Combination of multi-modal MRI radiomics and liquid biopsy technique for preoperatively non-invasive diagnosis of glioma based on deep learning: protocol for a double-center, ambispective, diagnostical observational study

Background

2021 World Health Organization (WHO) Central Nervous System (CNS) tumor classification increasingly emphasizes the important role of molecular markers in glioma diagnoses. Preoperatively non-invasive "integrated diagnosis" will bring great benefits to the treatment and prognosis of these patients with special tumor locations that cannot receive craniotomy or needle biopsy. Magnetic resonance imaging (MRI) radiomics and liquid biopsy (LB) have great potential for non-invasive diagnosis of molecular markers and grading since they are both easy to perform. This study aims to build a novel multi-task deep learning (DL) radiomic model to achieve preoperative non-invasive "integrated diagnosis" of glioma based on the 2021 WHO-CNS classification and explore whether the DL model with LB parameters can improve the performance of glioma diagnosis.

Methods

This is a double-center, ambispective, diagnostical observational study. One public database named the 2019 Brain Tumor Segmentation challenge dataset (BraTS) and two original datasets, including the Second Affiliated Hospital of Nanchang University, and Renmin Hospital of Wuhan University, will be used to develop the multi-task DL radiomic model. As one of the LB techniques, circulating tumor cell (CTC) parameters will be additionally applied in the DL radiomic model for assisting the "integrated diagnosis" of glioma. The segmentation model will be evaluated with the Dice index, and the performance of the DL model for WHO grading and all molecular subtype will be evaluated with the indicators of accuracy, precision, and recall.

Discussion

Simply relying on radiomics features to find the correlation with the molecular subtypes of gliomas can no longer meet the need for "precisely integrated prediction." CTC features are a promising biomarker that may provide new directions in the exploration of "precision integrated prediction" based on the radiomics, and this is the first original study that combination of radiomics and LB technology for glioma diagnosis. We firmly believe that this innovative work will surely lay a good foundation for the "precisely integrated prediction" of glioma and point out further directions for future research.

Clinical trail registration

This study was registered on ClinicalTrails.gov on 09/10/2022 with Identifier NCT05536024.

Efficacy and safety of vagus nerve stimulation on upper extremity motor function in patients with stroke: A meta-analysis of randomized controlled trials

BACKGROUND

In 2021, the U.S. Food and Drug Administration (FDA) approved paired vagus nerve stimulation (VNS) for patients with moderate-to-severe upper extremity motor impairments following chronic ischemic stroke.

OBJECTIVE

Previous meta-analyses have shown that VNS may impact stroke rehabilitation, but each has some limitations.

METHODS

PubMed, Ovid, Cochrane Library, ScienceDirect, Web of Science and WHO ICTRP databases were searched until July 14, 2022 for randomized controlled trials (RCTs). We defined primary outcomes as Fugl-Meyer Assessment for Upper Extremity (FMA-UE) and Wolf Motor Function Test (WMFT). Subgroup analyses included types of VNS, time since onset and long-term effects. Secondary outcomes included adverse events of VNS.

RESULTS

Eight RCTs involving 266 patients were analyzed, of which five used direct VNS and three transcutaneous auricular VNS. The results revealed that VNS enhanced upper extremity function via FMA-UE (SMD = 0.73; 95% CI: 0.48 to 0.99; P < 0.00001) and WMFT (SMD = 0.82; 95% CI:0.52 to 1.13; P < 0.00001) in comparison to the control group, but showed no significant change on long-term effects of FMA-UE (SMD = 0.69; 95% CI: - 0.06 to 1.44; P = 0.07). There was no difference in adverse events between the VNS and control groups (RR = 1.16; 95% CI: 0.46 to 2.92; P = 0.74).

CONCLUSION

For stroke victims with upper limb disabilities, VNS paired with rehabilitation was significantly safe and effective. More high-quality multicentric RCTs are needed to validate this conclusion.

[Status and related factors on the drinking behavior among primary and secondary students in China rural middle and western regions in 2019]

