Optimized Preload Leakage-Correction Methods to Improve the Diagnostic Accuracy of Dynamic Susceptibility-Weighted Contrast-Enhanced Perfusion MR Imaging in Posttreatment Gliomas

Graph shows cerebral blood volume (CBV) calculation methods. A, Uncorrected CBV calculation by trapezoidal integration without baseline subtraction (BLS). B, BLS integration excludes a triangular area (red) that estimates the T2/T2*WI residual effects. The remaining area under the curve (AUC) (blue lines) estimates the corrected CBV.

Graph shows CBV over an increasing PLD amount from normal-appearing gray matter (GM) and white matter (WM) regions of interest. CBV values are calculated by using BLS integration.

Graph shows distribution of rCBV values for tumor (red circles) and posttreatment radiation-effect (PTRE) (blue triangles) groups over the increasing PLD amounts without BLS (A) and with BLS (B). Gray boxes depict the range of overlapping rCBV values between groups under each condition. Asterisks in B denote conditions that demonstrated statistically significant increases in the rCBV receiver operator characteristic (ROC) AUC compared with the control acquisition (dotted circle). These are observed with BLS and PLD amounts of 0.05 (P = .045), 0.1 (P = .01), and 0.25 (P = .35) mmol/kg, with incubation times of 3, 6, and 15 minutes, respectively.

Graph shows the effects of PLD amount and the presence of BLS correction on DSC test accuracy (represented by ROC AUC). Asterisks (with P values) denote test conditions that demonstrate a statistically significant increase in rCBV ROC AUC compared with the control acquisition (red circle).