RESUMO
BACKGROUND: Because the early diagnosis of subclinical cardiac sarcoidosis (CS) remains difficult, we developed a deep learning algorithm to distinguish CS patients from healthy subjects using echocardiographic movies.MethodsâandâResults:Among the patients who underwent echocardiography from January 2015 to December 2019, we chose 151 echocardiographic movies from 50 CS patients and 151 from 149 healthy subjects. We trained two 3D convolutional neural networks (3D-CNN) to identify CS patients using a dataset of 212 echocardiographic movies with and without a transfer learning method (Pretrained algorithm and Non-pretrained algorithm). On an independent set of 41 echocardiographic movies, the area under the receiver-operating characteristic curve (AUC) of the Pretrained algorithm was greater than that of Non-pretrained algorithm (0.842, 95% confidence interval (CI): 0.722-0.962 vs. 0.724, 95% CI: 0.566-0.882, P=0.253). The AUC from the interpretation of the same set of 41 echocardiographic movies by 5 cardiologists was not significantly different from that of the Pretrained algorithm (0.855, 95% CI: 0.735-0.975 vs. 0.842, 95% CI: 0.722-0.962, P=0.885). A sensitivity map demonstrated that the Pretrained algorithm focused on the area of the mitral valve. CONCLUSIONS: A 3D-CNN with a transfer learning method may be a promising tool for detecting CS using an echocardiographic movie.
Assuntos
Aprendizado Profundo , Miocardite , Sarcoidose , Algoritmos , Ecocardiografia , Humanos , Filmes Cinematográficos , Sarcoidose/diagnóstico por imagemRESUMO
Deep learning models can be applied to electrocardiograms (ECGs) to detect left ventricular (LV) dysfunction. We hypothesized that applying a deep learning model may improve the diagnostic accuracy of cardiologists in predicting LV dysfunction from ECGs. We acquired 37,103 paired ECG and echocardiography data records of patients who underwent echocardiography between January 2015 and December 2019. We trained a convolutional neural network to identify the data records of patients with LV dysfunction (ejection fraction < 40%) using a dataset of 23,801 ECGs. When tested on an independent set of 7,196 ECGs, we found the area under the receiver operating characteristic curve was 0.945 (95% confidence interval: 0.936-0.954). When 7 cardiologists interpreted 50 randomly selected ECGs from the test dataset of 7,196 ECGs, their accuracy for predicting LV dysfunction was 78.0% ± 6.0%. By referring to the model's output, the cardiologist accuracy improved to 88.0% ± 3.7%, which indicates that model support significantly improved the cardiologist diagnostic accuracy (P = 0.02). A sensitivity map demonstrated that the model focused on the QRS complex when detecting LV dysfunction on ECGs. We developed a deep learning model that can detect LV dysfunction on ECGs with high accuracy. Furthermore, we demonstrated that support from a deep learning model can help cardiologists to identify LV dysfunction on ECGs.
Assuntos
Aprendizado Profundo , Eletrocardiografia , Disfunção Ventricular Esquerda/diagnóstico , Adulto , Idoso , Idoso de 80 Anos ou mais , Cardiologistas , Sistemas de Apoio a Decisões Clínicas , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Sensibilidade e Especificidade , SístoleRESUMO
Japan has implemented a cluster-based approach for coronavirus disease 2019 (COVID-19) from the pandemic's beginning based on the transmission heterogeneity (overdispersion) of severe acute respiratory coronavirus 2 (SARS-CoV-2). However, studies analyzing overdispersion of transmission among new variants of concerns (VOCs), especially for Omicron, were limited. Thus, we aimed to clarify how the transmission heterogeneity has changed with the emergence of VOCs (Alpha, Delta, and Omicron) using detailed contact tracing data in Yamagata Prefecture, Japan. We estimated the time-varying dispersion parameter ([Formula: see text]) by fitting a negative binomial distribution for each transmission generation. Our results showed that even after the emergence of VOCs, there was transmission heterogeneity of SARS-CoV-2, with changes in [Formula: see text] during each wave. Continuous monitoring of transmission dynamics is vital for implementing appropriate measures. However, a feasible and sustainable epidemiological analysis system should be established to make this possible.