Objective: To analyze the daily drinking behavior and related factors of primary and middle school students in the Nutrition Improvement Program for Rural Compulsory Education Students (NIPRCES) pilot regions. Methods: Multi-stage stratified random cluster sampling method was used to select one to three national pilot counties in 22 provinces in central and western China where the NIPRCES was implemented in 2019. According to different feeding patterns, two primary schools and two middle schools were selected as key monitoring schools. One or two classes were selected from grade 3 to grade 9. The student questionnaire was used to collect the basic information and daily drinking behavior. Taking whether the drinking water ≥5 cups every day as the dependent variable, multivariate logistic regression model was used to analyze the related factors of drinking behavior among students. Results: A total of 27 374 students were included. On average, primary and middle school students in the regions where NIPRCES was implemented had 3.9 cups of water every day. Logistic regression model showed that boys (OR=1.230, P<0.001), primary school students (OR=1.379, P<0.001), father worked outside the home (OR=1.169, P<0.001), both parents worked outside the home (OR=1.228, P<0.001), non-resident students (OR=1.142, P<0.001), the school in the village (OR=1.638, P<0.001) or township (OR=1.358, P<0.001), school feeding (OR=1.252, P<0.001), the school building with flush toilets (OR=1.384, P<0.001) and the central regions (OR=1.300, P<0.001) students were more likely to drink ≥5 cups water every day. Conclusion: The water consumption of primary and middle school students in the pilot regions of NIPRCES is low, and their drinking behaviors are affected by many factors.

Artificial neuromuscular fibers by multilayered coaxial integration with dynamic adaption

Integrating sense in a thin artificial muscle fiber for environmental adaption and actuation path tracing, as a snail tentacle does, is highly needed but still challenging because of the interfacing mismatch between the fiber's actuation and sensing components. Here, we report an artificial neuromuscular fiber by wrapping a carbon nanotube (CNT) fiber core in sequence with an elastomer layer, a nanofiber network, and an MXene/CNT thin sheath, achieving the ingenious sense-judge-act intelligent system in an elastic fiber. The CNT/elastomer components provide actuation, and the sheath enables touch/stretch perception and hysteresis-free cyclic actuation tracing due to its strain-dependent resistance. As a whole, the coaxial structure builds a dielectric capacitor that enables sensitive touchless perception. The key to seamless integration is to use a nanofiber interface that allows the sensing layer to adaptively trace but not restrict actuation. This work provides promising solutions for closed-loop control for future intelligent soft robots.

PET/CT-based radiomics analysis may help to predict neoadjuvant chemotherapy outcomes in breast cancer

Background

The aim of this study was to evaluate the clinical usefulness of radiomics signature-derived 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography–computed tomography (PET-CT) for the early prediction of neoadjuvant chemotherapy (NAC) outcomes in patients with (BC).

Methods

A total of 124 patients with BC who underwent pretreatment PET-CT scanning and received NAC between December 2016 and August 2019 were studied. The dataset was randomly assigned in a 7:3 ratio to either the training or validation cohort. Primary tumor segmentation was performed, and radiomics signatures were extracted from each PET-derived volume of interest (VOI) and CT-derived VOI. Radiomics signatures associated with pathological treatment response were selected from within a training cohort (n = 85), which were then applied to generate different classifiers to predict the probability of pathological complete response (pCR). Different models were then independently tested in the validation cohort (n = 39) regarding their accuracy, sensitivity, specificity, and area under the curve (AUC).

Results

Thirty-five patients (28.2%) had pCR to NAC. Twelve features consisting of five PET-derived signatures, four CT-derived signatures, and three clinicopathological variables were candidates for the model’s development. The random forest (RF), k-nearest neighbors (KNN), and decision tree (DT) classifiers were established, which could be utilized to predict pCR to NAC with AUC ranging from 0.819 to 0.849 in the validation cohort.

Conclusions

The PET/CT-based radiomics analysis might provide efficient predictors of pCR in patients with BC, which could potentially be applied in clinical practice for individualized treatment strategy formulation.

Stepwise Artificial Yarn Muscles with Energy-Free Catch States Driven by Aluminum-Ion Insertion

Present artificial muscles have been suffering from poor actuation step precision and the need of energy input to maintain actuated states due to weak interactions between guest and host materials or the unstable structural changes. Herein, these challenges are addressed by deploying a mechanism of reversible faradaic insertion and extraction reactions between tetrachloroaluminate ions and collapsed carbon nanotubes. This mechanism allows tetrachloroaluminate ions as a strong but dynamic "locker" to achieve an energy-free high-tension catch state and programmable stepwise actuation in the yarn muscle. When powered off, the muscle nearly 100% maintained any achieved contractile strokes even under loads up to 96,000 times the muscle weight. The actuation mechanism allowed the programmable control of stroke steps down to 1% during reversible actuation. The isometric stress generated by the yarn muscle (14.6 MPa in maximum, 40 times that of skeletal muscles) was also energy freely lockable and step controllable with high precision. Importantly, when fully charged, the muscle stored energy with a high capacity of 102 mAh g^-1, allowing the muscle as a battery to power secondary muscles or other devices.

Biomechanical evaluation of lumbar spondylolysis repair with various fixation options: A finite element analysis

Objective: This study was designed to compare the biomechanical properties of lumbar spondylolysis repairs using different fixation methods by using three-dimensional finite element analysis.

