ABSTRACT
Objective:To evaluate artificial intelligence constructed by deep convolutional neural network (DCNN) for the site identification in upper gastrointestinal endoscopy.Methods:A total of 21 310 images of esophagogastroduodenoscopy from the Cancer Hospital of Chinese Academy of Medical Sciences from January 2019 to June 2021 were collected. A total of 19 191 images of them were used to construct site identification model, and the remaining 2 119 images were used for verification. The performance differences of two models constructed by DCCN in the identification of 30 sites of the upper digestive tract were compared. One model was the traditional ResNetV2 model constructed by Inception-ResNetV2 (ResNetV2), the other was a hybrid neural network RESENet model constructed by Inception-ResNetV2 and Squeeze-Excitation Networks (RESENet). The main indices were the accuracy, the sensitivity, the specificity, positive predictive value (PPV) and negative predictive value (NPV).Results:The accuracy, the sensitivity, the specificity, PPV and NPV of ResNetV2 model in the identification of 30 sites of the upper digestive tract were 94.62%-99.10%, 30.61%-100.00%, 96.07%-99.56%, 42.26%-86.44% and 97.13%-99.75%, respectively. The corresponding values of RESENet model were 98.08%-99.95%, 92.86%-100.00%, 98.51%-100.00%, 74.51%-100.00% and 98.85%-100.00%, respectively. The mean accuracy, mean sensitivity, mean specificity, mean PPV and mean NPV of ResNetV2 model were 97.60%, 75.58%, 98.75%, 63.44% and 98.76%, respectively. The corresponding values of RESENet model were 99.34% ( P<0.001), 99.57% ( P<0.001), 99.66% ( P<0.001), 90.20% ( P<0.001) and 99.66% ( P<0.001). Conclusion:Compared with the traditional ResNetV2 model, the artificial intelligence-assisted site identification model constructed by RESENNet, a hybrid neural network, shows significantly improved performance. This model can be used to monitor the integrity of the esophagogastroduodenoscopic procedures and is expected to become an important assistant for standardizing and improving quality of the procedures, as well as an significant tool for quality control of esophagogastroduodenoscopy.
ABSTRACT
Objective@#To summarize the global research hotspots and trends of family intervention on autism spectrum disorder (ASD), to provide ideas for the development of localized research on family intervention for ASD in China.@*Methods@#This study used the bibliometric software CiteSpace to conduct visual analysis on emerging cutting edge documents in the Web of Science core database in the field of family intervention for autism in the past 20 years (2001-2020). Data collection used 7 subject terms related to family caregiver intervention in February 2021. A total of 2 342 literature was obtained after data sorting.@*Results@#The number of international publications has risen rapidly since 2013. The average annual number of international publications was 29 during 2001- 2012, then 250 during 2013-2020, which increased by 8 times. Among them, the United States ranked the top of publication with 1 295 articles, while publications from two universities in Canada and the United Kingdom showed higher impacts. The high frequency keywords that reflected the international hotspots included children, adolescent, mother, behavior, and stress; the high centrality keywords included follow up, intellectual disability, early recognition, young children and language.@*Conclusion@#The research hotspots in the field of ASD family intervention in the past 20 years included targets, strategies and forms, as well as the implementation effects of ASD family interventions. The trends of research in the past five years (2016-2020) mainly focus on the intervention studies of children, adolescents and adults with ASD and the randomized controlled trials of ASD family intervention focusing on caregivers.
ABSTRACT
Objective:To explore the value of multi-parametric MRI for thyroid gland in differentiating benign and malignant thyroid nodules.Methods:From December 2018 to May 2020, 78 patients with 91 post-surgically pathologically confirmed thyroid nodules were enrolled in Cancer Hospital & Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College. For each patient, the following MRI findings were obtained including the nodules′ location, size, shape, margin, signal intensity, cystic change, degree and pattern of contrast enhancement, involvement of surrounding structure, and ADC values. The time-intensity curve (TIC) were plotted and subtyped based on dynamic contrast enhancement MRI. The MRI findings between the benign and malignant thyroid nodules were compared using Mann-Whitney U test, χ 2 test or Fisher exact test. Multiple logistic regression analysis was used to select independent predictive variables and build a combined model, and the ROC curve was used to evaluate the diagnostic performance of each MRI finding and the combined model. Results:Between the benign and malignant thyroid nodules, the significant differences were found in size, shape, margin, presence of cystic changes, T 1WI signal intensity, ADC value, enhancement homogeneity, TIC subtypes and presence of thyroid capsule involvement ( P<0.05). Multivariate logistic analysis showed that ill-defined margin (OR=77.61), no presence of cystic changes (OR=36.11) and difference between TIC subtypes (OR=83.41) were independent predictive variables, and the area under the ROC curve (AUC) was 0.879, 0.788, and 0.751, respectively. The AUC, sensitivity and specificity of the combined model were 0.977, 0.986, and 0.904, respectively. Conclusions:Thyroid multi-parametric MRI derived findings can be used for the differential diagnosis of benign and malignant nodules. Combined with the independent risk factors with ill-defined margin, no presence of cystic changes, TIC of type plateau or washout, the diagnostic model has a higher diagnostic efficiency.
