Reply

I would like to respond to several issues raised by the letter of Jerosch-Herold et al. First, I do not purport that a lower temporal resolution of an image every six heart beats used in our study (1) confers an advantage. This is a clear compromise between having more slices and less data points on the first-pass transit curves, which was necessitated by the hardware constraints we had at the inception of this study. Although the use of an inversion-recovery sequence results in relatively prolonged image acquisition times even with the fastest magnetic resonance imaging (MRI) scanners, the images are stable and of good quality for analysis. This is in contrast to the saturation-recovery sequences used by Jerosch-Herold (2), which may allow more of the heart to be imaged in less time but in which the images are often of poor quality and subject to artefacts. However, the slower injection technique used in our study is an advantage over the power injections into the right subclavian vein reported in other studies (2); this is because a relatively invasive subclavian line is needed, which is less attractive for patient and operator. Also, a power injector is required, which is safe with magnetic fields and correspondingly expensive, whereas our technique can be administered manually through a peripheral vein.

Second, the question of whether the extraction efficiency (E) remains constant or varies with myocardial flow (F) is controversial and as yet remains a subject for further research. In support of using the EF product (K₁) for estimating flow, a recent study by Vallee et al. (3) in a canine occluded coronary artery model demonstrated that F measured with microspheres had a linear fit to K₁ for Gd-DTPA, (r = 0.88). However, as the behavior of E is uncertain at differing flow rates, the term “myocardial perfusion reserve index” was used in our study for the ratio of K₁ during stress to K₁ at rest.

Finally, the first-pass studies used by Jerosch-Herold’s team have yielded impressive results when compared with coronary flow reserve measurements using Doppler wires (2). In truth, the best comparison would be to compare the nutritive perfusion measurements obtained with positron emission tomography (PET), because coronary flow reserve and regional nutritive perfusion can differ. Nevertheless, although technically demanding, their first-pass technique looks promising, particularly if the imaging artefact problems are overcome and validation against PET can be achieved. The Kety model approach needs further research, but it may prove to have some advantages for the reasons already outlined.

• American College of Cardiology