Hye Jin Baek, Presenter: Nothing to Disclose

Ho Sung Kim, Abstract Co-Author: Nothing to Disclose

Choong Gon Choi MD, Abstract Co-Author: Nothing to Disclose

Ha Young Lee MD, Abstract Co-Author: Nothing to Disclose

Jee Won Park MD, Abstract Co-Author: Nothing to Disclose

Hyun Kyung Lim MD, Abstract Co-Author: Nothing to Disclose

To semiquantitatively differentiate pseudoprogression from early tumor recurrence using both skewness and kurtosis derived from MR perfusion histogram.

135 patients with newly diagnosed glioblatomas underwent chemoradiotherapy after surgical resection. New or enlarged contrast-enhancing lesions after chemoradiotherapy (CCRT) were prospectively assessed by dynamic susceptibility contrast-enhanced perfusion MR imaging. According to their histogram pattern of relative cerebral blood volume (CBV), study patients were classified into four groups: group 1, positive skewness and leptokurtic; group 2, positive skewness and platykurtic; group 3, negative skewness and leptokurtic; group 4, negative skewness and platykurtic. All patients were regularly followed up at an interval of 2 or 3 months by conventional and perfusion MR images. Predictors for pseudoprogression were determined by logistic regression analysis.

After CCRT, new or enlarged contrast-enhancing lesions were found in 89 of 135 patients (65.9%), which were subsequently classified as pseudoprogression (39 patients, 28.9%) and early tumor recurrence (50 patients, 37.0%). Pseudoprogression were developed in 18 (69.2%) of 26 patients of group 1, in 7 (38.9%) of 18 patients of group 2, in 9 (36.0%) of 25 patients of group 3, and in 5 (25%) of 20 patients of group 4, respectively (Chi-square test, P < 0.0148). The receiver operating characteristic (ROC) curve-derived threshold value (groups 1 and 2) differentiated pseudoprogression from early tumor recurrence with a sensitivity of 82.0 and a specificity of 89.7. Histogram pattern (skewness and kurtosis) of relative CBV was a independent predictor for pseudoprogression (Odds ratio 1.84 [1.23 – 2.76], P = 0.0020).

Semiquantitative histogram pattern analysis of relative CBV with both skewness and kurtosis may be a promising approach to differentiate pseudoprogression from early tumor recurrence in patients with posttreatment glioblastomas.

Skewness and kurtosis derived from relative CBV histogram can be used to differentiate pseudoprogression from early tumor recurrence and may affect therapeutic plans of posttreatment glioblastomas.

Baek, H, Kim, H, Choi, C, Lee, H, Park, J, Lim, H, Utility of Skewness and Kurtosis Derived from MR Perfusion Histogram to Predict Pseudoprogression in Patients with Glioblastomas.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11015019.html