Objective To evaluate the accuracy of a deep learning-based intelligent contouring system in delineating head and neck organs at risk (OAR) and to explore its potential application in clinical radiotherapy. Methods Forty-five cases of head and neck tumor patients' positioning CT images were randomly selected. The SmartContour.AI software was used to contour 12 important OARs in the head and neck region (brainstem, spinal cord, left and right optic nerves, optic chiasm, left and right eyeballs, left and right lenses, left and right parotid glands, thyroid gland), and compared with the manual contouring by senior physicians. The consistency between automatic and manual contouring was quantitatively analyzed using seven geometric evaluation metrics, including Hausdorff distance (HD), Dice similarity coefficient (DSC), Jaccard coefficient, centroid deviation (CMD), sensitivity coefficient (SI), inclusiveness coefficient (II), and volume difference (VD). Additionally, the clinical acceptance of the intelligent contouring system was assessed by subjective evaluation from senior physicians. Results In terms of geometric accuracy, the maximum mean Hausdorff distance was 10.73 mm for the right optic nerve, and the minimum was 1.03 mm for the left lens. The mean DSC index for the OARs ranged from a minimum of 0.68 for the right optic nerve to a maximum of 0.95 for the right eyeball. The mean Jaccard coefficient for the OARs was highest for the right eye at 0.91 and lowest for the right optic nerve at 0.57. The maximum mean centroid deviation (CMD) was 4.08 mm for the right optic nerve and the minimum was 0.48 mm for the left lens. The sensitivity coefficient (SI) was relatively low for the left and right optic nerves and optic chiasm, at 0.63, 0.59, and 0.67, respectively, The SI value of the right eyeball is the largest, approaching 1.00. The mean inclusiveness coefficient (II) for all organs was ≥0.84. In terms of volume difference (VD), the brainstem, spinal cord, left and right eyeballs, left and right lenses, and thyroid gland were close to 0, while the VD values for the parotid gland, optic chiasm, and optic nerve exceed 0.10. In terms of clinical acceptance, the parotid glands, optic nerves, and optic chiasm had lower clinical acceptance, while the acceptance for the other organs was relatively high. Conclusion Automatic contouring of the brainstem, spinal cord, eyeballs, lenses, and thyroid gland showed high geometric consistency with manual contouring and was well accepted clinically, while automatic contouring of the parotid glands, optic nerves, and optic chiasm required further physician review and substantial modification.

Objective Under the background of the normalization of orthopedic medical consumables collection, an early warning mechanism for the use of medical consumables is established to help medical institutions better adapt to the collection policy, provide data basis for managers, facilitate the timely adoption of management measures, guide the rational use of medical consumables, and make the management of consumables more scientific, intelligent and refined. Methods Systematic method and induction method were used to analyze the difficulties and risk points in the management of orthopedic medical consumables after the collection and mining of orthopedic medical consumables became normal. Based on information technology, the early warning mechanism for the use of medical consumables was established. The use monitoring and analysis were only conducted for the indicators such as the use quantity, cost and benefit of medical consumables, and multi-dimensional comprehensive monitoring was adopted to fully reflect the use of various consumables. Use big data analysis to give early warning to abnormal data and unreasonable use. Results By establishing an early warning mechanism, the hospital has improved its management mechanism for the use of medical consumables, reducing the proportion of key monitoring consumables. The proportion of 18 key monitoring consumables has decreased by 10.46% compared to 2023. The joint procurement task for the new year has been 100% completed, improving the hospital's management efficiency and effectiveness, and achieving an overall improvement in the management level of orthopedic medical consumables. Conclusion The establishment of the early warning mechanism for medical consumables puts forward clear management ideas and scientific management methods for monitoring the use of medical consumables after collection of orthopedic consumables, which can be properly promoted to help managers of large public hospitals make more scientific decisions, conduct preventive monitoring of medical consumables in hospitals, prevent the accumulation of minor problems, and constantly strengthen the connotation construction of hospital consumables monitoring system. Detect suspected problems through monitoring, control the source and continuous improvement.

Objective This study aimed to construct a maintenance quality evaluation index system for continuous renal replacement therapy devices based on the "structure-process-outcome" three-dimensional quality evaluation model to guide clinical practice. Methods Through the evidence-based method and the theoretical framework of the "structure-process-result" model, the first draft of the correspondence was constructed, and two rounds of expert correspondence were conducted by the Delphi method to determine the items at each level of the evaluation index, and finally the evaluation index system of maintenance quality of continuous renal replacement therapy device was formed. Results A total of 24 experts were selected for consultation. The active coefficients of experts in the two rounds of consultation were 82.35% and 100.00%, and the authority coefficients were 0.871 and 0.900. The coordination degree of expert opinions was good. The Kendall's coefficients of concordance in the two rounds of consultation were 0.160 and 0.166, P < 0.001. A maintenance quality evaluation index system for continuous renal replacement therapy devices was formed, including 3 first-level indicators, 12 second-level indicators and 52 third-level indicators. Conclusion The quality evaluation index system for the maintenance of continuous renal replacement therapy devices constructed based on the three-dimensional quality evaluation model has good consistency and scientific nature, and can provide a basis for the management of the maintenance quality of continuous renal replacement therapy devices.

Objective The next-generation integrated design of Positron Emission Tomography/Magnetic Resonance (PET/MR) combines PET detectors into the MR device, enabling simultaneous data collection from both detectors in the same space. However, the stability of the PET component in this new device structure during daily clinical operations requires further investigation. This study tracks nearly 16 months of source quality control data from the digital integrated uPMR 790 PET/MR system, five years after installation, to explore the stability of the digital integrated PET/MR system in clinical use and the influencing factors between PET source quality control parameters. Methods Under clinical application conditions, source quality control of the uPMR 790 system was conducted based on the daily quality assurance (DQA) guidelines issued by the equipment manufacturer. Data from 72 weeks of source quality control system tests from January 2023 to May 2024 were collected, including system temperature and voltage, Look-Up Table (LUT), collection system (CMap), Time of Flight (TOF), and system energy state drift. Spearman correlation analysis and Mann-Whitney U tests were used to analyze the system quality control parameters. Multiple linear regression analysis was conducted to explore the influencing factors of TOF drift. Results The overall pass rate for 72 PET/MR source quality control tests was 100%. TOF state drift showed a positive correlation with LUT state drift (r = 0.614, P < 0.001) and a negative correlation with the system's maximum temperature (r = -0.736, P < 0.001). There were distributional differences between TOF state drift and changes in the system's acquisition CMap state (Z = 4.872, P < 0.001). Multiple linear regression analysis revealed that LUT drift (β = 0.705, P < 0.001) and system acquisition CMap (β = 0.131, P = 0.031) had positive effects on the degree of TOF drift, while the system's maximum temperature (β = -0.263, P < 0.001) had a negative effect on the degree of TOF drift. Conclusion After five years of clinical use, Over the past 16 months, the testing parameters for PET source quality control have remained within normal ranges, and the equipment has demonstrated stable performance in daily clinical applications. However, attention should be paid to potential drifts in parameters related to the PET detectors during source quality control. Regular monitoring, combined with risk alerts, is essential to ensure the long-term stability of the equipment and the reliability of diagnostic outcomes.

