Objective:To investigate the predictive value of clinical factor model, deep learning model based on baseline plain CT images, and combination of both for predicting hematoma expansion in cerebral hemorrhage.Methods:The study was cross-sectional. Totally 471 cerebral hemorrhage patients who were firstly diagnosed in the Second Affiliated Hospital of Soochow University from January 2017 to December 2021 were collected retrospectively. These patients were randomly divided into a training dataset ( n=330) and a validation dataset ( n=141) at a ratio of 7∶3 by using the random function. All patients underwent two noncontrast CT examinations within 24 h and an increase in hematoma volume of >33% or an absolute increase in hematoma volume of >6 ml was considered hematoma enlargement. According to the presence or absence of hematoma enlargement, all patients were divided into hematoma enlargement group and hematoma non-enlargement group.Two-sample t test, Mann-Whitney U test or χ2 test were used for univariate analysis. The factors with statistically significant differences were included in multivariate logistic regression analysis, and independent influences related to hematoma enlargement were screened out to establish a clinical factor model. ITK-SNAP software was applied to manually label and segment the cerebral hemorrhage lesions on plain CT images to train and build a deep learning model based on ResNet50 architecture. A combination model for predicting hematoma expansion in cerebral hemorrhage was established by combining independent clinical influences with deep learning scores. The value of the clinical factor model, the deep learning model, and the combination model for predicting hematoma expansion in cerebral hemorrhage was evaluated using receiver operating characteristic (ROC) curves and decision curves in the training and validation datasets. Results:Among 471 cerebral hemorrhage patients, 136 cases were in the hematoma enlargement group and 335 cases were in the hematoma non-enlargement group. Regression analyses showed that male ( OR=1.790, 95% CI 1.136-2.819, P=0.012), time of occurrence ( OR=0.812, 95% CI 0.702-0.939, P=0.005), history of oral anticoagulants ( OR=2.157, 95% CI 1.100-4.229, P=0.025), admission Glasgow Coma Scale score ( OR=0.866, 95% CI 0.807-0.929, P<0.001) and red blood cell distribution width ( OR=1.045, 95% CI 1.010-1.081, P=0.011) were the independent factors for predicting hematoma expansion in cerebral hemorrhage. ROC curve analysis showed that in the training dataset, the area under the curve (AUC) of clinical factor model, deep learning model and combination model were 0.688 (95% CI 0.635-0.738), 0.695 (95% CI 0.642-0.744) and 0.747 (95% CI 0.697-0.793) respectively. The AUC of the combination model was better than that of the clinical model ( Z=0.54, P=0.011) and the deep learning model ( Z=2.44, P=0.015). In the validation dataset, the AUC of clinical factor model, deep learning model and combination model were 0.687 (95% CI 0.604-0.763), 0.683 (95% CI 0.599-0.759) and 0.736 (95% CI 0.655-0.806) respectively, with no statistical significance. Decision curves showed that the combination model had the highest net benefit rate and strong clinical practicability. Conclusions:Both the deep learning model and the clinical factor model established in this study have some predictive value for hematoma expansion in cerebral hemorrhage; the combination model established by the two together has the highest predictive value and can be applied to predict hematoma expansion.

