Parametric image of regional bone metabolism using F-18 PET using a multiple linear regression analysis method

Dynamic PET data can be quantified several biological parameter using Nonlinear Least Square (NLS) and compartment modeling. However, NLS is inappropriate to voxel based analysis because of initial problem and excessive calculation time. Graphical method is able to reduce the computation time, but sensitive to noise and time duration. In this study, we applied a novel multiple linear regression analysis (MLAIR) method for the estimator of fluoride bone influx rate (K_i) compared with patlak graphical method in minipig using [¹⁸F]Fluoride PET. In ROI analysis, estimated K_i values using MLAIR and Gjedde-Patlak graphical analysis (PGA) method was slightly higher than those of NLS, but the results of MLAIR and PGA were equivalent. However, Patlak slopes (K_i) were changed with different t* in low uptake region. The parametric image quality was considerably improved when we used a new MLAIR method compared with Patlak. MLAIR showed reliable and robust properties for the voxel-wise parameter estimation in [¹⁸F]Fluoride PET study. It is expected that this method will be a good alternative to PGA for the radiotracers with irreversible 2- tissue compartment model.