Methods: Five finite element models (A, B, C, D, and E) of L4-S1 vertebral body were reconstructed by CT images of a male patient (A: intact model; B: spondylolysis model; C: spondylolysis model with intrasegmental direct fixation by Buck screw; D: spondylolysis model with intersegmental indirect fixation by pedicle screw system; E: spondylolysis model with hybrid internal fixation). L5-S1 level was defined as the operative level. After the intact model was verified, six physiological motion states were simulated by applying 500 N concentrated force and 10 Nm torque on the upper surface of L4. The biomechanical properties of the three different internal fixation methods were evaluated by comparing the range of motion (ROM), maximum stress, and maximum displacement.

Results: Compared with Model B, the ROM and maximum displacement of Model C, D, and E decreased. The maximum stress on L5/S1 disc in models A, B, and C was much higher than that in Model D and E under extension and lateral bending conditions. Under axial rotation and lateral bending conditions, the maximum stress of interarticular muscle and internal fixation system in Model B and Model C was significantly higher than that in Model D and Model E. In contrast to Model D, the stress in Model E was distributed in two internal fixation systems.

Conclusion: In several mechanical comparisons, hybrid fixation had better biomechanical properties than other fixation methods. The experimental results show that hybrid fixation can stabilize the isthmus and reduce intervertebral disc stress, which making it the preferred treatment for lumbar spondylolysis.

Imbalance of Circulating Monocyte Subsets in Subjects with Newly Emerged and Recurrent Hospital-Acquired Pneumonia

Objective

Hospital-acquired pneumonia (HAP) is one of the most common diseases in the intensive care unit, where the development of disease is closely related with the host immune response. Monocytes play an important role in both innate and adaptive immune system. We aimed to investigate the changes of circulating monocyte subsets in subjects with HAP to explore its value in monitoring HAP.

Methods

In total, 60 HAP patients and 18 healthy individuals were enrolled in this study. Human monocyte subsets are classified into 3 groups: nonclassical (NC), intermediate (ITM), and classical (CL). Also, programmed death ligand 1 (PD-L1) expression on circulating monocyte subsets was measured by flow cytometry.

Results

Data showed that the ratio of NC, ITM, and CL among monocytes was comparable between HAP patients and healthy controls (P &gt; .05). There was a remarkable imbalance of NC and CL in newly emerged HAP compared to healthy controls (P &lt; .05), subsequently reaching normalization in recurrent HAP (P &gt; .05). Furthermore, although PD-L1 was seemly constitutively expressed by NC, ITM, and CL groups regardless of disease status, it was noted that PD-L1 was dominantly expressed in the CL group (P &lt; .05).

Conclusion

Given distinct PD-L1 expression, a shift of CL/NC in newly emerged HAP would constitute an inhibitory anti-pathogen immune response. Normalization of circulating monocyte subsets on recurrence of HAP might be the consequence of immune memory of bacterial infection.

Salvianolic acid B combined with bone marrow mesenchymal stem cells piggybacked on HAMA hydrogel re-transplantation improves intervertebral disc degeneration

Cell-based tissue engineering approaches have emerged as a realistic alternative for regenerative disc tissue repair. The multidirectional differentiation potential of bone marrow mesenchymal stem cells (BMSCs) to treat disc degeneration intervertebral disc degeneration has also become a viable option. We used 1% HAMA hydrogel as a carrier and co-encapsulated BMSCs and Salvianolic acid B (SalB) into the hydrogel to reduce the apoptosis of the transplanted cells. The protective effect of SalB on BMSCs was first verified in vitro using the CCK8 method, flow cytometry, and Western-Blotting, and the physical properties and biocompatibility of HAMA hydrogels were verified in vitro. The rat model was then established using the pinprick method and taken at 4 and 8 W, to examine the extent of disc degeneration by histology and immunohistochemistry, respectively. It was found that SalB could effectively reduce the apoptosis of BMSCs in vitro by activating the JAK2-STAT3 pathway. 1% HAMA hydrogels had larger pore size and better water retention, and the percentage of cell survival within the hydrogels was significantly higher after the addition of SalB to the HAMA hydrogels. In the in vivo setting, the HAMA + SalB + BMSCs group had a more pronounced delaying effect on the progression of disc degeneration compared to the other treatment groups. The method used in this study to encapsulate protective drugs with stem cells in a hydrogel for injection into the lesion has potential research value in the field of regenerative medicine.

