Hong Cheng, Pilgram Thomas K, Zhu Fang, Bae Kyongtae T
Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO, USA.
Acad Radiol. 2004 Mar;11(3):286-92. doi: 10.1016/s1076-6332(03)00714-1.
The objective of this study was to investigate the relationship between the coronary calcium mass and Agatston score measured on multidetector row computed tomography.
Eighty-three consecutive subjects (60 men and 23 women) referred for coronary screening were examined prospectively by electrocardiogram-triggered sequential multidetector row computed tomography scan (4 x 2.5-mm collimation). Their coronary calcium was quantitated by means of the Agatston scoring and mass method. The values of score and mass were transformed by taking the natural logarithm (ln(value + 1)) to reduce skewness. The relationship between the mass and score was analyzed with multiple regression analysis.
Fifty-one subjects had a total of 328 calcified lesions detected and measured. The relationship between the calcium mass and score in 51 subjects conformed well to a linear relationship (r2 = 0.96). When analyzed in the total of 328 lesions, the relationship had a good empiric fit with a nonlinear (quadratic) model (r2 = 0.96). The best-fit equation was ln(lesion mass + 1) = -0.87 + 0.67 x ln(lesion score + 1) + 0.10 x (ln(lesion score + 1) - 2.86)2. This relationship was consistent in different coronary vessels and at different heart levels (r2 = 0.96 - 0.99), although there was a higher image noise at lower heart levels (paired t-tests, P < .0001).
A nonlinear (quadratic) relationship existed strongly and consistently between coronary calcium mass and score, demonstrating a possible cross-sectional conversion between the two measurements.
本研究的目的是探讨在多排螺旋计算机断层扫描上测量的冠状动脉钙化质量与阿加斯顿评分之间的关系。
对83例连续接受冠状动脉筛查的受试者（60例男性和23例女性）进行前瞻性心电图触发的多排螺旋计算机断层扫描（4×2.5毫米准直）检查。通过阿加斯顿评分法和质量法对其冠状动脉钙化进行定量分析。对评分和质量值取自然对数（ln(值 + 1)）进行转换以减少偏态。采用多元回归分析分析质量与评分之间的关系。
51例受试者共检测到328个钙化病变并进行了测量。51例受试者的钙化质量与评分之间的关系与线性关系拟合良好（r2 = 0.96）。在总共328个病变中进行分析时，该关系与非线性（二次）模型具有良好的经验拟合（r2 = 0.96）。最佳拟合方程为ln(病变质量 + 1) = -0.87 + 0.67×ln(病变评分 + 1) + 0.10×(ln(病变评分 + 1) - 2.86)2。尽管在较低心脏层面图像噪声较高（配对t检验，P <.0001），但这种关系在不同冠状动脉血管和不同心脏层面是一致的（r2 = 0.96 - 0.99）。
冠状动脉钙化质量与评分之间存在强烈且一致的非线性（二次）关系，表明这两种测量方法之间可能存在横断面转换。
