Accurate estimation of biological age and its application in disease prediction using a multimodal image Transformer system.

Aging in an individual refers to the temporal change, mostly decline, in the body's ability to meet physiological demands. Biological age (BA) is a biomarker of chronological aging and can be used to stratify populations to predict certain age-related chronic diseases. BA can be predicted from biomedical features such as brain MRI, retinal, or facial images, but the inherent heterogeneity in the aging process limits the usefulness of BA predicted from individual body systems. In this paper, we developed a multimodal Transformer-based architecture with cross-attention which was able to combine facial, tongue, and retinal images to estimate BA. We trained our model using facial, tongue, and retinal images from 11,223 healthy subjects and demonstrated that using a fusion of the three image modalities achieved the most accurate BA predictions. We validated our approach on a test population of 2,840 individuals with six chronic diseases and obtained significant difference between chronological age and BA (AgeDiff) than that of healthy subjects. We showed that AgeDiff has the potential to be utilized as a standalone biomarker or conjunctively alongside other known factors for risk stratification and progression prediction of chronic diseases. Our results therefore highlight the feasibility of using multimodal images to estimate and interrogate the aging process.

[Value of soluble triggering receptor expression on myeloid cells-1 level of alveolar fluid in early diagnosis of ventilator-associated pneumonia: a Meta-analysis].

OBJECTIVE: Serum soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) is a useful biomarker of bacterial infection. However, the diagnostic value of sTREM-1 of alveolar fluid in pulmonary infection is still unclear. This article aimed to explore the value of sTREM-1 of alveolar fluid in the early diagnosis of ventilator-associated pneumonia (VAP) by systematic review of relevant literatures. METHODS: CNKI, Wanfang, VIP, PubMed/Medline and Embase databases were retrieved. Articles on diagnosis of VAP by sTREM-1 before June 30, 2019 were collected. QUADAS-2 scale provided by Cochrane Collaboration Network was used to evaluate the quality of diagnostic experiments. RevMan 5.3 and Stata 13.0 software were used to complete Meta-analysis. The levels of sTREM-1 between VAP and non-VAP patients were analyzed by Meta-analysis, and then diagnostic test Meta-analysis was conducted. Heterogeneity, sensitivity and publication bias were analyzed. RESULTS: A total of 24 articles were enrolled. QUADAS-2 scale indicated that the selected literature had low bias and high clinical adaptability. (1) In Meta-analysis of sTREM-1 level in alveolar fluid, 20 articles were selected and found to have high heterogeneity (I2 = 94.4%, P = 0.000). The random effects models were used for Meta-analysis. It was indicated that the sTREM-1 level in alveolar fluid of VAP group was significantly higher than that of non-VAP group with significant difference [standardized mean difference (SMD) was 1.47, 95% confidence interval (95%CI) was 1.00-1.95, Z = 6.14, P = 0.000]. By subgroup analysis and Meta-regression analysis, no source of heterogeneity was found. Sensitivity analysis suggested that the results of this Meta-analysis were robust and credible, and Begg funnel plot analysis showed that there was no significant publication bias (Z = 1.46, P = 0.143). (2) A total of 18 articles were included in the Meta-analysis of diagnostic experiments. Deek funnel plot showed publication bias (P = 0.012). The combined sensitivity was 0.87 (95%CI was 0.81-0.91), specificity was 0.80 (95%CI was 0.73-0.86), and diagnostic odds ratio (DOR) was 26 (95%CI was 13-50). Subgroup analysis of three different sources of alveolar fluid (bronchoalveolar lavage fluid, endotracheal aspiration fluid and exhaled ventilator condensate) showed that STREM-1 had a certain value in early diagnosis of VAP. The I2 of combined DOR was 35.4%, and I2 of sensitivity was 79.46%, I2 of specificity was 77.61%, suggesting heterogeneity in the selected literature. Subgroup analysis found that nationality, subject design, sample source, sample size and diagnostic "gold criteria" were related to heterogeneity, but not age. The area under synthetic receiver operating characteristic (SROC) curve (AUC) was 0.90 (95%CI was 0.87-0.92). CONCLUSIONS: The detection of sTREM-1 level in alveolar fluid can be used for the early diagnosis of VAP with high sensitivity and specificity. If combined with other biomarkers, it may have more diagnostic value.