Application of deep learning image assessment software VeriSee™ for diabetic retinopathy screening.

PURPOSE: To develop a deep learning image assessment software VeriSee™ and to validate its accuracy in grading the severity of diabetic retinopathy (DR). METHODS: Diabetic patients who underwent single-field, nonmydriatic, 45-degree color retinal fundus photography at National Taiwan University Hospital between July 2007 and June 2017 were retrospectively recruited. A total of 7524 judgeable color fundus images were collected and were graded for the severity of DR by ophthalmologists. Among these pictures, 5649 along with another 31,612 color fundus images from the EyePACS dataset were used for model training of VeriSee™. The other 1875 images were used for validation and were graded for the severity of DR by VeriSee™, ophthalmologists, and internal physicians. Area under the receiver operating characteristic curve (AUC) for VeriSee™, and the sensitivities and specificities for VeriSee™, ophthalmologists, and internal physicians in diagnosing DR were calculated. RESULTS: The AUCs for VeriSee™ in diagnosing any DR, referable DR and proliferative diabetic retinopathy (PDR) were 0.955, 0.955 and 0.984, respectively. VeriSee™ had better sensitivities in diagnosing any DR and PDR (92.2% and 90.9%, respectively) than internal physicians (64.3% and 20.6%, respectively) (P < 0.001 for both). VeriSee™ also had better sensitivities in diagnosing any DR and referable DR (92.2% and 89.2%, respectively) than ophthalmologists (86.9% and 71.1%, respectively) (P < 0.001 for both), while ophthalmologists had better specificities. CONCLUSION: VeriSee™ had good sensitivity and specificity in grading the severity of DR from color fundus images. It may offer clinical assistance to non-ophthalmologists in DR screening with nonmydriatic retinal fundus photography.

3-methyadenine inhibits lipopolysaccharides-induced pulmonary inflammation at the early stage of silicosis via blocking NF-κB signaling pathway.

Occupational exposure to silica dust is related to pulmonary inflammation and silicosis. Lipopolysaccharides (LPSs) could aggravate apoptosis in alveolar macrophages (AMs) of human silicosis through autophagy, yet how the reduction of autophagy attenuated LPS-induced lung injury and the related mechanisms need to be investigated. In the study, we aim to understand the role of 3-methyladenine (3-MA), an inhibitor of autophagy, in LPS-mediated inflammatory responses and fibrosis. We collected AMs from observers/silicosis patients. The results showed that LPS induced NF-κB-related pulmonary inflammation in observers and silicosis patients, as confirmed by an increase in the expression of IL-1ß, IL-6, TNF-α, and p65, which could be inhibited by 3-MA treatment. In mice models, at the early stage (7d) of silicosis, but not the late (28d) stage, blocking autophagy reversed the increased levels of IL-1ß, IL-6, TNF-α, and p65 caused by LPS. Mechanism study revealed that LPS triggered the expression of LC3 II, p62, and cleaved caspase-3 at the early stage exposed to silica, which could be restored by 3-MA, while there was no difference in the expression of LAMP1 either at the early or late stage of silicosis in different groups. Similarly, 3-MA treatment did not prevent fibrosis characterized by destroyed alveoli, collagen deposition, and increased expression of α-SMA and Col-1 induced by LPS at the late stage of silicosis. The results suggested that 3-MA has a role in the protection of lung injury at the early stage of silicosis and provided an experimental basis for preventive strategies of pulmonary inflammation and silicosis.

Utility of sTREM-1 and Presepsin (sCD14-ST) as Diagnostic and Prognostic Markers of Sepsis.

