Percent Change of Perfusion Skewness and Kurtosis: A Potential Imaging Biomarker for Early Treatment Response in Patients with Newly Diagnosed Glioblastomas

Purpose: To test the predictive value of skewness and kurtosis changes of normalized cerebral blood volume (nCBV) during the early treatment period for differentiating early tumor progression from pseudoprogression in patients with newly diagnosed glioblastomas. Materials and Methods: The institutional review board approved this retrospective study. The authors assessed 135 patients with newly diagnosed glioblastomas who underwent concurrent chemotherapy and radiation therapy (CCRT) after surgical resection. Patients who developed new or enlarged contrast material-enhanced lesions after CCRT were assessed by means of conventional and perfusion magnetic resonance (MR) imaging. The percent change of skewness and kurtosis on nCBV histograms between the first and second post-CCRT follow-up were classified into four categories. Independent predictors of early tumor progression were determined by means of logistic regression analysis. Results: Of 135 patients, 79 had new or enlarged contrast-enhanced lesions after CCRT, subsequently classified as early tumor progression (n = 42, 53.2%) and pseudoprogression (n = 37, 46.8%). Pseudoprogression was observed in 23 of 24 (95.8%) patients in category 1, 10 of 15 (66.7%) in category 2, four of 20 (20.0%) in category 3, and 0 of 20 (0%) in category 4 (chi(2) test, P < .0001). The histographic pattern of nCBV was the best independent predictor (odds ratio, 3.51; P = .0032) for early tumor progression, rather than each percent change of skewness or kurtosis; the histographic pattern of nCBV represented the largest area under the receiver operating characteristic curve (0.934; 95% confidence interval: 0.855, 0.977), with a sensitivity of 85.7% and a specificity of 89.2%. Conclusion: The percent change of skewness and kurtosis of nCBV may be a potential imaging biomarker for early treatment response in patients with newly diagnosed glioblastomas. (C) RSNA, 2012