RT Journal Article
SR Electronic
T1 Respective role of F18-FDOPA PET and perfusion MRI for staging brain tumors
JF Journal of Nuclear Medicine
JO J Nucl Med
FD Society of Nuclear Medicine
SP 1008
OP 1008
VO 50
IS supplement 2
A1 Nioche, Christophe
A1 Soret, Marine
A1 Gontier, Eric
A1 Lahutte, Marion
A1 Buvat, Irene
YR 2009
UL http://jnm.snmjournals.org/content/50/supplement_2/1008.abstract
AB 1008 Objectives Grading brain tumors from imaging is a challenge. We compared the accuracy of brain tumor grading based on Standardized Uptake Values (SUV) measured with FDOPA-PET and tumor relative cerebral blood volume (rCBV) measured with perfusion MRI. Methods 10 patients with oligodendroglioma or gliobastoma were considered, with 5 high grade (HG) and 5 low grade (LG) tumors based on histopathology. Each patient had a 40 min dynamic FDOPA PET (Philips Gemini TF) 3-min post injection and T1, T2-weighted EPI perfusion MRI (GE Signa HDx 3T). Two PET frames corresponding to [0-40min] and [2-20min] acquisition times were reconstructed. These reconstructed volumes were manually registered to the MR T1-weighted volume. Tumor volumes of interest (VOI) were drawn on the 40-min PET volume or on the MRI if the tumor was not visible on the PET. A VOI was also drawn in the white matter contralateral to the tumor. Using these VOIs, the mean tumor SUV was calculated on the [2-20 min] PET data (SUV20) and the rCBV was calculated using the MR perfusion data. Results The average (±1 sd) tumor volumes were 56±65 and 19±15 cm3 for HG and LG tumors (t-test, NS). The average SUV20 were 5.6±1.9 and 3.0±1.2 for HG and LG tumors respectively (p=0.02). The average rCBV were 2.0±0.9 and 1.3±0.5 for HG and LG tumors (p=0.1). 8/10 tumors were properly classified using an SUV20 threshold of 4 to separate LG and HG tumors, while 7/10 tumors were properly classified using a 1.7 rCBV threshold. A joint analysis of SUV20 and rCBV yielded only 1 misclassified tumor. Conclusions In our sample, high and low grade brain tumors were slightly better distinguished using SUV from FDOPA PET than using rCBV from perfusion MRI. Combining the two improved tumor classification.