Objective To address the issue of insufficient tissue anchoring performance in existing spiral barbed sutures and design a process equipment to enhance the tissue anchoring performance of spiral barbed sutures. Methods Using the self-designed process equipment, through the synergistic effect of process parameters such as heating temperature, rotation angle, and linear displacement, the spiral barbed sutures were processed. The morphological characteristics of the barbs were observed by an optical microscope. The pull-out force test and breaking strength test were carried out by a material testing machine with specific tensile speeds and test gauges, and the ex vivo tissue suturing experiment was performed on fresh porcine skin tissue to systematically evaluate the influence of different process parameters on the barb performance. Results Under the combination of process parameters of 60°C heating temperature, 40° rotation angle, and 1mm linear displacement (Group B), the barbs exhibited the best morphological characteristics. The pull-out force (2.398N) was approximately 5.7 times higher than that of the untreated group (Group A, 0.423N), and the ex vivo tissue suturing strength (3.338N) was approximately 3 times higher. Moreover, compared with other groups, there was no significant difference in the breaking strength of the suture body in this group (Group A: 37.586N, Group B: 37.119N, Group C: 37.434N, Group D: 36.975N, while the breaking strength of Group E: 29.130N was significantly lower than that of Group B). Conclusion This study provides important technical support for improving the safety and efficacy of spiral barbed sutures in clinical applications and theoretical basis for the process optimization of spiral barbed sutures.

Objective To construct a core evaluation index system for remote quality control of magnetic resonance imaging (MRI), and to use it for an empirical study of MRI remote quality control work. Methods Through literature analysis, unstructured interviews, and Delphi method, the key elements of MRI remote quality control evaluation were systematically analyzed, and a core evaluation index system was constructed. The data of various indicators in remote quality control of hospital MRI was collected during 2022 to 2023, and evaluated by using the entropy weight TOPSIS method, and then the evaluation results were compared with the clinical photo evaluation results in MRI quality control. Results The established MRI remote quality control core indicator system contained 3 primary indicators and 15 secondary indicators. The results evaluated by the entropy weight TOPSIS method were basically consistent with the clinical photo evaluation method (Pearson coefficient 0.78). Conclusion This evaluation system is comprehensive, objective, and has certain practicality. It can provide solutions for evaluating the remote quality control of MRI, help improve the quality of MRI application, and reduce medical risks.

Objective To effectively regulate the management of medical measurement devices in hospitals, ensure the accuracy of their measurements, reduce the clinical use risks of medical devices, and improve medical quality. According to the relevant regulatory requirements of "The Detailed Rules for the Implementation of the Metrology Law of the People's Republic of China," "Regulations on the Supervision and Administration of Medical Devices," and "Measures for the Clinical Use Management of Medical Devices,"The oversight-in-inspection rate (here referring to the rate of failure to undergo inspection) of measuring instruments in use in hospitals should be kept at 0%. Methods This study innovatively establishes a model that deeply integrates the Plan-Do-Check-Act (PDCA) cycle with digital management, breaking through the limitations of traditional, singular management approaches. By utilizing the Man, Machine, Material, Method, and Environment (5M) analysis and Pareto analysis, it summarizes the primary causes of low detection coverage. Based on these primary causes, corresponding countermeasures are proposed using the "5W1H" method. Results After implementation, the inspection coverage rate increased from 88.68% to 100%, and the missed inspection rate of measuring instruments in use at the hospital was successfully reduced from 11.32% to 0%. Conclusion It has been verified that the implementation of the PDCA cycle can significantly control the missed inspection rate of measuring instruments, while also contributing to strengthening the risk management of medical devices and ensuring their safety in clinical use.

Objective To provide regulatory basis for medical device supervision and management departments, adopt effective mechanisms to carry out risk control and prevention of adverse events, reduce the risk of patients and medical staff using medical devices, and ensure the safety of people's use of equipment and life and property. Methods Based on the Shandong Province Medical Device Adverse Event Monitoring Special Research Platform, conduct adverse event data analysis and determine the research category of disposable medical consumables;Use the Analytic Hierarchy Process (AHP) to analyze and determine the weights of the indicators causing adverse events of such devices, and conduct consistency testing analysis on them;Develop targeted measures from different perspectives of device regulation, production, and use, implement them, and continuously track and monitor them. Results The research category has been identified as urinary catheterization devices, and the AHP analysis results have determined that product quality issues are the most important cause of adverse events in urinary catheterization devices. The consistency test results of the two dimensions (CR value=0) have passed, which is consistent with the statistical results of the provincial platform data; Usage management and personnel management (medical staff, patients) are also important factors leading to adverse events of urinary catheterization medical devices;After the implementation of targeted measures, the incidence of adverse events related to urinary catheterization devices in medical institutions significantly decreased. Conclusion The use of Analytic Hierarchy Process for data analysis can effectively determine the root cause indicators of medical device adverse events, take targeted measures to carry out risk control of medical device adverse events, further improve the performance and functionality requirements of medical institutions for the products they use, provide feedback to promote the improvement of enterprise product quality, and effectively reduce or avoid the recurrence of similar adverse events.

Objective To construct a performance evaluation index system for high-end medical equipment in public hospitals, and apply the analytic hierarchy process (AHP) to study it, with the aim of forming a standard for evaluating the use of medical equipment, which can be used for the decision-making and allocation of high-end medical equipment. Methods The performance evaluation index system for high-end medical equipment was constructed from four aspects of clinical effects, discipline construction, scientific research achievements, and economic benefits, using the analytic hierarchy process. Results Two sets of performance evaluation index models for high-end medical equipment were constructed, including a model for diagnostic equipment with 4 first-level indicators, 12 second-level indicators, and 28 third-level indicators; and a model for therapeutic equipment with 4 first-level indicators, 12 second-level indicators, and 29 third-level indicators. Conclusion This study classifies high-end medical equipment into two categories of diagnostic equipment and therapeutic equipment, and forms two sets of performance evaluation index systems, which are highly applicable. The selection of indicators and weighted scoring are scientific and reasonable, and can provide scientific reference for improving the comprehensive quality of medical, teaching, and research in public hospitals and enhancing economic benefits.

