Assessment of kidney function in chronic kidney disease by combining diffusion tensor imaging and total kidney volume.

OBJECTIVE: This study aimed to investigate the value and feasibility of combining fractional anisotropy (FA) values from diffusion tensor imaging (DTI) and total kidney volume (TKV) for the assessment of kidney function in chronic kidney disease (CKD). MATERIALS AND METHODS: Fifty-one patients were included in this study. All MRI examinations were performed with a 3.0 T scanner. DTI was used to measure FA values, and TKV was obtained from DTI and T2-weighted imaging (T2WI). Patients were divided into three groups (mild, moderate, severe) according to eGFR, which was calculated with serum creatinine. Differences in the FA values of the cortex and medulla were analysed among the three groups, and the relationships of FA values, TKV, and the product of the FA values and TKV with eGFR were analysed. Receiver operating characteristic (ROC) curve analysis was used to compare the diagnostic efficiency of the FA values, TKV, and the product of the FA values and TKV for kidney function in different CKD stages. RESULTS: Medullary FA values (m-FA), TKV, and the product of the m-FA values and TKV (m-FA-TKV) were significantly correlated with eGFR (r = 0.653, 0.685, and 0.797, respectively; all P < 0.001). ROC curve analysis showed that m-FA-TKV exhibited better diagnostic performance than m-FA values (P = 0.022). CONCLUSION: m-FA-TKV obtained by DTI significantly improves the accuracy of kidney function assessment in CKD patients.

Precise anatomical localization and classification of rib fractures on CT using a convolutional neural network.

OBJECTIVE: To develop a convolutional neural network (CNN) model for the detection, precise anatomical localization (right 1-12th and left 1-12th) and classification (fresh, healing and old fractures) of rib fractures automatically, and to compare the performance with the experienced radiologists. MATERIALS AND METHODS: A total of 640 rib fracture patients with 340,501 annotations were retrospectively collected from three hospitals. They consisted of a classification training dataset (n = 482), a localization training dataset (n = 30), an internal testing dataset (n = 90) and an external testing dataset (n = 38). RetinaNet with rib localization postprocessing and the result merging technique were employed to structure the CNN model. ROC curve, free-response ROC curve, AUC, precision, recall, and F1-score were calculated to choose the better option between model I (training classification and localization data together) and model II (adding an additional classification model to model I). RESULTS: The detection and classification performance of rib fractures was better in model II than in model I. The sensitivity of localization reached 97.11% and 94.87% on the right and left ribs, respectively. In the external dataset with different CT scanner and slice thickness, model II showed better diagnostic performance. Moreover, the CNN model was superior in diagnosing fresh and healing fractures to 5 radiologists and consumed shorter diagnosis time. CONCLUSIONS: Our CNN model was capable of detection, precise anatomical localization, and classification of rib fractures automatically.

Automatic detection and classification of rib fractures based on patients' CT images and clinical information via convolutional neural network.

OBJECTIVE: To develop a convolutional neural network (CNN) model for the automatic detection and classification of rib fractures in actual clinical practice based on cross-modal data (clinical information and CT images). MATERIALS: In this retrospective study, CT images and clinical information (age, sex and medical history) from 1020 participants were collected and divided into a single-centre training set (n = 760; age: 55.8 ± 13.4 years; men: 500), a single-centre testing set (n = 134; age: 53.1 ± 14.3 years; men: 90), and two independent multicentre testing sets from two different hospitals (n = 62, age: 57.97 ± 11.88, men: 41; n = 64, age: 57.40 ± 13.36, men: 35). A Faster Region-based CNN (Faster R-CNN) model was applied to integrate CT images and clinical information. Then, a result merging technique was used to convert 2D inferences into 3D lesion results. The diagnostic performance was assessed on the basis of the receiver operating characteristic (ROC) curve, free-response ROC (fROC) curve, precision, recall (sensitivity), F1-score, and diagnosis time. The classification performance was evaluated in terms of the area under the ROC curve (AUC), sensitivity, and specificity. RESULTS: The CNN model showed improved performance on fresh, healing, and old fractures and yielded good classification performance for all three categories when both clinical information and CT images were used compared to the use of CT images alone. Compared with experienced radiologists, the CNN model achieved higher sensitivity (mean sensitivity: 0.95 > 0.77, 0.89 > 0.61 and 0.80 > 0.55), comparable precision (mean precision: 0.91 > 0.87, 0.84 > 0.77, and 0.95 > 0.70), and a shorter diagnosis time (average reduction of 126.15 s). CONCLUSIONS: A CNN model combining CT images and clinical information can automatically detect and classify rib fractures with good performance and feasibility in actual clinical practice. KEY POINTS: • The developed convolutional neural network (CNN) performed better in fresh, healing, and old fractures and yielded a good classification performance in three categories, if both (clinical information and CT images) were used compared to CT images alone. • The CNN model had a higher sensitivity and matched precision in three categories than experienced radiologists with a shorter diagnosis time in actual clinical practice.

