Fan Yang1, Dong Li1, Zhenwen Yang2, Zhang Zhang1, Dan Wang1, Tielian Yu1. 1. Department of Radiology, Tianjin Medical University General Hospital, Tianjin 300052, China. 2. Department of Cardiovascular Disease, Tianjin Medical University General Hospital, Tianjin 300052, China.
Abstract
BACKGROUND: Pulmonary hypertension (PH) often leads to dilatation of main pulmonary artery (MPA). MPA measurements can be used to predict PH. This aim of this study is to investigate power of MPA vessel indices, which are acquired from cardiovascular magnetic resonance, to evaluate PH. METHODS: Cardiovascular-magnetic-resonance-determined parameters of MPA were acquired and calculated in 83 PH patients, whose diagnosis were confirmed with right heart catheterization and 49 healthy volunteers; these parameters included MPA diameter (DPA), ratio of DPA and ascending aorta diameter (DPA/DAo), max mean diameter (MDmax), min mean diameter (MDmin), fraction transverse diameter (fTD), fraction longitudinal diameter (fLD), and distensibility. RESULTS: Compared with control group, DPA, DPA/DAo, MDmax, and MDmin were significantly higher in patients with PH (P<0.001); fTD, fLD, and distensibility significantly decreased (P<0.001). fTD was lesser than fLD in control group (P<0.001), whereas difference was not observed in PH (P=0.305). MPA indices were significantly correlated with mean pulmonary arterial pressure (mPAP) (P<0.05), and strongest correlation was observed for DPA/DAo (r=0.534, P<0.001). By receiver operating characteristic analysis, MDmin>28.4 mm, and MDmax>32.4 mm (area under the curve, AUC=0.979, 0.981) showed best performance in predicting PH, yielding highest specificity at 100%. CONCLUSIONS: Noninvasive cardiovascular-magnetic-resonance-derived MPA measurements provide excellent and practical reference in clinical settings for detecting PH.
BACKGROUND:Pulmonary hypertension (PH) often leads to dilatation of main pulmonary artery (MPA). MPA measurements can be used to predict PH. This aim of this study is to investigate power of MPA vessel indices, which are acquired from cardiovascular magnetic resonance, to evaluate PH. METHODS: Cardiovascular-magnetic-resonance-determined parameters of MPA were acquired and calculated in 83 PHpatients, whose diagnosis were confirmed with right heart catheterization and 49 healthy volunteers; these parameters included MPA diameter (DPA), ratio of DPA and ascending aorta diameter (DPA/DAo), max mean diameter (MDmax), min mean diameter (MDmin), fraction transverse diameter (fTD), fraction longitudinal diameter (fLD), and distensibility. RESULTS: Compared with control group, DPA, DPA/DAo, MDmax, and MDmin were significantly higher in patients with PH (P<0.001); fTD, fLD, and distensibility significantly decreased (P<0.001). fTD was lesser than fLD in control group (P<0.001), whereas difference was not observed in PH (P=0.305). MPA indices were significantly correlated with mean pulmonary arterial pressure (mPAP) (P<0.05), and strongest correlation was observed for DPA/DAo (r=0.534, P<0.001). By receiver operating characteristic analysis, MDmin>28.4 mm, and MDmax>32.4 mm (area under the curve, AUC=0.979, 0.981) showed best performance in predicting PH, yielding highest specificity at 100%. CONCLUSIONS: Noninvasive cardiovascular-magnetic-resonance-derived MPA measurements provide excellent and practical reference in clinical settings for detecting PH.
CMR、MPA图像采集及径线测量方法(A-D为同一PH患者)。A:CMR横轴位非门控FIESTA序列MPA横径最大的层面,分别测量DPA和DAo。B:由A图取平行于MPA走行方向定位,获得RV流出道和MPA图像。C、D:由B图于肺动脉瓣上方1.5 cm-2 cm处垂直于血流方向定位,获得MPA横截面Fast Cine PC序列单层多时相图像,于幅度图上手动勾画心动周期内各时相MPA内缘,在最大、最小截面积时相分别测量MPA横径(TD)和纵径(LD)。
The measurement methods of CMR MPA diameters (A-D from the same patient). A: DPA and DAo were measured from CMR non-gated FIESTA transversal image on the slice of the maximum MPA diameter. B: RV outflow tract with longitudinal MPA was acquired by localizing parallel to the direction of MPA in A. C, D: An image plane was prescribed perpendicular to the MPA flow direction and 1.5 cm-2 cm above the level of the pulmonary valve, and a single slice multi-phase cross-section images of MPA were obtained with Fast Cine PC sequence. Tracing the inner edge of MPA manually on each amplitude image during cardiac cycle was in order to measure the MPA TD and LD on the phases with the maximum and minimun area. CMR: cardiovascular magnetic resonance; MPA: main pulmonar y ar ter y; PH: pulmonary hypertension; FIESTA: fast imaging employing steady-state acquisition; DPA: main pulmonary artery diameter; DAo: ascending aorta diameter; RV: right ventricle; PC: phase contrast; TD: transverse diameter; LD: longitudinal diameter.