[Overweight and obesity status and its associated factors among primary and secondary school students in China rural middle and western regions]

Objective: To analyze the overweight and obesity status of students in the national pilot counties of the Nutrition Improvement Program for Rural Compulsory Education Students in 2019 and its associated factors. Methods: In 2019, a multi-stage cluster random sampling method was used to select about 40 students from each grade in primary and secondary schools in China's central and western regions where the Nutrition Improvement Program for Rural Compulsory Education Students was implemented. The height and weight of the children were measured using height or weight scales. The school questionnaire and county questionnaire were used to investigate the associated factors. A Chi-square test was used for comparison between groups. The logistic regression analysis was used to analyze the associated factors. Results: In 2019, the prevalence of overweight and obesity among rural primary and secondary school students aged 6-15 years in central and western China 2019 was 11.5%. It was higher for boys (13.1%) than that for girls (9.8%), higher in central (14.3%) than that in the west (9.9%) and higher for elementary school students (12.4%) than that for secondary school students (9.5%, all P<0.001). The logistic regression showed that boys (OR=1.388), primary school students (OR=1.271), students without other dietary subsidies(OR=1.037), schools in rural areas (OR=1.133), schools with enterprise-based feeding mode (OR=1.043), schools without the provision of lunch (OR=1.143), schools without the provision of dinner (OR=1.122), and schools without providing drinking water (OR=1.015) were positively associated with overweight and obesity among students (P<0.05). Schools with snack shops (OR=0.952) were negatively associated with overweight and obesity among students (P<0.001). Conclusion: A certain proportion of primary and secondary school students in rural areas of central and western China are overweight and obese. The prevalence is not only related to children's gender, school section and county area but also related to school meals, whether schools provide drinking water and other factors.

Comparison of the Endophytic Bacterial Microbiota of Asymptomatic and Symptomatic Ginger Rhizomes During the Activation of Adventitious Bud Development

Bacterial infections are the cause of rhizome rot in ginger (Zingiber officinale). Key members of the endophytic microbial community in ginger rhizomes have not been identified, and their impact on the decay of rhizomes during the activation of adventitious bud development has not been investigated. High-throughput, 16S rRNA amplicon sequencing and inoculation experiments were used to analyze the microbial diversity, community structure and composition, and pathogenicity of isolated bacteria. Our results indicated that the composition of the endophytic microbiota underwent a shift during the progression of rhizome rot disease. Enterobacteriaceae, Lachnospiraceae, and the bacterial genera Clostridium, Bacteroides, Acrobacter, Dysgonomonas, Anaerosinus, Pectobacterium, and Lactococcus were relatively abundant in the bacterial community of rhizomes exhibiting bacterial decay symptoms but were also present in asymptomatic rhizomes. The presence of Enterobacteriaceae and Pseudomonadaceae were positively correlated (ρ = 0.83) at the beginning of the sampling period in the symptomatic group, while a positive correlation (ρ = 0.89) was only observed after 20 days in the asymptomatic group. These data indicate that the co-occurrence of Enterobacteriaceae and Pseudomonadaceae may be associated with the development of ginger rot. Bacterial taxa isolated from ginger rhizomes, such as Enterobacter cloacae, E. hormaechei, and Pseudomonas putida, induced obvious rot symptoms when they were inoculated on ginger rhizomes. Notably, antibiotic-producing bacterial taxa in the Streptococcaceae and Flavobacteriaceae were also relatively abundant in rhizomes with rot and appeared to be linked to the onset of rhizome rot disease. Our results provide important information on the establishment and management of disease in ginger rhizomes.

PET/CT Radiomic Features: A Potential Biomarker for EGFR Mutation Status and Survival Outcome Prediction in NSCLC Patients Treated With TKIs

Backgrounds

Epidermal growth factor receptor (EGFR) mutation profiles play a vital role in treatment strategy decisions for non–small cell lung cancer (NSCLC). The purpose of this study was to evaluate the predictive efficacy of baseline ¹⁸F-FDG PET/CT-based radiomics analysis for EGFR mutation status, mutation site, and the survival benefit of targeted therapy.

Methods

A sum of 313 NSCLC patients with pre-treatment ¹⁸F-FDG PET/CT scans and genetic mutations detection were retrospectively studied. Clinical and PET metabolic parameters were incorporated into independent predictors of determining mutation status and mutation site. The dataset was randomly allocated into the training and the validation sets in a 7:3 ratio. Three-dimensional (3D) radiomics features were extracted from each PET- and CT-volume of interests (VOI) singularly, and then a radiomics signature (RS) associated with EGFR mutation profiles is built by feature selection. Three different prediction models based on support vector machine (SVM), decision tree (DT), and random forest (RF) classifiers were established. Furthermore, nomograms for estimation of overall survival (OS) and progression-free survival (PFS) were established by integrating PET/CT radiomics score (Rad-score), metabolic parameters, and clinical factors. Predictive performance was assessed by the receiver operating characteristic (ROC) analysis and the calibration curve analysis. The decision curve analysis (DCA) was applied to estimate and compare the clinical usefulness of nomograms.

