Diagnostic performance of regional DTI-derived tensor metrics in glioblastoma multiforme: simultaneous evaluation of p, q, L, Cl, Cp, Cs, RA, RD, AD, mean diffusivity and fractional anisotropy

Authors:
Publisher website
Google Scholar
Abstract
Almost a dozen diffusion tensor-imaging (DTI) variables have been used to evaluate brain tumours with scarce information about their diagnostic ability. We aimed to perform a comprehensive evaluation of tensor metrics reported in the last decade. Retrospective case control study performed in 14 patients with glioblastoma multiforme (GBM) and 28 controls. Conventional brain MR sequences and image postprocessing of DTI allowed the calculation of: MD, FA, p, q, L, Cl, Cp, Cs, RA, RD and AD, classified into five regions: normal appearance white matter (NAWM), immediate and distant oedema, enhancing rim and cystic cavity. ANOVA and AUROC analyses were performed. ANOVA depicted a significant difference among all metrics (p < 0.05). RA had the highest performance in the NAWM and cystic cavity; immediate and distant zones of oedema were best diagnosed by RD and Cp respectively; q was the best biomarker of the enhancing rim zone; p < 0.001 for all metrics. FA and MD, accepted biomarkers of brain injury, were surpassed by other metrics. RA, together with Cs, Cl and CP, might be the new leaders in the evaluation of brain tumours. DTI tensor metrics depict different clinical applicability at each tumour region. • DTI-derived tensor metrics can characterise the regional impairment of intraaxial brain tumours. • A performance evaluation of new diagnostic tests should follow the STARD initiative. • Each tumour region in GBM is detected by a different tensor metric. • FA and MD are not the best biomarkers of tumour infiltration. • The diagnostic performance of some tensor metrics allows them to be used interchangeably.

This publication has 39 references indexed in Scilit: