PubMed2000
BACKGROUND AND PURPOSE: Reconstruction of first-pass bolus information to derive regional cerebral blood volume (rCBV) maps is commonly performed in many centers; however, various protocols with different doses of paramagnetic contrast injections have been reported. We evaluated the dose dependency of rCBV maps in a brain tumor population by using three different doses of gadodiamide injection to evaluate their diagnostic accuracy in blinded reader sessions. METHODS: Eighty-three patients with intraaxial brain tumors (72 gliomas) were studied at three centers and randomized to receive a bolus injection of 0.1, 0.2, or 0.3 mmol/kg per body weight of gadodiamide. rCBV maps were generated from T2*-weighted gradient-echo echoplanar sequences at 1.5 T. Data processing was performed according to the indicator dilution theory. RESULTS: The mean contrast-to-noise ratio (CNR) was significantly different between gadodiamide doses of 0.1 and 0.2 mmol/kg (CNR = 8.7 and 15.7) and between 0.1 and 0.3 mmol/kg (CNR = 17.7). No significant difference was found between doses of 0.2 and 0.3 mmol/kg. Sensitivity for the differentiation of benign and malignant brain tumors was 80%, 95%, and 91%, and specificity was 45%, 54%, and 43% by blinded readings at 0.1, 0.2, and 0.3 mmol/ kg, respectively, as compared with histologic findings. Nonblinded readings had a sensitivity of 83%, 100%, and 90% and a specificity of 82%, 100%, and 73% at 0.1, 0.2, and 0.3 mmol/kg, respectively. CONCLUSION: A dose of 0.2 mmol/kg of gadodiamide is recommended for reconstruction of rCBV maps if data are acquired with the T2*-weighted protocol described.