OBJECTIVE:Based on different algorithms of machine learning,the prediction model of lumbar disc herniation has become a trend and hot spot in the development of precision medicine.However,there is limited evidence on the reporting quality and methodological quality of prediction models of lumbar disc herniation outcomes using machine learning.This article is aimed to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation by comprehensively analyzing the report quality and risk of bias of previous studies that developed and validated prognosis prediction models based on machine learning through a comprehensive literature search,in order to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation.METHODS:The databases of CNKI,WanFang,VIP,SinOMED,PubMed,Web of Science,Embase,and The Cochrane Library were searched by computer.Studies on the use of machine learning to develop(and/or validate)prognostic prediction models for lumbar disc herniation were collected from the inception of the database to December 31,2023.Two researchers independently screened the literature,extracted data,and assessed the risk of bias of the included studies.The reporting quality and risk of bias of the included studies were assessed by the Multivariable Transparent Reporting of Predictive Models(TRIPOD)statement and the Predictive Model Risk of Bias Assessment Tool(PROBAST).The results of the evaluation were analyzed using descriptive statistics and visual charts.RESULTS:(1)A total of 23 articles were included,and the TRIPOD compliance of each study ranged from 11％to 87％,with a median compliance of 54％.The quality of reporting of titles,detailed descriptions of treatment measures,blinding of predictors,handling of missing data,details of risk stratification,specific procedures for enrollment,model interpretation,and model performance was mostly poor,with TRIPOD adherence rates ranging from 4％to 35％.(2)Of all included studies,61％had a high risk of bias and 39％had an unclear overall risk of bias.The area under the curve,accuracy,sensitivity and specificity were used to evaluate the performance of the model.The areas under the curve of 20 models were reported,ranging from 0.561 to 0.999.Three models reported the accuracy of the model,ranging from 82.07％to 89.65％.(3)Among all included studies,the statistical analysis domain was most often assessed as having a high risk of bias,mainly due to the small number of valid samples,the selection of predictors based on univariate analysis and the lack of calibration and discrimination assessment of the model in the study.CONCLUSION:These results indicate that machine learning can achieve good predictive ability in the development and validation of prognostic models for lumbar disc herniation.The commonly used algorithms include regression algorithm,support vector machine,decision tree,random forest,artificial neural network,naive Bayes and other algorithms.Reasonable algorithms combined with clinical practice can improve the accuracy of prognosis prediction of lumbar disc herniation.However,the reporting and methodological quality of prognosis prediction models based on machine learning are poor,the prediction performance of different models varies greatly,and the generalization and extrapolation of research models are unclear.There is an urgent need to improve the design,implementation and reporting of such studies.To promote the application of machine learning in the clinical practice of lumbar disc herniation prediction models,it is necessary to comprehensively consider various predictors related to the prognosis of the disease before modeling,and strictly follow the relevant standards of PROBAST tool during modeling.

OBJECTIVE:Based on different algorithms of machine learning,the prediction model of lumbar disc herniation has become a trend and hot spot in the development of precision medicine.However,there is limited evidence on the reporting quality and methodological quality of prediction models of lumbar disc herniation outcomes using machine learning.This article is aimed to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation by comprehensively analyzing the report quality and risk of bias of previous studies that developed and validated prognosis prediction models based on machine learning through a comprehensive literature search,in order to explore the performance of machine learning algorithms in predicting the prognosis of lumbar disc herniation.METHODS:The databases of CNKI,WanFang,VIP,SinOMED,PubMed,Web of Science,Embase,and The Cochrane Library were searched by computer.Studies on the use of machine learning to develop(and/or validate)prognostic prediction models for lumbar disc herniation were collected from the inception of the database to December 31,2023.Two researchers independently screened the literature,extracted data,and assessed the risk of bias of the included studies.The reporting quality and risk of bias of the included studies were assessed by the Multivariable Transparent Reporting of Predictive Models(TRIPOD)statement and the Predictive Model Risk of Bias Assessment Tool(PROBAST).The results of the evaluation were analyzed using descriptive statistics and visual charts.RESULTS:(1)A total of 23 articles were included,and the TRIPOD compliance of each study ranged from 11％to 87％,with a median compliance of 54％.The quality of reporting of titles,detailed descriptions of treatment measures,blinding of predictors,handling of missing data,details of risk stratification,specific procedures for enrollment,model interpretation,and model performance was mostly poor,with TRIPOD adherence rates ranging from 4％to 35％.(2)Of all included studies,61％had a high risk of bias and 39％had an unclear overall risk of bias.The area under the curve,accuracy,sensitivity and specificity were used to evaluate the performance of the model.The areas under the curve of 20 models were reported,ranging from 0.561 to 0.999.Three models reported the accuracy of the model,ranging from 82.07％to 89.65％.(3)Among all included studies,the statistical analysis domain was most often assessed as having a high risk of bias,mainly due to the small number of valid samples,the selection of predictors based on univariate analysis and the lack of calibration and discrimination assessment of the model in the study.CONCLUSION:These results indicate that machine learning can achieve good predictive ability in the development and validation of prognostic models for lumbar disc herniation.The commonly used algorithms include regression algorithm,support vector machine,decision tree,random forest,artificial neural network,naive Bayes and other algorithms.Reasonable algorithms combined with clinical practice can improve the accuracy of prognosis prediction of lumbar disc herniation.However,the reporting and methodological quality of prognosis prediction models based on machine learning are poor,the prediction performance of different models varies greatly,and the generalization and extrapolation of research models are unclear.There is an urgent need to improve the design,implementation and reporting of such studies.To promote the application of machine learning in the clinical practice of lumbar disc herniation prediction models,it is necessary to comprehensively consider various predictors related to the prognosis of the disease before modeling,and strictly follow the relevant standards of PROBAST tool during modeling.

