This review summarizes the applications and advancements of artificial intelligence(AI)in the analysis of retinal vascular parameters. Retinal vascular parameters, including vessel diameter, fractal dimension, vascular tortuosity, branching angles, and vessel density, are important indicators for assessing changes in the retinal vascular network structure. These parameters are not only related to various ophthalmic diseases but also reflect the conditions of systemic diseases such as diabetes and Alzheimer's disease. This article provides a detailed discussion on the advantages of AI technology in the automated identification and quantification of retinal vascular parameters, particularly in improving measurement efficiency and accuracy, and enabling the early detection and monitoring of various diseases. Additionally, the challenges faced by AI in the analysis of retinal vascular parameters were discussed, such as data standardization and insufficient sample diversity, and proposes directions for future research. By thoroughly analyzing the application of AI in retinal vascular parameter analysis, this article aims to offer new perspectives and methods for clinical diagnosis and early intervention of diseases, holding significant clinical significance and application prospects.

Objective By analying the concentration degree of China's medical service market from 2017 to 2021,to analyze the current situation of competition in China's medical service market,and put forward relevant policy rec-ommendations to provide reference for optimizing resource allocation in China's medical service market.Methods The competition of medical service market in each province(region and city)was calculated by Herfindahl-Hirschman In-dex(HHI)and visualized by Stata18 software.Results From 2017 to 2021,the annual growth rate of the number of beds and health technicians in most regions of China was higher than 2.5％,but the absolute number was significant-ly different.In most provinces(regions and cities),the HHI calculated by the number of beds and the number of health technicians ranged from 0.07 to 0.14,and the HHI showed an upward trend but a small change in the five years.Conclusion In recent years,the degree of competition in China's medical service market is generally high,but with the expansion of the supply of material resources and human resources in each province(region and city),the medical service market in most areas is developing towards a monopoly trend.

Glucosidases are an indispensable class of enzymes in the sugar metabolism of organisms. To investigate the biological functions and expression patterns of α-glucosidases (AGLUs) and β-glucosidases (BGLUs), we identified the two family members in the genome of melon (Cucumis melo). The number, location on chromosomes, gene structure, subcellular localization, conserved motifs, and phylogenetic relationship of the two family members were analyzed. Based on the cis-acting elements in the promoter region and protein interaction models, their functions were preliminarily predicted. Furthermore, the gene expression of the two family members was determined by qRT-PCR. The results showed that the melon genome contained five AGLU family members on five chromosomes, and all of the five members were located in the extracellular matrix, with the amino acid sequence lengths ranging from 899 aa to 1 060 aa. The melon genome carried 18 BGLU family members on 8 chromosomes, and all the members were located in the cell membrane or cytoplasm, with the amino acid lengths ranging from 151 aa to 576 aa. The qRT-PCR results showed that the expression of about 50% of the genes was down-regulated upon cold stress. CmAGLU5 and CmBGLU7 may be key members of the two families, respectively, in response to cold stress. The expression of all members of the two families was up-regulated under abscisic acid (ABA), high salt, and drought stress. In the AGLU family, CmAGLU3 was the key gene in response to ABA and high salt stress, while CmAGLU4 was the key gene in response to drought stress. In the BGLU family, CmBGLU18 was the key gene in response to ABA, while CmBGLU6 was the key gene in response to high salt and drought stress.
beta-Glucosidase/metabolism* ; Phylogeny ; alpha-Glucosidases/metabolism* ; Gene Expression Regulation, Plant ; Cucurbitaceae/enzymology* ; Multigene Family ; Cucumis melo/enzymology* ; Stress, Physiological

Objective:To construct a machine learning classification prediction model using planning-omics (planomics) features to predict the γ pass rate of intensity-modulated radiotherapy (IMRT) plan dose verification in fixed-field thoracic tumors, and evaluate the application of planomics in radiotherapy quality assurance.Methods:The fixed-field IMRT plans of 240 patients with chest tumors admitted to Department of Radiotherapy, Henan Cancer Hospital from August 2022 to March 2023 were retrospectively analyzed. All plans underwent dose verification using the electronic portal imaging system detector on the Varian accelerator to collect field dose data. The dose verification results were analyzed through Portal Dosimetry in the treatment planning system of Eclipse. The γ pass rate standard was set at 2%/2 mm with a 10% dose threshold. From the planning documents, 48 conventional planning features, 2476 planomics features, and the combination of the previous two feature sets were extracted. Subsequently, an auto-encoder classification model was constructed. To evaluate the classification efficacy of various feature sets, 20 random train-test divisions were conducted by calculating the area under the receiver operating characteristic curve (AUC) values along with the accuracy rates.Results:After the feature selection, 2 conventional features and 16 planomics features were finally selected. In the testing set, the AUC values for the model using combined features, planomics features, and conventional planned features were 0.802±0.030, 0.740±0.069, and 0.673±0.083, respectively. In contrast, in the training set, these AUC values were 0.844±0.074, 0.816±0.047, and 0.687±0.036, respectively. The accuracy rates were 0.752±0.083, 0.703±0.110, and 0.648±0.081 in the testing set, and 0.753±0.098, 0.751±0.075, and 0.624±0.054 in the training set for the combined, planomics, and conventional planning feature sets, respectively.Conclusions:For thoracic fixed-field adjusted radiotherapy planning, the machine learning method based on planomics features can be utilized to build a classification model for predicting the γ pass rate. Combining planomics features with conventional planned features can enhance the predictive performance of the classification models.

