【摘要】 目的 从卫生系统角度评价冠状动脉CT血管造影（CCTA）智能辅助诊断系统应用于冠状动脉粥样硬化性心脏病（CAD）诊断的经济性。方法 以诊断准确度作为效果值，比较CCTA-AI与CCTA-人工诊断的成本-效果；通过Treeage Pro2011软件构建决策树模型，以质量调整生命年（QALYs）为效用指标，模拟CAD患者利用CCTA人工智能辅助诊断技术1年的成本投入及获益情况，比较CCTA-AI与CCTA-人工诊断的成本-效用；并进行单因素及概率敏感性分析验证模型结果的稳定性。结果 基于诊断准确度的成本-效果分析，AI组成本为1 074 900元，效果为855.05，人工组成本为1 266 610元，高年资医师组效果为815.068 5，低年资医师组为793.4,AI组具有绝对优势；决策树模型分析结果显示，相较于高年资医师组，AI组增量成本-效用比（ICUR）值为2969386.31832元/QALY，相较于低年资医师组，AI组ICUR值为3 682 121.485 41元/QALY，均大于阈值（268 074元），表明AI组每增加1单位效用值，其成本在此阈值下不可接受。单因素敏感性分析和概率敏感性分析结果显示基础分析结果较为稳健。结论 将诊断准确度作为效果指标进行CEA时，AI辅助诊断相较于人工具有绝对优势，但将诊断放入CAD短期诊治过程中，以QALYs作为效用指标，来考察1年内从CAD诊断到治疗的成本-效果时，得出的结论则是AI辅助诊断不经济。更多还原

【Abstract】 Objective To evaluate the economy of applying the intelligent auxiliary diagnosis system of coronary CT angiography (CCTA) in the diagnosis of coronary atherosclerotic heart disease (CAD) from the perspective of the health system.Methods Taking the diagnostic accuracy as the effect value,the cost-effectiveness of CCTA-AI and CCTA-manual diagnosis was compared.The decision tree model was constructed through Treeage Pro 2011 software.Taking the quality-adjusted life years(QALYs) as the utility index,the cost input and benefit of CAD patients using the CCTA artificial intelligence-assisted diagnosis technology for one year were simulated to compare the cost-utility of CCTA-AI and CCTA-manual diagnosis.And univariate and probabilistic sensitivity analyses were conducted to verify the stability of the model results.Results based on the cost-effectiveness analysis of diagnostic accuracy,the cost of the AI group was 1 074 900 yuan with an effect of 855.05,and the cost of the manual group was 1 266 610 yuan,with an effect of 815.068 5 for the high-year-old physician group,and 793.4 for the low-year-old physician group,which gave an absolute advantage to the AI group.The results of the decision-tree modeling analysis showed that,compared with the high-year-old physician group,the ICUR value of AI group was 2 969 386.318 32 yuan/QALY,and compared to the low seniority physician group,the ICUR value of AI group was 3 682 121.485 41 yuan/QALY,both of which were greater than the threshold value (268 074 yuan),indicating that for every unit of increase in the value of utility of the AI group,its cost was unacceptable at this threshold value.The results of both the one-way sensitivity analysis and the probabilistic sensitivity analysis showed that the results of the underlying analysis were robust.Conclusion When diagnostic accuracy is used as the effect indicator for CEA,AI-assisted diagnosis has an absolute advantage over manual diagnosis.However,when diagnosis is placed in the short-term diagnosis and treatment process of CAD and QALYs is used as the utility indicator to examine the cost-effectiveness from CAD diagnosis to treatment within one year,the conclusion drawn is that AI-assisted diagnosis is uneconomical.更多还原