Results

Three hundred thirteen NSCLC patients were classified into a training set (n=218) and a validation set (n=95). Multivariate analysis demonstrated that SUVmax and sex were independent indicators of EGFR mutation status and mutation site. Eight CT-derived RS, six PET-derived RS, and two clinical factors were retained to develop integrated models, which exhibited excellent ability to distinguish between EGFR wild type (EGFR-WT), EGFR 19 mutation type (EGFR-19-MT), and EGFR 21 mutation type (EGFR-21-MT). The SVM model outperformed the RF model and the DT model, yielding training area under the curves (AUC) of EGFR-WT, EGFR-19-WT, and EGFR-21-WT, with 0.881, 0.851, and 0.849, respectively, and validation AUCs of 0.926, 0.805 and 0.859, respectively. For prediction of OS, the integrated nomogram is superior to the clinical nomogram and the radiomics nomogram, with C-indexes of 0.80 in the training set and 0.83 in the validation set, respectively.

Conclusions

The PET/CT-based radiomics analysis might provide a novel approach to predict EGFR mutation status and mutation site in NSCLC patients and could serve as useful predictors for the patients’ survival outcome of targeted therapy in clinical practice.

Radiomics in Gastric Cancer: First Clinical Investigation to Predict Lymph Vascular Invasion and Survival Outcome Using 18F-FDG PET/CT Images

Background

Lymph vascular invasion (LVI) is an unfavorable prognostic indicator in gastric cancer (GC). However, there are no reliable clinical techniques for preoperative predictions of LVI. The aim of this study was to develop and validate PET/CT-based radiomics signatures for predicting LVI of GC preoperatively. Radiomics nomograms were also established to predict patient survival outcomes.

Methods

This retrospective study registered 148 GC patients with histopathological confirmation for LVI status, who underwent pre-operative PET/CT scans (Discovery VCT 64 PET/CT system) from December 2014 to June 2019. Clinic-pathological factors (age, gender, and tumor grade, etc.) and metabolic PET data (maximum and mean standardized uptake value, total lesion glycolysis and metabolic tumor volume) were analyzed to identify independent LVI predictors. The dataset was randomly assigned to either the training set or test set in a 7:3 ratios. Three-dimensional (3D) radiomics features were extracted from each PET- and CT-volume of interests (VOI) singularly, and then a radiomics signature (RS) associated with LVI status is built by feature selection. Four models with different modalities (PET-RS: only PET radiomics features; CT-RS: only CT radiomics features; PET/CT-RS: both PET and CT radiomics features; PET/CT-RS plus clinical data) were developed to predict LVI. Patients were postoperatively followed up with PET/CT every 6-12 months for the first two years and then annually up to five years after surgery. The PET/CT radiomics score (Rad-scores) was calculated to assess survival outcome, and corresponding nomograms with radiomics (NWR) or without radiomics (NWOR) were established.

Results

Tumor grade and maximum standardized uptake value (SUVmax) were the independent LVI predictor. 1037 CT and PET 3D radiomics features were extracted separately and reduced to 4 and 5 features to build CT-RS and PET-RS, respectively. PET/CT-RS and PET/CT-RS plus clinical data (tumor grade and SUVmax) were also developed. The ROC analysis demonstrated clinical usefulness of PET/CT-RS plus clinical data (AUC values for training and validation, respectively 0.936 and 0.914) and PET/CT-RS (AUC values for training and validation, respectively 0.881 and 0.854), which both are superior to CT-RS (0.838 and 0.824) and PET-RS (0.821 and 0.812). SUVmax and LVI were independent prognostic indicators of both OS and PFS. Decision curve analysis (DCA) demonstrated NWR outperformed NWOR and was established to assess survival outcomes. For estimation of OS and PFS, the C-indexes of the NWR were 0. 88 and 0.88 in the training set, respectively, while the C-indexes of the NWOR were 0. 82 and 0.85 in the training set, respectively.

Conclusions

The PET/CT-based radiomics analysis might serve as a non-invasive approach to predict LVI status in GC patients and provide effective predictors of patient survival outcomes.

