Value of multimodal MRI in evaluation of tissue factor expression in glioblastoma

Abstract

Objective To explore the correlation between multimodal magnetic resonance imaging (MRI) and expression of tissue factor (TF) in glioblastoma (GBM), and investigate the values of MRI parameters as imaging markers for TF. Methods The preoperative routine and functional MRI data of 60 patients with GBM confirmed by surgery and pathology in our hospital from August 2014 to November 2018 were retrospectively collected and analyzed for the parameters, including necrosis ratio, surface regularity, enhanced area volume, edema area volume, volume transfer constant Ktrans, relative cerebral blood volume (rCBV) and relative cerebral blood flow (rCBF) by post-processing. The expression of TF in the GBM tissue was detected with immunohistochemical staining, and then the correlation between above each parameter and TF expression was analyzed. Receiver operating characteristic (ROC) curve analysis was used to obtain the optimal threshold values, sensitivity and specificity of various imaging parameters in differentiation between GBM with TF of high expression and of low expression. Results The necrosis ratio, and Ktrans, rCBV and rCBF values were positively correlated with TF expression level (r=0.665, r=0.631, r=0.661, r=0.619, all P < 0.001). The necrosis ratio and Ktrans, rCBV and rCBF values were significantly higher in the highly-expressed TF group than the lowly-expressed TF group (P < 0.01). ROC analysis showed that all these MRI parameters presented TF level well, with area under curve (AUC) of rCBF reaching 0.907 (0.889, 0.857), that of necrosis ratio 0.869 (0.846, 0.676), that of Ktrans 0.854 (0.813, 0.833), and that of rCBV 0.804 (0.944, 0.571). The combination of conventional magnetic resonance and perfusion magnetic resonance parameters showed even better presention for TF expression level, with the AUC of necrosis ratio combined with Ktrans 0.982 (1, 0.875), necrosis ratio with rCBV 0.939 (0.889, 0.905), and necrosis ratio with rCBF 0.971 (0.944, 0.857). Conclusion Various magnetic resonance parameters (necrosis ratio, Ktrans, rCBV and rCBF) can better reflect TF expression in GBM patients, and can be used as imaging markers for TF expression in GBM.

Keywords

• magnetic resonance imaging
• tissue factor
• glioblastoma
• imaging marker