Objective:To evaluate the detection capability of the contrast-enhanced three-dimensional turbo spin-echo (CE-SPACE) sequence for brain metastases (BM), aiming to provide evidence for precise target delineation in stereotactic radiotherapy (SRT).Methods:A total of 123 BM patients who received radiotherapy at the Affiliated Cancer Hospital of Zhengzhou University from May to November 2024 were enrolled. All patients underwent contrast-enhanced (CE) MRI and CT scans in the same treatment position, with images rigidly registered in the Eclipse planning system. Two experienced radiation oncologists independently delineated BM lesions on CE-MPRAGE and CE-SPACE sequences in a blinded manner. Patients were divided into the delayed group (10 min, n=61) and a priority group (5 min, n=62) based on the time interval between gadolinium injection and CE-SPACE acquisition. The non-parametric Wilcoxon rank-sum test was used to compare the lesion counts and volume differences between the two imaging sequences. Point-biserial correlation analysis was performed to assess the correlation between the additional lesions identified by CE-SPACE and lesion volume. Results:The overall analysis demonstrated that CE-SPACE detected 421 BM lesions, achieving an 8.2% higher detection rate than CE-MPRAGE ( Z=3.78, P<0.001). In terms of lesion volume, the median BM lesions volume identified by CE-SPACE [0.30(0.07,1.53)cm 3] was 8.7% larger than that by CE-MPRAGE [0.23 (0.04, 1.34) cm 3] ( Z=12.88, P<0.001). CE-SPACE demonstrated superior sensitivity for lesions ≤ 0.06 cm3, with negative correlation between the number of additional lesions detected and lesion volume ( r=-0.104, P=0.034). Subgroup analysis revealed that in the delayed group, CE-SPACE detected significantly more lesions [median 2 (1, 3.5) vs. 2 (1, 3), P=0.002] and larger volumes [0.39 (0.08, 2.24) cm3 vs. 0.29 (0.05, 1.99) cm3, P<0.001] than CE-MPRAGE. In the priority group, CE-SPACE detected significantly larger lesion volumes [0.55 (0.09, 2.06) cm3 vs. 0.45 (0.08, 1.88) cm3, P<0.001], but no significant difference was observed in lesion counts between two sequences ( P=0.059). Conclusions:Three-dimensional CE-SPACE sequence offers superior detection sensitivity for small BM (≤ 0.06 cm3), providing crucial guidance for accurate target delineation in SRT.

Objective By analying the concentration degree of China's medical service market from 2017 to 2021,to analyze the current situation of competition in China's medical service market,and put forward relevant policy rec-ommendations to provide reference for optimizing resource allocation in China's medical service market.Methods The competition of medical service market in each province(region and city)was calculated by Herfindahl-Hirschman In-dex(HHI)and visualized by Stata18 software.Results From 2017 to 2021,the annual growth rate of the number of beds and health technicians in most regions of China was higher than 2.5％,but the absolute number was significant-ly different.In most provinces(regions and cities),the HHI calculated by the number of beds and the number of health technicians ranged from 0.07 to 0.14,and the HHI showed an upward trend but a small change in the five years.Conclusion In recent years,the degree of competition in China's medical service market is generally high,but with the expansion of the supply of material resources and human resources in each province(region and city),the medical service market in most areas is developing towards a monopoly trend.

