BACKGROUND:Rapid and accurate segmentation of brain tissue in medical images is of great significance for three-dimensional biomechanical modeling and diagnosis of craniocerebral injuries.Currently,artificial intelligence(AI)-based baseline models exhibit excellent generalization capabilities on large-scale datasets.However,due to the specificity and complexity of craniocerebral tissues,these models have certain limitations in their application to craniocerebral tissue segmentation.Additionally,the scarcity of craniocerebral tissue samples makes it difficult for baseline models to achieve precise segmentation results through fine-tuning.OBJECTIVE:To construct a craniocerebral tissue segmentation model based on texture feature retrieval enhancement to improve segmentation accuracy under a small number of samples.METHODS:Segment Anything in Medical Images(MedSAM)model was selected as the basic framework,and texture features were combined with deep learning to build a brain tissue segmentation model based on texture feature retrieval enhancement(DP-MedSAM).Dice Coefficient and mean intersection over union(MIoU)were selected to evaluate the efficiency of image segmentation results.In comparison with the original MedSAM model,the ablation experiment systematically evaluated the influence of key components on the model performance.The sensitivities of MedSAM,the Segment Anything Model(SAM)for medical image segmentation(SAM-Med2D)and DP-MedSAM in the mandible,left optic nerve,and left parotid gland were compared.RESULTS AND CONCLUSION:(1)By verifying the impact of the number of point prompts on segmentation results on the HaN-Seg dataset,the experimental results indicated that the optimal Dice score was achieved with the addition of three points.(2)DP-MedSAM demonstrated performance improvements compared with MedSAM and SAM-Med2D on two datasets(HaN and Public Domain Database for Computational Anatomy).Especially on the Public Domain Database for Computational Anatomy dataset,in terms of the MIoU metric,DP-MedSAM outperformed MedSAM by 6.59％and SAM-Med2D by 37.35％;in terms of the Dice metric,DP-MedSAM outperformed MedSAM and SAM-Med2D by 4.34％and 25.32％,respectively.(3)The ablation experiment results showed that removing the texture feature extraction module in the DP-MedSAM model,relying solely on original image features,led to a significant decrease in results on the test set.Furthermore,removing the vector cache database and its retrieval enhancement function from the model,which deprived the ability of the model to perform similarity retrieval using an external knowledge base,further reduced model performance.(4)Under conditions of limited data resources,the DP-MedSAM model outperformed the other two models in all evaluation metrics.The DP-MedSAM model performed excellently when processing simple and moderately difficult samples,demonstrating a clear advantage over the other two models and indicating good generalization ability.Processing the fine structures of difficult samples placed higher demands on the model's segmentation capabilities.Although the performance of the DP-MedSAM model declined slightly,it still outperformed the other two models.(5)This study proposes an innovative craniocerebral tissue segmentation model,DP-MedSAM,which improves the baseline model's performance in capturing local details and global structural information in medical images by introducing target region texture feature extraction.Through vector similarity retrieval technology,DP-MedSAM can retrieve the feature vector most similar to the current target region from a pre-constructed vector database,providing more precise guiding information for the segmentation process.

