Prediction of Hemorrhagic Transformation after Recanalization Therapy Using T2*- Permeability Magnetic Resonance Imaging Oh Young Bang, MD, PhD, 1,2 Brian H. Buck, MD, 1 Jeffrey L. Saver, MD, 1 Jeffry R. Alger, PhD, 1 Sa Rah Yoon, MD, 3 Sidney Starkman, MD, 1,4 Bruce Ovbiagele, MD, 1 Doojin Kim, MD, 1 Latisha K. Ali, MD, 1 Nerses Sanossian, MD, 1 Reza Jahan, MD, 3 Gary R. Duckwiler, MD, 3 Fernando Vin ˜uela, MD, 3 Noriko Salamon, MD, 3 J. Pablo Villablanca, MD, 3 and David S. Liebeskind, MD 1 Objective: Predicting hemorrhagic transformation (HT) is critical in the setting of recanalization therapy for acute stroke. Dedicated magnetic resonance imaging (MRI) sequences for detection of increased blood–brain barrier (BBB) permeability recently have been developed. We evaluated the ability of a novel MRI permeability technique to detect baseline derangements predictive of various forms of HT after recanalization therapy. Methods: We retrospectively analyzed the clinical and pretreatment MRI data on patients undergoing recanalization therapy for acute cerebral ischemia at a university medical center from January 2004 to November 2006. Pretreatment MRI permeability images derived from perfusion source data were compared with posttreatment imaging to evaluate whether baseline BBB per- meability derangements may predict HT after recanalization therapy. The use of a novel permeability technique to illustrate BBB derangements was based on the detection of decreased signal intensity at later time points in perfusion MRI acquisition, signi- fying continued local accumulation of contrast caused by leakage. Results: Among 32 patients, some degree of HT occurred in 12. Permeability image abnormalities at baseline were present in 7 of 12 patients with HT and none of the 20 patients without HT on follow-up images. The sensitivity of permeability abnormality for parenchymal hematoma was 83%. False-negative findings were noted in five cases, most commonly asymptom- atic or minor HT after mechanical clot retrieval. Interpretation: Permeability images derived from pretreatment perfusion MRI source data may identify patients at risk for HT with high specificity. Our preliminary demonstration of permeability imaging based on standard perfusion data for prediction of hemorrhage merits further study with dedicated MRI BBB permeability acquisitions and multicenter validation. Ann Neurol 2007;62:170 –176 Hemorrhagic transformation (HT) is a feared compli- cation of recanalization therapy for acute ischemic stroke. Although HT may occur spontaneously, 1 ther- apeutic approaches for recanalization of a proximal ar- terial occlusion increase this risk. Several studies have already identified clinical and laboratory predictors of HT after thrombolytic treatment for stroke. 2–5 Increasing use of multimodal magnetic resonance imaging (MRI) offers an additional source of data re- garding tissue status and potential propensity for HT. 6,7 Perfusion-diffusion characterization of the isch- emic territory may provide valuable information for the potential identification of patients at increased risk for HT after recanalization therapy. 8 –10 Additional MRI parameters have variably been reported in association with an increased risk for HT after thrombolysis: leu- koaraiosis, 11 prior cerebral microbleeds visualized with T2*-weighted MRI sequences, 10 early parenchymal en- hancement, 12 and early cerebrospinal fluid hyperinten- sity. 13 MRI permeability imaging is a promising potential marker to identify patients with an increased propen- sity to HT of infarction. In recent years, dedicated MRI acquisitions have been used to identify blood– brain barrier (BBB) permeability derangements sur- rounding brain tumors, for tumor grading. 14 –16 The From the 1 Department of Neurology, University of California, Los Angeles, Los Angeles, CA; 2 Department of Neurology, Samsung Medical Center, Sungkyunkwan University, Seoul, South Korea; and Departments of 3 Radiology and 4 Emergency Medicine, Univer- sity of California, Los Angeles, Los Angeles, CA. Received Feb 23, 2007, and in revised form May 15. Accepted for publication Jun 1, 2007. This article includes supplementary materials available via the Inter- net at http://www.interscience.wiley.com/jpages/0364-5134/supp- mat Published online August 7, 2007 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.21174 Address correspondence to Dr Liebeskind, UCLA Stroke Center, 710 Westwood Plaza, Los Angeles, CA 90095. E-mail: davidliebeskind@yahoo.com 170 © 2007 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services