Assuntos
COVID-19 , SARS-CoV-2 , Humanos , SARS-CoV-2/genética , Japão/epidemiologia , COVID-19/epidemiologia , Busca de Comunicante , Taxa RespiratóriaRESUMO
BACKGROUND: There are more than 7000 rare diseases, most of which have no specific treatment. Disease profiles, such as prevalence and natural history, among the population of a specific country are essential in determining for which disease to research and develop drugs. In Japan, disease profiles of fewer than 2000 rare diseases, called Nanbyo, have been investigated. However, non-Nanbyo rare diseases remain largely uninvestigated. Accordingly, we revealed the prevalence and natural history of rare diseases among the Japanese population. This cross-disease study is the first to analyze rare-disease epidemiology in Japan with high accuracy, disease coverage, and granularity. METHOD: We applied for permission to use the National Database of Health Insurance Claims and Specific Health Checkups of Japan (NDB), which covered 99.9% of public health insurance claims from hospitals and 97.9% from clinics as of May 2015. Then, we obtained 10 years of data on the number of patients of approx. 4500 rare diseases, by sex and age. We translated disease names and established correspondences between rare diseases in NDB and those in Orphanet. Accordingly, we compared the prevalence and natural history between them. RESULTS: About 3000 diseases in NDB are included in Orphanet and other medical databases. The data indicates that even if the Nanbyo systems do not cover a rare disease, its patients survive in many cases. Regarding natural history, genetic diseases tend to be diagnosed later in Japan than in the West. The data shown in this research are available in the Additional file 1 and the website of NanbyoData. CONCLUSIONS: Our research revealed the basic epidemiology and natural history of Japanese patients with some rare diseases using a health insurance claims database. The results imply that the coverage of the present Nanbyo systems is inadequate for rare diseases. Therefore, fundamental reform might be needed to reduce unfairness between rare diseases. Most diseases in Japan follow a tendency of natural history similar to those reported in Orphanet. However, some are detected later, partly because fewer clinical genetic tests are available in Japan than in the West. Finally, we hope that our data and analysis accelerate drug discovery for rare diseases in Japan.
Assuntos
Seguro Saúde , Doenças Raras , Estudos Epidemiológicos , Humanos , Japão/epidemiologia , Prevalência , Doenças Raras/epidemiologiaRESUMO
OBJECTIVES: Super-spreading events caused by overdispersed secondary transmission are crucial in the transmission of COVID-19. However, the exact level of overdispersion, demographics, and other factors associated with secondary transmission remain elusive. In this study, we aimed to elucidate the frequency and patterns of secondary transmission of SARS-CoV-2 in Japan. METHODS: We analyzed 16,471 cases between January 2020 and August 2020. We generated the number of secondary cases distribution and estimated the dispersion parameter (k) by fitting the negative binomial distribution in each phase. The frequencies of the secondary transmission were compared by demographic and clinical characteristics, calculating the odds ratio using logistic regression models. RESULTS: We observed that 76.7% of the primary cases did not generate secondary cases with an estimated dispersion parameter k of 0.23. The demographic patterns of primary-secondary cases differed between phases, with 20-69 years being the predominant age group. There were higher proportions of secondary transmissions among older individuals, symptomatic patients, and patients with 2 days or more between onset and confirmation. CONCLUSIONS: The study showed the estimation of the frequency of secondary transmission of SARS-CoV-2 and the characteristics of people who generated the secondary transmission.
Assuntos
COVID-19 , SARS-CoV-2 , COVID-19/epidemiologia , Demografia , Humanos , Japão/epidemiologiaRESUMO
In response to the outbreak of coronavirus disease 2019 (COVID-19) in Japan, a national COVID-19 cluster taskforce (comprising governmental and nongovernmental experts) was established to support the country's Ministry of Health, Labour and Welfare in conducting daily risk assessment. The assessment was carried out using established infectious disease surveillance systems; however, in the initial stages of the pandemic these were not sufficient for real-time risk assessment owing to limited accessibility, delay in data entry and inadequate case information. Also, local governments were publishing anonymized data on confirmed COVID-19 cases on their official web sites as daily press releases. We developed a unique database for nationwide real-time risk assessment that included these case lists from local government web sites and integrated all case data into a standardized format. The database was updated daily and checked systematically to ensure comprehensiveness and quality. Between 15 January 2020 and 15 June 2021, 776 459 cases were logged in the database, allowing for analysis of real-time risk from the pandemic. This semi-automated database was used in daily risk assessments, and to evaluate and update control measures to prevent community transmission of COVID-19 in Japan. The data were reported almost every week to the Japanese Government Advisory Panel on COVID-19 for public health responses.