Objective:To construct a machine learning classification prediction model using planning-omics (planomics) features to predict the γ pass rate of intensity-modulated radiotherapy (IMRT) plan dose verification in fixed-field thoracic tumors, and evaluate the application of planomics in radiotherapy quality assurance.Methods:The fixed-field IMRT plans of 240 patients with chest tumors admitted to Department of Radiotherapy, Henan Cancer Hospital from August 2022 to March 2023 were retrospectively analyzed. All plans underwent dose verification using the electronic portal imaging system detector on the Varian accelerator to collect field dose data. The dose verification results were analyzed through Portal Dosimetry in the treatment planning system of Eclipse. The γ pass rate standard was set at 2%/2 mm with a 10% dose threshold. From the planning documents, 48 conventional planning features, 2476 planomics features, and the combination of the previous two feature sets were extracted. Subsequently, an auto-encoder classification model was constructed. To evaluate the classification efficacy of various feature sets, 20 random train-test divisions were conducted by calculating the area under the receiver operating characteristic curve (AUC) values along with the accuracy rates.Results:After the feature selection, 2 conventional features and 16 planomics features were finally selected. In the testing set, the AUC values for the model using combined features, planomics features, and conventional planned features were 0.802±0.030, 0.740±0.069, and 0.673±0.083, respectively. In contrast, in the training set, these AUC values were 0.844±0.074, 0.816±0.047, and 0.687±0.036, respectively. The accuracy rates were 0.752±0.083, 0.703±0.110, and 0.648±0.081 in the testing set, and 0.753±0.098, 0.751±0.075, and 0.624±0.054 in the training set for the combined, planomics, and conventional planning feature sets, respectively.Conclusions:For thoracic fixed-field adjusted radiotherapy planning, the machine learning method based on planomics features can be utilized to build a classification model for predicting the γ pass rate. Combining planomics features with conventional planned features can enhance the predictive performance of the classification models.

Objective:To evaluate the detection capability of the contrast-enhanced three-dimensional turbo spin-echo (CE-SPACE) sequence for brain metastases (BM), aiming to provide evidence for precise target delineation in stereotactic radiotherapy (SRT).Methods:A total of 123 BM patients who received radiotherapy at the Affiliated Cancer Hospital of Zhengzhou University from May to November 2024 were enrolled. All patients underwent contrast-enhanced (CE) MRI and CT scans in the same treatment position, with images rigidly registered in the Eclipse planning system. Two experienced radiation oncologists independently delineated BM lesions on CE-MPRAGE and CE-SPACE sequences in a blinded manner. Patients were divided into the delayed group (10 min, n=61) and a priority group (5 min, n=62) based on the time interval between gadolinium injection and CE-SPACE acquisition. The non-parametric Wilcoxon rank-sum test was used to compare the lesion counts and volume differences between the two imaging sequences. Point-biserial correlation analysis was performed to assess the correlation between the additional lesions identified by CE-SPACE and lesion volume. Results:The overall analysis demonstrated that CE-SPACE detected 421 BM lesions, achieving an 8.2% higher detection rate than CE-MPRAGE ( Z=3.78, P<0.001). In terms of lesion volume, the median BM lesions volume identified by CE-SPACE [0.30(0.07,1.53)cm 3] was 8.7% larger than that by CE-MPRAGE [0.23 (0.04, 1.34) cm 3] ( Z=12.88, P<0.001). CE-SPACE demonstrated superior sensitivity for lesions ≤ 0.06 cm3, with negative correlation between the number of additional lesions detected and lesion volume ( r=-0.104, P=0.034). Subgroup analysis revealed that in the delayed group, CE-SPACE detected significantly more lesions [median 2 (1, 3.5) vs. 2 (1, 3), P=0.002] and larger volumes [0.39 (0.08, 2.24) cm3 vs. 0.29 (0.05, 1.99) cm3, P<0.001] than CE-MPRAGE. In the priority group, CE-SPACE detected significantly larger lesion volumes [0.55 (0.09, 2.06) cm3 vs. 0.45 (0.08, 1.88) cm3, P<0.001], but no significant difference was observed in lesion counts between two sequences ( P=0.059). Conclusions:Three-dimensional CE-SPACE sequence offers superior detection sensitivity for small BM (≤ 0.06 cm3), providing crucial guidance for accurate target delineation in SRT.