To improve the efficiency and accuracy of extracting unstructured information from clinical case texts and promote the development of medical intelligence, a case text structured model based on BERT fusion algorithm is proposed. This model comprehensively uses bidirectional encoders for semantic representation, extracts local dependencies between words using graph convolutional neural networks, integrates long short-term memory networks to establish temporal relationships, and introduces conditional random fields to optimize the consistency of label sequences. The experiment selected two authoritative clinical datasets, MIMIC-III and ClinicalSTS, for analysis, and constructed five types of medical text classification tasks to compare the performance of all models under different structural combinations. The results show that the accuracy, recall, and F1 score of the proposed model are 0.92, 0.90, and 0.91, respectively, all of which are about 10% higher than traditional bidirectional encoder models. When processing long text scenes exceeding 1000 words, the model's efficiency is improved by 12%, demonstrating good timeliness and scalability. The study verified the effectiveness of deep fusion of multiple structures in improving the ability to process case text structures, and provided a theoretical basis for intelligent medical text analysis.

Objective To study the dose distribution around the PET/CT automatic injection system in the working process, and to explore the dose distribution characteristics of each face of the PET/CT automatic injection equipment, so as to provide clinical data reference for reducing the radiation dose during the work of injection nurses. Methods LiF(Mg,Cu,P) thermoluminescent dosimeter was used to measure the dose equivalent at 5 cm around the Intego PET automated delivery system equipped with 18F-FDG when the equipment was working, and the dose contour map was drawn. Results The initial source activity was 9546 MBq. Thirteen patients underwent 18F-FDG PET/CT examination with a total administered dose of 3136.12 MBq, and the entire process took 133 minutes. Dose contour maps were drawn for the A, B, C, D, E, and F surfaces at 5 cm around the PET/CT automatic injection system. There was a significant high-dose area on the A, C, D, and E surfaces, while two high-dose areas were present on the B surface. The maximum dose equivalents for the A, B, C, D, E, and F surfaces were 28.6±1.4, 78±15, 85.8±3.5, 16.2±0.9, 41±2.6, and 2.7±0.4 μSv, respectively. The maximum dose equivalent in the second highest dose area of the B surface was 12±0.8 μSv. Conclusion The maximum dose equivalent on the C side is the highest. The C side is mainly exposed to the radioactive drugs in the infusion pipelines around the RP pump and in the dose calibrator. The dose distribution around the automatic injection equipment is uneven. The formation of the high-dose area within 5 cm around the equipment is mainly due to the radioactive drugs in the source tank and the infusion pipelines inside and outside the system. The estimated maximum annual exposure dose for the injection nurse at a distance of 30 cm from the equipment surface is 0.8 mSv, which does not exceed the occupational exposure dose constraint value (5 mSv/a). Certain protective strategies can be adopted to reduce the radiation dose received by the injection nurse.

Objective To sort out and analyze the supervision status of supervision and inspection of medical device manufacturers in Shaanxi Province, aiming to further improve the quality and efficiency of inspection, and provide reference for the improvement and promotion of the quality management system of production enterprises. Methods By systematically combing the inspection status, inspection results and distribution of defect items in the supervision and inspection of medical device manufacturers in Shaanxi Province in 2023, the outstanding problems existing in the quality management system of different types of enterprises were analyzed. Results The defects found in the quality management system of the enterprise were mainly concentrated in production management, equipment, quality control, plant and facilities. Among them, the average defect items of implant enterprises in the production management, plant and facilities chapters are higher than those of other types of production enterprises, in vitro diagnostic reagent enterprises have more defects in equipment and quality control, and the defects of independent software enterprises have institutions and personnel and quality control chapters. Conclusion In order to guide manufacturers to improve the quality management system and strengthen the management of medical device production, some regulatory suggestions are put forward, such as implementing the main responsibility of enterprises, implementing precise inspection, and improving the ability of inspectors.

Objective To improve the operation efficiency of the backup center. Methods Utilizing the Plan-Do-Check-Act (PDCA) cycle method, BP neural networks, and redundancy technology were employed to optimize the quantity of monitors in both the backup equipment center and clinical departments. Rental rules were revised to improve the turnover rate of equipment in the backup equipment center. Results The number of monitors in the backup center has been optimized, the operating cost of 22 monitors has been saved, the turnover rate of respiratory humidification therapy instruments and ear thermometers (P<0.05) has been improved, and the operation efficiency of the backup center has been improved. Conclusion The application of the PDCA cycle method in the operation and management of the backup equipment center contributes to optimizing the quantity allocation of equipment, increasing the equipment turnover rate, and improving the operational efficiency of the center.

Objective To innovate and optimize the patient service experience in response to the challenges of secure sharing brought about by the increasing amount of electronic medical record data in the process of medical informatization. Methods This paper designs an innovative medical data security sharing scheme based on blockchain and proxy re-encryption technology. Firstly, a blockchain network composed of hospital nodes is constructed, and smart contracts are used to achieve permission management and log synchronization. Then, a certificateless mechanism is adopted to simplify key management, and public and private keys are dynamically allocated through the key generation center. Finally, combined with proxy re-encryption technology, The blockchain node acts as a proxy to convert the original ciphertext into ciphertext that the target user can decrypt, ensuring that the plaintext is not exposed during the data sharing process. Results The experimental results show that the time required for the scheme proposed in this paper to perform one encryption is 1.80ms. The efficiency in the total key generation stage and the decryption stage is higher than that of other schemes. The key generation stage and the encryption stage are statistically significant (p<0.05). Although the p value is greater than 0.05 when compared with the one-person scheme in the decryption stage, However, in this scheme, the decryption time remains almost unchanged as the number of attributes increases. In the attack experiment, the number of blockchains reached more than 20, the probability of the outcome was 0, it could resist the attack and had good security, indicating the effectiveness of this scheme. Conclusion The scheme proposed in this paper integrates blockchain and proxy re-encryption technology, ensuring the secure sharing of electronic medical records among different users and institutions, effectively enhancing the security and sharing efficiency of electronic medical records, and improving the convenience and satisfaction of the public in seeking medical treatment.

