Glioma assessment using quantitative blood volume maps generated by T1-weighted dynamic contrast-enhanced magnetic resonance imaging: a receiver operating characteristic study

Abstract

Purpose: To investigate the use of blood volume maps in the non-invasive separation of glioma grades.

Material and methods: T1-weighted quantitative dynamic contrast-enhanced magnetic resonance imaging was used to quantify the fractional intratumoral blood volume of 41 gliomas (World Health Organization (WHO) grades II-IV). Two methods, mean fractional intratumoral blood volume determination and a system based on thresholds for extracting the tumor pixels with the highest vascularization from the blood volume maps, were investigated by means of receiver operating characteristic (ROC) analysis. The thresholds were adjusted using the ROC curve area calculated using the trapezoid method.

Results: The ability to separate grade II (WHO) gliomas from grades III-IV was nearly the same for both methods (ROC curve area 0.941 (threshold) versus 0.932 (mean value)) and significantly greater than the ability to separate grade IV (WHO) gliomas from grades II-III (ROC curve area 0.792 (threshold) versus 0.787 (mean value)). The best correspondence with WHO glioma grading was achieved using thresholds corresponding to only the 5.2% of tumor voxels with the largest blood volume for separating grade II gliomas and 4% for separating grade IV gliomas.

Conclusion: Use of the optimized threshold resulted in matching with the WHO grading system in 74% of cases.