Zhang J J, Kim S M, Andrews D, Intenzo C M
Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania 19107, USA.
Clin Nucl Med. 1998 Oct;23(10):691-4. doi: 10.1097/00003072-199810000-00009.
The objective of this study was to determine which semi-quantitative thallium-201 brain tumor index correlated best with clinical outcome. Fifty-two patients had Tl-201 brain single-photon emission computer tomography (SPECT) performed for the evaluation of recurrent brain tumor. A semi-quantitative tumor index was calculated utilizing four currently available techniques: 1) the ratio of average counts of the region of interest (ROI) in the lesion area and its mirror image in normal brain tissue, 2) the ratio of maximum counts of the region of interest in the lesion area and its average counts in the mirror image, 3) the ratio of maximum counts of the region of interest in the lesion area and its mirror image and 4) the ratio of the average counts of the region of interest in the lesion area and the region of interest of the scalp. A two-tailed Student's t test was performed. The mean value plus one standard deviation was used to assess the sensitivity and specificity correlated with clinical follow-up evaluation, defined as inactive or active at the time of brain SPECT by the referring neurosurgeons. Results of correlation coefficient, P value, mean, standard deviation, sensitivity, and specificity of each index using cutoff values were obtained. For those patients with recurrence or residual malignant tumor after therapy, the maximum count ratio index correlated best with clinical outcome. It provided a sensitivity of 92% and specificity of 88% in differentiating active from inactive brain tumors.