Accumulation of Ultrasmall Superparamagnetic Particles of Iron Oxide in Human Atherosclerotic Plaques
To the Editor:
We read with interest the recent article by Kooi et al,1 reporting their experience with ultrasmall superparamagnetic iron oxide (USPIO)–enhanced magnetic resonance (MR) imaging of carotid atheroma. This exploratory study holds promise for contrast-enhanced MR imaging of inflammatory lesions. We do express some reservations, however, regarding the authors’ methodology, conclusions, and the quality of the MR images on which they base their findings.
It is unclear exactly how many of the patients given USPIO had a positive effect (ie, area of new/increased signal-intensity reduction) and in how many MR images this was evident. The authors, having made an effort to correlate USPIO accumulation with ruptured/rupture-prone histological sections, failed to say whether this translated to a signal loss on corresponding MR images, which is an important consideration before advocating the use of this methodology to identify inflamed carotid plaques. The authors suggested that in one patient, no USPIOs had accumulated because of a short infusion-surgery interval of 3 days; yet they conclude that USPIO-induced signal intensity reduction was optimal at 24 hours. Additionally, it is difficult to determine why one patient was excluded, because the infusion-surgery interval was too long as USPIOs were still present in the plaque nonetheless.
The authors’ demonstration of USPIOs in the plaque, and their colocalization with macrophages, is commendable. We are concerned, however, that the authors have not shown the histological section corresponding to the axial MR images in which they claim to visualize the reduction in signal intensity (the images do not appear to be corresponding). Admittedly, the authors note that they were unable to adequately co-register histological sections with MR images, and they may have elected to provide a representative image. In light of the inadequacy of image correspondence, one must interpret the outcome of the relative signal intensity analysis with caution. The use of vessel wall quadrants as the user-defined area from which relative signal intensity measurements were taken is inherently susceptible to measurement error, and this is made worse by poor quality images. Perhaps greater weight could be added to their conclusions if the authors attempted to measure the agreement in user-defined areas between pre- and postinfusion images.
Overall, we are encouraged that the authors found no difficulty in the usage of this USPIO agent and agree that the way has been paved for larger studies evaluating the utility of USPIO-enhanced plaque imaging.
Kooi ME, Cappendijk VC, Cleutjens KB, et al. Accumulation of ultrasmall superparamagnetic particles of iron oxide in human atherosclerotic plaques can be detected by in vivo magnetic resonance imaging. Circulation. 2003; 107: 2453–2458.
We thank Trivedi et al for their comments on our recent article; however, we discussed most of their criticism in our paper. Our article concerns a feasibility study in which analysis was performed on a quadrant level in magnetic resonance imaging (MRI) and histological slices. Trivedi et al suggest a study on the patient level combined with a correlation between MRI and histological slices. On this point, we described the inability to match MRI with histology as a limitation and indicated that improvement of signal-to-noise ratio is needed for accurate detection of ultrasmall superparamagnetic particles of iron oxide (USPIO) accumulation in patients. After this first feasibility study, larger studies are warranted to investigate whether signal decreases correlate with the number of macrophages in corresponding histological sections and whether a signal decrease in the MRI corresponds with a vulnerable patient.
We did not wish to suggest that the short infusion-surgery interval of 3 days has anything to do with the lack of USPIO uptake in one patient. Another 3 patients with an interval of 3 days or less did show USPIO uptake. Apparently our formulation confused the respondents.
Ideally, the infusion-surgery interval should be constant for all patients. On average, this interval was 4.7±3.0 days. One patient, however, had an interval of 21 days and was therefore excluded.
Trivedi et al are correct to state that the presented MR images do not correspond to the presented histological section. Indeed, the corresponding histological section was not presented because of space limitations. However, for this particular section, we did not experience matching difficulties. The carotid artery segment contained the carotid bifurcation that could be used as a point of reference; this slice was adjacent to the bifurcation. Additionally, the shape of the histological section corresponded with the presented MR images.
Furthermore, Trivedi et al suggest that in light of the inadequacy of image correspondence, the outcome of the relative signal intensity should be interpreted with caution. However, we did not experience difficulties matching the pre- and post-USPIO MR images. The operator-defined regions of interest were positioned anatomically identical on the pre- and post-USPIO MR images. Therefore, we disagree with Trivedi et al that this method is susceptible to measurement error. Furthermore, in our paper, we already acknowledged the relatively low signal-to-noise ratio of T2*w gradient-echo images and gave some suggestions for improvements in future studies.