体质量指数结合迭代重建算法在螺旋 CT冠状动脉成像中的应用

Application of iterative reconstruction technique on dose reduction at body mass index-based CT coronary angiography

目的：探讨 BMI 结合迭代重建算法在64排128层螺旋 CT 中降低冠状动脉 CT 成像辐射剂量的可行性。方法选取2013年12月—2014年8月山东省聊城市东昌府人民医院90例 CT 冠状动脉成像（CTCA）资料进行前瞻性研究,受检者均由临床医生拟诊为冠心病,且均无碘对比剂过敏史及其他不适合做 CTCA 检查的疾病。根据不同 BMI 将其分为3组,每组30例。 A 组,28 kg/ m2≤BMI≤31 kg/ m2；B 组,24 kg/ m2〈 BMI 〈28 kg/ m2；C 组,21 kg/ m2≤ BMI ≤24 kg/ m2。扫描参数：管电压均为120 kV,管电流 A、B、C 3组分别设为800、600、400 mAs；应用回顾性心电门控技术进行扫描,之后采用标准滤波反投影（FBP）、迭代重建算法 iDose4-3、iDose4-5及 iDose4-7分别对各组图像进行重建,从客观噪声及主观图像质量两方面进行评估。主观图像质量由2名具有10年以上诊断经验的医师采用冠状动脉15分段法进行评价,评价血管直径〉1.5 mm 的节段。评分采用4分制方法。客观噪声方面,3组重建图像中于升主动脉前方脂肪密度区域选择感兴趣区（面积约14.5 mm2）,测量其标准差值作为噪声值,比较噪声大小。扫描长度及辐射剂量指数由机器自动给出。结果主观评价方面,在 A、B、C 3组中均表现为：3分值以上等级者以 iDose4-3、iDose4-5重建组血管节段数最多,2分值以下等级者以 FBP 与 iDose4-7重建组血管节段数最多；FBP 和 iDose4-7分别与 iDose4-3、iDose4-5重建方式差异有统计学意义（P 值均〈0.05）,iDose4-3与 iDose4-5、FBP 与 iDose4-7差异均无统计学意义（P 值均〉0.05）。客观评价方面,与 FBP 相比,iDose4-3、iDose4-5、iDose4-7各重建组平均客观噪声分别降低29.96%、51.56%、69.76%。各组患者不同重建图像噪声值差异均有统计学意义（P 值均〈0.05）,相同重建方式之间噪声差异均无统计学意义（P 值均〉0.05）。与 A 组比较,B 组、C 组有效辐射剂量分别降低了25.83%、49.89%。结论迭代重建算法与 BMI 结合能有效降低辐射剂量,提高图像质量。

Objective To explore the feasibility of iterative reconstruction technique on dose reduction at body mass index （BMI）-based CT coronary angiography （CTCA） with retrospective gating in 128 slice spiral CT. Methods From December 2013 to August 2014, 90 patients in Dongchangfu People＇s Hospital who were suspected coronary artery disease were selected. All subjects were diagnosed as coronary heart disease. The selected patients had no history of allergy to iodine contrast agent and the other disease that is not suitable for CTCA test. According to the different BMI of CTCA subjects, the subjects were divided into 3 groups： group A, 28 kg/ m2 ≤BMI≤31 kg/ m2; group B, 24 kg/ m2 〈 BMI 〈 28 kg/ m2;group C, 21 kg/ m2≤BMI≤24 kg/ m2 . The number of patients in each group was 30. The tube current was set as 800 mAs, 600 mAs, and 400 mAs for group A, B and C, respectively. The tube voltages of three groups are all 120 kV. Four series of images were reconstructed with filtered back project（FBP）, iDose4-3, iDose4-5, iDose4-7 for each subject after used retrospectively ECG-gated CTCA technology to scan. The objective noise and subjective image quality were assessed. Two doctors who have 10 years experience in diagnosis assessed subjective image quality （the vascular diameter 〉 1. 5 mm） using 15 sections method and the four point method to evaluate coronary artery. The region of interest of three group＇s reconstructive images chosen from the fat density area in front of the ascending aorta which covers an area of about 14. 5 mm2 was measured as noise value and compared between them. Scan length and radiation dose index＆amp;nbsp;were given by the machine automatically. Results All the groups were shown that the maximum number of segments over 3 scores level was in the iDose4-3 and iDose4-5 reconstructive image groups, and the maximum number of segments under 2 score level was in the FBP and iDose4-5 reconstructive image groups. The difference was statistical significance between the group FBP or iDose4-7 and the group iDose4-3 or iDose4-5（all P values 〈 0. 05）, and there was no statistical difference between neither group iDose4-3 and iDose4-5 nor group FBP and iDose4-7（all P values 〉 0. 05）. Compared with the group FBP, the average objective noise of the group iDose4-3, iDose4-5, and iDose4-7 reduced 29. 96% , 51. 56% and 69. 76% , respectively. The difference of noise value was all statistical significance between the groups of different patients reconstruction image （all P values 〈 0. 05）. And all of the difference was no statistical significance between the same reconstruction images （ all P values 〉 0. 05）. Compared with group A, the effective radiation dose of group B and group C was reduced by 25. 83% and 25. 83% , respectively. Conclusions The iterative reconstruction algorithm can effectively reduce the noise, combined with a body mass index can effectively reduce the radiation dose, improve the quality of image.