Objective Exploring the construction and application effectiveness of a refined management mode for medical consumables from the perspective of intelligent supply chains. Methods The medical consumables refined management system architecture under the intelligent supply chain is composed of base layer, access layer, data center and application layer. The base layer realizes the data collection of the whole life cycle of medical consumables circulation in hospital. The access layer realizes the access of various consumption-related business information systems. The data center realizes deep mining and scientific utilization of massive medical consumables business data; The application layer is the business application of all kinds of data, including the medical consumables supplier evaluation, warehouse management, consumption monitoring, usage assessment, and rationality analysis. The use data of medical consumables before and after implementing the refined management mode based on the perspective of smart supply chain in our hospital were selected and analyzed to verify the implementation effect of the intelligent mode.Results After the application of the refined management mode, the medical consumables management liquidation process time decreased from (70.3±3.5) hours to (18.1±2.1) hours，inventory turnover time decreased from (38.5±3.7) days to (26.2±2.1) days; request for replenishment time decreased from (122.3±3.4) hours to (23.9±7.2) hours; closed-loop traceability management time decreased from (30.6±4.1) minutes to (6.2±2.2) minutes; the conformity rate of accounts increased from (95.2±1.1%) to (99.1±0.3%); The satisfaction rate increased from (83.2±3.1%) to (98.5±0.9%) and the satisfaction of medical staff to the management mode increased from (80.6±2.87%) to (99.1±0.3%).Conclusion Medical consumables management mode from the perspective of smart supply chain can save medical costs, improve work efficiency, improve medical quality, and improve user satisfaction. It can provide reference for medical institutions.

Objective To explore the effect of deep learning reconstruction algorithm (Deep Learning Reconstruction, DLR) combined with metal artifact removal (Metal Artifact Removal, MAR) in CT imaging after lumbar fixation, and to evaluate its impact on image quality. Methods 34 patients were collected, Each patient underwent a weighted 50% adaptive iterative reconstruction (AV50), 50% adaptive iterative reconstruction (Adaptive Statistical Iterative Reconstruction-V, ASIR-V) combined with MAR (AV 50-MAR), moderate-intensity deep learning reconstruction (DM), moderate-intensity deep learning reconstruction combined with MAR (DM-MAR), high-intensity deep learning reconstruction (DH), high-intensity deep learning reconstruction combined with MAR (DH-MAR). Analyze and compare the noise values under 6 different reconstruction methods (Standard Deviation, SD), the signal-to-noise ratio (Signal to Noise Ratio, SNR), contrast to noise ratio (Contrast to Noise Ratio, CNR), the artifact index (Artifacts Index, AI), the uniformity index (Homogeneity Index, Differences in the results for HI), And scored image structural clarity, degree of metal artifacts and overall image quality by two experienced diagnostic physicians. Results The objective evaluation and subjective score of the reconstruction method with MAR technique are better than the individual reconstruction method in the psoas muscle and homogeneous area, and in the reconstruction method with MAR technique, DH-MAR showed the lowest noise and the best texture details. However, in the bone cancellous area, although the noise value of the MAR reconstruction method is reduced, the signal intensity (SNR, CNR) and homogeneity (HI) are lower than the individual reconstruction method. Conclusion MAR technology combined with MAR technology can provide smaller noise artifacts and better image quality.

The research aims to improve the robustness and accuracy of the lung ultrasound image detection system, in order to assist doctors in early diagnosis of lung diseases. On the basis of TransUNet, a hybrid model of self attention mechanism and convolution operation is introduced to improve the accuracy of lesion detection. At the same time, combining EfficientNet version 2 transfer learning technology to achieve accelerated convergence of the model. The results showed that the F1 score of the optimized model was 0.943, while the basic model TransUNet was 0.891, indicating that the classification performance of the optimized model was good. The optimized model has an intersection to union ratio of 0.924 and TransUNet of 0.862, indicating good segmentation performance. This model significantly improves the segmentation and classification performance of lung ultrasound images, with a 7.2% increase in lesion localization and segmentation accuracy. This can improve the efficiency of early detection and diagnosis of lung diseases, and provide intelligent technical support for ultrasound image detection.

Objective To solve the problem of low efficiency in extracting key clinical information from electronic medical record text data, support precision medical decision-making and disease research, and develop an efficient joint information extraction algorithm to achieve automated extraction of entities and relationships in medical record text. Methods Based on the differences in unstructured features of medical record text, a joint extraction model integrating multi task learning is proposed. Firstly, a Bi directional Long Short Term Memory Conditional Random Field (BiLSTM-CRF) benchmark model is constructed, and entity recognition is completed by combining the bidirectional encoder with the Conditional Random Field; Secondly, introducing a multi head attention mechanism to capture remote dependencies between entities; Finally, a multi task learning framework was used to address the issue of entity relationship overlap, resulting in the final joint model (Bidirectional Encoder Representation from Transformers Bi directional Long Short Term Memory Conditional Random Field, BERT-BiLSTM-CRF). Results Training and validation were conducted on a Chinese electronic medical record dataset to evaluate the extraction performance in the field of cerebrovascular diseases. The results show that the BERT-BiLSTM-CRF model has an accuracy of over 80% in text information entity recognition on the dataset, and the extraction error of entity relationships does not exceed 0.2, which is superior to other algorithm models. And the BERT-BiLSTM-CRF algorithm has a maximum recognition accuracy of 91.08% in extracting entity relationships in cerebrovascular diseases, which can effectively recognize relationships in medical text data. Conclusion The BERT-BiLSTM-CRF model can effectively overcome the technical bottleneck of entity relationship overlap, provide a new method for deep mining of electronic medical records, and provide research ideas for clinical medical decision-making and disease diagnosis.

With the rapid advancement of information technology, digital medical devices—leveraging their strengths in real-time monitoring, diagnostic support, and data-driven decision making—are profoundly reshaping traditional healthcare models. However, existing literature tends to focus on individual devices or isolated application scenarios, lacking a systematic integration and in-depth analysis of their technical implementation pathways and clinical workflows. This paper first defines the core concepts and technical framework of digital medical devices, systematically reviewing key implementation approaches—sensors, algorithms, and wireless communication. It then examines the typical application workflows and practical case studies of wearable and implantable devices in remote monitoring and smart healthcare. Building on this, the paper provides a detailed analysis of the primary challenges encountered during deployment, including standardization and interoperability, data privacy and security, ease of operation, and clinician‐and‐patient acceptance. Finally, by integrating cutting‐edge technologies such as 5G, edge computing, artificial intelligence, and the Internet of Things, it explores future development trends for digital medical devices. The study aims to establish a comprehensive research framework that uncovers the coupling mechanisms between technology and application scenarios, offering theoretical and practical guidance for device development, standards formulation, and clinical implementation.

