Objective Precise pose estimation of intraoperative X-ray images relative to preoperative computed tomography(CT)volumes is a fundamental aspect of 2D/3D registration,which is essential for improving the accuracy of medical analy-sis and integration in various fields,including surgical planning,radiation therapy,and neural navigation.Traditional methods can be divided into intensity-and feature-based approaches.In intensity-based methods,different ray-tracing techniques are typically used to project 3D volumes for generating simulated 2D X-rays,which are known as digitally recon-structed radiographs(DRR).An optimizer is then employed to find the optimal spatial transformation that maximizes the similarity between the DRR and the corresponding X-ray.Intensity-based methods can achieve high accuracy,but these methods encounter some challenges,such as long registration time due to iterative pose search and the need for generating numerous DRRs for similarity calculation.Moreover,iterative pose search relies on good initialization;otherwise,it may converge to local maximization,which results in registration failure.Meanwhile,feature-based methods typically utilize landmarks for matching to determine the spatial transformation,which can be specific points of anatomical structures or key points.Alternatively,feature detection operators such as Harris,or segmentation can also be employed to extract features.These methods extract features while filtering out a large amount of image information,which leads to higher computation efficiency but lower accuracy than intensity-based methods.Recently,deep learning-based models have emerged as power-ful tools for medical image registration.Their feature extraction capability has efficiently addressed the time-consuming issue of traditional methods.The suitable network is designed to learn the feature representation that can describe the com-plex mapping relationship between images and corresponding labels.This network directly regresses transformation param-eters and avoids the need for extensive searching and sampling in the pose space.The intraoperative planar images of com-plete spines cannot establish rigid correspondence with the preoperative CT,and existing registration algorithms commonly face issues such as low registration and insufficient robustness when dealing with the complete structures of the spine.To address the issue,we proposed a registration method that combines local detail features and global position features for 2D/3D rigid registration in a single vertebra manner.Method Convolutional neural networks can enhance the ability of the model to learn the shape,boundaries,and local structures of vertebra,while the Transformer uses the self-attention mecha-nism to effectively capture global dependencies between images and extract key features of the vertebra.Therefore,by com-bining the characteristics of both structures,we proposed a multi-stage dual-branch network to effectively extract local and global features from single vertebra images for learning the relationship between features and spatial transformations.The aim was to improve the performance of the regressor.In each stage,the local branch utilizes down-sampling operation and stacked convolution blocks to capture details,edges,and other local information more effectively while reducing computa-tional load.The global branch employs convolution-based patch embedding and multi-head self-attention mechanism to capture the positional relationship between various features and reduce the interference of background in images.The fea-ture fusion module,which is based on the channel and spatial attention mechanism,maps the features of different branches to the same feature space and adaptively fuses local details and global contextual features at different stages,which enhances the expressive ability of the model.Moreover,we progressively optimize feature representations in a coarse-to-fine manner within the network to better capture relevant information.This approach improves the perceptual capability of the network at various scales and levels.Finally,we incorporate auxiliary registration heads to predict pose parameters from the multi-stage fused features,which enhances the supervision information available to network and helps the network gradually optimize pose predictions during training.Ultimately,the final registration accuracy is improved.The input of the network is a single grayscale low-dose X-ray image,and the output is the predicted pose parameters of 6 degrees of free-dom(DoFs),which are used to obtain the registered image by the projecting operation.Our network is implemented using the PyTorch framework.The input images are resized to 128 × 128 pixels for training,the learning rate is set to 6e-5,and the weight decay is 0.05.The Adam learning procedure is accelerated using a NVIDIA GeForce RTX 3090 GPU device,which takes approximately 6 h for 320 iterations.Result We compared our model with 5 state-of-the-art models on 30 simu-lated datasets,including 3 iterative optimization-based methods(OPT-GO,OPT-GC,and OPT-NGI)and 2 deep learning-based methods(ResRegNet and EFbackbone-based).The quantitative evaluation metrics contained mean target registra-tion error(mTRE),mean absolute error(MAE),and the registration time,and we provided several visualized registration results of each method.Comparative experiments demonstrated that our model outperforms other methods on the single ver-tebra datasets.The visualized results also indicate that the registered images have minimal pose deviation from the corre-sponding target images and demonstrate good alignment,which prove that our proposed method can improve the accuracy in the single vertebra 2D/3D image registration tasks.The average mTRE is 1.40 mm,and the average MAE of 6DoF pose parameters is 0.008.Compared with the second-best method,our model reduces mTRE by approximately 2.70 mm and MAE by 0.02.Furthermore,we conducted a series of ablation experiments and provided corresponding quantitative met-rics.These experiments clearly demonstrate the effectiveness of each module in our model,including the local branch,global branch,dual-branch feature fusion module,and auxiliary registration heads.Conclusion In this study,we proposed a dual-branch single vertebra 2D/3D registration model that contains feature fusion modules to integrate local and global fea-tures,which improves the registration accuracy.Moreover,the auxiliary registration heads realized through the features at different layers can enhance supervision information,which increases the stability of the registration model.The experi-mental results show that our model outperforms several state-of-the-art registration methods and further enhances the regis-tration performance.
