Objective To observe the value of artificial intelligence(AI)knowledge graph and image classification for quality control(QC)of chest posterior-anterior position X-ray radiograph(abbreviated as chest film).Methods Totally 9 236 chest films from 595 medical institutions in Anhui province imaging cloud platform were retrospectively enrolled.QC knowledge graph containing 21 classification labels were constructed.Firstly,QC of chest films based on the above knowledge graph were performed by 10 technicians for 2 rounds of single person and 1 round of multi person,and the results were recorded as A,B and C,respectively.Then AI algorithms were used to classify and evaluate based on knowledge graph,and the result was recorded as D.Finally,a QC expert reviewed results C and D to determine the final QC results and taken those as references to analyze the efficiency of the above 4 QC.Results The area under the curve(AUC)of AI algorithm for QC of chest films were all ≥0.780,with an average value of 0.939.The average precision of QC for chest films of A,B,C and D was 81.15％,85.47％,91.65％and 92.21％,respectively.Conclusion AI knowledge graph and image classification technology could be effectively used for QC of chest posterior-anterior position X-ray radiograph.

Purpose To explore the effect of compressed sensing(CS)technique with different acceleration factors on the quality of cardiac magnetic resonance cine sequences and feasibility of clinical application.Materials and Methods From January to July 2021,40 healthy volunteers were prospectively recruited for cardiac magnetic resonance cine imaging in Hefei First People's Hospital.Four scan protocols of volunteers were designed:SENSE 2 and CS-cine sequences with acceleration factors of 3,4 and 8.The imaging slices include four chamber heart,short axis heart,left ventricular two chamber heart and three chamber heart.Subjective score of image quality,left ventricular cardiac function and 16-segment myocardial thickness of SENSE 2 and CS3,CS4 and CS8 were compared and analyzed.Results Taking SENSE 2 image as the standard score of 5,the scores of CS3 and CS4 were above 3,and the scores of CS8 were below 3.There was no significant difference between the scores of four chamber heart and short axis heart images of CS3 and SENSE 2 sequence(all P>0.05).There were statistically significant differences in the subjective scores of image quality between the left ventricular two chamber heart and three chamber heart groups of each sequence(all P<0.05).There was no statistically significant difference in left ventricular function(left ventricular ejection fraction,left ventricular end systolic volume,left ventricular end diastolic volume,left ventricular stroke volume,left ventricular end-diastolic mass)(F=0.027,0.182,0.057,0.140,0.545)and myocardial thickness(F=0.052-7.366)among the four imaging schemes(all P>0.05).Conclusion Cardiac cine based on compressed sensing technology have good application prospects.With the increase of acceleration factors,the scanning time gradually decreases,and the corresponding image quality will also decrease.When the acceleration factor is 4,that is,the scanning time of the conventional cine sequence is reduced by 50%,the left ventricular function and myocardial thickness can still be accurately measured,and the image quality basically meets the diagnostic requirements.

Purpose Develop deep learning models to assess the clarity of chest X-ray images and validate the model's effectiveness by comparing it with the subjective evaluations of radiologists.Materials and Methods A retrospective collection of 9 135 chest X-ray images from 590 hospitals in Anhui Province,spanning from June 2015 to August 2022,was organized involving multiple radiologists who repeatedly evaluated the clarity of the images using a five-level scoring system.Individual assessments were designated as A and B,whereas the collective result of multiple assessments was designated as C.By constructing a deep learning model based on ResNet-50,image clarity evaluations of chest X-ray images were performed,we used the result C as the training and testing data for the model.The model's evaluation results were denoted as D.A radiology quality control expert conducted an audit assessment of the model's evaluation results and the multi-person assessments of physicians,serving as the reference standard for image clarity.Their assessment results were labeled as E.Statistical analysis,including Spearman's rank correlation coefficient,root mean square error(RMSE)and accuracy was employed to evaluate the effectiveness of the model.Results Compared with the reference standard E,D achieved an average accuracy of 0.85,exceeding the accuracy of C,which stood at 0.84.The ρ values for A,B,C,D and E were 0.58(0.54,0.62),0.59(0.55,0.63),0.74(0.71,0.77)and 0.80(0.78,0.82),respectively.The model exhibited the highest correlation with E.The ρ between A and B was 0.45(0.41,0.49),indicating a lower correlation between two individual subjective assessments of image clarity.The RMSE values for A,B,C,D and E were 0.99,0.94,0.72,and 0.71,respectively.The model's RMSE was lower than that of manual assessments.Conclusion This research model is capable of accurately assessing the clarity of chest X-ray images,and reducing the subjective interference of manual evaluation through deep learning methods,thereby providing an effective and objective evaluation tool for the assessment of clarity in clinical radiographic images.

