Objective To comparatively analyze the difference between visual method and soft-ware program in calculating different types of myocardium and to evaluate the correlation and consistency of these two methods for PET/CT myocardial perfusion and metabolic imaging. Methods Fifty-one pa-tients whose LVEF were severely reduced(≤40%)received 13N-ammonia(NH3)/18F-fluorodeoxyglu-cose(FDG)PET/CT to evaluate the number of normal myocardium,hibernating myocardium and scar myocardium semi-quantitatively measured by visual method and ECToolbox software. According to the re-sults of perfusion and metabolic imaging,myocardium was divided into normal myocardium,hibernating myocardium and scar. Percent of each type of myocardium was measured by those two methods and com-pared. The number of hibernating myocardium was divided into three groups according to the result of vis-ual method. Group 1:the number of hibernating myocardium was less than 10%;group 2:the number of hibernating myocardium was more than 10% but less than 20%;group 3:the number of hibernating my- ocardium was more than 20%. The number of hibernating myocardium evaluated by different methods in each group was then differentiated. Results The number of hibernating myocardium(16.8 ± 11.8 vs. 12.1 ± 12.6,P＝0.002)and normal myocardium(68.7 ± 11.9 vs. 63.7 ± 15.4,P＝0.016)were dif-ferent between the two methods. The number of hibernating myocardium and normal myocardium calculat-ed by software was higher than that by visual method,and there was no significant difference in the num-ber of scar myocardium(P＞0.05). A positive correlation for hibernating myocardium(r＝0.767)and normal myocardium(r＝0.478)was found between visual method and software. Bland-Altman curve in-dicated that the number of hibernating myocardium and normal myocardium evaluated by these two meth-ods was inconsistent except for the number of scar myocardium. Hibernation subgroup analysis:the num-ber of hibernating myocardium evaluated by visual method was higher than that evaluated by software (2.6 ± 3.6 vs. 10.2 ± 6.2,P＝0.000). There was no significant difference in the number of hiberna-ting myocardium between group 2(P＝0.064)and group 3(P＝0.299). Conclusions Visual method correlates well with software program in assessing scar myocardium. The number of hibernating myocardi-um evaluated by software is higher than that by visual method. Visual method will underestimate the num-ber of hibernating myocardium,but with the range of mismatch (hibernating myocardium)increasing, the difference tends to be smaller.%目的 比较左心室功能不全的冠心病患者应用PET/CT心肌灌注-代谢显像软件自动定量法及目测法分析所获得的各类型心肌数量的异同,并检测两种方法的相关性和一致性.方法 收集符合入组标准的51例冠心病患者(左心室射血分数均≤40%).入组患者均行PET/CT心肌灌注-代谢显像评估存活心肌,并分别通过软件自动分析和观察者目测法分析正常心肌、冬眠心肌、瘢痕心肌数量.比较同一患者采用两种方法所得正常心肌、冬眠心肌、瘢痕心肌数量的异同,并评估两种方法结果的相关性和一致性.进一步按目测法分析得出的冬眠心肌数量将上述患者分为三组,组1:目测法冬眠心肌数量＜10%;组2:10%≤目测法冬眠心肌数量＜20%;组3:目测法冬眠心肌数量≥20%.对比分析上述各组患者冬眠心肌数量的差异.结果 两种方法计算所得冬眠心肌数量(16.8 ± 11.8比12.1 ± 12.6,P＝0.002)及正常心肌数量(68.7 ± 11.9比63.7 ± 15.4,P＝0.016)存在差异,软件自动定量方法计算所得冬眠心肌数量及正常心肌数量均高于目测法(P均＜0.05);而同一患者两种方法计算所得瘢痕心肌数量比较差异未见统计学意义(P＞0.05).软件自动定量法与目测评估方法计算所得瘢痕心肌数量成高度正相关(r＝0.835),冬眠心肌数量成中等程度正相关(r＝0.767);正常心肌数量呈低度正相关(r＝0.478).两种分析方法计算出的冬眠心肌及正常心肌数量结果不一致,瘢痕心肌数量结果一致.冬眠心肌分组分析:组1两种分析方法所得冬眠心肌数量差异有统计学意义(P＜0.05),目测法计算冬眠心肌数量比定量法计算冬眠心肌数量少(2.6 ± 3.6比10.2 ± 6.2);组2及组3两种分析方法所得冬眠心肌数量比较差异未见统计学意义(P＞0.05).结论 PET/CT心肌灌注-代谢显像应用软件自动定量计算与目测法评估各类心肌时,瘢痕心肌数量方面结果一致性好;在计算冬眠心肌和正常心肌数量方面前者所得结果较后者相对大.当冬眠心肌数量少时(＜10%),目测法评估数量会低估冬眠心肌数量,而随着灌注/代谢不匹配(冬眠心肌)范围的增大,二者无明显差异,趋于一致.
展开▼