CMR、MPA图像采集及径线测量方法(A-D为同一PH患者)。A:CMR横轴位非门控FIESTA序列MPA横径最大的层面,分别测量DPA和DAo。B:由A图取平行于MPA走行方向定位,获得RV流出道和MPA图像。C、D:由B图于肺动脉瓣上方1.5 cm-2 cm处垂直于血流方向定位,获得MPA横截面Fast Cine PC序列单层多时相图像,于幅度图上手动勾画心动周期内各时相MPA内缘,在最大、最小截面积时相分别测量MPA横径(TD)和纵径(LD)。The measurement methods of CMR MPA diameters (A-D from the same patient). A: DPA and DAo were measured from CMR non-gated FIESTA transversal image on the slice of the maximum MPA diameter. B: RV outflow tract with longitudinal MPA was acquired by localizing parallel to the direction of MPA in A. C, D: An image plane was prescribed perpendicular to the MPA flow direction and 1.5 cm-2 cm above the level of the pulmonary valve, and a single slice multi-phase cross-section images of MPA were obtained with Fast Cine PC sequence. Tracing the inner edge of MPA manually on each amplitude image during cardiac cycle was in order to measure the MPA TD and LD on the phases with the maximum and minimun area. CMR: cardiovascular magnetic resonance; MPA: main pulmonar y ar ter y; PH: pulmonary hypertension; FIESTA: fast imaging employing steady-state acquisition; DPA: main pulmonary artery diameter; DAo: ascending aorta diameter; RV: right ventricle; PC: phase contrast; TD: transverse diameter; LD: longitudinal diameter.
CMR图像分析和数据测量
将CMR扫描图像传输至GE AW 4.3 MRI工作站,应用Report Card 3.7软件进行数据分析。在CMR的横轴位非门控FIESTA图像上,选取MPA横径最大的层面分别测量MPA横径值DPA和升主动脉横径值DAo,并计算DPA/DAo。在MPA横截面的PC序列幅度图上手动描点勾画心动周期各时相MPA内缘,软件自动计算各个时相MPA的截面积,在MPA最大、最小截面积的时相图上分别测量MPA横径(transverse diameter, TD)和纵径(longitudinal diameter, LD)(图 1C,图 1D),计算最大平均直径(max mean diameter, MDmax)、最小平均直径(min mean diameter, MDmin)、横径变化分数(fraction transverse diameter, fTD)、纵径变化分数(fraction longitudinal diameter, fLD)、MPA顺应性(distensibility)。计算公式如下:MDmax=(TDmax+LDmax)/2MDmin=(TDmin+LDmin)/2fTD=100×[(TDmax-TDmin)/TDmin]fLD=100×[(LDmax-LDmin)/LDmin]Distensibility=100×[(最大截面积–最小截面积)/最小截面积]
采用SPSS 18.0统计软件对数据进行统计。计量资料以均数±标准差(Mean±SD)表示。PH组与对照组的MPA和主动脉各相关测量结果采用多变量一般线性模型(multivariate general linear model)进行比较,以排除年龄、性别及BSA的影响。分别比较PH组及对照组的fTD和fLD。对CMR获取的MPA各参数及DPA/DAo与RHC的mPAP进行偏相关分析,以控制协变量年龄、性别及BSA的影响。采用受试者工作特征(receiver operating characteristic, ROC)曲线计算并综合评估各指标曲线下面积(area under the curve, AUC),比较各指标评价PH的敏感度、特异度,计算约登指数(Youden index)以寻找评估PH状态的临界值(cut-off value)。以P < 0.05为差异有统计学意义。
Comparison of MPA vessel diameters between PH and control group. DPA, MDmax and MDmin were significantly increased in PH. MDmax: max mean diameter; MDmin: min mean diameter.
1
PH组与对照组CMR MPA、Ao各参数比较
Comparation of parameters of MPA and Ao between PH and control group
Parameters
Control group (n=49)
PH group (n=83)
P value
Ao: ascending aorta; DPA/DAo: ratio of the DPA and DAo.
Comparison between the fTD and fLD. FTD was smaller than fLD in control group, whereas there was no significant difference in PH. fTD: fraction transverse diameter; fLD: fraction longitudinal diameter.