Objective To explore CTA and MRA characteristics of persistent trigeminal artery (PTA)with vascular disease. Methods 54 patients diagnosed as PTA by CTA and MRA were collected retrospectively,whose classification,variation and associated vascular disease were analyzed.Results There were more females than males in 54 patients,whose average age was 59.72±16.32,and the incidence of left ones were higher than that of the right ones.Type Ⅰ of Weon classification was 16 cases (29.63%),type Ⅱ 7 cases (12.96%),typeⅢ 16 cases (29.63%),type Ⅳ 1 1 cases (20.37%),type Ⅴ 4 cases (7.41%).The lateral type of Salas classification was significantly higher than that in the middle type.There were 16 cases (29.63%)with aneurysm,in which 3 cases (18.75%)were rupture.3 cases were with moyamoya disease (5.56%).39 cases (72.22%)were with basilar artery dysplasia.There were 33 cases (61.11%)with the postembryonic cerebral artery.There were 29 cases with cerebral infarction (53.7%),4 cases with cerebral hemorrhage (7.4%),3 cases with subarachnoid hemorrhage (5.5 6%).Cerebral arteriosclerosis with different degrees of stenosis was showed in 1 7 cases (31.48%). 2 cases were accompanied with trigeminal neuralgia (3.7%),and only 1 case was with ocular nerve palsy (1.85%).Conclusion CTA and MRA could be used to display the anatomical features and radiographic classification of PTA ,and also clearly display and evaluate PTA associated with vascular disease.

Purpose To investigate the value of arterial spin labeling (ASL) in evaluating renal function in patients with renal cell carcinoma (RCC) after laparoscopic partial nephrectomy.Materials and Methods Fifteen patients with RCC undergoing laparoscopic partial nephrectomy were studied prospectively.The patients were performed ASL scan one week before and three months after operation.The correlation between renal blood flow (RBF) value measured by ASL and the glomerular filtration rate (GFR) measured by radionuclide method in the renal cortex of healthy side was analyzed.The RBF values in the kidney of affected side or healthy side were measured,the difference of which between before operation and three months after operation was compared.Results The RBF value and GFR data in the renal cortex of healthy side had positive correlation (r=0.638,P＜0.05).In the affected side of kidney,the RBF value of remaining renal tissue [(291.5 ± 37.3) ml/(100g·min)] compared with that of preoperative renal tissue [(237.8 ± 46.2) ml/(100g·min)]increased about 53.7 ml/(100g · min) (P＜0.05).In the healthy side of kidney,the RBF value of renal tissue [(241.1 ± 50.3) ml/(100 g · min)] compared with that of preoperative renal tissue [(290.4 ± 51.8) ml/(100 g·min)] decreased about 49.3 ml/(100 g·min) (P＜0.05).Conclusion ASL can be used to evaluate renal function,and it is valuable to evaluate renal perfusion function after laparoscopic partial nephrectomy of RCC.

