Search

Objective To explore the feasibility of object detection and deep learning model applied on the localization and classification of cervical precancerous lesions based on identifying the epithelial and vascular features in colposcopy images. Methods A total of 28 975 colposcopic images were collected from Mar 2018 to Jul 2019 in the Obstetrics and Gynecology Hospital of Fudan University, including cervical low-grade lesion (5 708 patients), high-grade lesion (2 206 patients) and cervical cancer (514 patients). According to the colposcopy standardized terminology of the International Federation for Cervical Pathology and Colposcopy and American Society for Colposcopy and Cervical Pathology，39 858 valid labels were obtained after pixel-level labeling based on 16 types of cervical epithelial and vascular signs. In order to reduce the error of fine-grained labeling, labels were further classified into three categories: low-grade, high-grade and cancer. Using ResNet101 pre-training network after secondary transfer learning as feature extractor, the models of high-grade lesion object detection and three categories (low-grade, high-grade and cancer) object detection based on Faster-RCNN network structure were constructed respectively. Results Based on the ResNet101 model of ImageNet pre-training, the first transfer learning was performed through the open source colposcopy data of cervical transformation zone classification, and the second transfer learning was carried out based on our own data of lesion classification to obtain the feature extractor. The average recognition accuracy mAP@IOU=0.5 obtained on the high-grade lesion model and the three categories model were 0.82 and 0.67，respectively. Conclusion Using the large sample data of the largest colposcopy center in China, the deep learning model can make a good performance in the detection of cervical precancerous lesions based on the fine labeling of epithelial and vascular features. Although there was still room for improvement on accuracy, these models were shown to be potential for cervical cancer screening, especially on guidance for location. [ABSTRACT FROM AUTHOR]

Objective To discuss the association between the different cut-off point of FRAX and BMD, and the clinical value in Wuxi region. Methods We collected the medical examination of hospital staff, we measured the BMD through DXA, simultaneously we evaluated the 10-year risk of bone fragility by FRAX (Fracture Risk Assessment Tool). The gold standard is BMD, We explored age, sex, height, weight to affect the results of FRAX by multiple linear regression analysis. Results Compared with DXA, the cut-off point of America, the Sensitivity is 3. 51%, the Specific is 99. 79%, the positive predictive value is 66. 67%, the negative predictive value is 89. 95%. Then England, the Sensitivity is 3. 51%, the Specific is 98. 58%, the positive predictive value is 22. 22%, the negative predictive value is 89. 83%; Compared with Japan, the Sensitivity is 8. 77%, the Specific is98. 38%, the positive predictive value is 38. 46%, the negative predictive value is 90. 32%. then Urumqi the Sensitivity is 5. 26%, the Specific 97. 97%, the positive predictive value is 23. 08%, the negative predictive value is 89. 94%. the greatest impact on the FRAX results is age and femoral neck T-score. Conclusion The clinical value of the different intervention cut-off point of FRAX in Wuxi region is meaningless, we should establish an appropriate cut-off point of Local region. [ABSTRACT FROM AUTHOR]

目的探讨银杏内酯B(GB)对大鼠胚胎中脑神经细胞损伤的保护作用。方法健康成年雌雄Wistar大鼠各30只,雌、雄鼠合笼受孕,取30只受孕15 d的孕鼠,无菌条件下取出胚胎中脑组织,培养中脑神经细胞。将中脑神经细胞分为对照组、神经生长因子(NGF)组(NGF组)、GB组、海人藻酸(KA)组、NGF+KA组、GB+KA组。对照组仅在培养基中培养;NGF组在培养基中加入50 mg/L NGF;GB组在培养基中加入50 mg/L GB;KA组在培养基中加入100μmol/L KA;NGF+KA组在培养基中加入50 mg/L NGF及100μmol/L KA;GB+KA组在培养基中加入50 mg/L GB及100μmol/L KA。各组均培养24、48、72、96 h。采用MTT法测定细胞存活率。采用Giemsa染色法测定中脑神经细胞集落分化率。结果 KA组48、72、96 h时细胞存活率均低于对照组(P均<0.05);NGF+KA组、GB+KA组在48、72、96 h时,细胞存活率均高于KA组(P均<0.05);NGF组、GB组在72、96 h时,细胞存活率均高于对照组(P均<0.05)。NGF组、GB组集落分化率均高于对照组(P均<0.05),KA组、NGF+KA组、GB+KA组集落分化率均低于对照组(P均<0.05),NGF+KA组、GB+KA组集落分化率均高于KA组(P均<0.05)。结论 GB能有效对抗KA对大鼠中脑神经细胞的毒性作用,其作用机制可能与GB促进中脑神经细胞的生长与分化有关。 [ABSTRACT FROM AUTHOR]