Use of exponential diffusion imaging to determine the age of ischemic infarcts

Objective. Diffusion-weighted magnetic resonance imaging (DWI) detects acut e ischemic infarcts with high lesion conspicuity. Determination of infarct age is difficult on DWI alone because infarct signal intensity (SIinfarct) on DWI is influenced by T-2 properties ("T-2 shine-through"). Maps of the a pparent diffusion coefficient (ADC) reflect pure diffusion characteristics without T-2 effects but have low lesion conspicuity. Thus, in clinical prac tice, combined use of DWI and ADC maps is required. Exponential DWI (eDWI) is an innovative means of MRI-diffusion data analysis that merges the advan tages of DWI and ADC maps. The authors hypothesized that SIinfarct on eDWI would correlate with infarct age. The authors studied 114 consecutive patie nts who had 120 ischemic strokes with clearly determined onset times and wh o underwent echo-planar DWI. The eDWI were generated by dividing the signal intensity on DWI by that on the corresponding T-2 image on a pixel-by-pixe l basis. SIinfarct on eDWI was measured in the lesion core and expressed as a percentage of contralateral control tissue. On eDWI, relative SIinfarct changed significantly with infarct age (P < .0001). When patients were sort ed in infarct-age groups, no significant differences were found within the first 120 hours. However, for patients studied within 5 days, the mean rela tive SIinfarct was significantly higher compared with patients studied grea ter than or equal to8 days after stroke (P < .05). For all infarcts up to 5 days old, the eDWI signal intensity was higher than control tissue (hyperi ntense appearance). All infarcts > 10 days old had an eDWI signal intensity lower than control tissue (hypointense appearance). The authors concluded that the use of eDWI, as a single set of images, reliably differentiates ac ute infarcts (less than or equal to5 days old) from infarcts > 10 days old. This feature would be expected to be helpful when the distinction between acute and nonacute infarction cannot be determined on clinical grounds.