Objective: To observe the expression of long non-coding RNA myocardial infarction associated transcript (LncRNA-MIAT) in tumor necrosis factor-α (TNF-α) induced endothelial cells (ECs) inflammation in vitroand to study the impact of LncRNA-MIAT on inflammatory regulation. Methods: LncRNA-MIAT expression in ECs was induced by TNF-α at different time and concentration. Expressions of intercellular adhesion molecule-1 (ICAM-1) and LncRNA-MIAT in inflammatory ECs were examined by quantitative real time polymerase chain reaction (qRT-PCR) and Western blot analysis. Moreover, ECs was transfected by siRNAMIAT to observe the effect of LncRNA-MIAT knock-down on ICAM-1 expression. Results: LncRNA-MIAT expression showed the increasing trend by elevated time and concentration of TNF-stimulation. Compared with TNF-α stimulation at 0h, 6h and 12h, LncRNA-MIAT expressions were increased at 24h and 48h of TNF-αstimulation respectively, allP<0.05; compared with TNF-α concentration at 0ng/ml and 0.125ng/ml, LncRNA-MIAT expressions were elevated by TNF-α stimulation at 1.000ng/ml and 10.000ng/ml respectively, allP<0.05. With siRNAMIAT knock-down, TNF-α induced ICAM-1 protein expression was significantly reduced in ECs,P<0.05. Conclusion: LncRNA-MIAT might be involved in ECs inflammatory response and it may play a role to promote inflammation.

Objective To explore the correlationships between microperfusion diffusion indexes derived from intravoxel incoherent motion(IVIM)and perfusion values measured by arterial spin labeling (ASL)in renal allograft. Methods A total of 76 renal allograft recipients and 26 age-matched volunteers (group 0)were included in this prospective study. All subjects were underwent conventional MRI, IVIM and ASL MRI which were performed in the oblique-sagittal plane. Seventy-six recipients were divided into two groups based on the estimated glomerular filtration rate(eGFR):recipients with good allograft function(group 1, eGFR≥ 60 ml · min-1 · 1.73m-2,n=44)and recipients with impaired allograft function(group 2, eGFR<60 ml · min-1 · 1.73m-2,n=32). Three IVIM indexes values, including true diffusion coefficient(ADCslow), pseudo-diffusion coef fi cient(ADCfast), perfusion fraction(PF), and one ASL index value of renal cortex(renal blood flow, RBF)were measured. One-way analysis of variance and the least significant difference were used to compare the different of each cortical index values among three groups. Correlations between the ADCslow, ADCfast, PF, RBF and eGFR as well as the correlation among the indexes were evaluated using Pearson correlation coefficients. Results For cortical ADCslow, ADCfast, PF and RBF values, allografts with good function and impaired function showed significantly differences compared healthy controls(all P<0.01). In allografts with good function, cortical ADCslow,ADCfast,PF showed no significantly differences compared with controls(all P>0.05), but RBF value was significantly lower(P<0.05). The ADCslow, ADCfast, PF and RBF values of renal cortex were significantly lower in allografts with impaired function compared to allografts with good function(all P<0.01). In renal allografts, there were significant positive correlations between cortical ADCslow, ADCfast, PF, RBF value and eGFR(r values were 0.604, 0.552, 0.579 and 0.673, all P<0.01). Cortical ADCfast and PF value exhibited a significant correlation with RBF for recipients(r values were 0.501 and 0.423, all P<0.01). Conclusion Cortical ADCfast and PF values derived from IVIM and RBF measured by ASL show a significant positive correlation in renal allografts.

Objective To explore the clinical application of material decomposition technique on spectral CT imaging and evaluate hemodynamic changes in different liver lobes with liver cirrhosis.Methods 30 patients with liver cirrhosis diagnosed clinically in our hospital were collected and underwent enhanced scanning of abdomen with spectral CT protocol.The monochromatic energy images and iodine-based material decomposition (MD) images were reconstructed after scanning.The iodine concentration (IC) was measured in five liver lobes (the caudate, left lateral, left inner, right anterior and right posterior lobes) and the abdominal aorta of the same axial slice in both the arterial phase (AP) and portal venous phase (VP) on the iodine-water based material decomposition images.The arterial iodine fraction (AIF) and the portal venous iodine concentration (PVIC) as well as the normalized iodine concentration (NIC) during the AP and VP were calculated.The differences of IC,the NIC,the AIF,and the PVIC in five liver lobes in AP and VP were compared by using single factor analysis of variance.Results The IC,the NIC in both AP and VP and the AIF of the caudate liver lobe were higher than those of other four liver lobes, with statistically significant difference (P<0.05), while these values in the other four lobes showed no statistically significant difference (P>0.05).The PVIC of the caudate liver lobe was slightly lower than that of the other four liver lobes, however, the difference was not statistically significant (P=0.929).Conclusion The quantitative iodine concentration measurement of liver lobes on spectral CT material decomposition technique can evaluate the hemodynamic changes in liver lobes with liver cirrhosis,and provide more information about the change of blood flow in liver cirrhosis.