PH组与对照组间MPA径线参数比较。PH组DPA、MDmax、MDmin均显著升高。Comparison of MPA vessel diameters between PH and control group. DPA, MDmax and MDmin were significantly increased in PH. MDmax: max mean diameter; MDmin: min mean diameter.PH组与对照组CMR MPA、Ao各参数比较Comparation of parameters of MPA and Ao between PH and control groupPH组及对照组各自fTD与fLD相比,对照组fTD小于fLD(P < 0.001),而PH组两参数间无显著性差异(P=0.305)。Comparison between the fTD and fLD. FTD was smaller than fLD in control group, whereas there was no significant difference in PH. fTD: fraction transverse diameter; fLD: fraction longitudinal diameter.
PH组CMR MPA各参数及DPA/DAo与mPAP间相关性
DPA、DPA/DAo、MDmax、MDmin分别与mPAP呈正相关,其中DPA/DAo相关性最强(r=0.534, P < 0.001),其次为MDmin(r=0.362, P=0.001),DPA与mPAP相关性较上述两参数稍弱(r=0.326, P=0.003),MDmax表现相对较弱(r=0.315, P=0.004)。顺应性、fLD、fTD分别与mPAP呈负相关,对应相关系数r值依次为-0.298(P=0.007)、-0.290(P=0.009)和-0.273(P=0.014)(图 4)。
Scatterplots showed the relationships between MPA indices and mPAP. DPA, DPA/DAo, MDmax, and MDmin were positively correlated with mPAP respectively, and the strongest was found for DPA/DAo (r=0.534, P < 0.001). Disitensibility, fLD, and fTD had negative correlation with mPAP. MPA: main pulmonary artery; mPAP: mean pulmonary artery pressure.
PH组MPA各参数与mPAP间的相关性的散点图。DPA、DPA/DAo、MDmax、MDmin分别与mPAP呈显著正相关,其中DPA/DAo相关性最强;顺应性、fLD、fTD分别与mPAP呈负相关。Scatterplots showed the relationships between MPA indices and mPAP. DPA, DPA/DAo, MDmax, and MDmin were positively correlated with mPAP respectively, and the strongest was found for DPA/DAo (r=0.534, P < 0.001). Disitensibility, fLD, and fTD had negative correlation with mPAP. MPA: main pulmonary artery; mPAP: mean pulmonary artery pressure.
ROC curve showed the ability of MPA indices to predict PH. MDmin>28.4 mm and MDmax>32.4 mm performed relative better than others in predicting PH (AUC=0.979, 0.981, respectively). ROC: Receiveroperating characteristic.
MPA各参数预测PH的效能Diagnostic performance of MPA parameters for predicting PHROC曲线显示MPA各参数预测PH的效能。MDmin> 28.4 mm和MDmax>32.4 mm预测PH效能较其他参数更高(AUC=0.979, 0.981)。ROC curve showed the ability of MPA indices to predict PH. MDmin>28.4 mm and MDmax>32.4 mm performed relative better than others in predicting PH (AUC=0.979, 0.981, respectively). ROC: Receiveroperating characteristic.
Authors: J B Haimovici; B Trotman-Dickenson; E F Halpern; G W Dec; L C Ginns; J A Shepard; T C McLoud Journal: Acad Radiol Date: 1997-05 Impact factor: 3.173
Authors: Marius M Hoeper; Harm Jan Bogaard; Robin Condliffe; Robert Frantz; Dinesh Khanna; Marcin Kurzyna; David Langleben; Alessandra Manes; Toru Satoh; Fernando Torres; Martin R Wilkins; David B Badesch Journal: J Am Coll Cardiol Date: 2013-12-24 Impact factor: 24.094
Authors: Anand Devaraj; Athol U Wells; Mark G Meister; Tamera J Corte; Stephen J Wort; David M Hansell Journal: Radiology Date: 2010-02 Impact factor: 11.105
Authors: Nicolas Creuzé; Susana Hoette; David Montani; Sven Günther; Edmund Lau; Julien Ternacle; Laurent Savale; Xavier Jaïs; Florence Parent; Barbara Girerd; Olivier Sitbon; Gérald Simonneau; Carlos E Rochitte; Rogerio Souza; Marc Humbert; Denis Chemla Journal: Can J Cardiol Date: 2015-04-25 Impact factor: 5.223
Authors: Andrew J Swift; Smitha Rajaram; Robin Condliffe; Dave Capener; Judith Hurdman; Charlie Elliot; David G Kiely; Jim M Wild Journal: Invest Radiol Date: 2012-10 Impact factor: 6.016