Objective Enhance the management level of hospital cold chain to ensure the quality and safety of medical supplies, Methods reduce manual intervention through intelligent monitoring, and enable intelligent prediction for early warning and timely response. The system integrates IoT, big data, and cloud computing technologies, utilizing temperature and humidity sensors to establish a browser/server architecture-based cold chain monitoring platform. Three predictive modeling approaches were sequentially implemented: a standard BP neural network, a genetic algorithm-optimized BP neural network, and an improved genetic algorithm-enhanced BP neural network. These models were trained using complete 2022-2023 temperature monitoring datasets from a 2-8°C medical refrigerator, subsequently generating three distinct sets of temperature predictions for January 2024. Results Establish a cold chain monitoring system to achieve comprehensive temperature and humidity monitoring of equipment, and develop a predictive model for equipment failures. Compared to pre-implementation, the system significantly reduced equipment alarm response time, repair time, and spare parts application time, the variance of the error values in the equipment failure prediction model was significantly reduced compared to the pre-improvement version, while improving clinical satisfaction. These differences were statistically significant (P < 0.05). Conclusion The IoT-based cold chain monitoring system enables comprehensive, continuous, and intelligent surveillance of the cold chain environment, significantly enhancing the storage quality and safety of pharmaceuticals and biological agents. By employing an improved genetic algorithm to optimize the BP neural network, the system establishes an equipment failure prediction model capable of forecasting malfunctions and issuing early warnings, thereby preventing large-scale drug spoilage caused by temperature deviations. This approach provides a developmental roadmap for sustainable cold chain monitoring systems, with profound implications for hospital operations, patient safety, and public health.

Objective The drug coated balloon can inhibit the abnormal proliferation of vascular intima by filling and releasing the drug in the coating in the position of intravascular lesions, so as to improve the blood flow. In this process, the drug coating as insoluble particles is released in a large amount in a short time, which has the risk of causing acute vascular embolism. In this paper, a standard vascular model was used to evaluate the shedding performance of insoluble particles from drug coated balloon in vitro. Methods A set of vascular model system was established by using standard vascular model, guide catheter and guide wire to simulate the release process of drug coated balloon in vitro. Particle analyzer, ultra depth of field microscope and particle size analyzer were used to evaluate the particle shedding performance of drug coated balloon in three stages of delivery, release and withdrawal, as well as in simulated blood flow at different flow rates. Results Comparing the test results of the two different evaluation methods, the number of particles shed by the drug coated balloon in the simulated blood flow at different flow rates was significantly higher than that in the staged test. Compared with the delivery and withdrawal process, the particle shedding of drug coated balloon mainly occurred in the release process of the vascular model, and the number of particles shed accounted for the highest proportion in the whole process. When using the method of simulating blood flow to evaluate the particle shedding performance, it is recommended to use 100 ml/min circulating water flow to meet the requirements of particle recovery. The morphological characteristics of the shed particles were analyzed. Most of the particles were in sheet shape, and a few were in rod or fiber shape. In addition, the roundness and sphericity of the exfoliated particles are low, and the length diameter ratio is high, so the evaluation of the particle size of the exfoliated particles by the photoresist method is more one-sided. Conclusion A vascular model system was established, and the particle shedding performance of drug-coated balloons was evaluated by collecting particles at different release links or using a simulated blood flow method.

Objective To explore the key points of biological evaluation for medical aesthetic devices, ensuring their safety and efficacy in clinical applications, providing scientific evidence for industry regulation and technical review, and promoting standardized industry development. Methods Systematically analyze market-approved medical aesthetic devices, integrate domestic and international regulations and technical standards, and further investigate common issues and solutions in biocompatibility testing. Results The biological evaluation of medical aesthetic devices requires focused attention on factors such as the selection of evaluation pathways and preparation of test solutions. Specific biocompatibility testing must fully consider the devices' physicochemical properties and clinical application scenarios. Conclusion Biological evaluation is a critical safeguard for the safe clinical application of medical aesthetic devices. It must be conducted scientifically and rationally based on device characteristics to ensure compliance with safety requirements.

Objective Structured knowledge is the limitation of conventional maintenance models. Based on the vertical knowledge bases and large language models (LLMs), a medical equipment maintenance expert model solution is explored to address the problem. Methods A case study of Tyco PB840 ventilator was used in this study. The vertical knowledge base was constructed by integrating CNKI literature, in-house maintenance records, and equipment manuals. The DeepSeek R1 framework was employed to build the knowledge base expert model. And the model was compared with DeepSeek R1(on-prem) and DeepSeek R1(internet-connected) models. The models were evaluated through assessments of fault diagnosis and component-disassemble individually. Results The knowledge base expert model demonstrated superior performance in fault troubleshooting (8.44 points) with device-specific advantages. And all three models showed high reliability in root cause analysis (7.97–8.01 points). The vertical knowledge base significantly improved the professional specificity, especially in component-disassemble. Literature primarily supported fault diagnosis, while equipment manuals contributed substantially to disassembly. In-house maintenance records were underutilized due to their brevity. Conclusion LLMs exhibit generalization capabilities in medical equipment maintenance. By integrating of multi-source complementary data, the vertical knowledge bases enhances model specificity and professionalism. This study provides a knowledge-enhanced technical pathway and empirical reference for developing intelligent medical equipment maintenance systems.

Objective To explore the current applications, research hotspots, and future trends of functional magnetic resonance imaging (fMRI) in migraine research using bibliometric methods. Methods A comprehensive search was conducted for Chinese and English literature on fMRI studies in migraine published in the China National Knowledge Infrastructure (CNKI) and Web of Science (WoS) Core Collection databases up to February 28, 2025. VOSviewer 1.6.18 and CiteSpace 6.1.R6 software were employed to visualize and analyze authors, institutions, and keywords. Results A total of 163 Chinese articles and 463 English articles were included. Annual publications demonstrated a steady upward trend. Core research teams led by authors such as Yi Ren and Fanrong Liang were identified internationally. Keyword frequency and cluster analysis revealed research hotspots related to "acupuncture," "functional connectivity," "brain networks," and "resting-state." Timeline and burst analysis highlighted emerging frontiers such as "functional connectivity" and "machine learning." Conclusion fMRI is widely utilized in migraine research. Future efforts should focus on strengthening interdisciplinary collaboration, addressing research hotspots and frontiers, and promoting sustainable development in this field.

SPECT/CT is an important method for detecting myocardial blood flow perfusion of the examiners under resting and stressing states,and can accurately evaluate the myocardial ischemia in patients with coronary heart disease.In recent years,SPECT/CT technology has achieved rapid development and been increasing widely used in the diagnosis of myocardial ischemia in coronary heart disease,gaining universal recognition.However,insufficient understanding by some personnel has limited its development.Based on thorough study and research,this paper reviews the applications of SPECT/CT imaging agents,advancements and optimization in imaging methods,etc,in the diagnosis of myocardial ischemia in coronary heart disease,predicts future relevant development directions,and improves medical staff`s understanding in this field,with a view to expanding the application of SPECT/CT in myocardial ischemia of coronary heart disease.

