Application of deep residual neural network in detecting small bowel obstruction by abdominal plain film / 西安交通大学学报(医学版)

Objective: To study whether a deep residual neural network can detect small bowel obstruction patterns on upright abdominal radiographs. Methods:The data of training set and test set used in this study were obtained from The First Affiliated Hospital of Xi'an Jiaotong University and No.215 Hospital of Shaanxi Nuclear Industry; the data of validation set came from No.215 Hospital of Shaanxi Nuclear Industry. Totally 3 298 clinical upright abdominal radiographs obtained from two hospitals were classified into obstructive and non-obstructive categories independently by two radiologists on the basis of the four signs on upright abdominal radiographs, who discussed and reached consensus when disagreements arose. Among them, 569(17.3%) images were found to be consistent with small bowel obstruction, and 2 729 (82.7%) images had no small bowel obstruction. A total of 2 305 training sets and 993 test sets (training set: test set = 2.3:1) were composed of data from the two groups, including 405 cases (17.6%) of small bowel obstruction, 1 900 cases (82.4%) of non-small bowel obstruction, 164 cases (16.5%) of small bowel obstruction, and 829 cases (83.5%) of non-small bowel obstruction. The diagnosis of small bowel obstruction in training and testing sets was based on experienced radiologists' evaluation. Totally 861 abdominal upright abdominal radiographs constituted the validation set (99 with small bowel obstruction and 762 with no small bowel obstruction); the surgical results and clinical diagnosis were set as the gold standard. In this study, the image 2012 large-scale visual recognition challenge data set (ILSVRC2012) was used for pre-training the deep residual neural network (ResNet38). The retraining of deep residual network (ResNet38) with training set data was used to establish the diagnostic model. The test set was mainly used in the learning algorithm process to adjust the algorithm parameters to modify the network, so as to make the network model more efficient. Results: After training, the deep residual neural network achieved an AUC of 0.83 on the test set (95% CI 0.82-0.92). The sensitivity of the system for small bowel obstruction was 84.1%, with a specificity of 65.0%. And on validation set it achieved an AUC of 0.87 (95% CI 0.82-0.92), the sensitivity of the system for small bowel obstruction was 89.9%, with a specificity of 68.0%. Conclusion: Transfer learning with deep residual neural network may be used to train a detector for small bowel obstruction on upright abdominal radiographs even with limited training data.

Relationship between effect of lamivudine in the treatment of cirrhotic patients with uncompensated hepatitis B with HBV genotypes and cytotoxic T lymphocyte / 中华实验和临床病毒学杂志

<p><b>OBJECTIVE</b>To explore relationship between effect of Lamivudine in the treatment of cirrhotic patients with uncompensated hepatitis B with hepatitis B virus (HBV)genotypes and HBV specific cytotoxic T lymphocytes (CTL).</p><p><b>METHODS</b>80 cases of uncompensated cirrhotic hepatitis B (40 cases with genotype B and 40 with genotype C), HBV DNA positive, HBeAg positive and human leukocyte antigen (HLA)-A2 positive,were treated with Lamivudine 100 mg/d, one year later, its effect and relationship with HBV genotypes and HBV specific CTL were observed.</p><p><b>RESULTS</b>HBV DNA turned negative:40 cases with genotype B turned negative (100%). In the 9th and 10th month of treatment, there was one case with genotype C had YMDD variation respectively and Adefovir dipivoxil was used for treatment, of the rest 38 cases, HBV DNA of 26 cases (68.42%) turned negative,HBV DNA negative rate of patients with genotype is lower than that of patients with genotype B, chi2 = 14.91, P < 0.01. HBeAg turned negative: 18 cases with genotype B (45%) turned negative, more than that of patients with genotype C (7 cases, 18.42%), chi2 = 6.32, P < 0.05. Peripheral blood HBV specific CTL level: before treatment, it was (0.33 +/- 0.03)% of patients with genotype B,higher than that of patients with genotype C [(0.11 +/- 0.02)%], t = 8.12, P < 0.001. 1 year after treatment: it was (0.44 +/- 0.04)% of patients with genotype B, higher than that before treatment, t = 4.01, P < 0.001, it was also higher than that of patients with genotype C 1 year after treatment [(0.23 +/- 0.03)%], t = 5.63, P < 0.01, alanine amino-transferase (ALT) returned to normal: 38 cases with genotype B (95%) returned to normal, more than that of patients with genotype C (28 cases, 73.68%), X2 = 6.79, P < 0.01.</p><p><b>CONCLUSION</b>Effect of Lamivudinein the treatment of cirrhotic patients with uncompensated hepatitis B is better in patients with genotype B than patients with genotype C, its mechanism may be related to lower level of HBV specific CTL in patients with genotype C than patients with genotype B.</p>