BACKGROUND:Rapid and accurate segmentation of brain tissue in medical images is of great significance for three-dimensional biomechanical modeling and diagnosis of craniocerebral injuries.Currently,artificial intelligence(AI)-based baseline models exhibit excellent generalization capabilities on large-scale datasets.However,due to the specificity and complexity of craniocerebral tissues,these models have certain limitations in their application to craniocerebral tissue segmentation.Additionally,the scarcity of craniocerebral tissue samples makes it difficult for baseline models to achieve precise segmentation results through fine-tuning.OBJECTIVE:To construct a craniocerebral tissue segmentation model based on texture feature retrieval enhancement to improve segmentation accuracy under a small number of samples.METHODS:Segment Anything in Medical Images(MedSAM)model was selected as the basic framework,and texture features were combined with deep learning to build a brain tissue segmentation model based on texture feature retrieval enhancement(DP-MedSAM).Dice Coefficient and mean intersection over union(MIoU)were selected to evaluate the efficiency of image segmentation results.In comparison with the original MedSAM model,the ablation experiment systematically evaluated the influence of key components on the model performance.The sensitivities of MedSAM,the Segment Anything Model(SAM)for medical image segmentation(SAM-Med2D)and DP-MedSAM in the mandible,left optic nerve,and left parotid gland were compared.RESULTS AND CONCLUSION:(1)By verifying the impact of the number of point prompts on segmentation results on the HaN-Seg dataset,the experimental results indicated that the optimal Dice score was achieved with the addition of three points.(2)DP-MedSAM demonstrated performance improvements compared with MedSAM and SAM-Med2D on two datasets(HaN and Public Domain Database for Computational Anatomy).Especially on the Public Domain Database for Computational Anatomy dataset,in terms of the MIoU metric,DP-MedSAM outperformed MedSAM by 6.59％and SAM-Med2D by 37.35％;in terms of the Dice metric,DP-MedSAM outperformed MedSAM and SAM-Med2D by 4.34％and 25.32％,respectively.(3)The ablation experiment results showed that removing the texture feature extraction module in the DP-MedSAM model,relying solely on original image features,led to a significant decrease in results on the test set.Furthermore,removing the vector cache database and its retrieval enhancement function from the model,which deprived the ability of the model to perform similarity retrieval using an external knowledge base,further reduced model performance.(4)Under conditions of limited data resources,the DP-MedSAM model outperformed the other two models in all evaluation metrics.The DP-MedSAM model performed excellently when processing simple and moderately difficult samples,demonstrating a clear advantage over the other two models and indicating good generalization ability.Processing the fine structures of difficult samples placed higher demands on the model's segmentation capabilities.Although the performance of the DP-MedSAM model declined slightly,it still outperformed the other two models.(5)This study proposes an innovative craniocerebral tissue segmentation model,DP-MedSAM,which improves the baseline model's performance in capturing local details and global structural information in medical images by introducing target region texture feature extraction.Through vector similarity retrieval technology,DP-MedSAM can retrieve the feature vector most similar to the current target region from a pre-constructed vector database,providing more precise guiding information for the segmentation process.

Objective:To observe the early clinical efficacy of 3D-printed modular supporting prosthesis for reconstruction of lacunar bone defect caused by giant cell tumor of the distal femur.Methods:From May 2018 to July 2023, a total of 9 patients with giant cell tumor of the distal femur were treated with 3D-printed modular supporting prosthesis to reconstruct lacunar bone defects in the Department of Orthopedics, West China Hospital, Sichuan University. There were 4 males and 5 females, aged 30.8±6.1 years (range, 24-44 years), 5 cases on the left side, 4 cases on the right side, 2 cases of Campanacci grade I, 7 cases of Campanacci grade II. The anteroposterior and lateral X-ray films and T-SMART tomosynthesis imaging of the knee joint were taken to observe the bone graft healing and osseointegration after operation. Musculoskeletal Tumor Society (MSTS)-93 was used to evaluate knee function, and visual analogue scale (VAS) was used to evaluate knee pain.Results:All patients were successfully operated and followed up for an average of 30.8±7.5 months (range, 18-42 months). The operation time was 124.2±23.6 min, and the intraoperative blood loss was 105.6±17.4 ml. All autografts showed bony union at the graft-host junction, and the healing time was 3.3±0.4 months (range, 3.0-4.0 months). At 6 months after surgery, T-SMART tomosynthesis imaging showed that the gap between the prosthesis-bone interface was less than 1 mm in all patients. At the last follow-up, the thickness of residual subchondral bone was 5.7±1.3 mm, which was greater than that before operation 2.2±0.8 mm, and the difference was statistically significant ( t=10.823, P<0.001). At the last follow-up, the score of MSTS-93 was 26.7±2.4, which was higher than that before operation 18.8±3.7, and the difference was statistically significant ( t=5.367, P<0.001). At the last follow-up, the range of motion of the knee joint was 122.8°±9.1°, which was higher than that before operation 108.3°±6.1°, and the difference was statistically significant ( t=3.970, P<0.001). All patients were able to walk normally, go up and down stairs and other daily activities, and 7 patients were able to complete squats. At the last follow-up, there was no local tumor recurrence, distant metastasis, death, joint infection, pain (VAS score was 0), delayed wound healing, joint degeneration, prosthesis loosening or articular surface collapse. Conclusion:Reconstruction of lacunar bone defect caused by giant cell tumor of distal femur with 3D-printed modular supporting prosthesis can effectively improve knee joint function and osseointegration, and the short-term clinical results are satisfactory.

