T2' imaging predicts infarct growth beyond the acute diffusion-weighted imaging lesion in acute stroke.

PURPOSE To show that measurement of the transverse relaxation time that characterizes signal loss caused by local susceptibilities (T2') is sensitive to an increased deoxyhemoglobin concentration in the brain, indicating tissue at risk for infarction. MATERIALS AND METHODS The study was approved by the local institutional review board; patients or their guardians provided informed consent. Magnetic resonance (MR) imaging was performed within 6 hours of symptom onset and again 1-11 days thereafter in 100 consecutive stroke patients, all of whom received intravenous thrombolytic therapy (mean age, 67 years). The MR imaging protocol included diffusion- and perfusion-weighted imaging for determination of apparent diffusion coefficient (ADC) and time to peak (TTP), along with quantitative T2 and T2* imaging. T2' maps were calculated and visually compared with ADC and TTP lesions by two independent observers. RESULTS A T2'>ADC mismatch was observed by reader 1 in 73 (73%) of 100 patients, and by reader 2 in 65 (65%) patients. Respective sensitivities of T2'>ADC and of TTP>ADC mismatches for later infarct growth were 0.87 and 0.98 for reader 1 and 0.78 and 0.98 for reader 2, with respective specificities of 0.42 and 0.04 for reader 1 and 0.46 and 0.04 for reader 2. The odds ratios for infarct growth in the presence of a T2'>ADC mismatch were 4.59 (reader 1 P = .002) and 3.10 (reader 2 P = .012), while the odds ratios for TTP>ADC mismatch were 2.22 (reader 1 P = .606) and 1.73 (reader 2 P > .999). CONCLUSION The presence of a T2'>ADC mismatch is a more specific predictor of infarct growth than is TTP>ADC mismatch and hence may be of clinical value in patient selection for acute stroke therapies in the future.