Deep Neural Networks Offer Morphologic Classification and Diagnosis of Bacterial Vaginosis.

Bacterial vaginosis (BV) is caused by the excessive and imbalanced growth of bacteria in vagina, affecting 30 to 50% of women. Gram staining followed by Nugent scoring based on bacterial morphotypes under the microscope is considered the gold standard for BV diagnosis; this method is often labor-intensive and time-consuming, and results vary from person to person. We developed and optimized a convolutional neural network (CNN) model and evaluated its ability to automatically identify and classify three categories of Nugent scores from microscope images. The CNN model was first established with a panel of microscopic images with Nugent scores determined by experts. The model was trained by minimizing the cross-entropy loss function and optimized by using a momentum optimizer. The separate test sets of images collected from three hospitals were evaluated by the CNN model. The CNN model consisted of 25 convolutional layers, 2 pooling layers, and a fully connected layer. The model obtained 82.4% sensitivity and 96.6% specificity with the 5,815 validation images when altered vaginal flora and BV were considered the positive samples, which was better than the rates achieved by top-level technologists and obstetricians in China. The capability of our model for generalization was so strong that it exhibited 75.1% accuracy in three categories of Nugent scores on the independent test set of 1,082 images, which was 6.6% higher than the average of three technologists, who are hold bachelor's degrees in medicine and are qualified to make diagnostic decisions. When three technologists ran one specimen in triplicate, the precision of three categories of Nugent scores was 54.0%. One hundred three samples diagnosed by two technologists on different days showed a repeatability of 90.3%. The CNN model outperformed human health care practitioners in terms of accuracy and stability for three categories of Nugent score diagnosis. The deep learning model may offer translational applications in automating diagnosis of bacterial vaginosis with proper supporting hardware.

Development and evaluation of a monoclonal antibody-based blocking ELISA to detect antibodies against the E2 protein of bovine viral diarrhea virus-1.

With the rapid development of cattle industry, bovine viral diarrhea virus (BVDV) is becoming widespread in China, which causes serious economic losses to the industry. Effective vaccination and viral surveillance are critical for the prevent and control of BVDV infection. In the present study, the immunogenic domain of E2 protein of BVDV-1 was expressed by prokaryotic pET-28a vector. Monoclonal antibodies (mAbs) against E2 protein were prepared and systemically examined by western blot, immunofluorescence assay, blocking ELISA (bELISA) and virus neutralization test (VNT). The mAb 1E2B3, which showed good reactivity and neutralizing activity to BVDV-1 strains, was selected for ELISA establishment. After a series of screening and optimization, a novel bELISA for highly sensitive and specific detection of BVDV-1 antibodies was established, using HRP-labeled 1E2B3 and recombinant E2 protein. ROC analysis of 91 positive and 84 negative reference bovine serum samples yielded the area under the curve (AUC) of 0.9903. A diagnostic specificity of 96.43 % and a sensitivity of 95.6 % were achieved when the cutoff value was set at 24.31 %. There was no cross reaction to the positive sera of classical swine fever virus (CSFV), BVDV-2, border disease virus (BDV), bovine parainfluenza virus type 3 (BPIV3), infectious bovine rhinotracheitis virus (IBRV), foot-and-mouth disease virus (FMDV), Mycoplasma bovis (M.bovis) and Brucella. The total agreement rate of bELISA with VNT was 93.96 % (249/265). In addition, the result of bELISA was positively correlated with neutralizing antibody titer, and the bELISA could well distinguish the serum samples before and after BVDV vaccination. These results indicate that the established bELISA in this study is specific, sensitive, simple and convenient, which provides technical support for the vaccine efficacy evaluation, prevention and control of BVD in the future.