2.Retrospectively analysis of the vagus nerve reflex in bronchial artery embolization
Zhigang FU ; Chengxin YU ; Xiaolin ZHANG ; Haitao LI ; Qiang HAN ; Xiaohu QI ; Wenjiang ZHAO
Journal of Practical Radiology 2016;32(3):429-431
Objective To evaluate the cause and the treatment of the vagus nerve reflex in patients with hemoptysis during bron-chial artery embolization (BAE).Methods 1 12 patients with much hemoptysis were enrolled,9 of whom represented vagus nerve reflex in the process of interventional embolization.Results In 9 patients with mixed vagal reflex,5 occurred in the process of bron-chial artery embolization,1 in removing of sheath,1 in hemostasis by compression and 2 in returning to the ward.The intraoperative vagus reflex during BAE was related to over tension and unnormolized operation,and it improved by block of vagus nerve,raising blood pressure and fluid expansion without serious complications.Conclusion Vagus nerve reflex during BAE should be noticed, and early detection and timely intervention may improve its prognosis.
3.Research progress on target delineation for radiotherapy in hepatocellular carcinoma with
Haimin LIN ; Chengxin LIU ; Dali HAN ; Jinming YU
Chinese Journal of Radiation Oncology 2019;28(7):551-554
Modern medical imaging techniques,such as computed tomorgraphy (CT),magnetic resonance imaging (MRI) and position emission tomorgraphy/computed tomorgraphy (PET-CT) can accurately delineate the gross target volume (GTV) of hepatocellular carcinoma (HCC).Comparison of postoperative pathological subclinical lesions,imaging and clinical parameters contributes to the precise delineation of clinical target volume (CTV).Moreover,radiotherapy-assisted techniques,such as fourdimensional computed tomography (4DCT),compression of abdomen,active breathing control and respiratory gating,can minimize the internal target volume (ITV).In addition,immobilization with vacuum cushion and body membrane can reduce the set-up error,minimize the planning target volume (PTV) and avoid or decrease the irradiation error or missing irradiation.All these approach can minimize the target volume,elevate the dose and reduce the complications during radiotherapy for HCC.In this article,the research progress on the target delineation for external beam radiotherapy in HCC patients was reviewed.
4.Deep learning for classification of multi?sequence MR images of the prostate
Junhua FANG ; Qiubai LI ; Chengxin YU ; Xinggang WANG ; Zhihua FANG ; Tao LIU ; Liang WANG
Chinese Journal of Radiology 2019;53(10):839-843
Objective To develop a convolution neural network (CNN) model to classify multi?sequence MR images of the prostate. Methods ResNet18 convolution neural network (CNN) model was developed to classify multi?sequence MR images of the prostate. A deep residual network was used to improve training accuracy and test accuracy. The dataset used in this experiment included 19 146 7?sequence prostate MR images (transverse T1WI, transverse T2WI, coronal T2WI, sagittal T2WI, transverse DWI, transverse ADC, transverse PWI), from which a total of 2 800 7?sequence MR images was selected as a training set. Three hundred and eighty eight 7?sequence MR images were selected as test sets. Accuracy was used to evaluate the effectiveness of ResNet18 CNN model. Results The classification accuracy of the model for transverse DWI, sagittal T2WI, transverse ADC, transverse T1WI, and transverse T2WI was as high as 100.0% (44/44,52/52), and the accuracy for transverse PWI was also as high as 96.7% (116/120). The accuracy for coronal T2WI was 77.5% (31/40). 0.8% (1/120) of transverse PWI was incorrectly assigned to transverse T2WI, and 2.5% (3/120) incorrectly assigned to sagittal T2WI. 15.0% (6/40) of coronal T2WI was incorrectly assigned to transverse T2WI, and 7.5% (3/40) to sagittal T2WI. Conclusion The experimental results show the effectiveness of our deep learning method regarding accuracy in the prostate multi?sequence MR images detection.