[Anemia prevalence and its influencing factors among students involved in the Nutrition Improvement Program for Rural Compulsory Education Students in 2019]

Objective: To analyze anemia prevalence and its influencing factors of students involved in the Nutritional Improvement Program for Rural Compulsory Education Students in 2019. Methods: From the 2019 surveillance system of the Nutrition Improvement Program for Rural Compulsory Education Students, 47 297 primary and middle school students aged 6-17 were included in the study. Hemoglobin level was tested according to the criteria of WHO 2011. Anemia prevalence of different genders, ages, and regions was analyzed. Results: The average hemoglobin level was 135.19 g/L, with the prevalence of anemia as 8.7% in the children aged 6-17. The prevalence of anemia was 10.0% in girls, higher than that in boys (7.4%). The prevalence rates in western and central areas were 9.8% and 7.1%, respectively. From northwest, southwest, central and south, east, north to northeast areas of China, the anemia rate appeared gradually decreasing (10.2%, 9.7%, 8.3%, 7.5%, 5.7% and 3.5%). The anemia prevalence rates were 8.0%, 8.3%, and 10.9% in children from the 6-, 11-, and 14-17 years age groups, respectively. Logistic regression models revealed that students from schools not using catering software (OR=1.482, 95%CI:1.296-1.694,P<0.001), schools not serving lunch (OR=1.241, 95%CI:1.103-1.395,P<0.001), and from relatively low-income families (OR=1.297, 95%CI:1.211-1.389, P<0.001) showed as risk factors for anemia. After supplementing students' dietary factors, the results showed that students who ate meat three or more times a week had a lower risk of anemia (OR=0.907, 95%CI:0.832-0.989, P=0.026). Conclusions: The Nutritional Improvement Program for Rural Compulsory Education Students had an essential impact on improving the anemia prevalence of primary and middle school students. Family income, school location, economic factors, school feeding, and students' diet programs all impacted the prevalence of anemia.

[Growth retardation of children and its influencing factors in the Nutrition Improvement Program for Rural Compulsory Education Students in 2019]

Objective: To understand the growth retardation among primary and secondary school students in areas covered by the Nutrition Improvement Program for Rural Compulsory Education Students and its influencing factors to provide evidence for improving the nutrition status of rural students in China. Methods: The multi-stage cluster random sampling method selected 1 550 969 primary and secondary school students aged 6-15 years from China's central and western regions. The ratio of male and female students was balanced. The height was measured, and the growth retardation of students was determined according to the Screening Criteria for School-age Children and Adolescents malnutrition (WS/T 456-2014), from the school and county questionnaire survey related factors. The number of cases and percentages described the growth retardation of students, and the χ² test was used for comparison between groups. Binary logistic regression was used to analyze students' growth retardation factors. Results: In 2019, the growth retardation rate of primary and secondary school students in areas covered by the Nutrition Improvement Program for Rural Compulsory Education Students was 5.7% (88 631/1 550 969), the growth retardation rate in the western part (7.1%, 66 167/927 954) was higher than that in the central part (3.7%,19 511/533 973) with difference statistically significant (P<0.001). The growth retardation rate of the boys (6.3%,50 665/803 851) were higher than that of girls (5.1%, 37 966/747 118), the difference was statistically significant (P<0.001). The growth retardation rate of primary school students in central China was 3.9%(14 914/380 598), higher than that of junior middle school students (3.0%,4 597/153 375, P<0.001). In contrast, the growth retardation rate of the western junior high school students (7.2%, 21 494/297 217) were higher than that of elementary school students (7.1%, 44 673/630 737), with a difference statistically significant (all P=0.009). Multi-factor logistic regression results showed that, in high income area (OR=0.829, 95%CI: 0.816-0.842, P<0.001), parents providing part of the meal cost (OR=0.948, 95%CI: 0.931-0.965, P<0.001), enterprises providing meals (OR=0.845, 95%CI: 0.805-0.887, P<0.001), schools providing milk (OR=0.780, 95%CI: 0.767-0.793, P<0.001), health education courses (OR=0.702, 95%CI: 0.682-0.723, P<0.001) and other local nutrition improvement efforts (OR=0.739, 95%CI: 0.720-0.758, P<0.001) were negatively correlated with the occurrence of growth retardation, The growth retardation rate of the students was lower. Conclusions: There appeared significant regional, gender, and age differences in the growth retardation rate of primary and middle school students in areas covered by the Nutrition Improvement Program for Rural Compulsory Education Students. Appropriate food supply in schools, health education courses, and parental participation in nutritional improvement was related to children's lower growth retardation rate.