Purpose To explore the changes in biventricular volume and mass in chronic kidney disease(CKD)patients with preserved ejection fraction using cardiac magnetic resonance(CMR),and evaluate cardiac function using tissue-tracking strain analysis.Materials and Methods The retrospective analysis of 40 CKD patients without symptoms of cardiovascular disease,diabetes and the estimated glomerular filtration rate<60 ml/(min·1.73 m2)who were treated at the First Affiliated Hospital of Anhui Medical University from October 2020 to June 2023 and 25 age-and gender-matched healthy volunteers.All participants underwent 1.5T CMR scans to calculate left and right ventricular function,myocardial strain and native T1 and T2 values,and differences in various imaging parameters between the two groups were compared.Results The native T1 value[(1 109.6±35.9)ms]and T2 value[(58.1±2.9)ms]of CKD patients were significantly higher than those in control group[(1 046.3±15.9)ms,(53.3±2.3)ms](t=8.296,6.916,both P<0.001).The left ventricular end-diastolic volume index,left ventricular end-systolic volume index,left ventricular mass index,right ventricular end-diastolic volume index,right ventricular end-systolic volume index and right ventricular mass index of CKD patients were significantly increased compared with the control group(t=3.233,3.350,5.751,7.746,5.937,2.363,all P<0.05),while the left and right ventricular strain parameters,left ventricular global longitudinal strain,left ventricular global circumferential strain,left ventricular global radial strain,right ventricular global longitudinal strain,right ventricular global circumferential strain and right ventricular global radial strain were significantly decreased(t=9.506,9.072,-10.603,10.496,11.574,-4.018,all P<0.001),and the estimated glomerular filtration rate exhibited significant correlations with left ventricular global longitudinal strain(r=-0.636),left ventricular global radial strain(r=0.688),right ventricular global longitudinal strain(r=-0.513),right ventricular global circumferential strain(r=-0.827)and right ventricular global radial strain(r=0.514)(all P<0.001).Conclusion CMR can quantitatively assess myocardial fibrosis and edema in patients with CKD,and tissue tracking strain analysis can detect myocardial motion abnormalities before changes in ventricular ejection fraction,offering value in the early detection of cardiac involvement in CKD patients..

With inherent sparse spike-based coding and asynchronous event-driven computation, spiking neural network (SNN) is naturally suitable for processing event stream data of event cameras. In order to improve the feature extraction and classification performance of bio-inspired hierarchical SNNs, in this paper an event camera object recognition system based on biological synaptic plasticity is proposed. In our system input event streams were firstly segmented adaptively using spiking neuron potential to improve computational efficiency of the system. Multi-layer feature learning and classification are implemented by our bio-inspired hierarchical SNN with synaptic plasticity. After Gabor filter-based event-driven convolution layer which extracted primary visual features of event streams, we used a feature learning layer with unsupervised spiking timing dependent plasticity (STDP) rule to help the network extract frequent salient features, and a feature learning layer with reward-modulated STDP rule to help the network learn diagnostic features. The classification accuracies of the network proposed in this paper on the four benchmark event stream datasets were better than the existing bio-inspired hierarchical SNNs. Moreover, our method showed good classification ability for short event stream input data, and was robust to input event stream noise. The results show that our method can improve the feature extraction and classification performance of this kind of SNNs for event camera object recognition.
Visual Perception ; Learning ; Action Potentials ; Neural Networks, Computer ; Neuronal Plasticity

