Looking through the imaging perspective: the importance of imaging necrosis in glioma diagnosis and prognostic prediction

A real-world study

Abstract

Background. To investigate the diagnostic value of imaging necrosis (Im_necrosis) in grading, predict the genotype and prognosis of gliomas, and further assess tumor necrosis by dynamic contrast-enhanced MR perfusion imaging (DCE-MRI).

Patients and methods. We retrospectively included 150 patients (104 males, mean age: 46 years old) pathologically proved as adult diffuse gliomas and all diagnosis was based on the 2021 WHO CNS classification. The pathological necrosis (Pa_necrosis) and gene mutation information were collected. All patients underwent conventional and DCE-MRI examinations and had been followed until May 31, 2021. The Im_necrosis was determined by two experienced neuroradiologists. DCE-MRI derived metric maps have been post-processed, and the mean value of each metric in the tumor parenchyma, peritumoral and contralateral area were recorded.

Results. There was a strong degree of inter-observer agreement in defining Im_necrosis (Kappa=0.668, p<0.001) and a strong degree of agreement between Im_necrosis and Pa_necrosis (Kappa=0.767, p<0.001). Compared to low-grade gliomas, high-grade gliomas had more Im_necrosis (85.37%, p<0.001), and Im_necrosis significantly increased with the grade of gliomas increasing. And Im_necrosis was significantly more identified in IDH-wildtype, 1p19q-non-codeletion, and CDKN2A/B-homozygous-deletion gliomas. Using multivariate Cox regression analysis, Im_necrosis was an independent and unfavorable prognosis factor (Hazard Ratio=2.113, p=0.046) in gliomas. Additionally, ve in tumor parenchyma derived from DCE-MRI demonstrated the highest diagnostic efficiency in identifying Pa_necrosis and Im_necrosis with high specificity (83.3% and 91.9%, respectively).

Conclusions. Im_necrosis can provide supplementary evidence beyond Pa_necrosis in grading, predicting the genotype and prognosis of gliomas, and ve in tumor parenchyma can help to predict tumor necrosis with high specificity.