5.Mining and Analysis of the Formulation Rules of Chinese Patent Medicine for Cold Based on 2015 Edition of Chinese Pharmacopeia(Part Ⅰ)
Haidu HONG ; Chengxin LIU ; Yu HONG ; Huiting HUANG ; Dongting LI ; Yue PAN ; Si CHEN ; Chuangrong CHEN ; Xiaohong LIU
China Pharmacy 2019;30(13):1812-1816
OBJECTIVE: To mine and analyze the formulation rules of Chinese patent medicine for cold, and to provide reference for clinical dialectical medication and R&D of new medicine for cold. METHODS: The name, dosage form, formulation, function of curing of Chinese patent medicines for cold were collected from 2015 edition of Chinese Pharmacopeia (part Ⅰ) and then input into TCM Inheritance Assistant Platform V 2.5; use frequency of TCM were counted. Apriori algorithm and association rules were used to analyze the core medicinal material combination (10% support and 0.65 confidence). New formulation combinations were extracted by unsupervised entropy hierarchical clustering method. RESULTS: A total of 130 kinds of Chinese patent medicine (196 Chinese patent medicine with the same prescription and different dosage forms) for treating cold were collected, including granules (47), pills (32), tablets (31), mixtures (31), etc. 264 medicinal materials were involved. The cold syndromes contained wind-heat syndrome, wind-cold syndrome, summer-dampness syndrome and Qi deficiency syndrome. Top 3 medicinal materials in the list of use frequency were Glycyrrhiza uralensis (45.38%), Scutellaria baicalensis (32.31%) and Platycodon grandiflorus (31.54%). There were 28 core medicinal material combinations, among which the top 3 were G. uralensis-P. grandiflorus, Mentha haplocalyx-P. grandiflorus and S. baicalensis-Forsythia suspensa. New combinations were excavated, including Nepeta cataria-P. grandiflorus-M. haplocalyx-Citrus reticulate-Folium Perillae-Citrus aurantium- Poria cocos, F. suspense-S. baicalensis-Lonicera japonica- Arctium lappa-Fermented soybean. CONCLUSIONS: This study analyzed the formulation rules of Chinese patent medicine for treating cold by using the TCM inheritance assistant platform V2.5, which can provide reference for clinical dialectical medication and R&D of new medicines for cold.
6. Robotic and endoscopic cooperative surgery in the third space for the resection of gastric submucosal tumors
Chengxin SHI ; Yingchao LI ; Qi SUN ; Feiyu SHI ; Yaguang LI ; Tianyu YU ; Qian QIN ; Hong WU ; Guanghui WANG ; Junjun SHE
Chinese Journal of General Surgery 2019;34(11):952-955
Objective:
To evaluate combined robotic and endoscopic surgery in the third space for gastric submucosal tumors(SMTs).
Methods:
Combined surgery in 4 patients were compared with 19 patients who underwent laparoscopic wedge resection between Aug 2017 and Feb 2018.
Results:
R0 resection was achieved in all combined surgery patients. The operation time was longer (112±29 )min
7. Deep learning for classification of multi-sequence MR images of the prostate
Junhua FANG ; Qiubai LI ; Chengxin YU ; Xinggang WANG ; Zhihua FANG ; Tao LIU ; Liang WANG
Chinese Journal of Radiology 2019;53(10):839-843
Objective:
To develop a convolution neural network (CNN) model to classify multi-sequence MR images of the prostate.
Methods:
ResNet18 convolution neural network (CNN) model was developed to classify multi-sequence MR images of the prostate. A deep residual network was used to improve training accuracy and test accuracy. The dataset used in this experiment included 19 146 7-sequence prostate MR images (transverse T1WI, transverse T2WI, coronal T2WI, sagittal T2WI, transverse DWI, transverse ADC, transverse PWI), from which a total of 2 800 7-sequence MR images was selected as a training set. Three hundred and eighty eight 7-sequence MR images were selected as test sets. Accuracy was used to evaluate the effectiveness of ResNet18 CNN model.
Results:
The classification accuracy of the model for transverse DWI, sagittal T2WI, transverse ADC, transverse T1WI, and transverse T2WI was as high as 100.0% (44/44,52/52), and the accuracy for transverse PWI was also as high as 96.7% (116/120). The accuracy for coronal T2WI was 77.5% (31/40). 0.8% (1/120) of transverse PWI was incorrectly assigned to transverse T2WI, and 2.5% (3/120) incorrectly assigned to sagittal T2WI. 15.0% (6/40) of coronal T2WI was incorrectly assigned to transverse T2WI, and 7.5% (3/40) to sagittal T2WI.
Conclusion
The experimental results show the effectiveness of our deep learning method regarding accuracy in the prostate multi-sequence MR images detection.