ABSTRACT
Objective@#The random forest algorithm was used to construct a rapid screening diagnostic prediction model for children with autism spectrum disorder, to provide the references for early detection, early diagnosis of ASD children, and to reduce the pressure of ASD clinical diagnosis and assessment.@*Methods@#The random forest algorithm of machine learning was applied to build the auxiliary diagnosis model. Totally 346 ASD children and 90 normal children were evaluated by Social Responsiveness Scale and Vineland Adaptive Behavior Scales. ROC curve, and accuracy was used to evaluate the models.@*Results@#Among the models, the accuracy of 13 feature factors and 7 feature factors were above 0.9, the sensitivity was up to 0.927, the specificity was up to 0.936 and the AUC was up to 0.979. The accuracy, sensitivity, specificity and AUC of the model were 0.943,0.959,0.931 and 0.978 respectively. The fitting and generalization effects of the three models were all satisfactory.@*Conclusion@#A random forest model based on the SRS Scales and Vineland Adaptive Behavior Scales can be used to diagnose ASD accurately and provide scientific basis for the development of rapid screening and diagnosis tools.
ABSTRACT
Objective@#There is no effective therapy for patients with advanced medullary thyroid carcinoma (MTC). Vandetanib,a novel multitargeted receptor tyrosine kinase inhibitor, has previously shown antitumor activity in phase Ⅱ studies of patients with advanced MTC. This study was to evaluate the efficacy and the safety of vandetanib on advanced MTC.@*Methods@#This study was an open, international multi-center phase Ⅲ clinical trial and the study number was NCT01298323. The single-center study was a sub-group analysis of the international study, which was conducted on 9 pathologically confirmed advanced MTC patients by Cancer Hospital Chinese Academy of Medical Sciences between March 2012 and October 2017. Vandetanib (300 mg) was orally administered daily till death or withdrawal. The efficacy was evaluated according to RECIST criteria and the adverse events were evaluated according to NCI criteria.@*Results@#The objective response rate was 3/9,and the disease control rate was 4/9. The median progression-free survival was 44 months. All patients who had the elevated levels of calcitonin (CTN) and carcino-embryonic antigen (CEA) before treatment began to show the decreases in the level of CTN and CEA after 3 months and later showed again the increases in the levels of both tumor markers with tumor progression. By ROC curve analysis, CTN was of statistically significance(P<0.05, 95%CI 0.558-0.834), but CEA was not(P>0.05). Adverse events were generally mild (grade 1 or 2),including hypertension (9 cases),skin rash (9 cases), and diarrhea (6 cases). Two patients developed grade 3 elevation of serum glutamate pyruvate transaminase and one patient developed grade 3 elevation of drug-related bowel disease. No grade 4 drug-related adverse event occurred.@*Conclusions@#Vandetanib is effective and well tolerated for patients with locally advanced or metastatic MTC who have no chance for surgery. This indicates the increase of CTN is clinically relevant to disease progression, but the number of patients are extremely low, and, therefore further research is needed. Long-term use of vandetanib may cause resistance.
ABSTRACT
Objective@#To investigate the value of jugulo-omohyoid lymph nodes (JOHLN) in predicting occult lateral cervical lymph node metastasis in patients with papillary thyroid carcinoma (PTC).@*Methods@#The clinicopathological data of 136 out of 2 100 PTC patients, who had a high risk of lateral neck lymph node metastasis and treated by us from January 2010 to December 2015, were retrospectively analyzed. Super selective neck dissection (SSND, level Ⅲ and Ⅳ)was performed and JOHLNs were sent for frozen section in all the 136 cases. The clinicopathological data was analyzed and the significance of JOHLN in predicting lateral cervical LNM was calculated using the SPSS software package.@*Results@#Of the 136 patients, total thyroidectomy was performed in 76 cases (55.9%) and unilateral lobectomy plus isthmus was performed in the other 60 cases (44.1%). SSND was performed in 72 patients (52.9%), level Ⅱ-Ⅳ dissection in 15 (11.0%), and level Ⅱ-Ⅴ dissection in 49 (36.0%). According to the pathological results, 38 patients were pN0(27.9%), 18 (13.2%) were pN1a and 80 (58.8%) were pN1b. The lymph node metastasis(LNM) rates at level Ⅱ-Ⅵ were 19.9%, 43.4%, 42.6%, 2.9%, and 59.6%, respectively. The sensitivity, specificity and accuracy of JOHLN in predicting lateral neck metastasis were 58.8%, 62.9%, and 76.7%, respectively. The rates for predicting level Ⅱ metastasis were 81.5%, 43.2%, and 59.4%, respectively. None of the patients died in the follow-up. Only 1 recurrence was found in level Ⅱ and regional control was achieved after level Ⅱ and Ⅴ dissection.@*Conclusions@#JOHLN has a high accuracy for predicting lateral cervical lymph node metastasis and high sensitivity for level Ⅱ metastasis. For patients with high risk of lateral cervival metastasis, super-selective neck dissection including level Ⅲ and Ⅳ can confirm the stage and reduce the risk of reoperation. Dissection for level Ⅱ, Ⅲ, and Ⅳ is recommended.