BACKGROUND:Radiotherapy for head and neck tumors can easily cause xerostomia,seriously affecting the quality of life of patients.In recent years,engineered stem cells and their paracrine factors have shown therapeutic potential in the repair of salivary gland injury.However,there is currently no experimental study on the application of amniotic mesenchymal stem cell-derived exosome in radiation-induced salivary gland injury.OBJECTIVE:To preliminarily explore the repair effect of exosome derived from human amniotic mesenchymal stem cells on radiation-induced submandibular gland injury.METHODS:H uman amniotic mesenchymal stem cell exosomes were extracted and identified by ultrafiltration and ultracentrifugation.SD rats were randomly divided into a control group,a radiation injury group,and a radiation injury+exosome group.An in vitro model of radiation-induced submandibular gland injury was constructed using the submandibular gland tissue of SD rats irradiated with 18 Gy of radiation.One day after radiation modeling,exosome derived from human amniotic mesenchymal stem cells was injected into the submandibular gland in situ.Samples are taken at 1,3,7,and 14 days to detect the resting salivary flow rate.The structure of the submandibular gland tissue was observed by hematoxylin-eosin staining.The expression of glycogen particles in the submandibular gland tissue was observed by Periodic Acid-Schiff staining.Fibrosis in the submandibular gland tissue was observed by Masson staining.The secretion of salivary amylase was detected by enzyme-linked immunosorbent assay.The expression of aquaporin and tight junction proteins in submandibular gland tissue was observed by immunofluorescence staining.Real-time fluorescence quantitative PCR was used to detect the relative expression levels of aquaporins and salivary amylase mRNA in submandibular gland tissue.TUNEL assay was used to detect the apoptosis rate of submandibular gland tissues in each group.RESULTS AND CONCLUSION:After radiomodeling,compared with the radiation injury group,(1)hematoxylin-eosin staining observed that the submandibular gland tissue structure in the radiation injury+exosome group was restored,the nucleoli increased,the number of acinus increased,and the acinar atrophy improved.(2)Glycogen staining observed that the number and density of positive zymogen granules in the acinar cytoplasm of the radiation injury+exosome group gradually increased.(3)Masson staining results observed that the number and density of positive collagen fibers in the interstitium and around the ducts in the radiation injury+exosome group gradually decreased,the degree of fibrosis decreased,and the collagen deposition decreased.(4)The salivary flow rate in the radiation injury+exosome group increased(P＜0.05).The fluorescence intensity of aquaporin-5 was enhanced(P＜0.05)and the gene expression was significantly enhanced(P＜0.01).The fluorescence distribution of tight junction protein 4 was weakened and the fluorescence intensity decreased(P＜0.05,P＜0.01).The content of salivary amylase increased(P＜0.05)and gene expression were significantly increased(P＜0.01).The number of positive apoptotic cells decreased(P＜0.05,P＜0.01).It is indicated that local injection of exosome derived from human amniotic mesenchymal stem cells could improve the pathological morphology of submandibular gland tissue,promote saliva flow rate and amylase expression,and may play a functional repair role in radioactive submandibular gland injury by inhibiting acinar apoptosis.