[Status and influencing factors on the leftover school meals among students the Nutrition Improvement Program for Rural Compulsory Education Students in 2019]

Objective: To analyze the situation and influencing factors of school meals leftover among primary and secondary school students in the area of the Nutrition Improvement Program for Rural Compulsory Education Students, improve the quality of school meals, develop healthy dietary behavior, and reduce food waste. Methods: In 2019, among the 50 monitoring counties that implemented the Compulsory Education Student Nutrition Improvement Program, two primary schools and two junior schools were randomly selected according to different food supply patterns.This study randomly selected one or two classes from grade 3 to grade 9. Basic information and school meals of 26 778 students were collected by using a student questionnaire. Multivariate logistic regression was used to analyze the influencing factors of leftovers rate. Results: 54.93% (14 709) of students wasted school meals, in which the highest rate was the staple food, with the main reason as "not in favor". 11.87% (1 743) of the students wasted school meals 6-7 days a week, with 54.20% (7 957) of students wasted but in less amount. The leftover rate of staple food was the highest (29.78%), followed by vegetables and meat. The main reason of leftovers was that they didn't like this kind of food (33.52%). The rate of school meal waste was higher for girls (OR=1.19,95%CI:1.13-1.25), junior high school students (OR=1.17, 95%CI: 1.11-1.25), resident students (OR=1.06, 95%CI: 1.00-1.12), lower economic level (OR=1.06, 95%CI: 1.00-1.12), parents working outside their houses (OR=1.22, 95%CI: 1.13-1.30), health education classes (OR=1.70, 95%CI: 1.40-2.06), company-based meals (OR=1.89, 95%CI: 1.71-2.07) and school meals were not as good as home food(OR=1.89, 95%CI: 1.78-2.00)(P<0.05). Conclusions: It is common for poor rural primary and middle school students in central and western China to waste school meals, and the reasons were affected by many factors. Reducing food waste requires the joint efforts of individuals, families, schools and society.

A deep learning radiomics model may help to improve the prediction performance of preoperative grading in meningioma

PURPOSE

This study aimed to investigate the clinical usefulness of the enhanced-T1WI-based deep learning radiomics model (DLRM) in differentiating low- and high-grade meningiomas.

METHODS

A total of 132 patients with pathologically confirmed meningiomas were consecutively enrolled (105 in the training cohort and 27 in the test cohort). Radiomics features and deep learning features were extracted from T1 weighted images (T1WI) (both axial and sagittal) and the maximum slice of the axial tumor lesion, respectively. Then, the synthetic minority oversampling technique (SMOTE) was utilized to balance the sample numbers. The optimal discriminative features were selected for model building. LightGBM algorithm was used to develop DLRM by a combination of radiomics features and deep learning features. For comparison, a radiomics model (RM) and a deep learning model (DLM) were constructed using a similar method as well. Differentiating efficacy was determined by using the receiver operating characteristic (ROC) analysis.

RESULTS

A total of 15 features were selected to construct the DLRM with SMOTE, which showed good discrimination performance in both the training and test cohorts. The DLRM outperformed RM and DLM for differentiating low- and high-grade meningiomas (training AUC: 0.988 vs. 0.980 vs. 0.892; test AUC: 0.935 vs. 0.918 vs. 0.718). The accuracy, sensitivity, and specificity of the DLRM with SMOTE were 0.926, 0.900, and 0.924 in the test cohort, respectively.

CONCLUSION

The DLRM with SMOTE based on enhanced T1WI images has favorable performance for noninvasively individualized prediction of meningioma grades, which exhibited favorable clinical usefulness superior over the radiomics features.

Human-in-the-loop Extraction of Interpretable Concepts in Deep Learning Models

The interpretation of deep neural networks (DNNs) has become a key topic as more and more people apply them to solve various problems and making critical decisions. Concept-based explanations have recently become a popular approach for post-hoc interpretation of DNNs. However, identifying human-understandable visual concepts that affect model decisions is a challenging task that is not easily addressed with automatic approaches. We present a novel human-in-the-Ioop approach to generate user-defined concepts for model interpretation and diagnostics. Central to our proposal is the use of active learning, where human knowledge and feedback are combined to train a concept extractor with very little human labeling effort. We integrate this process into an interactive system, ConceptExtract. Through two case studies, we show how our approach helps analyze model behavior and extract human-friendly concepts for different machine learning tasks and datasets and how to use these concepts to understand the predictions, compare model performance and make suggestions for model refinement. Quantitative experiments show that our active learning approach can accurately extract meaningful visual concepts. More importantly, by identifying visual concepts that negatively affect model performance, we develop the corresponding data augmentation strategy that consistently improves model performance.

