Parametric imaging and statistical mapping of brain tumor in Ga-68 EDTA dynamic PET studies

To improve the reliability and sensitivity of quantitative analysis in the study and evaluation of brain tumor using Ga-68 EDTA dynamic PET, a linear parametric imaging algorithm was developed in this study for estimation of both distribution volume (DV) and blood brain barrier permeability. F statistics was used for separating tumor from normal tissue. A two-compartmental model was used to describe the tracer kinetics. The operational equations: C_pet=((K₁+k₂)V_p)∫C_pds-k₂∫C_petds+V_pC_p and ∫C_pet=(DV+V_p)∫C_pds-(1/k₂)C_pet+(V_p/k₂)C_p, were used to estimate K₁ (permeability) and DV(=K₁/k₂), respectively. A reliable and robust linear regression with spatial constraint parametric imaging algorithm was developed to generate the K₁ and DV images. Pixel-wise F statistics with 2 and k-2 degree of freedom was calculated as: F=(((k-2)k/(2(k²-1)))D² with D²=(x-μ)´S^-1(x-μ), where the sample is from two dimensional sample space {(K₁, DV)} of reference regions in normal brain tissue pixels, the sample size k is the number of pixels within the normal reference ROIs, μ and S are, respectively, the sample mean vector (K₁, DV) and covariance matrix. By setting critical a values at 0.2, 0.05, and 0.001, statistical significance level images were generated, and its pixel values can be, 0 if F0.2, 1 if F_0.2=0.05, 2 if F_0.05=0.01, and 3 if F_0.01<=F. The methods were applied to eleven brain tumor Ga-68 EDAT dynamic PET studies. Results shown that the DV, K₁, and F images are of good image quality. A highly correlated linear relationship (R²>0.92) was found between the values of K₁ and DV estimated by model fitting to ROI time activity curve and ones calculated from parametric images. The method for generating K₁, DV, F, and significance level images is of high computation efficiency and is easy to be implemented. The statistical model developed in the current study provided a tool to integrate the multi-dimensional physiological information. The normal reference region method and the integration of multi-phy- siological images may improve the sensitivity and specificity of brain tumor detection and evaluation of treatment.