Objective:To explore the consistency of MRI-based ovarian-adnexal report and data system (O-RADS) score and its diagnostic value for ovarian adnexal masses.Methods:The MRI data of 309 patients with ovarian adnexal masses confirmed by pathology were retrospectively collected from January 2017 to August 2021 in the Second Affiliated Hospital of Soochow University, including 327 lesions consisted of 250 benign lesions, 21 borderline lesions, and 56 malignant lesions confirmed by pathology. Borderline and malignant lesions were classified into the malignant group ( n=77) and benign lesions were classified as benign group ( n=250). Two radiologists scored all lesions according to the MRI-based O-RADS, and scored again after 6 months. The proportion of borderline/malignant lesions in each MRI-based O-RADS score was calculated. The weighted Kappa test was used to assess the intra-reader and inter-reader consistency of the image interpretation results. The receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic efficacy of MRI-based O-RADS classification for distinguishing benign and malignant ovarian adnexal masses. Results:The weighted Kappa value of the MRI-based O-RADS score between the two radiologists was 0.810 (95%CI 0.764-0.855), and the weighted Kappa values of the two radiologists′ scores at different times were 0.848 (95%CI 0.806-0.889) and 0.875 (95%CI 0.835-0.914), respectively. The borderline/malignant lesions accounted for 0/16, 0.8% (1/127), 10.1% (10/99), 76.0% (57/75), 9/10 and 0/17, 0 (0/122), 8.0% (8/100), 76.2% (48/63), and 84.0% (21/25) of the lesions in the two radiologists based on the MRI O-RADS score of 1, 2, 3, 4, and 5, respectively. When adopting O-RADS score>3 as a cut-off value, the area under the ROC curve of the two radiologists for distinguishing benign and malignant ovarian adnexal masses was 0.928 (95%CI 0.895-0.954) and 0.942 (95%CI 0.911-0.965), respectively. The sensitivity was 0.857 and 0.896, the specificity was 0.924 and 0.924, and the accuracy was 0.908 and 0.917 respectively.Conclusion:The MRI-based O-RADS yields high diagnostic efficiency in the evaluation of benign and malignant ovarian adnexal masses, and the intra-reader and inter-reader consistency of the image interpretation is strong.

Objective:To develop a grading prediction model of traumatic hemorrhage volume based on deep learning and assist in predicting traumatic hemorrhage volume.Methods:A retrospective observational study was conducted based on the experimental data of pig gunshot wounds in the time-effect assessment database for experiments on war-traumatized animals constructed by the General Hospital of the Chinese People's Liberation Army. The hemorrhage volume data of the study population were extracted, and the animals were divided into 0-300 mL, 301-600 mL, and > 600 mL groups according to the hemorrhage volume. Using vital signs indexes as the predictive variables and hemorrhage volume grading as the outcome variable, trauma hemorrhage volume grading prediction models were developed based on four traditional machine learning and ten deep learning methods. Using laboratory test indexes as predictive variables and hemorrhage volume grading as outcome variables, trauma hemorrhage volume grading prediction models were developed based on the above fourteen methods. The effect of the two groups of models was evaluated by accuracy and area under the receiver operator characteristic curve (AUC), and the optimal models in the two groups were mixed to obtain hybrid model 1. Feature selection was conducted according to the genetic algorithm, and hybrid model 2 was constructed according to the best feature combination. Finally, hybrid model 2 was deployed in the animal experiment database system.Results:Ninety-six traumatic animals in the database were enrolled, including 27 pigs in the 0-300 mL group, 40 in the 301-600 mL group, and 29 in the > 600 mL group. Among the fourteen models based on vital signs indexes, fully convolutional network (FCN) model was the best [accuracy: 60.0%, AUC and 95% confidence interval (95% CI) was 0.699 (0.671-0.727)]. Among the fourteen models based on laboratory test indexes, recurrent neural network (RNN) model was the best [accuracy: 68.9%, AUC (95% CI) was 0.845 (0.829-0.860)]. After mixing the FCN and RNN models, the hybrid model 1, namely RNN-FCN model was obtained, and the performance of the model was improved [accuracy: 74.2%, AUC (95% CI) was 0.847 (0.833-0.862)]. Feature selection was carried out by genetic algorithm, and the hybrid model 2, namely RNN-FCN* model, was constructed according to the selected feature combination, which further improved the model performance [accuracy: 80.5%, AUC (95% CI) was 0.880 (0.868-0.893)]. The hybrid model 2 contained ten indexes, including mean arterial pressure (MAP), hematocrit (HCT), platelet count (PLT), lactic acid, arterial partial pressure of carbon dioxide (PaCO 2), Total CO 2, blood sodium, anion gap (AG), fibrinogen (FIB), international normalized ratio (INR). Finally, the RNN-FCN* model was deployed in the database system, which realized automatic, continuous, efficient, intelligent, and grading prediction of hemorrhage volume in traumatic animals. Conclusion:Based on deep learning, a grading prediction model of traumatic hemorrhage volume was developed and deployed in the information system to realize the intelligent grading prediction of traumatic animal hemorrhage volume.