Purpose To investigate the clinical value of artificial intelligence(AI)quantitative parameters in predicting the invasion degree of lung adenocarcinoma with diameter≤2 cm of ground-glass density.Materials and Methods A total of 80 patients with lung adenocarcinoma with diameter≤2 cm ground-glass density confirmed by pathology from March 2019 to April 2022 were retrospectively analyzed.A total of 90 nodules were rerolled,including 8 adenocarcinomas in situ(AIS),34 minimally invasive adenocarcinomas(MIA)and 48 invasive adenocarcinomas(IAC).They were divided into the experimental group(IAC)and the control group(AIS and MIA).The differences of the AI quantitative parameters such as volume,three-dimensional length diameter,maximum area,maximum CT value,minimum CT value and average CT value were compared between two groups,and the predictive values of AI quantitative parameters for the invasion degree of lung adenocarcinoma was evaluated.Results There were statistically significant differences with age,volume,three-dimensional length diameter,maximum area,maximum CT value and average CT value between the two groups(all P<0.05),but no statistically significant differences in gender and minimum CT value(both P>0.05).Binary Logistic regression analysis showed that the three-dimensional length diameter(odd ratio=2.020,P=0.034)and the maximum CT value(odd ratio=1.008,P=0.013)were independent predictors for lung adenocarcinoma with diameter≤2 cm of ground-glass density.The regression model based on the three-dimensional length diameter and the maximum CT value had the best predictive performance,and its AUC was 0.901.When the critical value was 2.432,its sensitivity and specificity were 93.75%and 71.43%,respectively.Conclusion AI quantitative parameters have a high value in predicting the degree of invasion of lung adenocarcinoma with diameter≤2 cm of ground-glass density,and the combined model with three dimensional long diameter and maximum CT value has the highest diagnostic efficiency.

Objective:To explore whether left ventricular interstitial fibrosis is associated with left atrial enlargement and left atrial dysfunction in patients of hypertrophic cardiomyopathy(HCM) with preserved ejection fraction.Methods:From October 2018 to September 2021, 59 HCM including 30 with enlarged maximal left artrial volume index (LAVI max), 29 with normal LAVI max and 28 age-and gender-matched controls were retrospectively enrolled. Imaging protocol included cine sequence, late gadolinium enhancement and T 1 mapping.The relationships between left ventricular mass index (LVMI), quantitative myocardial fibrosis and left atrial-related indexes were analyzed. One-way analysis of variance with Bonferroni post hoc correction or Kruskal-Wallis was performed for continuous variables. Categorical variables were assessed using the Chi-square test or Fisher′s exact test. Pearson or Spearman analysis was used for linear or monotonic nonlinear correlations. Results:The left ventricular end-diastolic volume index, left ventricular end-systolic volume index, left ventricular cardiac output and LVMI of HCM with enlarged LAVI max group were higher than HCM with normal LAVI max group and control group( P<0.05).Correlation analysis showed that LVMI correlated positively with LAVI max( r=0.780, P<0.001) and minimal left artrial volume index (LAVI min) ( r=0.816, P<0.001), extracellular volume correlated positively with LAVI max( r=0.462, P<0.001) and LAVI min( r=0.483, P<0.001),%LGE was correlated positively with LAVI max( r=0.311, P<0.05) and LAVI min( r=0.327, P<0.05),left ventricular index interstitial volume was correlated negatively with left atrial ejection fraction of reservoir ( r=-0.669, P<0.001),left atrial ejection fraction of conduit ( r=-0.472, P<0.001),left atrial ejection fraction of pump ( r=-0.518, P<0.001)and left atrial expansion index( r=-0.626, P<0.001). Conclusion:There is association between LVMI and fibrosis and left atrial enlargement and phases dysfunction in HCM with preserved ejection fraction.

Objective To investigate the dosimetric feasibility of volumetric modulated arc therapy (VMAT)with a simultaneous integrated boost(SIB-VMAT58.75 Gy)for preoperative chemoradiotherapy in patients with locally advanced rectal cancer(LARC),and to provide a basis for clinical practice.Methods Nine patients with stage Ⅱ-Ⅲ rectal cancer who underwent preoperative concurrent chemoradiotherapy were involved in the study,and two plans were performed for each patient:SIB-VMAT58.75 Gy and VMAT50.00 Gy. For the SIB-VMAT58.75 Gy plan,the prescribed dose was 58.75 Gy(2.35 Gy/fraction)for the local rectal tumor and positive lymph nodes(GTV 58.75 Gy),and 50 Gy(2 Gy/fraction)for the regions at high risk of harboring microscopic disease(pelvic lymphatic drainage area)(PTV 50Gy).For the VMAT50.00 Gy plan,the prescribed dose was 50 Gy(2 Gy/fraction)for the regions at high risk of harboring microscopic disease(pelvic lymphatic drainage area)without a boost. The conformity index(CI),homogeneity index (HI),and dose for target areas and organs at risk(OAR)were assessed according to the dose-volume histogram. The paired t-test or nonparametric rank test was used to compare the differences between the two plans. Results Both plans met the prescription goal for PTV dose coverage. There was no significant difference in CI for the PTV between the two plans(1.0±0.0 vs. 1.0±0.0,P>0.05).The SIB-VMAT58.75 Gy plan had a worse HI than the VMAT50.00 Gy plan(0.2± 0.2 vs. 0.1± 0.0,P<0.05).There was no significant difference in V10-V50of the small intestine,bladder,femoral heads,and pelvis between the two plans(P>0.05),but D 2 cm3of the small intestine was significantly higher in the SIB-VMAT58.75 Gy plan than in the VMAT50.00 Gy plan(P=0.038). Conclusions The SIB-VMAT58.75 Gy plan for LARC achieves required target volume dose coverage and OAR dose constraints,which is safe and feasible in terms of dosimetry,and its clinical efficacy and adverse effects need further evaluation.

