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.

Precancerous lesions of gastric cancer is a key stage in the development of gastric cancer.The reprogramming of glucose metabolism is a prominent feature of precancerous lesions of gastric cancer.Hypoxic microenvironment and hypoxia-inducible factors are important factors influencing the occurrence of glucose metabolic reprogramming.This article summarized the relationship between hypoxic microenvironment and the reprogramming of glucose metabolism in precancerous lesions of gastric cancer,and concluded the relevant research on TCM compounds and effective components to improve hypoxic microenvironment and further regulate glycolysis for the treatment of this disease.It was concluded that the mechanism may be the inhibition of angiogenesis,regulation of signaling pathways and key proteins of glycolysis,expression of multiple enzymes,reduction of lactate secretion,inhibition of cell malignant proliferation and invasion.It explored the mechanism of Chinese materia medica in improving hypoxic microenvironment and regulating glycolysis,so as to provide reference for the prevention and treatment of precancerous lesions of gastric cancer.

The angiogenic microenvironment is a new blood vessel with different molecular and functional characteristics that sprouts on the original blood vessels through different mechanisms,which directly affects the process of tumor cell growth,proliferation,and migration and has an important impact on the occurrence and development of precancerous lesions of gastric cancer.Correa mode has shown that precancerous lesions of gastric cancer is the key pathological stage before the occurrence of gastric cancer,and it is of great significance to advance the prevention and treatment strategy to this stage.TCM believes that qi deficiency and blood stasis is the key pathogenesis of precancerous lesions of gastric cancer,and its basic treatment is to replenish qi and remove blood stasis,and based on the syndrome differentiation,drugs with the efficacy of nourishing yin and tonifying stomach,soothing the liver and regulating qi,resolving phlegm and dispersing lumps,and clearing heat and dampness for treatment.This article discussed the correlation between precancerous lesions of gastric cancer and angiogenic microenvironment and its regulatory pathways,and summarized the methods and mechanisms of TCM in the treatment of precancerous lesions of gastric cancer from the perspective of regulating angiogenic microenvironment-related pathways,in order to provide a reference for the treatment of precancerous lesions of gastric cancer with TCM.

To investigate the effect of Zhiwei Fuwei Pills (ZWFW) on the expression of mammalian target of rapamycin (mTOR)/autophagy key molecule yeast Atg6 homologue (Beclin1)/microtubuleassociated protein 1 light chain 3 (LC3) signaling axis key molecules in gastric antrum tissue of rats with precancerous gastric lesions (PLGC). METHODS: SPF SD rats were randomly divided into normal group, model group, folic acid group, ZWFW low-dose, medium-dose, high-dose group. In addition to the normal group, the model group, folic acid group, ZWFW low-dose, medium-dose and high-dose groups, were used to establish the PLGC rat model by five factors compound modeling methods: N-methyl-N ' - nitro-n-nitroguanidine (MNNG) combined with hunger and satiation, ethanol intragastric administration, free drinking of ammonia and ranitidine feed. The rats were treated with normal saline, folic acid tablet aqueous solution (0.002 g/kg), ZWFW low-dose, medium-dose, high-dose aqueous solution (0.42, 0.84, 1.67 g/kg) for 4 weeks, and the stomach was removed by laparotomy. Hematoxylineosin (HE) staining was used to observe the histopathological changes in the antrum of rats, and real-time polymerase chain reaction (real-time PCR), Western blot (WB) and immunohistochemistry (IHC) were used to detect the expression of mammalian target of rapamycin mTOR, yeast Atg6 homologue 1 (Beclin1), microtubule-associated protein 1 light chain 3β (LC3B) mRNA and protein in the antrum of rats. RESULTS: Compared with the normal group, the Gastric antrum tissue of the model group was distended, thinner gastric wall, palegastric mucosa, atrophic and flat folds, disordered course and nodules and vegetations were visible. HE staining showed that compared with the normal group, the gastric mucosal glands in the model group were crowded and disordered, and the cell morphology was different, including a large number of goblet cells, basophilic cytoplasm, large, hyper-chromatic and irregular nuclei, and mucosal muscle infiltration and destruction. Compared with the model group, treated by ZWFW can significantly improve the pathological manifestations of gastric mucosal gland structure disorder and cell atypia. Compared with the normal group, mTOR mRNA and protein expression were significantly increased (P< 0.05) and Beclin1 and LC3B mRNA and protein expression were significantly decreased (P<0.05) in the antral tissue of rats in the model group; compared with the model group, mTOR mRNA and protein expression were decreased (P<0.05) in the medium and high dose groups of ZWFW, Beclin1 and LC3B protein expression in the antral tissue of rats in the low dose group of ZWFW and Beclin1 and LC3B mRNA and protein expression were increased (P<0.05) in the medium and high dose groups. CONCLUSION: Zhiwei Fuwei Pills can significantly improve the abnormal histopathological findings of gastric mucosa in PLGC model rats, and the mechanism may be related to the down-regulation of mTOR expression, up-regulation of Beclin1 and LC3B expression and then promoting autophagy.