![]() The automatic identification of cardiac sources of embolism from the echocardiography reports will lighten the burden of neurologists at a certain extent, and it will also reduce the erroneous diagnosis caused by the misinterpretation of the reports. The echocardiography reports reflect the cardiac sources of embolism related to the abnormal structure and function of the heart, while the electrocardiogram reports reflect the cardiac sources of embolism related to abnormal cardiac rhythm. In the clinical practice, neurologists make decisions mainly depending on the interpretation of echocardiography reports and electrocardiogram reports. Cardiac sources of embolism are diagnosis evidences of cardiogenic stroke. Among them, cardiogenic stroke is one of the most common type of acute ischemic stroke, which is caused by a variety of cardiac sources of embolism and accounts for 20% of stroke. In CISS the Ischemic stroke is divided into five categories: large artery atherosclerosis (LAA), cardiogenic stroke (CS), penetrating artery disease (PAD), other etiology (OE) and undetermined etiology (UE). proposed the Chinese Ischemic Stroke Subclassification (CISS), which is suitable for stroke classification in China. Referring to the international stroke classification, Gao et al. The accurate classification of ischemic stroke has significant impact on the treatment of patients and stroke-related studies such as clinical trial, epidemiology and gene study. The National Health Commission of the Peoples’ Republic of China has adopted a series of policies and methods for stoke prevention and control. With the continuous acceleration of the aging population and urbanization, the unhealthy lifestyle of residents is becoming popular, which results in the sharply rising incidence of stroke and brings heavy burden on families and societies in China. Ischemic stroke is the most common type of stroke, which accounts for 69.6 to 70.8% of stroke in China. The method of automatically identifying diagnosis evidence of cardiogenic stroke proposed in this study will be further refined in the practice. The model trained based on the corpus also has a good performance on the testing set. We use the phrases to generate an annotated corpus automatically, which greatly reduces the cost of manual annotation. In this study, we analyze the structure of the echocardiograph reports and summarized 149 phrases on diagnosis evidence of cardiogenic stroke. In addition, our method is capable to identify the novel diagnosis evidence of cardiogenic stroke description such as “二尖瓣中-重度狭窄” (mitral stenosis), “主动脉瓣退行性病变” (aortic valve calcification) et al. Our machine learning method achieved the average performance on the diagnosis evidence identification is 98.03, 90.17 and 93.94% respectively. The generated corpus is divided into training set and testing set in the ratio of 8:2, which is used to train and validate a machine learning model to identify the evidence of cardiogenic stroke using BiLSTM-CRF algorithm. ![]() We selected 11 most frequent diagnosis evidence types such as “二尖瓣狭窄” (mitral stenosis) for further identifying. Furthermore, we developed an annotated corpus via mapping 149 phrases to the 4188 reports. ![]() Collaborating with neurologists and sonographers, we summarized 149 phrases on diagnosis evidence of cardiogenic stroke such as “二尖瓣重度狭窄” (severe mitral stenosis), “主动脉瓣退行性变” (aortic valve degeneration) and so on. We collected 4188 Chinese echocardiograph reports of 4018 patients, with average length 177 Chinese characters in free-text style. In this study, we developed a machine learning model to automatically identify diagnosis evidences of cardiogenic stroke providing to neurologist for clinical decision making. Sonographers will investigate patients’ heart via echocardiograph, and describe them in the echocardiograph reports. In cardiogenic stroke diagnosis, echocardiograph examination is one of the most important examinations. Cardiogenic stroke has increasing morbidity in China and brought economic burden to patient families. ![]()
0 Comments
Leave a Reply. |