Predicting visual acuity with machine learning in treated ocular trauma patients.

AIM: To predict best-corrected visual acuity (BCVA) by machine learning in patients with ocular trauma who were treated for at least 6mo. METHODS: The internal dataset consisted of 850 patients with 1589 eyes and an average age of 44.29y. The initial visual acuity was 0.99 logMAR. The test dataset consisted of 60 patients with 100 eyes collected while the model was optimized. Four different machine-learning algorithms (Extreme Gradient Boosting, support vector regression, Bayesian ridge, and random forest regressor) were used to predict BCVA, and four algorithms (Extreme Gradient Boosting, support vector machine, logistic regression, and random forest classifier) were used to classify BCVA in patients with ocular trauma after treatment for 6mo or longer. Clinical features were obtained from outpatient records, and ocular parameters were extracted from optical coherence tomography images and fundus photographs. These features were put into different machine-learning models, and the obtained predicted values were compared with the actual BCVA values. The best-performing model and the best variable selected were further evaluated in the test dataset. RESULTS: There was a significant correlation between the predicted and actual values [all Pearson correlation coefficient (PCC)>0.6]. Considering only the data from the traumatic group (group A) into account, the lowest mean absolute error (MAE) and root mean square error (RMSE) were 0.30 and 0.40 logMAR, respectively. In the traumatic and healthy groups (group B), the lowest MAE and RMSE were 0.20 and 0.33 logMAR, respectively. The sensitivity was always higher than the specificity in group A, in contrast to the results in group B. The classification accuracy and precision were above 0.80 in both groups. The MAE, RMSE, and PCC of the test dataset were 0.20, 0.29, and 0.96, respectively. The sensitivity, precision, specificity, and accuracy of the test dataset were 0.83, 0.92, 0.95, and 0.90, respectively. CONCLUSION: Predicting BCVA using machine-learning models in patients with treated ocular trauma is accurate and helpful in the identification of visual dysfunction.

Construction and Characterization of a Recombinant Human Respiratory Syncytial Virus Encoding Enhanced Green Fluorescence Protein for Antiviral Drug Screening Assay.

Human respiratory syncytial virus (RSV) is the single most important cause of lower respiratory tract disease in infants and young children and a major viral agent responsible for respiratory tract disease in immunosuppressed individuals and the elderly, but no vaccines and antiviral drugs are available. Herein the recombinant RSV (rRSV) encoding enhanced green fluorescence protein (EGFP, rRSV-EGFP) was constructed and the potential for screening anti-RSV drugs was investigated. The recombinant plasmid of pBRATm-rRSV-EGFP, containing T7 transcription cassette composed of T7 promoter, RSV antigenomic cDNA with EGFP gene, HDV ribozyme (δ), and T7 terminator in the order of 5' to 3', was constructed and cotransfected into BHK/T7-9 cells together with helper plasmids encoding N, P, L, and M2-1 gene, respectively. The rescued rRSV-EGFP was confirmed by increasing expression of EGFP over blind passages and by RT-PCR. rRSV-EGFP was comparable to the other two recombinant RSVs encoding red fluorescent protein (RFP, rRSV-RFP) or luciferase (Luc, rRSV-Luc) in the growth kinetic, and there was a difference in sensitivity between them for screening anti-RSV agents based on infection of HEp-2 cells. The EGFP-encoding rRSV has been constructed and rescued successfully and has the potential for high-throughput anti-RSV drug screening in vitro.

A single intranasal administration of virus-like particle vaccine induces an efficient protection for mice against human respiratory syncytial virus.

