RESUMEN
AIMS: Machine learning models can use image and text data to predict the number of years since diabetes diagnosis; such model can be applied to new patients to predict, approximately, how long the new patient may have lived with diabetes unknowingly. We aimed to develop a model to predict self-reported diabetes duration. METHODS: We used the Brazilian Multilabel Ophthalmological Dataset. Unit of analysis was the fundus image and its meta-data, regardless of the patient. We included people 40â¯+ years and fundus images without diabetic retinopathy. Fundus images and meta-data (sex, age, comorbidities and taking insulin) were passed to the MedCLIP model to extract the embedding representation. The embedding representation was passed to an Extra Tree Classifier to predict: 0-4, 5-9, 10-14 and 15â¯+ years with self-reported diabetes. RESULTS: There were 988 images from 563 people (mean ageâ¯=â¯67 years; 64â¯% were women). Overall, the F1 score was 57â¯%. The group 15â¯+ years of self-reported diabetes had the highest precision (64â¯%) and F1 score (63â¯%), while the highest recall (69â¯%) was observed in the group 0-4 years. The proportion of correctly classified observations was 55â¯% for the group 0-4 years, 51â¯% for 5-9 years, 58â¯% for 10-14 years, and 64â¯% for 15â¯+ years with self-reported diabetes. CONCLUSIONS: The machine learning model had acceptable accuracy and F1 score, and correctly classified more than half of the patients according to diabetes duration. Using large foundational models to extract image and text embeddings seems a feasible and efficient approach to predict years living with self-reported diabetes.
Asunto(s)
Diabetes Mellitus , Fondo de Ojo , Aprendizaje Automático , Valor Predictivo de las Pruebas , Autoinforme , Humanos , Femenino , Masculino , Anciano , Persona de Mediana Edad , Factores de Tiempo , Diabetes Mellitus/diagnóstico , Diabetes Mellitus/epidemiología , Brasil/epidemiología , Adulto , Bases de Datos Factuales , Retinopatía Diabética/diagnóstico , Retinopatía Diabética/epidemiología , Minería de Datos/métodos , Reproducibilidad de los Resultados , Interpretación de Imagen Asistida por ComputadorRESUMEN
Abstract Angiotensin II (AngII) causes endothelial dysfunction. Eucommia ulmoides extract (EUE) is documented to manipulate AngII, but its impact on cardiac microvascular endothelial cell (CMVEC) function remains unknown. This study determines the effects of EUE on AngII-treated CMVECs. CMVECs were treated with different concentrations of AngII or EUE alone and/or the p53 protein activator, WR-1065, before AngII treatment, followed by examinations of the apoptotic, migratory, proliferative, and angiogenic capacities and nitric oxide (NO), p53, von Willebrand factor (vWF), endothelin (ET)-1, endothelial NO synthase (eNOS), manganese superoxide dismutase (MnSOD), hypoxia-inducible factor (HIF)-1α, and vascular endothelial growth factor (VEGF) levels. AngII induced CMVEC dysfunction in a concentration-dependent manner. EUE enhanced the proliferative, migratory, and angiogenic capacities and NO, MnSOD, and eNOS levels but repressed apoptosis and vWF and ET-1 levels in AngII-induced dysfunctional CMVECs. Moreover, AngII increased p53 mRNA levels, p-p53 levels in the nucleus, and p53 protein levels in the cytoplasm and diminishes HIF-1α and VEGF levels in CMVECs; however, these effects were counteracted by EUE treatment. Moreover, WR-1065 abrogated the mitigating effects of EUE on AngII-induced CMVEC dysfunction by activating p53 and decreasing HIF-1α and VEGF expression. In conclusion, EUE attenuates AngII-induced CMVEC dysfunction by upregulating HIF-1α and VEGF levels via p53 inactivation
Asunto(s)
Eucommiaceae/efectos adversos , Extractos Vegetales/efectos adversos , Células Endoteliales/clasificación , Factor A de Crecimiento Endotelial Vascular/análisisRESUMEN
The global impact of coronavirus disease 2019 (COVID-19) is unprecedented, and many control and prevention measures have been implemented to test for and trace COVID-19. However, invisible-spreaders, who are associated with nucleic acid detection and asymptomatic infections, have received insufficient attention in the current COVID-19 control efforts. In this paper, we analyze the time series infection data for Italy, Germany, Brazil, India and Sweden since the first wave outbreak to address the following issues through a series of experiments. We conclude that: 1) As of June 1, 2020, the proportion of invisible-spreaders is close to 0.4% in Sweden, 0.8% in early Italy and Germany, and 0.4% in the middle and late stages. However, in Brazil and India, the proportion still shows a gradual upward trend; 2) During the spread of this pandemic, even a slight increase in the proportion of invisible-spreaders could have large implications for the health of the community; and 3) On resuming work, the pandemic intervention measures will be relaxed, and invisible-spreaders will cause a new round of outbreaks.
Asunto(s)
Infecciones Asintomáticas/epidemiología , COVID-19/transmisión , Brasil/epidemiología , COVID-19/diagnóstico , COVID-19/epidemiología , COVID-19/virología , Alemania/epidemiología , Humanos , India/epidemiología , Italia/epidemiología , Modelos Teóricos , Pandemias , SARS-CoV-2/aislamiento & purificación , Suecia/epidemiologíaRESUMEN
BACKGROUND: Currently available tools for assessing high cardiovascular risk (HCR) often require measurements not available in resource-limited settings in low- and middle-income countries (LMIC). There is a need to assess HCR using a pragmatic evidence-based approach. OBJECTIVES: This study sought to report the prevalence of HCR in 10 LMIC areas in Africa, Asia, and South America and to investigate the profiles and correlates of HCR. METHODS: Cross-sectional analysis using data from the National Heart, Lung, and Blood Institute-UnitedHealth Group Centers of Excellence. HCR was defined as history of heart disease/heart attack, history of stroke, older age (≥50 years for men and ≥60 for women) with history of diabetes, or older age with systolic blood pressure ≥160 mm Hg. Prevalence estimates were standardized to the World Health Organization's World Standard Population. RESULTS: A total of 37,067 subjects ages ≥35 years were included; 53.7% were women and mean age was 53.5 ± 12.1 years. The overall age-standardized prevalence of HCR was 15.4% (95% confidence interval: 15.0% to 15.7%), ranging from 8.3% (India, Bangalore) to 23.4% (Bangladesh). Among men, the prevalence was 1.7% for the younger age group (35 to 49 years) and 29.1% for the older group (≥50); among women, 3.8% for the younger group (35 to 59 years) and 40.7% for the older group (≥60). Among the older group, measured systolic blood pressure ≥160 mm Hg (with or without other conditions) was the most common criterion for having HCR, followed by diabetes. The proportion of having met more than 1 criterion was nearly 20%. Age, education, and body mass index were significantly associated with HCR. Cross-site differences existed and were attenuated after adjusting for age, sex, education, smoking, and body mass index. CONCLUSIONS: The prevalence of HCR in 10 LMIC areas was generally high. This study provides a starting point to define targeted populations that may benefit from interventions combining both primary and secondary prevention strategies.