Objective To explore the mechanism of Marsdenia tenacissima in the treatment of breast cancer through network pharmacology and experimental verification.Methods Literature retrieval was conducted to obtain the active components of Marsdenia tenacissima.The SwissTargetPrediction database was used to predict the potential targets of these active components.Targets of breast cancer were obtained from GeneCards,GEPIA2,OMIM,PharmGKB and TTD databases.The intersection targets were obtained,and a Marsdenia tenacissima-breast cancer-targets network was constructed using Cytoscape 3.9.0 software.The core targets were identified through protein-protein interaction(PPI)network analysis,followed by GO and KEGG pathway enrichment analysis to screen relevant signaling pathways.Molecular docking validation was performed for the top 10 key targets and major active components.The human breast cancer cell line MDA-MB-231 was treated with Marsdenia tenacissima injection in vitro.Cell proliferation ability was detected by CCK-8 assay and colony formation assay.Cell apoptosis was detected by Calcein-AM/PI staining and flow cytometry.Cell migration ability was detected by Transwell assay.Western blot experiment was used to validate the PI3K-AKT signaling pathway.Results Totally 37 active components and 276 potential targets against breast cancer were screened from Marsdenia tenacissima,including 11alpha-O-Benzoyl-12beta-O-acetyl tenacigenin B,Caffeic acid,Drevogenin A and Kaempferol.25 core targets were screened by PPI network such as AKT1,EGFR,TNF,CTNNB1 and IL-6,which mainly affected the estrogen signaling pathway,ErbB signaling pathway,HIF-1 signaling pathway and PI3K-AKT signaling pathway,etc.The molecular docking results showed that the main active components of Marsdenia tenacissima exhibited good binding activities with the core targets AKT1,ALB,CASP3,ESR1 and TNF.The results of in vitro experiments showed that Marsdenia tenacissima injection could inhibit the proliferation and migration ability of MDA-MB-231 cells(P<0.01,P<0.001)and induce apoptosis(P<0.001),as well as inhibit the activation of PI3K-AKT signaling pathway(P<0.05,P<0.01).Conclusion Marsdenia tenacissima may exert its anti-breast cancer effects through multiple targets and pathways,and the mechanism may be related to the inhibition of PI3K-AKT signaling pathway.

Objective:To investigate the accuracy of magnetic resonance imaging (MRI) cine sequences in determining the range of tumor motion in radiotherapy, providing a basis for the precise delineation of the target volume in motion for radiation therapy.Methods:A modified chest motion phantom was placed in a MRI scanner, and a water-filled sphere was used to simulate a tumor. True fast imaging with steady precession (TrueFISP) MRI cine sequences from Siemens were used to capture the two-dimensional motion images of the simulated tumor. The phantom experiments were divided into three modes: head-foot motion mode, rotation motion mode, and actual respiratory waveform mode. In the head-foot motion mode, respiratory motion period (3, 4, 5, 6, 7 and 8 s), amplitude (5, 10 and 15 mm), and respiratory waveform of the simulated tumor (sin and cos4) were set, resulting in a total of 36 motion combinations. In the rotation motion mode, a cos4 waveform was used for respiration, with respiratory periods of 3, 4, 5, 6, 7 and 8 s, head-foot motion set amplitudes of 5, 10 and 15 mm, and anterior-posterior (AP) and left-right (LR) motion set amplitudes in three combinations ([2.5, 2.5] mm, [2.5, 5.0] mm, [5.0, 5.0] mm), resulting in a total of 54 motion combinations. In the actual respiratory waveform mode, respiratory waveforms of 5 randomly selected patients from Affiliated Cancer Hospital of Zhengzhou University were obtained. Under each motion combination, TrueFISP cine images (30 frames, with an acquisition time of 11 s per frame) were obtained. The code was used to automatically identify the two-dimensional coordinates of the center of the simulated tumor in each image, and sin and cos4 functions were separately employed to fit the tumor position in the motion direction, thereby obtaining the fitted motion period and amplitude. The difference between the maximum and minimum values of the tumor's center coordinates in the head-to-foot direction is taken as the range of movement, referred to as the calculated amplitude. For the actual respiratory waveform, the distance between the measured maximum and minimum positions is used to calculate the amplitude.Results:In the head-foot motion mode, the fitted amplitudes of both sin and cos4 waveforms deviated from the set amplitudes by 0-0.51 mm, with relative deviations of 0%-4.2%. The deviation range between the calculated amplitudes and the set amplitudes of the two waveforms were 0.08-0.94 mm, with relative deviations of 1.1%-6.3%. In the rotation motion mode, the fitted amplitudes deviated from the set amplitudes by 0-0.61 mm, with relative deviations of 0%-6.2%. And the deviation range between the calculated amplitudes and the set amplitudes were 0.16-0.94 mm, with relative deviations of 0%-6.3%. In the actual respiratory waveform motion mode, the deviation range between the calculated amplitudes and the set amplitudes were 0.10-0.48 mm, with relative deviations of 2.2%-8.6%.Conclusion:TrueFISP cine sequences show minimal deviations in determining the range of tumor head-foot motion and effectively captures the tumor's movement state, thereby providing important support for the precise definition of the tumor movement target area during radiotherapy .