The development of anti-angiogenic therapies for tumors has led to a demand for imaging-based surrogate markers of the angiogenic process. The utility of such markers is highly dependent on their test-retest reproducibility. This paper presents a formal assessment of the reproducibility of measurements of relative blood volume (rBV), normalized rBV (rBVnorm), and vascular tortuosity as estimated by measurement of relative recirculation (rR). The study was conducted in 11 patients with glioma who were scanned on two occasions 36-56 hours apart. The observed reliability estimates were used to calculate 95% confidence limits for detection of differences between groups and for changes in individual cases. The results show that measurement of rBV or rBVnorm in consecutive studies is statistically capable of reliably detecting changes in excess of 15% in between group studies and 25% in individual patients. Measurement of vascular tortuosity using is less reproducible but is able to confidently identify changes in excess of 30% in group studies and 35% in individuals.
Copyright 2001 Wiley-Liss, Inc.