Introduction:
Perfusion computed tomography (PCT) allows to quantitatively assess haemodynamic characteristics of brain tissue. We investigated if different brain tumor types can be distinguished from each other using Patlak analysis of PCT data.

Methods:
PCT data from 43 patients with brain tumours were analysed with a commercial implementation of the Patlak method. Four patients had low-grade glioma (WHO II), 31 patients had glioblastoma (WHO IV) and eight patients had intracerebral lymphoma. Tumour regions of interest (ROIs) were drawn in a morphological image and automatically transferred to maps of cerebral blood flow (CBF), cerebral blood volume (CBV) and permeability (K Trans). Mean values were calculated, group differences were tested using Wilcoxon and Mann Whitney U-tests.

Results:
In comparison with normal parenchyma, low-grade gliomas showed no significant difference of perfusion parameters (p > 0.05) , whereas high-grade gliomas demonstrated significantly higher values (p < 0.0001 for K Trans, p < 0.0001 for CBV and p = 0.0002 for CBF). Lymphomas displayed significantly increased mean KTrans values compared with unaffected cerebral parenchyma (p = 0.0078) but no elevation of CBV. High-grade gliomas show significant higher CBV values than lymphomas (p = 0.0078).

Discussion:
PCT allows to reliably classify gliomas and lymphomas based on quantitative measurements of CBV and K Trans.