesting that deep learning techniques can potentially change the design paradigm of the CADx systems without the need of explicit design and selection of problem-oriented features.

Automatic and accurate whole-heart and great vessel segmentation from 3D cardiac magnetic resonance (MR) images plays an important role in the computer-assisted diagnosis and treatment of cardiovascular disease. However, this task is very challenging due to ambiguous cardiac borders and large anatomical variations among different subjects. In this paper, we propose a novel densely-connected volumetric convolutional neural network, referred as DenseVoxNet, to automatically segment the cardiac and vascular structures from 3D cardiac MR images. The DenseVoxNet adopts the 3D fully convolutional architecture for effective volume-to-volume prediction. From the learning perspective, our DenseVoxNet has three compelling advantages. First, it preserves the maximum information flow between layers by a densely-connected mechanism and hence eases the network training. Second, it avoids learning redundant feature maps by encouraging feature reuse and hence requires fewer parameters to achieve high performance, which is essential for medical applications with limited training data. Third, we add auxiliary side paths to strengthen the gradient propagation and stabilize the learning process. We demonstrate the effectiveness of DenseVoxNet by comparing it with the state-of-the-art approaches from HVSMR 2016 challenge in conjunction with MICCAI, and our network achieves the best dice coefficient. We also show that our network can achieve better performance than other 3D ConvNets but with fewer parameters.

BackgroundConnectivity map data and associated methodologies have become a valuable tool in understanding drug mechanism of action (MOA) and discovering new indications for drugs. One of the key ideas of connectivity map (CMAP) is to measure the connectivity between disease gene expression signatures and compound-induced gene expression profiles. Despite multiple impressive anecdotal validations, only a few systematic evaluations have assessed the accuracy of this aspect of CMAP, and most of these utilize drug-to-drug matching to transfer indications across the two drugs.MethodsTo assess CMAP methodologies in a more direct setting, namely the power of classifying known drug-disease relationships, we evaluated three CMAP-based methods on their prediction performance against a curated dataset of 890 true drug-indication pairs. The disease signatures were generated using Gene Logic BioExpress™ system and the compound profiles were derived from the Connectivity Map database (CMAP, build 02, http://www.broadinstitute.org/CMAP/).ResultsThe similarity scoring algorithm called eXtreme Sum (XSum) performs better than the standard Kolmogorov-Smirnov (KS) statistic in terms of the area under curve and can achieve a four-fold enrichment at 0.01 false positive rate level, with AUC = 2.2E-4, P value = 0.0035.ConclusionConnectivity map can significantly enrich true positive drug-indication pairs given an effective matching algorithm.Electronic supplementary materialThe online version of this article (doi:10.1186/s13073-014-0095-1) contains supplementary material, which is available to authorized users.

Background Coronavirus disease 2019 (COVID-19) is an emerging infectious disease that first appeared in Wuhan, China, and quickly spread throughout the world. We aimed to understand the relationship between diabetes mellitus and the prognosis of COVID-19. Methods Demographic, clinical, laboratory, radiologic, treatments, complications, and clinical outcomes data were extracted from electronic medical records and compared between diabetes (n = 84) and nondiabetes (n = 500) groups. Kaplan-Meier method and multivariate Cox analysis were applied to determine the risk factors for the prognosis of COVID-19. Results Compared with nondiabetic patients, diabetic patients had higher levels of neutrophils ( P = .014), C-reactive protein ( P = .008), procalcitonin ( P < .01), and D-dimer ( P = .033), and lower levels of lymphocytes ( P = .032) and albumin ( P = .035). Furthermore, diabetic patients had a significantly higher incidence of bilateral pneumonia (86.9%, P = .020). In terms of complications and clinical outcomes, the incidence of respiratory failure (36.9% vs 24.2%, P = .022), acute cardiac injury (47.4% vs 21.2%, P < .01), and death (20.2% vs 8.0%, P = .001) in the diabetes group was significantly higher than that in the nondiabetes group. Kaplan-Meier survival curve showed that COVID-19 patients with diabetes had a shorter overall survival time. Multivariate Cox analysis indicated that diabetes (hazard ratio 2.180, P = .031) was an independent risk factor for COVID-19 prognosis. In subgroup analysis, we divided diabetic patients into insulin-required and non-insulin-required groups according to whether they needed insulin, and found that diabetic patients requiring insulin may have a higher risk of disease progression and worse prognosis after the infection of severe acute respiratory syndrome coronavirus 2. Conclusions Diabetes is an independent risk factor for the prognosis of COVID-19. More attention should be paid to the prevention and treatment for diabetic patients, especially those who require insulin therapy.

Connectivity map data and associated methodologies have become a valuable tool in understanding drug mechanism of action (MOA) and discovering new indications for drugs. However, few systematic evaluations have been done to assess the accuracy of these methodologies. One of the difficulties has been the lack of benchmarking data sets. Iskar et al. (PLoS. Comput. Biol. 6, 2010) predicted the Anatomical Therapeutic Chemical (ATC) drug classification based on drug-induced gene expression profile similarity (DIPS), and quantified the accuracy of their method by computing the area under the curve (AUC) of the Receiver Operating Characteristic (ROC) curve. We adopt the same data and extend the methodology, by using a simpler eXtreme cosine (XCos) method, and find it does better in this limited setting than the Kolmogorov-Smirnov (KS) statistic. In fact, for partial AUC (a more relevant statistic for actual application to repositioning) XCos does 17% better than the DIPS method (p=1.2e-7). We also observe that smaller gene signatures (with 100 probes) do better than larger ones (with 500 probes), and that DMSO controls from within the same batch obviate the need for mean centering. As expected there is heterogeneity in the prediction accuracy amongst the various ATC codes. We find that good transcriptional response to drug treatment appears necessary but not sufficient to achieve high AUCs. Certain ATC codes, such as those corresponding to corticosteroids, had much higher AUCs possibly due to strong transcriptional responses and consistency in MOA.