超声-CT多模态融合成像构建器质性二尖瓣关闭不全个体化计算流体动力学模型的可行性研究

Feasibility study on constructing individualized computational fluid dynamics models for organic mitral valve regurgitation using ultrasound⁃CT multimodal fusion imaging

目的:探索超声和CT多模态融合成像结合计算流体动力学(CFD)模型在评估器质性二尖瓣关闭不全患者中的可行性和有效性,旨在为临床决策提供更精确和全面的诊断信息。方法:研究纳入2022年3月至2023年9月期间武汉大学人民医院入院治疗的15例器质性二尖瓣关闭不全患者,年龄范围58~72岁。对超声和心脏CT扫描图像进行后处理和图像融合,基于这些融合图像构建个体化CFD模型。在模型中进行解剖和流场测量并与超声数据进行比对以评估模型的解剖精确性和流场精确性。解剖参数包括二尖瓣前瓣长度(AL)、后瓣长度(PL)、瓣环的前后径(ANN_AP)、二尖瓣脱垂的宽度(PW)以及二尖瓣脱垂缝隙宽度(GAP);流场参数包括PISA半径(PISAR)、最大反流速度(VMAX)、反流速度时间积分(VTI)、缩流颈(VC)、有效反流口面积(EROA)和反流容积(RV)。结果:超声-CT融合成像与CFD模型结合显示出高度的解剖结构重建精确性和血流动力学特性模拟能力。融合模型在解剖参数除GAP(P=0.03)外均与超声数据差异无统计学意义(P>0.05);血流动力学参数除VC(P=0.002)外均与超声测量值差异无统计学意义(P>0.05)。Bland-Altman分析显示超声-CT融合模型与超声测量值之间存在良好的一致性。结论:基于超声-CT多模态融合的解剖模型可以构建较为精确的器质性二尖瓣关闭不全反流的计算机流体模型,对二尖瓣疾病的评估和治疗具有潜在的临床价值。

Objective To explore the feasibility and effectiveness of combining ultrasound and CT multimodal fusion imaging with computational fluid dynamics(CFD)models in evaluating patients with organic mitral valve regurgitation,aiming to provide more accurate and comprehensive diagnostic information for clinical decision⁃making.Methods The study included 15 patients with organic mitral valve regurgitation treated in Renmin Hospital of Wuhan University from March 2022 to September 2023,aged from 58 to 72 years.Ultrasound and cardiac CT scan images were post⁃processed and fused to construct individualized CFD models.Anatomical and flow field measurements were taken within the model and compared with ultrasound data to assess the model's anatomical accuracy and flow field accuracy.Anatomical parameters included mitral valve anterior leaflet length(AL),posterior leaflet length(PL),anterior⁃posterior diameter of the annulus(ANN_AP),width of mitral valve prolapse(PW),and width of the prolapse gap(GAP);flow field parameters included PISA radius(PISAR),maximum regurgitation velocity(VMAX),regurgitation velocity⁃time integral(VTI),contraction neck(VC),effective regurgitation orifice area(EROA),and regurgitation volume(RV).Results The combination of ultrasound⁃CT fusion imaging and CFD modeling demonstrated high accuracy in anatomical reconstruction and hemodynamic characteristics modeling.The fusion model showed no statistical difference in anatomical parameters except for GAP(P=0.03)as compared with ultrasound data(P>0.05);hemodynamic parameters showed no statistical difference except for VC(P=0.002)as compared with ultrasound measurements(P>0.05).Bland⁃Altman analysis indicated good consistency between the ultrasound⁃CT fusion model and ultrasound measurements.Conclusion The anatomical model based on ultrasound⁃CT multimodal fusion can construct a relatively accurate computational fluid model of organic mitral valve regurgitation reflux,offering potential clinical value in the assessment and treatment of mitral valve diseases.