OBJECTIVE: To evaluate the feasibility and short-term effectiveness of three-dimensional (3D)-printed customized porous acetabular components for reconstruction of extensive acetabular bone defects during primary total hip arthroplasty (THA). METHODS: The clinical data of 8 patients with extensive acetabular bone defects, who were treated with 3D-printed individualized porous acetabular components between July 2018 and January 2022, were retrospectively analyzed. The cohort comprised 4 males and 4 females with an average age of 48 years ranging from 34 to 56 years. Acetabular bone defects were classified as Paprosky type ⅢA in 3 cases and type ⅢB in 5 cases. The causes of acetabular destruction were hip tuberculosis (5 cases), pigmented villonodular synovitis (2 cases), and syphilitic arthritis (1 case). Visual analogue scale (VAS) score and Harris hip score (HHS) were used to evaluate the pain relief and hip function before and after operation. Reconstruction outcomes were further assessed by imaging results [X-ray film and Tomosynthesis Shimadzumetal artefact reduction technology (T-SMART)], and the mechanical properties were evaluated by finite element analysis. RESULTS: The operation time ranged from 174 to 195 minutes (mean, 187 minutes), and intraoperative blood loss ranged from 390 to 530 mL (mean, 465 mL). All 8 patients were follow-up 26-74 months (mean, 44 months). Among the 5 patients with tuberculosis, none experienced postoperative recurrence. At last follow-up, the VAS score was 0.3±0.5 and the HHS score was 87.9±3.7, both significantly improved compared to preoperative values ( t=25.170, P<0.001; t=-28.322, P<0.001). X-ray films at 2 years after operation demonstrated satisfactory matching between the 3D-printed customized acetabular component and the acetabulum. The postoperative center of rotation of the operated hip was shifted by (2.1±0.5) mm horizontally and (2.0±0.7) mm vertically relative to the contralateral side, with both offsets showing significant differences compared to preoperative values ( t=24.700, P<0.001; t=55.230, P<0.001). T-SMART imaging showed satisfactory osseointegration at the implant-host bone interface. No complications such as aseptic loosening or screw breakage was observed during follow-up. Finite element analysis showed that the acetabular component had good mechanical properties. CONCLUSION The application of 3D-printed individualized porous acetabular components in the reconstruction of extensive acetabular bone defects demonstrated precise anatomical reconstruction, stable mechanical support, and good functional performance in short-term follow-up, offering a potential alternative for acetabular defect reconstruction in primary THA.
Humans ; Middle Aged ; Male ; Female ; Printing, Three-Dimensional ; Arthroplasty, Replacement, Hip/instrumentation* ; Acetabulum/diagnostic imaging* ; Adult ; Follow-Up Studies ; Retrospective Studies ; Hip Prosthesis ; Prosthesis Design ; Porosity ; Treatment Outcome ; Plastic Surgery Procedures/methods*

Retinal artery occlusion(RAO)is a vascular neuro-ophthalmological emergency and a rare but serious complication of carotid artery stenting(CAS).It is characterized by a sudden decrease in vision or even vision loss in one eye.This article reports a case of central RAO caused by emboli through a branch of the external carotid artery during CAS,aiming to provide insights for optimizing CAS procedures,and preventing and controlling this complication.

Objective:To observe the early clinical efficacy of 3D-printed modular supporting prosthesis for reconstruction of lacunar bone defect caused by giant cell tumor of the distal femur.Methods:From May 2018 to July 2023, a total of 9 patients with giant cell tumor of the distal femur were treated with 3D-printed modular supporting prosthesis to reconstruct lacunar bone defects in the Department of Orthopedics, West China Hospital, Sichuan University. There were 4 males and 5 females, aged 30.8±6.1 years (range, 24-44 years), 5 cases on the left side, 4 cases on the right side, 2 cases of Campanacci grade I, 7 cases of Campanacci grade II. The anteroposterior and lateral X-ray films and T-SMART tomosynthesis imaging of the knee joint were taken to observe the bone graft healing and osseointegration after operation. Musculoskeletal Tumor Society (MSTS)-93 was used to evaluate knee function, and visual analogue scale (VAS) was used to evaluate knee pain.Results:All patients were successfully operated and followed up for an average of 30.8±7.5 months (range, 18-42 months). The operation time was 124.2±23.6 min, and the intraoperative blood loss was 105.6±17.4 ml. All autografts showed bony union at the graft-host junction, and the healing time was 3.3±0.4 months (range, 3.0-4.0 months). At 6 months after surgery, T-SMART tomosynthesis imaging showed that the gap between the prosthesis-bone interface was less than 1 mm in all patients. At the last follow-up, the thickness of residual subchondral bone was 5.7±1.3 mm, which was greater than that before operation 2.2±0.8 mm, and the difference was statistically significant ( t=10.823, P<0.001). At the last follow-up, the score of MSTS-93 was 26.7±2.4, which was higher than that before operation 18.8±3.7, and the difference was statistically significant ( t=5.367, P<0.001). At the last follow-up, the range of motion of the knee joint was 122.8°±9.1°, which was higher than that before operation 108.3°±6.1°, and the difference was statistically significant ( t=3.970, P<0.001). All patients were able to walk normally, go up and down stairs and other daily activities, and 7 patients were able to complete squats. At the last follow-up, there was no local tumor recurrence, distant metastasis, death, joint infection, pain (VAS score was 0), delayed wound healing, joint degeneration, prosthesis loosening or articular surface collapse. Conclusion:Reconstruction of lacunar bone defect caused by giant cell tumor of distal femur with 3D-printed modular supporting prosthesis can effectively improve knee joint function and osseointegration, and the short-term clinical results are satisfactory.