Interactive Visual Pattern Search on Graph Data via Graph Representation Learning

Graphs are a ubiquitous data structure to model processes and relations in a wide range of domains. Examples include control-flow graphs in programs and semantic scene graphs in images. Identifying subgraph patterns in graphs is an important approach to understand their structural properties. We propose a visual analytics system GraphQ to support human-in-the-loop, example-based, subgraph pattern search in a database containing many individual graphs. To support fast, interactive queries, we use graph neural networks (GNNs) to encode a graph as fixed-length latent vector representation, and perform subgraph matching in the latent space. Due to the complexity of the problem, it is still difficult to obtain accurate one-to-one node correspondences in the matching results that are crucial for visualization and interpretation. We, therefore, propose a novel GNN for node-alignment called NeuroAlign, to facilitate easy validation and interpretation of the query results. GraphQ provides a visual query interface with a query editor and a multi-scale visualization of the results, as well as a user feedback mechanism for refining the results with additional constraints. We demonstrate GraphQ through two example usage scenarios: analyzing reusable subroutines in program workflows and semantic scene graph search in images. Quantitative experiments show that NeuroAlign achieves 19%-29% improvement in node-alignment accuracy compared to baseline GNN and provides up to 100× speedup compared to combinatorial algorithms. Our qualitative study with domain experts confirms the effectiveness for both usage scenarios.

[Effect of measurement sites on coronary computed tomography angiography-derived instantaneous wave-free ratio in the diagnosis of coronary stenosis ischemia]

Objective: To explore the influence of measurement sites on the diagnostic performance of coronary computed tomography angiography (CCTA) -derived instantaneous wave-free ratio (iFR_CT) for coronary stenosis ischemia. Methods: A total of 44 patients, including 27 males and 17 females aged from 44 to 83 (59±90) years, who underwent coronary computed tomography angiography (CCTA) and invasive fractional flow reserve (FFR) in the Department of Radiology, Eastern Theater Command General Hospital from February 2009 to May 2018 were retrospectively collected. Based on CCTA images, iFR_CT values at different locations (stenosis, proximal downstream of stenosis, 2 cm downstream of stenosis, 3 cm downstream of stenosis, and 4 cm downstream of stenosis) were calculated through fluid mechanics simulation. With invasive FFR≤0.80 as the reference standard, receiver operating characteristic (ROC) was used to analyze and evaluate the diagnostic performance of iFR_CT in different sites. Bland-Altman and spearman correlation coefficient were used to analyze the consistency between iFR_CT and invasive FFR. Intraclass correlation coefficient (ICC) and spearman correlation coefficient were used to test the repeatability of iFR_CT. Results: iFR_CT values at the proximal downstream of the stenosis (approximately 1 cm downstream of the stenosis) was superior to other measurement sites in the diagnosis of coronary ischemic specific stenosis. The threshold, area under the curve, sensitivity, specificity and accuracy of iFR_CT at the stenosis were 0.91, 0.87 (95%CI:0.76~0.96), 82% (95%CI:48%~97%), 76% (95%CI:57%~88%) and 77% (95%CI: 63%~87%), respectively. Bland-Altman analysis showed that the mean differences between iFR_CT and FFR at the proximal downstream of the stenosis was 0.07 (95%LoA: 0.06~0.09) and Spearman correlation coefficient was 0.53 (P<0.001). At the proximal downstream of the stenosis, the intraobserver ICC of iFR_CT was 0.92 (95%CI: 0.85 ~0.95), the correlation coefficient was 0.85 (P<0.001); the interobserver ICC was 0.84 (95%CI: 0.60~0.94), correlation coefficient was 0.75 (P<0.001), demonstrating good repeatability between iFR_CT and FFR. Conclusion: With the invasive FFR value as reference standard, iFR_CT had a high diagnostic performance in detecting myocardial ischemia and the best measurement site was the proximal downstream of the stenosis.

Whole Genome Sequence of Alternaria alternata, the Causal Agent of Black Spot of Kiwifruit

Alternaria alternata is a pathogen in a wide range of agriculture crops and causes significant economic losses. A strain of A. alternata (Y784-BC03) was isolated and identified from “Hongyang” kiwifruit and demonstrated to cause black spot infections on fruits. The genome sequence of Y784-BC03 was obtained using Nanopore MinION technology. The assembled genome is composed of 33,869,130bp (32.30Mb) comprising 10 chromosomes and 11,954 genes. A total of 2,180 virulence factors were predicted to be present in the obtained genome sequence. The virulence factors comprised genes encoding secondary metabolites, including non-host-specific toxins, cell wall-degrading enzymes, and major transcriptional regulators. The predicted gene clusters encoding genes for the biosynthesis and export of secondary metabolites in the genome of Y784-BC03 were associated with non-host-specific toxins, including cercosporin, dothistromin, and versicolorin B. Major transcriptional regulators of different mycotoxin biosynthesis pathways were identified, including the transcriptional regulators, polyketide synthase, P450 monooxygenase, and major facilitator superfamily transporters.