With the continuous advancement and deepening of the construction of hospital information management, the application of information technology to achieve the management of disinfection supply center (CSSD) has become the inevitable trend of the development of hospital CSSD. The CSSD information quality traceability system can record and track the whole process of reusable instruments, appliances and articles in CSSD from recycling, cleaning, disinfection, packaging, sterilization, storage, distribution to clinical departments. Through the application of information traceability system, the entire process of disinfection and sterilization can be dynamically monitored, accurately tracked, responsible to people, and ensure the safety of aseptic supply, which is the key to the quality control and management of hospital CSSD. Through extensive literature review, this article elaborates the development status of information-based quality traceability system in CSSD, specifically analyzes the advantages and disadvantages of information-based quality traceability system in the application process, and provides new ideas for the development of a more professional and intelligent information-based quality traceability management system, further improves the quality of hospital equipment management, and promotes the construction of intelligent CSSD in hospitals.

Objective To explore the main injuries, types of malfunctions, causes, and response measures of adverse events in hemodialysis medical devices. Methods Through statistical analysis of adverse event reports of blood dialysis medical devices in Zhejiang Province from January 2019 to January 2024 in the National Medical Device Adverse Event Monitoring Information System, the risk types were sorted and studied, the risk factors were analyzed, and corresponding measures were proposed. Results 636 adverse events occurred with blood dialyzer products, with an average age of 58.68 years, mainly using imported dialyzers. The main harm to patients is blood leakage, followed by allergic reactions such as chest tightness and difficulty breathing. The most common equipment malfunction is membrane rupture in dialyzers. The tertiary hospitals have the highest number of reported cases of adverse events, especially serious injuries. Conclusion Blood dialyzer products have known and unexpected risks that may cause harm to the human body. Risk monitoring and management should be strengthened in the production, storage, use, education, and supervision of such products to minimize the risk of adverse events caused by medical devices.

Objective To develop an automatic analysis model of comprehensive performance of medical equipment based on data platform, and explore the realization and application value of comprehensive evaluation of equipment performance, so as to improve the efficiency of equipment management. Methods Using POWER BI series of big data analysis and management tools, using high-configuration virtual server, through the interface, collecting multi-service system data, integrating social benefit data calculated by fuzzy analytic hierarchy process, established a data quality management standard, efficient data platform. On the basis of the data platform, 205 measurements and 68 visual reports in 8 sections are designed to complete data analysis and equipment performance evaluation. Results With the help of RPA, the model realized the automatic process of equipment performance evaluation from data collection to visual report, and realized the multi-angle evaluation and analysis of equipment performance and real-time early warning monitoring. After the implementation of the system, the number of performance indicators gradually increased from the previous 4 to nearly 50, and the accuracy rate after source data cleaning was nearly 100%. The paired T-test was used to evaluate the difference in the time required by the two systems to calculate the month-on-month and year-on-year indicators for three consecutive years. The results showed that there was a significant difference in the analysis time between the traditional software and the big data model (t = 89.579, p < 0.01). The average analysis time of big data model is significantly shorter than that of traditional software, and the average time difference is 1196.50 seconds. At the same time, the standard deviation of data model analysis (1.08) is significantly lower than that of traditional software (41.85), which verifies the significant advantages of big data model in data query analysis tasks. A questionnaire was used to systematically evaluate the performance difference between traditional software and big data model from 10 technical indicators such as functionality, user interface (UI) and user experience (UX). Anova results showed that, in addition to compatibility (F(1,10)=0.24, P=0.634) and maintainability (F(1,10)=0.58, P=0.463), The big data model showed significant advantages (P<0.001) in eight indexes including functionality (F(1,10)=198.00, η²=0.94) and user experience (F(1,10)=244.00, η²=0.95), which reflected that users had higher comprehensive recognition of the big data model. Conclusion The application of this model realizes the effective transformation of equipment performance analysis from theoretical research to practical application, provides new tools and methods for hospital equipment management and economic operation, greatly improves the management efficiency and risk perception ability of medical equipment, and is a beneficial attempt of new quality productivity in hospital management.

Objective To develop an AI-based prone position facial dynamic monitoring system, aiming to automatically detect dangerous situations for surgical patients, such as unexpected facial movement, poor eyelid closure, and tracheal tube dislodgement. This system seeks to enhance automated monitoring levels for patients undergoing prone position surgery. Methods This system utilizes AI technology to intelligently analyze the patient’s facial condition. It employs the MTCNN algorithm for face detection, combined with the ERT algorithm for feature point recognition and precise localization. By analyzing changes in the nose tip position, the system identifies unexpected facial movements. The degree of eye-opening is determined using the EAR calculation, with grayscale variance analysis to detect incomplete eyelid closure, while MAR calculations of the mouth opening between consecutive frames help detect potential tracheal tube dislodgement. An experimental environment was established, recruiting 10 healthy volunteers to simulate prone position surgery patients for a clinical application study. Results The system achieved real-time monitoring and automatic alerting functions. Medical staff could observe the patient’s facial conditions on a computer screen at any time, and the system could automatically recognize dangerous situations such as facial movement, incomplete eyelid closure, and tracheal tube dislodgement, issuing voice warnings when needed. In the experimental setup, the accuracy of the algorithms for detecting facial movement, incomplete eyelid closure, and tracheal tube dislodgement was 100%, 95.7%, and 90.9%, respectively, with recall rates of 100%, 90%, and 80%. Conclusion The AI-based prone position facial dynamic monitoring system successfully achieves automatic monitoring and alerting for dangerous conditions like facial movement, incomplete eyelid closure, and tracheal tube dislodgement through intelligent algorithm analysis of camera-acquired image data.

Objective To analyze the differences between the new and old versions of GB/T 7247.1-2024 and GB 7247.1-2012 of "Safety of laser products--Part 1: Equipment classification and requirements," and to facilitate manufacturers and inspectors of laser medical devices in understanding the new version of the standard. Methods By combining common parameters and conditions in practical inspections, the significant changes in the new version of the standard were analyzed. Results In aspects of great concern for inspections, compared to the old version, the new version has undergone major changes in laser classification, measurement and evaluation conditions, accessible emission limits, maximum on-axis angle, external marking, and other aspects. Conclusion The new version of the standard incorporates more scientific data and has made more detailed revisions and improvements to multiple parameters, making it more aligned with actual situations. This paper analyzes the key changes in the new version of the standard, which is conducive to promoting the development of standardization in this field.