BACKGROUND:Radiotherapy for head and neck tumors can easily cause xerostomia,seriously affecting the quality of life of patients.In recent years,engineered stem cells and their paracrine factors have shown therapeutic potential in the repair of salivary gland injury.However,there is currently no experimental study on the application of amniotic mesenchymal stem cell-derived exosome in radiation-induced salivary gland injury.OBJECTIVE:To preliminarily explore the repair effect of exosome derived from human amniotic mesenchymal stem cells on radiation-induced submandibular gland injury.METHODS:H uman amniotic mesenchymal stem cell exosomes were extracted and identified by ultrafiltration and ultracentrifugation.SD rats were randomly divided into a control group,a radiation injury group,and a radiation injury+exosome group.An in vitro model of radiation-induced submandibular gland injury was constructed using the submandibular gland tissue of SD rats irradiated with 18 Gy of radiation.One day after radiation modeling,exosome derived from human amniotic mesenchymal stem cells was injected into the submandibular gland in situ.Samples are taken at 1,3,7,and 14 days to detect the resting salivary flow rate.The structure of the submandibular gland tissue was observed by hematoxylin-eosin staining.The expression of glycogen particles in the submandibular gland tissue was observed by Periodic Acid-Schiff staining.Fibrosis in the submandibular gland tissue was observed by Masson staining.The secretion of salivary amylase was detected by enzyme-linked immunosorbent assay.The expression of aquaporin and tight junction proteins in submandibular gland tissue was observed by immunofluorescence staining.Real-time fluorescence quantitative PCR was used to detect the relative expression levels of aquaporins and salivary amylase mRNA in submandibular gland tissue.TUNEL assay was used to detect the apoptosis rate of submandibular gland tissues in each group.RESULTS AND CONCLUSION:After radiomodeling,compared with the radiation injury group,(1)hematoxylin-eosin staining observed that the submandibular gland tissue structure in the radiation injury+exosome group was restored,the nucleoli increased,the number of acinus increased,and the acinar atrophy improved.(2)Glycogen staining observed that the number and density of positive zymogen granules in the acinar cytoplasm of the radiation injury+exosome group gradually increased.(3)Masson staining results observed that the number and density of positive collagen fibers in the interstitium and around the ducts in the radiation injury+exosome group gradually decreased,the degree of fibrosis decreased,and the collagen deposition decreased.(4)The salivary flow rate in the radiation injury+exosome group increased(P＜0.05).The fluorescence intensity of aquaporin-5 was enhanced(P＜0.05)and the gene expression was significantly enhanced(P＜0.01).The fluorescence distribution of tight junction protein 4 was weakened and the fluorescence intensity decreased(P＜0.05,P＜0.01).The content of salivary amylase increased(P＜0.05)and gene expression were significantly increased(P＜0.01).The number of positive apoptotic cells decreased(P＜0.05,P＜0.01).It is indicated that local injection of exosome derived from human amniotic mesenchymal stem cells could improve the pathological morphology of submandibular gland tissue,promote saliva flow rate and amylase expression,and may play a functional repair role in radioactive submandibular gland injury by inhibiting acinar apoptosis.

Retinal artery occlusion(RAO)is a vascular neuro-ophthalmological emergency and a rare but serious complication of carotid artery stenting(CAS).It is characterized by a sudden decrease in vision or even vision loss in one eye.This article reports a case of central RAO caused by emboli through a branch of the external carotid artery during CAS,aiming to provide insights for optimizing CAS procedures,and preventing and controlling this complication.