A method for reconstructing the arterial input function during helical CT: implications for myocardial perfusion distribution imaging

• Published in: Radiology Vol. 255; no. 2; pp. 396 - 404
• Main Authors: George, Richard T, Ichihara, Takashi, Lima, João A C, Lardo, Albert C
• Format: Journal Article
• Language: English
• Published: United States 01.05.2010
• Subjects: Adenosine - pharmacology; Animals; Area Under Curve; Contrast Media - pharmacokinetics; Coronary Angiography; Coronary Circulation; Coronary Stenosis - diagnostic imaging; Coronary Stenosis - physiopathology; Dogs; Linear Models; Radiographic Image Interpretation, Computer-Assisted - methods; Tomography, Spiral Computed - methods; Triiodobenzoic Acids - pharmacokinetics; Vasodilator Agents - pharmacology
• ISSN: 0033-8419; 1527-1315
• DOI: 10.1148/radiol.10081121

To determine if a multidetector computed tomographic (CT) image acquisition and analysis method can enable accurate measurement of the arterial input function (AIF) during first-pass adenosine stress helical multidetector CT angiography and to test the effect of using this method on the semiquantitative assessment of myocardial perfusion distribution. The animal care and use committee of Johns Hopkins University approved the use of all procedures. The AIF was reconstructed by using a combination of bolus-tracking and time-registered helical multidetector CT data. After the AIF reconstruction method was validated in healthy animals, coronary stenosis was induced in seven dogs and contrast material-enhanced multidetector CT was performed during adenosine infusion (0.14-0.21 mg per kilogram of body weight per minute). Myocardial attenuation density (AD) parameters normalized to portions of the AIF were compared with microsphere myocardial blood flow (MBF) measurements at linear regression analysis. There was no significant difference between the area under the curve (AUC) for dynamic multidetector CT-derived AIF (3108 + or - 1250 [standard deviation]) and that for combined bolus-tracking and time-registered multidetector helical CT-derived AIF (3086 + or - 941) (P = .90). When AIF analysis was applied to helical multidetector CT myocardial perfusion measurements, the correlation between MBF and mean myocardial AD normalized to the AUC for the entire AIF was significant (R(2) = 0.82, P <.001). Myocardial AD normalized to the AUC for the AIF measured during helical multidetector CT correlated best with MBF (R(2) = 0.86, P <.001). The combination of bolus tracking and time-registered helical imaging enables reconstruction of the AIF during multidetector CT perfusion imaging. The helical CT AIF can be used to improve the semiquantitative assessment of myocardial perfusion distribution.