Endometrial injury disease is a kind of uterine disease of thin endometrium and intrauterine adhesion caused by endometrial basal layer damage.The current clinical use of drug therapy and hysteroscopy surgery can not have good curative effects.Mesenchymal stem cells(MSCs)and platelet-rich plasma(PRP)can secrete a variety of active factors to regulate the inflammatory response and immune response of endometrial,promote endometrial proliferation and repair,and there is a mutual promoting effect between the two.Combined use can provide a new idea for the clinical treatment of endometrial injury diseases.
Objective To evaluate the preventive effect and mechanism of Yigan Powder(YGP)on the development of acute stress dis-order(ASD)to post-traumatic stress disorder(PTSD)in mice.Methods Male C57BL/6J mice were randomly divided into the nor-mal,model,YGP low-dose,YGP high-dose,and sertraline hydrochloride groups.A mouse model of ASD progressing to PTSD was establishedusing foot shock.The YGP low and high-dose groups were administered0.5 and1g·kg-1 of YGP respectively,while the sertraline hydrochloride group received0.01g·kg-1 of sertraline hydrochloride.The normal and model groups were given normal saline.All treatments were gavaged once daily for 27 consecutive days.Behavioral changes in mice were evaluated in the acute and chronic pha-ses using the open-field test,trauma cue avoidance test,contextual fear test,and behavioral characteristic analysis.Hematoxylin-eo-sin(HE)staining was used to observe morphological changes in the mouse hippocampus.Immunofluorescence double-labeling staining was employed to detect the proliferation and differentiation of hippocampal stem cells in mice.Results Compared with the normal group,the model group mice showed a decreased percentage of central area movement distance in the open-field test,reduced time spent explo-ring objects,and increased immobility time during both the acute and chronic phases(P<0.05,P<0.01).There was a positive corre-lation between the time spent exploring objects and immobility time during the acute and chronic phases(P<0.05).The hippocampal tissue cells exhibited abnormal morphology,loose structure,and reduced cell numbers.The proliferation and differentiation of hippocam-pal stem cells were also reduced(P<0.01).After drug administration,compared with the model group,the YGP low-dose,YGP high-dose,and sertraline hydrochloride groups showed an increased percentage of central area movement distance in the open-field test(P<0.05,P<0.01).The YGP low-dose group exhibited increased time spent exploring objects(P<0.05),while the YGP high-dose group showed reduced immobility time(P<0.05).The hippocampal tissue cells exhibited morphological and structural recovery with in-creased cell numbers.The YGP high-dose group and sertraline hydrochloride group showed increased proliferation and differentiation of hippocampal stem cells,with the YGP high-dose group exhibiting larger increase compared to the sertraline hydrochloride group(P<0.01).Conclusion YGP promotes hippocampal neuronal regeneration in mice with ASD progressing to PTSD,protects neuronal struc-ture and function,relieves anxiety,avoidance behavior and re-experiencing behavior in the mice,thus preventing the progression of ASD to PTSD.