Objective:To observe the changes of arterial blood gas indexes in pigs with the free-field primary blast lung injury (PBLI) model, and to explore the value of arterial blood gas indexes in predicting moderate to severe PBLI.Methods:Nine adult healthy Landrace pigs were selected to construct the pig free-field PBLI model. Arterial blood samples were taken 15 minutes before the explosion (before injury) and 10, 30, 60, 120, and 180 minutes after the explosion (after injury). Arterial blood gas indexes and pulse oxygen saturation (SpO 2) were measured, compare the changes of blood gas analysis indexes and SpO 2 levels at different time points, and observe the changes of gross injury scores and pathological injury scores of lung tissue. Analyze the correlation between the blood gas indicators. Results:As time prolonged, at each time point, pH, arterial partial pressure of oxygen (PaO 2), and SpO 2 were lower than those before the injury, and blood lactic acid (Lac) and arterial partial pressure of carbon dioxide (PaCO 2) were higher than those before the injury. Compared with that before the injury, the pH value in the blood decreased significantly 10 minutes after the injury (7.39±0.06 vs. 7.46±0.02, P < 0.05), and the Lac increased significantly (mmol/L: 3.61±2.89 vs. 1.10±0.28, P < 0.05), and lasts until 180 minutes after injury (pH value: 7.37±0.07 vs. 7.46±0.02, Lac (mmol/L): 2.40±0.79 vs. 1.10±0.28, both P < 0.05); while PaO 2 and SpO 2 decreased significantly at 180 minutes after injury [PaO 2 (mmHg, 1 mmHg = 0.133 kPa): 59.40±10.94 vs. 74.81±9.39, P < 0.05; SpO 2: 0.75±0.11 vs. 0.89±0.08, P < 0.05], PaCO 2 increased significantly (mmHg: 56.17±5.38 vs. 48.42±4.93, P < 0.05). Correlation analysis showed that the gross injury score of lung blast injury animals was positively correlated with the pathological injury score ( r = 0.866, P = 0.005); PaO 2 and SpO 2 were positively correlated ( r = 0.703, P = 0.000); pH value and Lac were negative Correlation ( r = -0.400, P = 0.006); pH value is negatively correlated with PaCO 2 ( r = -0.844, P = 0.000). Conclusion:This study successfully established a large mammalian free-field PBLI model, arterial blood gas analysis is helpful for the early diagnosis of PBLI, whether SpO 2 can be used to evaluate the severity of lung injury remains to be further verified.

Objective To evaluate the imaging features of nigrosome-1 in Parkinson's disease (PD) with a 3 T scanner by susceptibility weighted imaging (SWI),and to explore its clinical relevance.Methods Thirty-two patients with primary PD diagnosed by neurologists were collected.Healthy controls matched to their age and gender were recruited during the same period (n=20).All subjects underwent routine brain magnetic resonance imaging (MRI) and sensitive weighted imaging (SWI).The SWI images of the subjects were evaluated to evaluate nigrosome-1 by blinded investigators.Then,the correlation between imaging features and clinical data was analyzed.Results In the PD group,21 cases of bilateral "absent swallow-tail sign",five cases of bilateral "indecisive swallow-tail sign",five cases of "absent swallow-tail sign" on one side and "indecisive swallow-tail sign" on the other side,and one case of bilateral "clear swallow-tail sign" were found.The course of the "absent swallow-tail sign" group (56 (54) months) was significantly longer than the "non-absent swallow-tail sign" group (18 (18) months;U=-2.47,P=0.01).The Hoehn-Yahr stage was significantly higher in the "absent swallow-tail sign" group (2.0 (0.5)) than in the "non-absent swallow-tail sign" group (1.5 (0.5),U=-2.21,P=0.03).There was also a statistically significant difference in the Unified Parkinson's Disease Rating Scale score (24 (8),13 (14)) between the two groups (U=-2.91,P=0.01).However,there were no statistically significant differences between the two groups in the Hamilton Depression Scale score (5 (2) vs 5 (7),U=-0.10,P=0.94) and the Hamilton Anxiety Scale score (3.0 (2.5) vs 3.0 (3.0),U=-0.02,P=1.00).Conclusion The images of nigrosome-1 by SWI are closely related to the severity of the condition and motor symptoms of patients with PD,which can reflect the severity of the disease.