Objective To assess the value of arterial spin labeling(ASL) MRI in the staging of early renal allograft function. Methods Sixty two renal allograft recipients (2 to 4 weeks after kidney transplantation) and 20 age match volunteers were included in this study. All subjects underwent conventional MRI and ASL MRI which was performed in the oblique-sagittal plane. Recipients were divided into two groups according to the estimated glomerular filtration rate (eGFR), recipients with good allograft function (eGFR≥60 ml · min-1 · 1.73 m-2,n=37) and recipients with impaired allograft function (eGFR<60 ml · min - 1 · 1.73 m - 2,n=25). Renal blood flow (RBF) was measured and an intra-class correlation coefficient (ICC) was calculated to confirm the reproducibility of the measured results from two doctors. One-way analysis of variance (ANOVA) and Bonferroni were used to compare the different cortical RBF among three groups. Correlation of RBF with eGFR was evaluated using Pearson correlation coefficients. The receiver operating characteristic (ROC) curve was performed to assess the diagnostic efficacy of using cortical RBF to discriminate allografts with impaired function from good function. Results RBF values showed good reproducibility between doctors with an ICC larger than 0.90 in different group. Mean cortical RBF were (390 ± 61),(290 ± 69),(201 ± 86) ml · 100 g-1 · min-1 for healthy controls, recipients with good and impaired allograft function, respectively(F=37.313,P<0.01). RBF exhibited a significant correlation with renal function as determined by eGFR for recipients (r=0.60,P<0.01). Mean cortical RBF showed a high area under the ROC curve (0.773) to discriminate renal allografts with different function, with a sensitivity of 56.0% (14/25) and a specificity of 89.2% (33/37). Conclusion ASL MRI can assess the early renal allografts perfusion, and provide valuable information in the staging of renal function. It could be a useful method for evaluating renal function noninvasively.

Objective To analyse the efficacy of microvascular decompression for hemifacial spasm (HFS) caused by vertebral basilar artery compression. Methods A total of 141 patients with HFS treated by microvascular decompression in our hospital were collected in this study. The improvement of the symptoms after operation was compared between patients with HFS caused by vertebral basilar artery compression (28 cases) and patients with HFS caused by non-vertebral basilar artery compression (113 cases). Results There was no significant difference in the effective rate between the two groups of HFS (96.43%vs. 98.23%,P=0.49) with mean following-up 13.81 ± 1.57 months. And there was no significant difference in the delayed cure rate after surgery between two groups (37.04%vs. 20.72%,χ2=1.38, P>0.05). Conclusion Microvascular decompression is a safe and effective method for the treatment of HFS caused by compressed vertebral basilar artery.

Objectives To investigate the ability of T1 mapping and intravoxel incoherent motion imaging (IVIM) parameters for evaluating renal allografts at the early stage after renal transplantation. Methods This prospective study protocol was approved by the local ethics committee, and written informed consent was obtained from all subjects. Sixty two recipients 2 to 4 weeks after kidney transplantation and 20 healthy volunteers (control group) underwent routine MRI, T1 mapping, and IVIM imaging (11 b values, 0 to 700 s/mm2). Recipients were divided into two groups base on their estimated glomerular filtration rate (eGFR):37 recipients with good allograft function (eGFR≥60 ml·min-1·1.73 m-2) and 25 recipients with impaired allograft function (eGFR<60 ml·min-1·1.73 m-2). The ADC, true diffusion coefficient (ADCslow), pseudo-diffusion coefficient (ADCfast), perfusion fraction (f) and T1 values were measured on both cortex and medulla. Differences among groups were compared using the one-way analysis of variance. Correlations between eGFR and the parameters in renal allografts were assessed by using Pearson correlation analysis. ROC was performed to assess the diagnostic utilities of using these parameters to discriminate allografts with impaired function from good function. Results Excepting for cortical T1, ADCfast and medullary T1, f values, allografts with good function showed no differences in other parameters compared with healthy control. Excepting for medullary T1 and ADCfast,the other values showed significantly differences in allografts with impaired function compared to allografts with good function (all P<0.05). Excepting for medullary f and ADCfast values, allografts with impaired function showed significantly differences in the parameters compared with good function group(all P<0.05). In renal allografts, excepting for medullary T1, ADCfast, and f values, cortical T1 exhibited a negative correlation with renal function, and there was a significant positive correlation between eGFR and other parameters. Cortical T1 value showed high sensitivity(91.9%) to discriminate renal allografts with different function, with the threshold of 17.36 × 102 ms, and ADC value showed high specificity(96.0%)with the threshold of 1.98 × 10-3 mm2/s. Conclusion T1 mapping and IVIM technique may be useful for detecting renal allograft dysfunction, and be a reliable imaging for evaluating and monitoring allograft function.