Assuntos
COVID-19 , COVID-19/epidemiologia , Humanos , Japão/epidemiologia , Pandemias , Saúde Pública , Medição de RiscoRESUMO
BACKGROUND: Aortic regurgitation (AR) is a common heart disease, with a relatively high prevalence of 4.9% in the Framingham Heart Study. Because the prevalence increases with advancing age, an upward shift in the age distribution may increase the burden of AR. To provide an effective screening method for AR, we developed a deep learning-based artificial intelligence algorithm for the diagnosis of significant AR using electrocardiography (ECG). METHODS: Our dataset comprised 29,859 paired data of ECG and echocardiography, including 412 AR cases, from January 2015 to December 2019. This dataset was divided into training, validation, and test datasets. We developed a multi-input neural network model, which comprised a two-dimensional convolutional neural network (2D-CNN) using raw ECG data and a fully connected deep neural network (FC-DNN) using ECG features, and compared its performance with the performances of a 2D-CNN model and other machine learning models. In addition, we used gradient-weighted class activation mapping (Grad-CAM) to identify which parts of ECG waveforms had the most effect on algorithm decision making. RESULTS: The area under the receiver operating characteristic curve of the multi-input model (0.802; 95% CI, 0.762-0.837) was significantly greater than that of the 2D-CNN model alone (0.734; 95% CI, 0.679-0.783; p<0.001) and those of other machine learning models. Grad-CAM demonstrated that the multi-input model tended to focus on the QRS complex in leads I and aVL when detecting AR. CONCLUSIONS: The multi-input deep learning model using 12-lead ECG data could detect significant AR with modest predictive value.
Assuntos
Insuficiência da Valva Aórtica , Aprendizado Profundo , Algoritmos , Insuficiência da Valva Aórtica/diagnóstico , Inteligência Artificial , Eletrocardiografia/métodos , Humanos , Estudos RetrospectivosRESUMO
Background and Aims: Radiofrequency ablation (RFA) is a widely accepted, minimally invasive treatment for hepatocellular carcinoma (HCC). This study aimed to develop a machine learning (ML) model to predict the risk of HCC recurrence after RFA treatment for individual patients. Methods: We included a total of 1778 patients with treatment-naïve HCC who underwent RFA. The cumulative probability of overall recurrence after the initial RFA treatment was 78.9% and 88.0% at 5 and 10 years, respectively. We developed a conventional Cox proportional hazard model and 6 ML models-including the deep learning-based DeepSurv model. Model performance was evaluated using Harrel's c-index and was validated externally using the split-sample method. Results: The gradient boosting decision tree (GBDT) model achieved the best performance with a c-index of 0.67 from external validation, and it showed a high discriminative ability in stratifying the external validation sample into 2, 3, and 4 different risk groups (P < .001 among all risk groups). The c-index of DeepSurv was 0.64. In order of significance, the tumor number, serum albumin level, and des-gamma-carboxyprothrombin level were the most important variables for the prediction of HCC recurrence in the GBDT model. Also, the current GBDT model enabled the output of a personalized cumulative recurrence prediction curve for each patient. Conclusion: We developed a novel ML model for the personalized risk prediction of HCC recurrence after RFA treatment. The current model may lead to the personalization of effective follow-up strategies after RFA treatment according to the risk stratification of HCC recurrence.
RESUMO
Detection and isolation of infected people are believed to play an important role in the control of the COVID-19 pandemic. Some countries conduct large-scale screenings for testing, whereas others test mainly people with high prior probability of infection such as showing severe symptoms and/or having an epidemiological link with a known or suspected case or cluster of cases. However, what a good testing strategy is and whether the difference in testing strategy shows a meaningful, measurable impact on the COVID-19 epidemic remain unknown. Here, we showed that patterns of association between effective reproduction number (Rt) and test positivity rate can illuminate differences in testing situation among different areas, using global and local data from Japan. This association can also evaluate the adequacy of current testing systems and what information is captured in COVID-19 surveillance. The differences in testing systems alone cannot predict the results of epidemic containment efforts. Furthermore, monitoring test positivity rates and severe case proportions among the nonelderly can predict imminent case count increases. Monitoring test positivity rates in conjunction with the concurrent Rt could be useful to assess and strengthen public health management and testing systems and deepen understanding of COVID-19 epidemic dynamics.