Objective:To compare the pregnancy outcome before the implementation of management specification of cervical ripening with dinoprostone suppository with that after the implementation and explore its safety for cervical ripening .Methods:The clinical data of 612 puerperants ,who received cervical ripening with dinoprostone suppository and delivered in First Mater‐nal and Infant Hospital Affiliated to Tongji University during Sep 2012 and Feb 2014 ,were retrospectively analyzed .The man‐agement specification of cervical ripening with dinoprostone had been implemented since Apr 2013 .The maternal and neonatal outcomes ,as well as the rate of adverse reactions ,before and after the implementation of management specification of cervical ripening with dinoprostone suppository were compared with each other ,so as to explore its effects on pregnancy outcome .Re‐sults:There were 449 puerperants receiving cervical ripening with suppository before the implementation of management speci‐fication in Apr 2013 and 163 cases after that .Totally 413 cases(67 .5% ) were induced labour successfully .The success rates of inducting labor within 48 h ,24 h and 12 h after the implementation of management specification were lower than that before the implementation of management specification ,but there was no significant difference .The cesarean rate was 27 .1% and there was no statistically significant difference between the cesarean rate before the implementation of management specification and that after the implementation .There was no statistically significant difference in Apgar score and admission rate to ICU between the two groups .The incidence rates of tachysystole and hypertonus ,meconium‐stained amniotic fluid after the implementation of specification were significantly lower than that before the implementation ,and the difference was statistically significant(P<0 .05) .There was no statistically significant difference regarding the incidence rates of uterine hyperstimulation and the clinical chorioamnionitis .There was no occurrence of complications as amniotic fluid embolism and severe perineal laceration .Conclusions:Di‐noprostone suppository is a safe and effective method for cervical ripening and labor induction .The implementation of management speci‐fication is conducive to protecting maternal and infant and reducing the adverse pregnancy outcomes .

OBJECTIVE: The prior functional MRI studies have demonstrated significantly abnormal activity in the bilateral superior temporal gyrus (STG) of anxiety patients. The purpose of the current investigation was to determine whether the abnormal activity in these regions was related to a loss of functional connectivity between these regions. MATERIALS AND METHODS: Ten healthy controls and 10 anxiety patients underwent noninvasive fMRI while actively listening to emotionally neutral words alternated by silence (Task 1) or threat-related words (Task 2). The participants were instructed to silently make a judgment of each word's valence (i.e., unpleasant, pleasant, or neutral). A coherence analysis was applied to the functional MRI data to examine the functional connectivity between the left and the right STG, which was selected as the primary region of interest on the basis of our prior results. RESULTS: The data demonstrated that the anxiety patients exhibited significantly increased activation in the bilateral STG than the normal controls. The functional connectivity analysis indicated that the patient group showed significantly decreased degree of connectivity between the bilateral STG during processing Task 2 compared to Task 1 (t = 2.588, p = 0.029). In addition, a significantly decreased connectivity was also observed in the patient group compared to the control group during processing Task 2 (t = 2.810, p = 0.012). CONCLUSION: Anxiety patients may exhibit increased activity of the STG but decreased functional connectivity between the left and right STG, which may reflect the underlying neural abnormality of anxiety disorder, and this will provide new insights into this disease.
Adult ; Anxiety Disorders/pathology/*physiopathology ; Case-Control Studies ; Emotions/physiology ; Female ; Humans ; Magnetic Resonance Imaging/*methods ; Male ; Middle Aged ; Temporal Lobe/pathology/*physiopathology