BACKGROUND: Sepsis is a condition prevalent among hospitalized patients which carries a high risk of morbidity and mortality. Rapid recognition of sepsis as the cause of deterioration is desirable, and then effective treatment can be initiated rapidly. Traditionally, diagnosis was based on the presence of two or more positive SIRS criteria due to infection. However, recently published sepsis-3 criteria put more emphasis on organ dysfunction caused by infection in the definition of sepsis. Regardless of this, no gold standard for diagnosis exists, and clinicians still rely on a number of traditional and novel biomarkers to discriminate between patients with and without infection, as the cause of deterioration. The present study aims to observe the changes of soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) and presepsin (sCD14-ST) 1evels in plasma of sepsis patients and explore the diagnosis and prognosis of sepsis. METHODS: Sixty patients with sepsis admitted to the Department of Critical Care medicine in Hainan General Hospital from October 2013 to March 2019 were selected as the experimental group. Also chosen in the same period were 60 cases of hospitalized non-sepsis patients as a control group. Plasma levels of sTREM-1 and presepsin were determined by enzyme-linked immunosorbent assay (ELISA) in the 60 patients with sepsis and the 60 non-sepsis patients. Changes of sTREM-1 and presepsin plasma 1evels were observed in the survival subgroup and non-survival subgroup of the patients with sepsis on days 1, 4, 7, and the day of discharge or death. RESULTS: Plasma levels of sTREM-1 and presepsin in the patients with sepsis were higher than those in the control group on the day of admission (p < 0.01). The levels of sTREM-1 and presepsin in the septic shock group were significantly higher than those in the sepsis group (p < 0.01). Plasma levels of sTREM-1 and presepsin showed a decreasing trend in the survival subgroup of the patients with sepsis, while maintaining high 1evels or increased in the subgroup of non-survivors. At different time points, the plasma levels of sTREM-1 and presepsin of the subgroup of non-survivors were all significantly higher than the subgroup of survivors. There was a significant positive correlation between plasma 1evels of sTREM-1 and presepsin (r = 0.596, p < 0.01). According to the plasma sTREM-1, presepsin, CRP, and PCT levels on the first day of enrollment in patients with sepsis, ROC curve analysis was performed and AUC was calculated. The results showed that the AUC values of sTREM-1 and presepsin were relatively high, which was 0.925 and 0.910, respectively. The AUC of PCT and CRP were slightly lower, which was 0.861 and 0.816, respectively (p < 0.01). CONCLUSIONS: ROC curve was used to study the value of sTREM-1 and presepsin in the diagnosis of sepsis, which suggested that sTREM-1 and presepsin should be significantly superior to CRP and PCT levels. The sTREM-1 combined with presepsin had the highest AUC. sTREM-1 has been shown to have an optimal threshold of 125.00 pg/mL for the diagnosis of sepsis, the specificity was 86.0% and the sensitivity was 87.0%. Presepsin has been shown to have an optimal threshold of 1,025.00 pg/mL for the diagnosis of sepsis, the specificity was 83.0% and the sensitivity was 85.0%. The sTREM-1 and presepsin plasma levels have great reference value for the diagnosis of sepsis, and the sTREM-1 and presepsin plasma levels are relative to the severity of sepsis. It is helpful to evaluate treatment effect and prognosis of sepsis by dynamically monitoring the plasma 1evels of sTREM-1 and presepsin.

Cross-Camera External Validation for Artificial Intelligence Software in Diagnosis of Diabetic Retinopathy.

Aims: To investigate the applicability of deep learning image assessment software VeriSee DR to different color fundus cameras for the screening of diabetic retinopathy (DR). Methods: Color fundus images of diabetes patients taken with three different nonmydriatic fundus cameras, including 477 Topcon TRC-NW400, 459 Topcon TRC-NW8 series, and 471 Kowa nonmyd 8 series that were judged as "gradable" by one ophthalmologist were enrolled for validation. VeriSee DR was then used for the diagnosis of referable DR according to the International Clinical Diabetic Retinopathy Disease Severity Scale. Gradability, sensitivity, and specificity were calculated for each camera model. Results: All images (100%) from the three camera models were gradable for VeriSee DR. The sensitivity for diagnosing referable DR in the TRC-NW400, TRC-NW8, and non-myd 8 series was 89.3%, 94.6%, and 95.7%, respectively, while the specificity was 94.2%, 90.4%, and 89.3%, respectively. Neither the sensitivity nor the specificity differed significantly between these camera models and the original camera model used for VeriSee DR development (p = 0.40, p = 0.065, respectively). Conclusions: VeriSee DR was applicable to a variety of color fundus cameras with 100% agreement with ophthalmologists in terms of gradability and good sensitivity and specificity for the diagnosis of referable DR.

Lipopolysaccharides promote pulmonary fibrosis in silicosis through the aggravation of apoptosis and inflammation in alveolar macrophages.

Alveolar macrophages (AMs) play an important defensive role by removing dust and bacteria from alveoli. Apoptosis of AMs is associated with lung fibrosis; however, the relationship between this apoptotic event and environmental factors, such as the presence of lipopolysaccharides (LPSs) in the workplace, has not yet been addressed. To investigate whether exposure to LPS can exacerbate fibrosis, we collected AMs from 12 male workers exposed to silica and incubated them in the presence and absence of LPS for 24 h. We show that the levels of cleaved caspase-3 and pro-inflammatory cytokines interleukin (IL)-1ß, IL-6, and tumor necrosis factor-alpha were increased in these AMs following LPS treatment. Moreover, we demonstrate that LPS exposure aggravated apoptosis and the release of inflammatory factors in AMs in a mouse model of silicosis, which eventually promoted pulmonary fibrosis. These results suggest that exposure to LPS may accelerate the progression of pulmonary fibrosis in silicosis by increasing apoptosis and inflammation in AMs.