Recognition of eye diseases based on deep neural networks for transfer learning and improved D-S evidence theory.

BACKGROUND: Human vision has inspired significant advancements in computer vision, yet the human eye is prone to various silent eye diseases. With the advent of deep learning, computer vision for detecting human eye diseases has gained prominence, but most studies have focused only on a limited number of eye diseases. RESULTS: Our model demonstrated a reduction in inherent bias and enhanced robustness. The fused network achieved an Accuracy of 0.9237, Kappa of 0.878, F1 Score of 0.914 (95% CI [0.875-0.954]), Precision of 0.945 (95% CI [0.928-0.963]), Recall of 0.89 (95% CI [0.821-0.958]), and an AUC value of ROC at 0.987. These metrics are notably higher than those of comparable studies. CONCLUSIONS: Our deep neural network-based model exhibited improvements in eye disease recognition metrics over models from peer research, highlighting its potential application in this field. METHODS: In deep learning-based eye recognition, to improve the learning efficiency of the model, we train and fine-tune the network by transfer learning. In order to eliminate the decision bias of the models and improve the credibility of the decisions, we propose a model decision fusion method based on the D-S theory. However, D-S theory is an incomplete and conflicting theory, we improve and eliminate the existed paradoxes, propose the improved D-S evidence theory(ID-SET), and apply it to the decision fusion of eye disease recognition models.

Annexin A1-mediated inhibition of inflammatory cytokines may facilitate the resolution of inflammation in acute radiation-induced lung injury.

The present study evaluated the role of annexin A1 (ANXA1) in the treatment of acute radiation-induced lung injury (RILI) and investigated the mechanism of its action. The expression of ANXA1, interleukin-6 (IL-6) and myeloperoxidase (MPO) in the plasma of patients with RILI prior to and following hormonotherapy was assessed by enzyme-linked immunosorbent assay. The association of plasma ANXA1 concentration with clinical effect, and the correlation between the expression of ANXA1 and that of IL-6 and MPO were evaluated. ANXA1 was overexpressed or knocked down in a macrophage cell line, and its impact on IL-6 and MPO expression was measured. Following glucocorticoid hormonotherapy, patients with RILI exhibited a higher plasma concentration of ANXA1 compared with that prior to treatment, while IL-6 and MPO levels were lower. The concentration of ANXA1 in plasma was negatively correlated with IL-6 and MPO levels, with a correlation coefficient of -0.492 and -0.437, respectively (P<0.001). The increasing concentration of ANXA1 in plasma following treatment was associated with the clinical effect in patients with RILI (P=0.007). The expression levels of of IL-6 and MPO were inhibited both in the cytoplasm and in the culture solution, when ANXA1 expression was upregulated in a macrophage cell line. In conclusion, ANXA1 inhibited the synthesis and secretion of IL-6 and MPO inflammatory cytokines, indicating that ANXA1 may have therapeutic potential as a treatment target for RILI.

The diagnostic value of 99TcM-2-(2-methyl-5-nitro-1H-imidazol-1-yl) ethyl dihydrogen phosphate hypoxia imaging and its evaluation performance for radiotherapy efficacy in non-small-cell lung cancer.

BACKGROUND AND AIM: This study was designated to assess the diagnostic value of 99TcM-2-(2-methyl-5-nitro-1H-imidazol-1-yl) ethyl dihydrogen phosphate (99TcM-MNLS) hypoxia imaging and its evaluation performance for radiotherapy efficacy in patients with non-small-cell lung cancer (NSCLC). PATIENTS AND METHODS: A total of 61 patients with NSCLC were selected for this study. All patients were injected with 99TcM-MNLS within 1 week prior to radiotherapy and they were injected with 99TcM-MNLS again 3 months after radiotherapy. Qualitative analysis along with semiquantitative analysis results were obtained from hypoxia imaging. Meanwhile, the effect of radiotherapy on patients with NSCLC was evaluated based on the solid tumor curative effect evaluation standard. Finally, SPSS 19.0 statistical software was implemented for statistical analysis. RESULTS: There was no significant difference in age or sex between the NSCLC patient group and benign patient group (P>0.05). 99TcM-MNLS was selectively concentrated in tumor tissues with a clear imaging in 24 hours. Results from both qualitative analysis and semiquantitative analysis indicated that the sensitivity and specificity of 99TcM-MNLS hypoxia imaging in diagnosing NSCLC were 93.8% and 84.6% and 72.9% and 100%, respectively. Moreover, the receiver operating characteristic curve provided evidence that 99TcM-MNLS hypoxia imaging was a powerful diagnostic tool in distinguishing malignant lung cancer from benign lesions. As suggested by 24-hour imaging, the tumor-to-normal ratio of patients in the 99TcM-MNLS high-intake group and low-intake group had a decline of 24.7% and 14.4% after radiotherapy, respectively. The decline in the tumor-to-normal ratio between the two dosage groups was significantly different (P<0.05). CONCLUSION: 99TcM-MNLS hypoxia imaging had reliable values in both diagnosing NSCLC and evaluating therapeutic effects of radiotherapy on patients with NSCLC.