Automatic Detection and Classification of Rib Fractures on Thoracic CT Using Convolutional Neural Network: Accuracy and Feasibility.

OBJECTIVE: To evaluate the performance of a convolutional neural network (CNN) model that can automatically detect and classify rib fractures, and output structured reports from computed tomography (CT) images. MATERIALS AND METHODS: This study included 1079 patients (median age, 55 years; men, 718) from three hospitals, between January 2011 and January 2019, who were divided into a monocentric training set (n = 876; median age, 55 years; men, 582), five multicenter/multiparameter validation sets (n = 173; median age, 59 years; men, 118) with different slice thicknesses and image pixels, and a normal control set (n = 30; median age, 53 years; men, 18). Three classifications (fresh, healing, and old fracture) combined with fracture location (corresponding CT layers) were detected automatically and delivered in a structured report. Precision, recall, and F1-score were selected as metrics to measure the optimum CNN model. Detection/diagnosis time, precision, and sensitivity were employed to compare the diagnostic efficiency of the structured report and that of experienced radiologists. RESULTS: A total of 25054 annotations (fresh fracture, 10089; healing fracture, 10922; old fracture, 4043) were labelled for training (18584) and validation (6470). The detection efficiency was higher for fresh fractures and healing fractures than for old fractures (F1-scores, 0.849, 0.856, 0.770, respectively, p = 0.023 for each), and the robustness of the model was good in the five multicenter/multiparameter validation sets (all mean F1-scores > 0.8 except validation set 5 [512 × 512 pixels; F1-score = 0.757]). The precision of the five radiologists improved from 80.3% to 91.1%, and the sensitivity increased from 62.4% to 86.3% with artificial intelligence-assisted diagnosis. On average, the diagnosis time of the radiologists was reduced by 73.9 seconds. CONCLUSION: Our CNN model for automatic rib fracture detection could assist radiologists in improving diagnostic efficiency, reducing diagnosis time and radiologists' workload.

Forecasting and predicting intussusception in children younger than 48 months in Suzhou using a seasonal autoregressive integrated moving average model.

OBJECTIVE: The aims of this study were to highlight some epidemiological aspects of intussusception cases younger than 48 months and to develop a forecasting model for the occurrence of intussusception in children younger than 48 months in Suzhou. DESIGN: A retrospective study of intussusception cases that occurred between January 2007 and December 2017. SETTING: Retrospective chart reviews of intussusception paediatric patients in a large Children's hospital in South-East China were performed. PARTICIPANTS: The hospital records of 13 887 intussusception cases in patients younger than 48 months were included in this study. INTERVENTIONS: The modelling process was conducted using the appropriate module in SPSS V.23.0. METHODS: The Box-Jenkins approach was used to fit a seasonal autoregressive integrated moving average (ARIMA) model to the monthly recorded intussusception cases in patients younger than 48 months in Suzhou from 2007 to 2016. RESULTS: Epidemiological analysis revealed that intussusception younger than 48 months was reported continuously throughout the year, with peaks in the late spring and early summer months. The most affected age group was younger than 36 months. The time-series analysis showed that an ARIMA (1,0,1 1,1,1)12 model offered the best fit for surveillance data of intussusception younger than 48 months. This model was used to predict intussusception younger than 48 months for the year 2017, and the fitted data showed considerable agreement with the actual data. CONCLUSION: ARIMA models are useful for monitoring intussusception in patients younger than 48 months and provide an estimate of the variability to be expected in future cases in Suzhou. The models are helpful for predicting intussusception cases in Suzhou and could be useful for developing early warning systems. They may also play a key role in early detection, timely treatment and prevention of serious complications in cases of intussusception.