Objective To explore the difference in efficacy between multiparametric MRI (Mp-MRI) based on prostate imaging reporting and data system version 2 (PI-RADS v2) and abbreviated biparametric MRI (Bp-MRI) in detecting prostate cancer (PCa) and clinically significant prostate cancer (csPCa), and to evaluate the consistency of image interpretation between different readers. Methods The imaging, pathological and clinical data of patients with prostatic Mp-MRI in our hospital from February 2015 to June 2018 were retrospectively analyzed. At the beginning, 250 patients were randomly selected. Two radiologists visually evaluated the images of those patients using two 5-point scoring schemes based on Mp-MRI and Bp-MRI. The remaining cases were independently proceeded by one of the radiologists using two schemes respectively. Weighted Kappa test was used to assess the consistency of the results interpreted by the two radiologists. The receiver operating characteristic (ROC) curve was used to evaluate the efficiency of the two scoring schemes in detecting PCa and csPCa, and with Z test to investigate whether there was any difference in detection efficiency between the two schemes. Results Nine hundred and seventy eight patients were eventually enrolled in the study. The results of the consistency assessment showed that there was good agreement between the two radiologists, whether using Mp-MRI or Bp-MRI, with the weighted Kappa coefficient of 0.800 and 0.812, respectively. The ROC curve analysis showed that the area under the curve (AUC) of PCa detected by Mp-MRI and Bp-MRI was 0.873 and 0.879, respectively, and the AUC of csPCa detected was 0.922 and 0.932, respectively. In addition, there was no statistically significant difference between the AUC of PCa and csPCa detected by the two schemes (P>0.05). Conclusion The Bp-MRI scoring scheme has good stability in the evaluation of benign and malignant prostate, and its detection efficiency of PCa or csPCa is not lower than that of standard Mp-MRI based on PI-RADS v2.

Objective To investigate the value and diagnostic efficiency of the quantitative dynamic contrast enhanced-magnetic resonance imaging (DCE-MRI) and intravoxel incoherent motion (IVIM) parameters using three dimention (3D)-histogram analysis for discriminating the Gleason score (GS) of prostate cancer. Methods A total of 53 patients pathologically confirmed as prostate cancer by systemic prostate biopsy who had routine , DCE and DWI-MRI scans were retrospectively analyzed. There were 15 cases for low-risk and 38 cases for intermediate/high-risk prostate cancer. The 3D ROI of all lesions based on T2WI was achieved by image registration to get the quantitative parameters of DCE-MRI and DWI-IVIM. The parameters of DCE-MRI contains: transfer constant (Ktrans), rate constant (Kep) and extracellular-extravascular volume fraction (Ve).The DWI-IVIM related quantitative parameters were ADC, diffusion coefficient (D), diffusion coefficient related to perfusion (D*) and perfusion fraction (f). Then the histogram analysis of these quantitative parameters was performed to get the mean, median, 25th percentile, 75th percentile, Skewness and Kurtosis. Using the Spearman rank correlation analysis to evaluate the correlation of these parameters and GS of prostate cancer. The diagnostic performance of these quantitative histogram parameters related to the GS in identifying low-risk and intermediate/high-risk of prostate cancer was carried by ROC. Results The Kep and Ktrans (mean, median, 25th, 75th) of DCE-MRI were positively correlated with GS (r value was 0.346 to 0.696, P<0.05). The ADC (mean, median, 25th, 75th), D (mean, median, 25th, 75th, Skewness, Kurtosis) and D*(25th) of DWI-IVIM were correlated with GS (r value was-0.544 to 0.428, P<0.05). The DCE-MRI quantitative parameters Kep (25th) had the highest area under curve (AUC, 0.961); The ADC (median) and D (25th) had higher AUC( 0.832, 0.888) in the quantitative parameters of DWI-IVIM, the difference between Kep(25th) and ADC (median) was statistically significant (Z value was 2.212, P value was 0.027). The difference of AUC between Kep (25th) and D (25th), D (25th) and ADC (median) was not statistically significant (Z values were 1.027 and 1.398, P values were 0.162 and 0.304, respectively).Conclusion DCE and IVIM quantitative parameters (Kep, Ktrans, ADC, D) histogram analysis results are correlated with GS, and can be used for distinguishing low-risk from intermediate/high-risk prostate cancer.