[Objective]To explore the correlation between general anesthesia and mild cognitive impairment in older adults so as to provide new ideas for early prevention and timely intervention of mild cognitive impairment(MCI).[Methods]Based on the baseline survey of the Hubei memory and aging cohort study(2018-2023),the participants completed a thorough neuropsychological assessment and physical examination,and self-reported a history of general anesthesia and surgery.The association of general anesthesia and MCI in the elderly was analyzed using the logistic regression model.In addition,the stratification and interaction analysis of anesthesia history,anesthesia number and physical intellectual exercise were conducted separately.[Results]A total of 5 069 older adults aged 65 and above were included in this study,including 3 692 city dwellers and 1 377 rural people,among whom were 2 584 women(51%).Out of the 1 472 participants with history of general anesthesia,249 people(17.4%)had MCI.After controlling for confounding factors,there was a 39.6%increased risk of MCI in older adults who underwent general anesthesia[OR=1.396,95%CI(1.169,1.668),P<0.001],suggesting that general anesthesia may be an independent influence on MCI.For the older adults who had one general anesthesia[OR=1.235,95%CI(1.001,1.523),P=0.049],two general anesthesia[OR=1.779,95%CI(1.292,2.450),P<0.001],and three OR more general anesthesia[OR=2.395,95%CI(1.589,3.610),P<0.001],their risks of MCI were increased by 23.5%,77.9%,and 139.5%,respectively.Compared with the older adults without a history of general anesthesia who did not exercise,the risk of developing MCI was significantly negatively correlated with the exercise group,cognitive exercise group,and combined exercise and cognitive exercise groups(all P<0.001).The risk of developing MCI in the exercise group was 60.2%of that in the no exercise group[OR=0.602,95%CI(0.456,0.795)],the risk in the cognitive exercise group was 42.4%of that in the no exercise group[OR=0.424,95%CI(0.294,0.613)],and the risk in the combined exercise and cognitive exercise group was 27.0%of that in the no exercise group[OR=0.270,95%CI(0.208,0.353)].In the older adults with a history of general anesthesia,compared with the no exercise group,the risk of developing MCI was significantly negatively correlated with the cognitive exercise group and the combined exercise and cognitive exercise group(all P<0.05).The risk of developing MCI in the cognitive exercise group was 47.7%of that in the no exercise group[OR=0.477,95%CI(0.256,0.892)],the risk in the combined exercise and cognitive exercise group was 34.5%of that in the no exercise group[OR=0.345,95%CI(0.220,0.540)],while the risk in the exercise-only group did not show a significant difference.[Conclusion]The risk of MCI increased significantly in older adults with a history of general anesthesia,and this risk increased with the times of anesthesia.Physical and mental exercise reduces the risk of MCI.it is recommended that older adults with a history of anesthesia incorporate physical and mental exercise into their daily lives to prevent mild cognitive impairment.
Objective To investigate the factors influencing the efficacy of intravesical Bacille Calmette-Guérin(BCG)instillation after transurethral resection of bladder tumor(TURBT)in patients with intermediate-and high-risk non-muscle invasive bladder cancer(NMIBC),and to construct a prediction model for recurrence after BCG treatment.Methods A retrospective cohort study was conducted on the subjected patients diagnosed with intermediate-and high-risk NMIBC undergoing TURBT followed by standard BCG instillation.The 110 patients treated in Department of Urology of Army Medical Center of PLA from January 2018 to December 2023 were assigned into a training set,while the 52 patients treated at Department of Urology of General Hospital of Central Theater Command from January 2015 to December 2020 were into an external validation set.A total of 17 variables were included and analyzed.Univariate and multivariate Cox regression analyses were performed to identify factors associated with recurrence after BCG instillation,and nomograms were plotted to predict 1-year,3-year,and 5-year recurrence-free survival(RFS).Calibration curve,decision curve analysis(DCA),and receiver operating characteristic(ROC)curve analysis were conducted for internal and external validation to evaluate the predictive performance and clinical utility of the model.Results In the training set,26 patients(23.64%)experienced recurrence during the follow-up period,with a median RFS of 32.00(18.00~50.50)months.Univariate Cox regression analysis suggested that platelet count,eosinophil to lymphocyte ratio(ELR),neutrophil to lymphocyte ratio(NLR),platelet to lymphocyte ratio(PLR),systemic immune inflammation(SII)index,and neutrophil-monocyte to lymphocyte ratio(NMLR),pathological T1 stage(pT1)tumor and hemoglobin,albumin,lymphocyte,and platelet(HALP)score were potential factors influencing recurrence after BCG instillation.Multivariate Cox regression analysis identified high HALP score(HR=0.185,95%CI:0.046~0.736,P=0.017)as an independent protective factor,while high ELR(HR=3.599,95%CI:1.505~8.608,P=0.004)and pT1 stage(HR=3.240,95%CI:1.191~8.818,P=0.021)were independent risk factors for recurrence.Based on this,a nomogram prediction model was constructed.The calibration curves demonstrated good agreement between predicted and actual 1-,3-,and 5-year recurrence risks.Decision curve analysis indicated clinical utility across a wide threshold probability range.In the training set,the model showed strong predictive performance for 1-(AUC=0.842),3-(AUC=0.847),and 5-year(AUC=0.887)recurrence risks,which was further validated in the external cohort.Conclusion Higher HALP score prior to BCG instillation therapy is a protective factor against tumor recurrence,while higher ELR and pT1 stage are risk factors.Our nomogram prediction model based on HALP score,ELR and pathological T stage,can identify individuals at high risk of recurrence after BCG instillation therapy.