基于最优规划模型的磁共振成像设备质量控制研究

Research on quality control of magnetic resonance imaging equipment based on optimal planning model

目的:构建基于反向传播(BP)神经网络算法的最优规划模型,探讨其在磁共振成像(MRI)设备质量控制中的应用价值。方法:以MRI设备图像质量、质量控制成本和故障处理时长为控制目标,以环境因素、人为因素、设备因素及使用频率等方面的13项指标为决策因子,构建基于BP神经网络算法的最优规划模型。选取2021年5月31日至2023年6月4日中山市人民医院临床在用的1台1.5T磁共振设备的运行数据,将2021年5月31日至2022年5月29日(52周)的设备运行数据进行模型训练,作为传统质量控制方案数据,反向计算进行最优方案演化动态优化,运用动态优化方案在2022年6月6日至2023年6月4日进行应用实践,其运行数据作为优化质量控制方案数据,对比两种方案的MRI设备图像质量评分、质量控制成本和故障处理时长。结果:采用最优规划模型优化质量控制方案的MRI设备图像质量评分为(4.15±0.35)分,高于传统质量控制方案,质量控制成本和故障处理时长分别为(5247.44±1711.39)元和(4.34±2.31)h,低于传统质量控制方案,差异均有统计学意义(t=4.084、6.442、10.776,P<0.05)。结论:基于BP神经网络算法构建最优规划模型,对MRI设备质量控制方案进行优化,可有效提升MRI设备质量管理水平,保障图像质量,提高设备稳定性,降低设备故障率和质量控制成本。

Objective:To construct an optimal planning model based on backpropagation(BP)neural network algorithm,and to explore its application value in the quality control of magnetic resonance imaging(MRI)equipment.Methods:Taking image quality,quality control cost and troubleshooting time as control objectives,and 13 indicators such as environmental factors,human factors,equipment factors,and use frequency as decision factors,an optimal planning model based on BP neural network algorithm is constructed.The operation data of a 1.5T magnetic resonance device in clinical use in Zhongshan People's Hospital from 31 May 2021 to 4 June 2023 were selected.The equipment operation data for 52 weeks from 31 May 2021 to 29 May 2022 was used for model training,which was used as the data of the conventional quality control scheme,and the optimal scheme evolution and dynamic optimization were carried out by reverse calculation.The dynamic optimization scheme was used to apply the practice from 6 June 2022 to 4 June 2023,and its operation data was used as the data of the optimization quality control scheme.The equipment image quality score,quality control cost and troubleshooting time of the two schemes were compared.Results:The image quality score of MRI equipment optimized using the optimal planning model for quality control scheme was(4.15±0.35)points,which was higher than that of conventional quality control scheme,the quality control cost and troubleshooting time were CNY(5247.44±1711.39)and(4.34±2.31)h,respectively,which were lower than those of conventional quality control scheme,the differences were statistically significant(t=4.084,6.442,10.776,P<0.05).Conclusion:The optimal planning model was constructed based on the BP neural network algorithm and the quality control scheme of MRI equipment was optimized,which can effectively improve the quality management level of MRI equipment,ensure image quality,improve equipment stability,reduce failure rates and quality control costs.