RESUMO
Objective To develop a diagnostic model based on deep neural network for intelligent discrimination of thyroid function. Methods A total of 1616 patients ( 283 males, 1333 females, average age:52 years) who underwent thyroid imaging between May 2016 and June 2018 were selected. According to the clinical diagnosis, the 1616 cases included 299 normal thyroid cases, 876 hyperthyroidism cases and 441 hypothyroidism cases. Feature extraction and learning training were performed on 1000 training set sam-ples by two deep neural network models ( AlexNet;deep convolution generative adversarial networks ( DCGAN) ) using deep learning algorithm. Performance verifications were implemented on 616 test set samples. The con-sistency between the verification results of the two models and the clinical diagnosis was analyzed by Kappa test. Meanwhile, the time advantage of the intelligent diagnosis models was analyzed. Results The average diagnostic time of AlexNet model was 1 s/case, and the classification accuracy for normal thyroid, hyperthy-roidism, hypothyroidism were 82.29％(79/96), 94.62％(369/390), 100％(130/130), respectively. The Kappa value between results of AlexNet model and clinical diagnosis was 0.886 ( P<0.05) . The average di-agnostic time of DCGAN model was 1 s/case, and the classification accuracy for normal thyroid, hyperthy-roidism, hypothyroidism were 85.42％(82/96), 95.64％(373/390), 99.23％(129/130), respectively. The Kappa value between results of DCGAN model and clinical diagnosis was 0.904 ( P<0.05) . Conclusion The deep neural network intelligent diagnosis model can quickly determine the functional status of thyroid gland in thyroid imaging, and it has a high recognition accuracy, thus providing a new method for thyroid image review.
RESUMO
Objective To analyze the clinical and imaging data of patients with cerebral microbleeds after acute ischemic stroke (AIS) and to explore the related risk factors of cerebral microbleeds. Methods Seventy-six patients with AIS hospitalized in the Department of Neurology of Binhai Hospital, General Hospital of Tianjin Medical University from May 2015 to October 2016 were collected. All patients were divided into a positive group (32 cases) and a negative group (44 cases) according to whether there were micro-bleeds in the magnetic resonant imaging. The clinical and imaging data of the two groups were statistically analyzed, and the differences in risk factors related to microbleeds between the two groups were compared, and receiver operating characteristic curve (ROC) was drawn, the area under the ROC curve (AUC) was calculated to evaluate the predictive value of age, history of hypertension, and leukoaraiosis for cerebral microbleeds in patients with AIS. Results The age (years: 69.5±10.1 vs. 61.3±8.7), proportion of leukoaraiosis [59.4% (19/32) vs. 34.1% (15/44)] and history of hypertension [75.0% (24/32) vs. 29.5% (13/44)] in the positive group were all higher than those in the negative group (all P < 0.05). Correlation analysis showed that the more severe lacunar infarction, the higher the incidence of cerebral microbleeds (r = 0.278, P = 0.012). Logistic regression analyses showed that age [odds ratio (OR) = 5.11, 95% confidence interval (95%CI) = 3.25 - 12.20, P = 0.001), leukoaraiosis (OR = 4.62, 95%CI = 1.08 - 16.89, P = 0.019) and history of hypertension (OR = 9.28, 95%CI = 2.09 - 38.67, P = 0.003) were the related risk factors of cerebral microbleeds in patients with AIS. ROC curve analysis showed that age, history of hypertension could predict cerebral microbleeds in patients with AIS, with the area under ROC curve (AUC) were 0.751, 0.727 (all P < 0.05), 95%CI respectively was 0.634 - 0.868, 0.610 - 0.845, the sensitivity respectively was 59.4%, 75.0%, and specificity respectively was 84.1%, 70.5%. Conclusion The age, leukoaraiosis, and history of hypertension were related to the occurrence of cerebral microbleeds; the severity degree of lacunar infarction was positively correlated with the incidence of cerebral microbleeds; age and history of hypertension have certain predictive value for AIS.