The ceMRI images were compared to the histopathology in animal models of ischemic imagery. Reversible and irreversible ischemic injury was studied using a special preparation to produce both types of injury in the same heart. Obvious hyperenhancement was present in the infarcted zone, whereas no hyperenhancement was present in the area that suffered severe but reversible ischemic injury, despite myocardial stunning.

To determine whether the area of enhancement exactly matches the infarct size, the ceMRI was compared to the histopathology. In an animal who had sustained a 3-day old MI they found that the infarcted area with a very complex border was precisely matched to the area of enhancement on the ceMRI.

Their studies have extended to include infarct ages, from 4 hours to 8 weeks post MI, including reperfused and non-reperfused pathophysiologies. In each case they found that the spatial extent of hyperenhancement appears to match very closely with infarct size as determined by histopathology.

Acute and chronic infarcts are hyperenhanced in animals, a finding that would not have been expected a priori as the biochemical milieu is quite different. Previous studies in patients have shown that acute infarctions hyperenhance, however it had not been established whether chronic infarcts hyperenhance in humans.

CeMRI and cineMR were performed in patients with documented prior MI to address this question. Thirty-eight patients with peak CK-MB levels greater than 9.0 were enrolled and a ceMRI obtained more than 4 weeks, and in general several months, following their infarction. As a control, 14 patients with idiopathic non-ischemic cardiomyopathy were studied. All had significant ventricular dysfunction with ejection fractions less than 35%, but had normal coronary arteries as shown by their coronary angiograms.

The ischemic and non-ischemic images from both cohorts were randomized. Using a 14-segment model, three parameters were analyzed. One, regional hyperenhancement using the ceMRI images. Two, regional wall motion using the cineMR images. Three, the infarct related artery (IRA) by analyzing the coronary angiograms.

Sample images from three patients in the chronic MI group clearly show discrete hyperenhancement in the IRA-related territory. A subgroup was studied early after infarction. Sample images showed that when hyperenhancement was seen early after small or large infarctions, it was also seen late following infarction in the same myocardial territory. For the most part the study patients had small-sized infarcts as determined by cardiac enzyme levels. No areas of hyperenhancement were seen on ceMRI images in patients with non-ischemic cardiomyopathy despite significant ventricular dysfunction.

Of the 38 patients with documented chronic myocardial infarction, 37 had evidence of hyperenhancement. The one patient who did not had a peak CK-MB level of 10, just barely above the normal range. A significant discordance was found between wall motion and contrast-enhancement in the chronic MI patients. Of the 201 segments with hyperenhancement, more than 25% had normal wall motion. Of this 25%, predominantly it was only subendocardial hyperenhancement. Conversely, more than 30% of the segments with abnormal wall motion had no hyperenhancement. These two parameters index different pathophysiologic phenomena.

In 50 patients scheduled to undergo revascularization with resting wall motion abnormalities found on routine clinical workup, cine MRI for wall motion and ceMRI for regional viability assessment was performed. Cine MRI was repeated post-revascularization to document changes, if any, in wall motion. The age range of the study patients was 40-85 years; 21 patients had a history of prior MI, 34 underwent bypass surgery and 16 angioplasty. Follow-up cine MRI was done in 41 of the 50 patients at about 3 months following revascularization.

Image analysis

Interestingly, 40 of the 50 patients demonstrated some hyperenhancement, although only 21 of the patients had a history of MI. In total, 2083 segments were analyzed, 38% of which had a wall motion abnormality at baseline. Of this 38%, 43% had mild to moderate hyperkinesis, 38% severe hyperkinesis, and 20% akinesis or dyskinesis. After revascularization, 59% of segments with mild to moderate or severe hyperkinesis improved after revascularization, whereas about 30% of segments that were akinetic or dyskinetic improved.

Images from two patients were reviewed. In the first patient, the cine MR end diastolic still frame taken before revascularization showed impaired wall thickening, particularly in the anteroseptal region. However, the ceMRI performed before revascularization did not show any hyperenhancement in that territory. Therefore, based on the preliminary data from the animal studies, it would be predicted that contractility would improve after revascularization. Comparing the still frames taken after revascularization showed significant improvement in wall thickening. In the second patient, the baseline wall motion abnormality was in the anterolateral wall. However, the ceMR image showed a significant hyperenhancement of this territory before revascularization, predicting that this area should not improve in contractility after revascularization. Comparing the post-revascularization images showed significant impairment of wall thickening remains, and appeared to be to the same degree as pre-revascularization. Notably, the contours and location of the anatomical details were nearly the same before and after revascularization. An advantage of MR is true 3-D imaging that removes the registration issues found in non-MR imaging techniques.

Prediction of wall motion improvement

A subgroup analysis restricted to segments with severe hyperkinesis or akinesis at baseline showed the same inverse relation between the likelihood of wall motion improvement and the transmural extent of hyperenhancement.

A patient-by-patient analysis calculating the amount of the ventricle that was dysfunctional and non-hyperenhanced compared to the change in mean wall motion score after revascularization was performed. The analysis showed that patients with significant portions of dysfunctional ventricle and with less than 25% hyperenhancement were more likely to have wall motion improvement after revascularization (r=0.75; p< 0.001). A very good relation between this same group of patients and improvement in ejection fraction (r=0.70; p<0.005) was also found.

Delayed CeMRI is a new technique that can detect and delineate prior MI independent of wall motion and infarct age. Delayed ceMRI can predict quite accurately whether or not regions of myocardial dysfunction will improve following coronary revascularization.