Clinical Value of Gated 99Tcm-Mibi SPECT for Assessment of Left Ventricular Function in Patients with Left Ventricular Aneurysms
TIAN Cong-na,LU Min-jie,WEI Hong-xing,TIAN Yue-qin,SHEN Rui,YANG Min-fu,WANG Dao-yu,FANG Wei,ZHAO Shi-hua,WANG Yue-tao,HE Zuo-xiang,LIU Xiu-jie,ZHANG Xiao-li
DOI: https://doi.org/10.3760/cma.j.issn.2095-2848.2012.06.001
2012-01-01
Abstract:Objective To validate the accuracy of quantitative gated SPECT (QGS) in assessing EDV,ESV and LVEF in patients with left ventricular aneurysms (LVA),using MRI as a reference.Furthermore,the correlations between the number of segments with severe perfusion defects,size of LVA and the differences of left ventricular volumes and LVEF were explored.Methods Sixty-nine patients (62 males and 7 females; mean age,(56 ± 10) years) with LVA were examined using gated 99Tcm-MIBI SPECT (G-SPECT) (8 frames/cardiac cycle) at rest and MRI (25 frames/cardiac cycle) at (4 ± 3) d.All the data were retrospectively reviewed.EDV,ESV,LVEF and wall thickening (WT) scores were calculated using QGS.The size of LVA was defined by the number of akinetic or dyskinetic segments with WT =3,and the segments with severe perfusion defects (activity ≤ 30％) were analyzed using quantitative perfusion SPECT (QPS) with a 17-segment model.EDV,ESV and LVEF were obtained using the Simpson's rule from MR images.A paired t test,analysis of variance and x2 test were used to analyze the data obtained from G-SPECT and MRI.Pearson correlation coefficients (r) and Bland-Altman analysis were applied to assess the strength of the agreement,and standard error of estimate (SEE) was calculated.Results According to the number of segments with severe perfusion defects,patients were divided into Group 1 (≤ 2 segments,n =25),Group 2 (3-4 segments,n =16) and Group 3 (≥ 5 segnents,n =28).Correlation was good for all parameters (r =0.83-0.95,all P ＜ 0.001) in all groups,except for LVEF in Group 3 (r =0.64,SEE =8％,P ＜ 0.001).Compared with MRI,EDV in Group 1 was underestimated by G-SPECT ((174±44) ml vs (161 ±43) ml,t =-2.60,P＜0.05),while the other parameters in Groups 1 and 2 had no significant difference (Group 1,t =-1.63,-1.26 for ESV and LVEF ; Group 2,t =-0.62,0.48,-2.07 for EDV,ESV and LVEF,respectively,all P ＞ 0.05).In Group 3,EDV ((217 ± 68) ml vs (196±55) ml,t=2.99,P＜0.01) and ESV ((163 ±56) ml vs (141 ±50) ml,t =3.80,P ＜0.01)were overestimated,and LVEF was underestimated ((26 ± 6) ％ vs (29 ± 9) ％,t =-2.69,P ＜ 0.05) by G-SPECT.The rates of difference in Group 3 between G-SPECT and MRI for ≥ 30 ml in EDV,≥ 20 ml in ESV,≥ 5％ or ≥ 10％ in LVEF were 50％ (14/28),57％ (16/28),54％ (15/28),21％ (6/28),respectively,which were significantly higher than those in Groups 1 and 2.In addition,the number of segments with severe perfusion defects was correlated with dEDV (r =0.37,P ＜ 0.01) and dESV (r =0.39,P ＜0.01) between G-SPECT and MRI.Moreover,the size of LVA was correlated with dEDV (r =0.27,P ＜0.05),dEsv(r =0.41,P ＜0.001 and dLVEF (r =-0.33,P ＜ 0.01) between G-SPECT and MRI.Conclusions In LVA patients with mild to moderate extent of severe perfusion defects (≤ 4 segments),GSPECT measurements of EDV,ESV and LVEF show high correlation and agreement with MRI measurements,which suggests that G-SPECT could be used to evaluate left ventricular function.However,in LVA patients with large severe perfusion defects,LV volumes were overestimated and LVEF was underestimated by GSPECT.Moreover,the accuracy of LV function assessed by G-SPECT may be affected by the size of LVA.