Molecular imaging of Toll-like receptor 4 detects ischemia-reperfusion injury during intussusception.

We investigated the expression of Toll-like receptor 4 (TLR4) in the acute phase of intestinal I/R injury during intussusception and evaluated whether anti-TLR4 antibody-conjugated lead sulfide quantum dots (TLR4-PbS QDs) could be used to detect and monitor the injury. We first established a mouse model of I/R injury during intussusception. TLR-PbS QDs were then intravenously administered to intestinal I/R injured mice and visualized using whole-body fluorescence imaging in the second near-infrared window (NIR-II). Immunohistochemical analysis of intestinal tissue from the mice revealed that TLR4 expression was higher in the I/R injury group than the control and TAK-242 groups (5.189 ± 2.482, 1.186 ± 1.171, and 2.400 ± 0.857, respectively, P < 0.05). NIR-II fluorescence intensity was also higher in the I/R injury group than in the control and TAK-242 groups (86.415 ± 10.955, 38.975 ± 8.619, and 71.977 ± 3.838, respectively; P < 0.05). Thus, anti-TLR4-PbS QDs bound to TLR4 on the cell membranes of intestinal epithelial cells with high specificity in vitro and in vivo. These results indicate that TLR4 promotes intestinal I/R injury during intussusception and that the injury can be noninvasively imaged using TLR4-PbS QDs.

Risk factors for recurrent intussusception in children: a retrospective cohort study.

OBJECTIVE: The aim of this study was to assess the frequency of clinical features and pathological lead points in recurrent intussusception, with a special focus on the risk factors that lead to recurrent intussusception. DESIGN: This is a retrospective cohort study. A 5-year retrospective study was performed between January 2012 and July 2016 in the Children's Hospital of Soochow University, Suzhou, China, to determine the clinical features and pathological lead points of recurrent intussusception. SETTING: This is a retrospective chart review of recurrent intussusception cases in a large university teaching hospital. PARTICIPANTS: The medical records were obtained for 1007 cases with intussusception, including demographics, clinical signs and symptoms, imaging and recurrence times if available. INTERVENTIONS: Univariate and multivariate logistic regression analyses were used to measure significant factors affecting recurrent intussusception and recurrent intussusception with pathological lead points. RESULTS: There were 481 total episodes of recurrence in 191 patients. Among these, 87 had one recurrence and 104 had multiple recurrences. After comparing recurrent and non-recurrent intussusception cases using univariate analysis, it was determined that the factors associated with recurrent intussusception were age (>1 year), duration of symptoms (≤12 hours), the lack of bloody stool, paroxysmal crying or vomiting, the mass location (right abdomen) and pathological lead point (P<0.05). Age (>1 year), duration of symptoms (≤12 hours), the absence of vomiting, mass location (right abdomen) and pathological lead point were significantly independently predictive of recurrent intussusception. The factors associated with recurrent intussusception with lead points present were vomiting and mass location in the right abdomen (P<0.05). Vomiting and mass location (left abdomen) were significantly predictive of recurrent intussusception with lead points. CONCLUSIONS: Age (>1 year), symptom duration (≤12 hours), the absence of vomiting, mass location (right abdomen) and pathological lead points were significantly predictive of recurrent intussusception. Vomiting and mass location (left abdomen) were significantly predictive of recurrent intussusception with lead points.