Study on SDF-1α and CD44v6 expression in multiple myeloma patients / 白血病·淋巴瘤

Objective To explore the relationship of stromal derived factor 1α (SDF-1α) and CD44variant isoforms (CD44v6) with progress of multiple myeloma (MM). Methods Bone marrow mononuclear cells(MNCs) and bone marrow stromal cells (BMSCs) from 24 cases of MM patients (14 cases of untreated and relapsed and 10 cases of stable MM patients) and 15 cases of subjects were investigated as potential SDF-1αand CD44v6 product. The level of SDF-1α and CD44v6 of the conditioned media from MM patients and subjects were analyzed by ELISA. Results The level of SDF-1α and CD44v6 from MNCs in untreated and relapsed MM patients [(7232.41 ± 2644.97) pg/ml and (34.34 ± 13.20) ng/ml] were significantly higher than stable MM patients [(2315.49 ± 748.29) pg/ml and (15.69 ± 5.28) ng/ml] (t =6.25, t= 7.82, P ＜0.05) and 15 subjects [(1149.52 ± 636.06) pg/ml and (4.85 ± 3.62) ng/ml] (t= 4.60, t = 7.61, P＜ 0.05). The level of SDF-1α in stable MM patients was different from healthy subjects (P ＜0.05), but the level of CD44v6 in stable MM patients was not different from controls. The level of SDF-1α and CD44v6 in stable MM patients were significantly higher than health subjects (t = 2.99, t= 4.87, P ＜0.05). The level of SDF-1α was also detected from BMSCs of MM patients. When human MM cell lines U266 were adhered to BMSCs of 9 untreated and relapsed MM patients,and added rhIL-6 to it, there was significant increase of SDF-1α, compared with BMSCs in subjects and MM patients. The expression level of SDF-lα was correlated with the level of CD44v6 (r =0.51, P =0.03). Conclusion The increase of SDF-1α (may be produced by myeloma cells and BMSCs) and CD44v6 (may be produced by myeloma cells) activity is associated with the progress or pathogenesis of MM, and may be involved with tumor invasion. The completion of these processes in vivo may need participation of myeloma cells, BMSCs, IL-6,SDF-lα and CD44v6.

A study on the treatment of chronic hepatitis B with YMDD mutation / 中华肝脏病杂志

<p><b>OBJECTIVE</b>To explore the strategy for the treatment of chronic hepatitis B with YMDD mutation.</p><p><b>METHODS</b>A total of 120 chronic hepatitis B patients with YMDD mutation were randomly assigned into four groups. In group A, patients received adefovir dipivoxil for 48 weeks. In group B, patients received adefovir dipivoxil in combination with lamivudine during the first 12 weeks and adefovir dipivoxil only for the following 36 weeks. In group C, patients received adefovir dipivoxil in combination with lamivudine for 48 weeks. In group D, patients received entecavir for 48 weeks.</p><p><b>RESULTS</b>The rate of rebound of alanine aminotransferase (ALT) was 30.0% (9/30), 10.0% (3/30), 6.7% (2/30), 10.0% (3/30) (P < 0.05) during the first 12 weeks, and one patient with severe hepatitis was found in group A. The positive rate of YMDD mutation was 17.9%, 0, 0, 0 at week 12. There was no significant difference in the level of ALT and the rate of HBeAg seroconversion after 48-week treatment (P > 0.05). At week 48, there was significant difference in the ALT normalization rate and undetectable HBV DNA rate between group C and group A, and also between group D and group A, and the rate of drug resistant genotype was 6.9%, 6.7%, 0, 0. Two patients had rtN236T mutation in group A, and one patient had rtN236T mutation and another one had rtA181V mutation in group B.</p><p><b>CONCLUSION</b>Adefovir dipivoxil in combination with lamivudine or entecavir are safe and effective therapies for chronic hepatitis B patients with YMDD mutation.</p>