ABSTRACT
PURPOSE: Different centers and vendors use different sequences and contrast agent application schemes for MR myocardial perfusion imaging. The purpose of this study was to evaluate the role of different sequences, dosages, and injection speeds of contrast media for semiquantitative MR-perfusion assessment. METHODS: In a pilot study with 58 consecutive patients three of the most commonly used sequences for MR myocardial perfusion imaging (T1-GrE, GrE-EPI or SSFP) were compared to each other in terms of peak myocardial enhancement and image quality. For the main part of the study dynamic first pass MR perfusion imaging (Philips Intera CV, Best, Tthe Netherlands) was performed in 24 patients using the most favorable sequence from the pilot study (SSFP) after peripheral i.v. administration of Gd-BOPTA during adenosine stress. Two doses (0.05 mmol/kg bw and 0.025 mmol/kg bw) and four different injection speeds (8, 4, 3, 2 ml/s) were used. Signal intensity time curves were determined in the LV and myocardial segments supplied by normal coronary arteries and correlation between LV and myocardial upslope as well as peak enhancement were noted. RESULTS: The SSFP-sequence showed a higher peak enhancement when using 0.05 mmol/kg bw of Gd-BOPTA and a superior image quality for both dosage regimen compared with the other sequences and was consequently applied for the main study. A significant correlation was found between the upslopes in the LV and the myocardium (r square = 0.85, p < 0.001). However, LV and myocardial upslopes were largely independent of the dosage. Myocardial upslope was significantly slower at an injection rate of 2 ml/s compared to 3 and 4 ml/s. Higher Gd-doses led to significantly higher enhancement (p < 0.001). CONCLUSION: In healthy myocardial segments, the myocardial upslope is mainly determined from the LV upslope. Both myocardial enhancement and upslope are largely independent from the injection rate of a contrast agent bolus as long as the injection speed is not below 3 ml/s. Myocardial enhancement, however, is dose dependent. Thus, a simple correction for LV upslope allows to normalize a wide variety of input parameters. Differences of myocardial upslope or peak signal intensity after correction should be mainly dependent on blood flow.