Assuntos
COVID-19 , Número Básico de Reprodução , Humanos , Japão/epidemiologia , Pandemias , SARS-CoV-2RESUMO
Background: Roles of children and adolescents in spreading coronavirus disease 2019 (COVID-19) in the community is not fully understood. Methods: We analyzed the data of 7,758 children and adolescents with COVID-19 and characteristics of secondary transmission generated by these cases using case information published by local governments. Ratio of pediatric and adolescent cases generating secondary transmission was calculated for various social settings. Results: The incidence of COVID-19 was 24.8 cases per 105 population aged between 0 and 9 years, and 59.2 among those aged between 10 and 19 years, which was lower than that among individuals of all age groups (79.6 per 105 population) between January 15 and October 31, 2020. The proportion of cases generating secondary cases was 8.3% among infants and young children in nursery schools and kindergartens, 16% among children and adolescents attending primary schools, 34% among those attending junior high schools, 43% among those attending high schools, 31% among those attending professional training colleges, and 24% in those attending universities. Households were the most common setting for secondary transmission. Conclusion: The risk of generating secondary cases might be limited among pediatric and adolescent cases with COVID-19, especially in settings outside households. Effectiveness of traditional mitigation measures (e.g., school closures) to suppress COVID-19 transmissions should be carefully evaluated.
RESUMO
Intravascular ultrasound (IVUS) is a diagnostic modality used during percutaneous coronary intervention. However, specialist skills are required to interpret IVUS images. To address this issue, we developed a new artificial intelligence (AI) program that categorizes vessel components, including calcification and stents, seen in IVUS images of complex lesions. When developing our AI using U-Net, IVUS images were taken from patients with angina pectoris and were manually segmented into the following categories: lumen area, medial plus plaque area, calcification, and stent. To evaluate our AI's performance, we calculated the classification accuracy of vessel components in IVUS images of vessels with clinically significantly narrowed lumina (< 4 mm2) and those with severe calcification. Additionally, we assessed the correlation between lumen areas in manually-labeled ground truth images and those in AI-predicted images, the mean intersection over union (IoU) of a test set, and the recall score for detecting stent struts in each IVUS image in which a stent was present in the test set. Among 3738 labeled images, 323 were randomly selected for use as a test set. The remaining 3415 images were used for training. The classification accuracies for vessels with significantly narrowed lumina and those with severe calcification were 0.97 and 0.98, respectively. Additionally, there was a significant correlation in the lumen area between the ground truth images and the predicted images (ρ = 0.97, R2 = 0.97, p < 0.001). However, the mean IoU of the test set was 0.66 and the recall score for detecting stent struts was 0.64. Our AI program accurately classified vessels requiring treatment and vessel components, except for stents in IVUS images of complex lesions. AI may be a powerful tool for assisting in the interpretation of IVUS imaging and could promote the popularization of IVUS-guided percutaneous coronary intervention in a clinical setting.
Assuntos
Algoritmos , Inteligência Artificial , Angiografia Coronária/métodos , Doença da Artéria Coronariana/patologia , Vasos Coronários/patologia , Aprendizado Profundo , Ultrassonografia/métodos , Doença da Artéria Coronariana/diagnóstico por imagem , Vasos Coronários/diagnóstico por imagem , Humanos , Processamento de Imagem Assistida por ComputadorRESUMO
The Human Phenotype Ontology (HPO) is the de facto standard ontology to describe human phenotypes in detail, and it is actively used, particularly in the field of rare disease diagnoses. For clinicians who are not fluent in English, the HPO has been translated into many languages, and there have been four initiatives to develop Japanese translations. At the Biomedical Linked Annotation Hackathon 6 (BLAH6), a rule-based approach was attempted to determine the preferable Japanese translation for each HPO term among the candidates developed by the four approaches. The relationship between the HPO and Mammalian Phenotype translations was also investigated, with the eventual goal of harmonizing the two translations to facilitate phenotype-based comparisons of species in Japanese through cross-species phenotype matching. In order to deal with the increase in the number of HPO terms and the need for manual curation, it would be useful to have a dictionary containing word-by-word correspondences and fixed translation phrases for English word order. These considerations seem applicable to HPO localization into other languages.