Objective Asthma is one of the most common chronic respiratory diseases in children. Establishing a specialized platform for pediatric asthma is of great significance for clinical standardized diagnosis, treatment, management, and scientific research. Methods Utilizing data warehousing technology to collect data from various systems within the hospital, then using standardized management to process text and image data, forming a specialized database for childhood asthma, and constructing a specialized application platform for childhood asthma that meets clinical patient management and research needs. Results The asthma specific platform covered over 2.31 million data of 115,032 children with asthma, and has incubated 7 retrospective or prospective clinical research projects; Compared to before the platform went online, the efficiency of case screening, case data collection, data cleaning, and data analysis has significantly improved, and the difference is statistically significant (P<0.001); The medication adherence rate has increased to 72.2%, the follow-up rate has increased to 74.50%, the asthma attack rate has decreased to 22.7%. The overall diagnosis and treatment level has been improved, and the difference is statistically significant (P<0.05). Conclusion The children's asthma specialized disease platform effectively improves the management of asthma patients, enhances diagnosis and treatment efficiency. At the same time, the standardized asthma data boosts the scientific research on asthma.

Objective To analyze the overall trend in the development of innovative medical devices over the decade from 2014 to 2023 and to investigate the registration status of innovative medical devices in 2023 for the reference of medical device registrants. Methods This paper collates relevant data on innovative medical devices from 2014 to 2023 and analyzes 61 publicly available evaluation reports on innovative medical devices in 2023. The research covers development trends, preclinical study materials, clinical evaluation materials from the 2023 evaluation reports, and supplements content related to innovation certification and post-market surveillance. Results From the perspective of development trends, the field of innovative medical devices is growing rapidly. The number of innovative medical devices registered and launched in the past two years accounts for 46% of the total. Among these, more than 62% concentrate in the following four areas: "Passive Implantable Medical Devices," "Active Surgical Medical Devices," "Medical Imaging Devices," and "Active Implantable Medical Devices." Based on the preclinical study materials of innovative medical devices in 2023, 23 innovative medical device systems conducted animal experiments, primarily using large animals. In clinical evaluation, clinical trials are the predominant method. Among them, 46 clinical evaluations were completed through clinical trials (comprising 37 individual clinical trials), 8 clinical evaluation reports were completed through a combination of clinical trials and comparison with similar products, and the remainder were completed solely through comparison with similar products. Conclusion The trend for innovative medical devices is positive, with a certain degree of concentration in specific fields but also gradual expansion. Registrants of innovative medical devices should prioritize "risk management," plan research validation and clinical evaluation methods based on the risk-benefit assessment of the product, and continuously collect scientific evidence related to product risks and benefits throughout the full lifecycle of the medical device to implement risk management.

Brain tumors are caused by abnormal growth of brain tissue cells and pose a significant threat to human life. Magnetic resonance imaging is a typical non-invasive imaging technique that can generate high-resolution, non-invasive, and skull artifact free brain images. With the continuous development of biomedical technology, the use of magnetic resonance imaging (MRI) technology for the diagnosis and treatment of brain tumors has become the main technical means to improve patient survival rates and reduce computational costs. Objective: This article systematically reviews the latest research progress and achievements of deep learning in multimodal MRI brain tumor segmentation. Firstly, the evaluation criteria and publicly available datasets involved in multimodal MRI brain tumor image segmentation were introduced. Secondly, an in-depth exploration was conducted on the ability of processing multimodal brain tumor image segmentation based on U-Net, Transformer, and SAM. The advantages and limitations of these technologies were summarized, and the performance of various models was compared. Finally, the problems and challenges faced by current multimodal MRI brain tumor image segmentation were discussed, and future research directions were also discussed. Contribution: This article analyzes the performance of different deep learning models on multimodal MRI brain tumor images, providing reference for researchers and clinical doctors in the field of medical image analysis, and improving the accuracy and efficiency of brain tumor diagnosis.

Objective To establish the multi-modality images automatic contouring model of clinical target volume (CTV) and organs at risk (OARs) of cervical cancer by using CT and magnetic resonance images (MRI) based on deep learning neural network. Methods A retrospective collection of 150 cases of localization CT and corresponding MRI (T2) from cervical cancer radiotherapy patients treated at Northern Jiangsu People's Hospital from January 2022 to September 2023 was performed. After preprocessing, the data was randomly divided into a training set (100 cases), a validation set (25 cases), and a test set (25 cases). An automatic contour model based on deep learning neural network was constructed and trained with the training and validation sets,and then the test set was contoured automatically. The accuracy of the automatic contour model of CTV and OARs (small intestine, bladder, rectum, left and right kidneys, left and right femoral heads) were calculated with manual contouring by clinical physicians, and the time taken would also be recorded. Results The Dice similarity coefficients (DSC) for the automatic contour model of CTV and OARs (small intestine, bladder, rectum, left and right kidneys, left and right femoral heads) were 0.87±0.03, 0.79±0.04, 0.95±0.04, 0.88±0.04, 0.96±0.02, 0.96±0.03, 0.92±0.03, and 0.93±0.03, respectively. The 95% Hausdorff distances (HD) (mm) were 5.12±1.45, 22.37±15.68, 1.27±0.31, 5.45±1.56, 1.15±0.21, 1.22±0.25, 4.51±2.38, and 4.56±2.77, respectively. The Overlap Index (OI) were 0.88±0.05, 0.83±0.04, 0.97±0.02, 0.91±0.04, 0.97±0.02, 0.97±0.03, 0.98±0.02 and 0.98±0.01, respectively. The time required to fully contour a case was 1.19±0.22 minutes. Conclusion The automatic contour model based on multi-modality images can realize the automatic contouring of CTV and OARs of cervical cancer accurately, and provide some useful reference for clinical physicians, and saved a lot of time.

Mild cognitive impairment (MCI) is a transitional state characterized by mild cognitive decline and is considered a precursor to Alzheimer's disease (AD). Resting-state functional magnetic resonance imaging (rs-fMRI), as a significant tool for studying brain functional connectivity (FC), plays a crucial role in understanding the functional connectivity changes in different subtypes of MCI. This review focuses on the application of rs-fMRI in revealing the functional connectivity changes in amnestic MCI (aMCI) and nonamnestic MCI (naMCI) patients. Studies have found significant differences in functional connectivity within key brain networks such as the default mode network (DMN), executive control network (ECN), and salience network (SN) between aMCI and naMCI patients, which may serve as potential biomarkers for early diagnosis of MCI. The article also discusses the clinical challenges in diagnosing MCI and how rs-fMRI can be utilized to enhance the accuracy of MCI diagnosis and understanding of its underlying neurobiological mechanisms.