Human respiratory syncytial virus (RSV) is an important pediatric pathogen causing acute viral respiratory disease in infants and young children. However, no licensed vaccines are currently available. Virus-like particles (VLPs) may bring new hope to producing RSV VLP vaccine with high immunogenicity and safety. Here, we constructed the recombinants of matrix protein (M) and fusion glycoprotein (F) of RSV, respectively into a replication-deficient first-generation adenoviral vector (FGAd), which were used to co-infect Vero cells to assemble RSV VLPs successfully. The resulting VLPs showed similar immunoreactivity and function to RSV virion in vitro. Moreover, Th1 polarized response, and effective mucosal virus-neutralizing antibody and CD8+ T-cell responses were induced by a single intranasal (i.n.) administration of RSV VLPs rather than intramuscular (i.m.) inoculation, although the comparable RSV F-specific serum IgG and long-lasting RSV-specific neutralizing antibody were detected in the mice immunized by both routes. Upon RSV challenge, VLP-immunized mice showed increased viral clearance but decreased signs of enhanced lung pathology and fewer eosinophils compared to mice immunized with formalin-inactivated RSV (FI-RSV). In addition, a single i.n. RSV VLP vaccine has the capability to induce RSV-specific long-lasting neutralizing antibody responses observable up to 15 months. Our results demonstrate that the long-term and memory immune responses in mice against RSV were induced by a single i.n. administration of RSV VLP vaccine, suggesting a successful approach of RSV VLPs as an effective and safe mucosal vaccine against RSV infection, and an applicable and qualified platform of FGAd-infected Vero cells for VLP production.

Sublingual administration of a helper-dependent adenoviral vector expressing the codon-optimized soluble fusion glycoprotein of human respiratory syncytial virus elicits protective immunity in mice.

Sublingual (s.l.) immunization has been described as a convenient and safe way to induce mucosal immune responses in the respiratory and genital tracts. We constructed a helper-dependent adenoviral (HDAd) vector expressing a condon-optimized soluble fusion glycoprotein (sFsyn) of respiratory syncytial virus (HDAd-sFsyn) and explored the potential of s.l. immunization with HDAd-sFsyn to stimulate immune responses in the respiratory mucosa. The RSV specific systemic and mucosal immune responses were generated in BALB/c mice, and the serum IgG with neutralizing activity was significantly elevated after homologous boost with s.l. application of HDAd-sFsyn. Humoral immune responses could be measured even 14weeks after a single immunization. Upon challenge, s.l. immunization with HDAd-sFsyn displayed an effective protection against RSV infection. These findings suggest that s.l. administration of HDAd-sFsyn acts as an effective and safe mucosal vaccine against RSV infection, and may be a useful tool in the prevention of RSV infection.

Effects of allogeneic blood transfusion in patients with stage II colon cancer.

The aim of the present study was to determine whether allogeneic red blood cell transfusions showed a deleterious effect and what might be preoperative risk factors for blood transfusion in patients with TNM stage II colon cancer. Total 470 patients who fulfilled inclusion criteria were selected for a further 10-year follow-up study. We found that there were statistical significance between non-transfused and transfused group in mortality (P=0.018), local recurrence (P=0.000) and distant metastasis (P=0.040). Local recurrence and distant metastasis between 1 to 3 units and more than 3 units group did not show any significant differences. There was no difference in survival rate between non-transfused and 1 to 3 units group (log rank =0.031, P=0.860). The difference between different blood transfusion volume in transfused patients was found (78.77% vs 63.83%, P=0.006). Meanwhile, the significant difference of survival rate was existed between non-transfused group and more than 3 units group (84.83% vs 63.83%, P=0.002 ). Univariate analysis showed the following 3 variables to be associated with an increased risk of allogeneic blood transfusions: preoperative CEA level (P<0.05), location of tumor (P<0.01) and diameter of tumor (P<0.01). Multivariate analysis revealed that location of tumor and diameter of tumor are two independent factors for requirement of perioperative transfusions. Therefore, allogeneic transfusion increase the postoperative tumor mortality, local recurrence and distant metastasis in patients with stage II colon cancer. The postoperative tumor mortality, local recurrence and distant metastasis were not associated with the blood transfusion volume. The blood transfusion volume was associated with the survival rate. Location of tumor and diameter of tumor were the independent preoperative risk factors for blood transfusion.