Medical imaging is an interdisciplinary subject closely related to clinical and pathological subject. Its clinical reading skills' training has become the focus of postgraduate teaching. In the process of clinical teaching, the interactive reading mode of problem-based learning (PBL) combined with multi-disci-plinary team (MDT) was introduced into clinical reading meeting. The tutors chose the reading cases proved by pathology; designed in-depth issues step by step for execution of PBL teaching; guided postgraduates to delineate imaging signs and propose the diagnostic results, evidences and differential diagnoses according to the step from localizing to qualitative and then to pathological diagnosis;then guided postgraduates to attend in-depth case analysis of MDT and analyze the correlation or inconsistency between the imaging diagnosis and clinical and pathological diagnosis; exercise document retrieval and verbalization, multimedia design, and writing level of the records of the reading cases and papers. The interactive reading mode of PBL com-bined with MDT has achieved significant effects, which is worthy of further exploration and promotion.

Objective Using MRI compatible OxyLiteTM fiber-optic microprobes to verify the reliability of detecting the oxygenation changes in rats C6 glioma by BOLD fMRI with non- hemodynamic response function (non-HRF) post-processing algorithm. Methods A total of 20 male SD rats were used to establish the subcutaneous C6 glioma model. GRE-EPI BOLD fMRI scans were performed in the tumor-bearing rats with Carbogen inhalation after anatomic scans using 1.5 T MR imaging system with
Micro-47 microscopic coil. Fiber-optic microprobes were implanted in tumor to acquire the dynamic pO2 indications during BOLD fMRI scan.“Oxy-localization map”and“oxy-amplitude map”were generated from BOLD functional image data by non-HRF post-processing algorithm analysis. A ROI about 1.5 mm on a side centered to the tip of microprobe was defined on the MRI morphological image, and then was copied onto the“oxy-localization map”and“oxy-amplitude map”to extract the values of significant re-oxygenation (T), percent BOLD signal change (ΔPSC). The mean difference of pO2(ΔpO2) measured by fiber-optic microprobes before(pO2-Air)and after (pO2-Car)Carbogen inhalation in the ROI areas was calculated. Correlation analysis was madebetween cov (T value, Δ pO2) and cov (ΔPSC value, Δ pO2). The difference between pO2-Air and pO2-Car were tested by Mann Whitney U test. Results pO2 was successfully measured and recorded from 23 points in tumor using fiber-optic microprobe during the BOLD fMRI scan. The analysis results both of physiological and functional imaging parameters were as follows: pO2-Air=2.285(19.056) mmHg,pO2-Car=14.701(48.390)mmHg,ΔpO2=8.107(33.557)mmHg,ΔPSC=0.402(2.192)%,T=2.025 (8.293). (1) 10 points were identified clearly in parenchyma area of tumor. The mean value of pO2 during air inhalation [19.462(21.511)mmHg] significantly increased after Carbogen inhalation [59.904(56.710)mmHg] (U=14.000,P=0.007). (2) 5 points were identified in tumor necrosis area. The mean value of pO2 during air inhalation [0.149(0.479)mmHg] showed no significant change comparing with Carbogen inhalation[0.273 (8.050)mmHg](U=9.000,P=0.465). (3) 8 points were identified in the boundary of tumor parenchyma and necrosis areas. Among which, 5 showed the similar pO2 change to that located in tumor necrosis area, 2 showed the similar to the tumor parenchyma. However, the pO2 showed continuously decrease after Carbogen inhalation in the last 1 point. TheΔpO2 measured from the total of 23 points correlated positively toΔPSC and T value extracted from the corresponding ROI (r=0.660,0.576,P<0.01). TheΔpO2 measured from 10 points in tumor parenchyma correlated positively to ΔPSC(r=0.717,P=0.020). Conclusion“Oxy-localization map”and“oxy-amplitude map”generated from BOLD fMRI combined with non-HRF post-processing algorithm could show reliably not only the location but also the extent where the re-oxygenation occurred within tumor.