Objective To analyze the research status, hotspots and development trends in the field of hospital bidding and procurement in China, and to provide references and suggestions for the subsequent implementation of bidding and procurement work and related research. Methods Based on the literature in the CNKI database, CiteSpace 6.2.R7 software was used to bibliometricly and visually analyze the number of publications, authors, institutions, and keywords of the literature related to hospital bidding and procurement from 2004 to 2024. Results A total of 683 valid literatures were included, and the search results showed a general upward trend in the number of publications on bidding and purchasing research; besides the search terms , keywords such as “medical equipment” “drugs” “management” “problems” “countermeasures” “medical consumables” “internal management” “supervision” and “informatization” had a relatively high frequency of occurrence; through the analysis of specific literature, combined with the co-occurrence and clustering results of keywords, it was found that the hotspots of hospital bidding and purchasing research mainly included four aspects: research on problems and processes of bidding and procurement of medical equipment and medical consumables, research on audit and risk management of hospital bidding and procurement, research on volume-based procurement of hospital drugs, and research on government procurement of hospitals. Conclusion The policy documents issued by the state have a guiding effect on research related to hospital bidding and procurement, and future research should focus on the internal management and risk prevention and control of hospital bidding and procurement, strengthen the management of government procurement, and give play to the key role of informatization and intelligent means.

Objective To explore the application value of image enhanced reconstruction technology based on generative artificial intelligence (SupMR) in fast pituitary MRI in children. Materials and Methods Twenty-six children volunteers underwent fast sequences and conventional sequences scanning of the pituitary. The images of fast sequence were transferred to SupMR post-processing system to generate SupMR images automatically. Three groups of MRI images were independently scored by two radiologists for the clarity of pituitary, artifacts and overall image quality. Peak signal-to-noise ratio (PSNR) and structural similarity (SSIM) were used as quantitative image quality assessment indexes, and the two indexes were based on FSE-T1WI conventional images and FSE-T2WI conventional images, respectively. Results The scanning time of FSE-T1WISupMR sequence was 49.4% shorter than that of conventional FSE-T1WI sequence. The scanning time of FSE-T2WISupMR sequence was 47.6% shorter than that of conventional FSE-T2WI sequence. Qualitative analysis: The image quality scores of FSE-T1WISupMR were higher than those of fast FSE-T1WI and conventional FSE-T1WI in terms of the clarity of pituitary, artifacts and overall image quality (the clarity of pituitary：H=58.754，P＜0.001；artifact：H=30.188，P＜0.001；overall image quality.：H=30.302，P＜0.001); The image quality scores of FSE-T2WISupMR were higher than those of fast FSE-T2WI and conventional FSE-T2WI in terms of the clarity of pituitary, artifacts and overall image quality (the clarity of pituitary：H=38.567， P＜0.001；artifact：H=24.034，P＜0.001；overall image quality.：H=26.016，P＜0.001). Quantitative analysis: Both PSNR and SSIM between FSE-T1WISupMR sequence and FSE-T1WIconventional sequence images were superior to those between FSE-T1WIfast sequence and FSE-T1WIconventional sequence images (PSNR: t=-2.416, P=0.023; SSIM: t=-2.172, P=0.040); Both PSNR and SSIM between FSE-T2WISupMR sequence and FSE-T2WIconventional sequence images were superior to those between FSE-T2WIfast sequence and FSE-T2WIconventional sequence images (PSNR：t=-5.684，P＜0.001；SSIM：t=-4.614，P＜0.001). Conclusion s: Based on SupMR technology, MRI image quality can be guaranteed or improved to a certain extent, which helps to significantly shorten the scanning time and realize the fast scanning of children pituitary MRI.

Objective To establish a set of usability evaluation indicator system for magnetocardiograms based on user perspective, which can be used to support the selection decision of magnetocardiograms in medical institutions; and to explore post-market usability evaluation of innovative medical devices. Methods A technical route of deductive reasoning was adopted, which firstly developed the general usability evaluation indicator system of innovative medical devices with methods of literature analysis, brainstorming and expert consultation, and secondly established the usability evaluation indicator system of magnetocardiograms with methods of system analysis, field investigation and interview, and Delphi expert consultation, combined with the characteristics of magnetocardiograms. Results A usability evaluation index system for innovative medical devices including 4 first-level indicators, 10 second-level indicators and 40 third-level indicators and one for magnetocardiograms including 4 first-level indicators and 20 second-level indicators were established; the Delphi expert consultation got authority coefficients≥0.95, variation coefficients≤0.17. Conclusion The evaluation indicator system proposed is scientific, feasible and authoritative.

Objective Comprehensively considering the benefits of large medical equipment configuration and use, along with the specific characteristics of pediatric diagnosis and treatment services. The study aims to develop a standardized workload tool for imaging diagnostics in pediatric patients and construct a comprehensive performance evaluation model for large medical equipment in hospitals. This model seeks to establish an evaluation method applicable to both pediatric specialized hospitals and general hospitals, providing a reference for the management and evaluation of large medical equipment in public hospitals. Methods Through a literature review and expert consultation, a comprehensive performance evaluation index system for large medical equipment was constructed. Using the fuzzy analytic hierarchy process (FAHP), indicator weights were calculated, and data analysis was conducted to examine MRI treatment duration in pediatric patients across different age groups, quantifying the impact of pediatric diagnostic specificity. Results The comprehensive performance evaluation index system for large medical equipment includes three primary indicators—economic benefits, social benefits, and research value—and 13 secondary indicators. Secondary indicators with higher weights include dynamic investment recovery period, number of serviced patients, and total number of research outcomes. Compared to the adolescent group, the average diagnostic duration for preschool children increased by 16.32%, while for school-age children, it increased by 8.45%. For sedated patients, the duration was largely consistent, but sedation procedures required approximately an additional 83 seconds of workload per medical staff member. Conclusion The comprehensive performance evaluation model for large medical equipment, based on the specific characteristics of pediatric diagnosis and treatment, provides pediatric specialized hospitals with a refined approach to equipment management evaluation. This model enables precise identification of the current strengths and weaknesses of large medical equipment, helping hospitals to improve equipment utilization efficiency, reduce patient wait times for examinations, and continuously enhance patient satisfaction and healthcare experience.

Objective To improve the utilization rate of medical equipment and the level of fine, digital, intelligent and scientific management of medical equipment by constructing a benefit evaluation system for the use of large medical equipment. Methods Through literature analysis and focus group interviews, the evaluation index pool of the use efficiency of large medical equipment was established, and the evaluation index system and the weight of each index were determined by Delphi method and analytic hierarchy process. Results An evaluation index system for the use efficiency of large medical equipment was established, which included 4 first-level indicators of social benefit, economic benefit, management benefit, science and education benefit and 17 second-level indicators. Among them, the weight coefficients of the first-level indicators were 43.93 %, 31.07 %, 14.64 %, and 10.36 %, respectively. The weight coefficient interval of the secondary index is 0.82 % ~ 19.90 %. Conclusion This index system can objectively evaluate the use efficiency of medical equipment in public hospitals, and provide data support for equipment procurement, use and operation and maintenance.