Impact of Model Based Spillover and Partial Volume Correction on Myocardial ¹⁸F-FDG Kinetics Analysis in Rat Hearts

Objectives High-resolution dedicated small animal PET is an established tool for imaging mice and rats, but the size of the rodents´ hearts is still close to its spatial resolution which can lead to significant estimation error in quantification. Recently, a model-based correction approach using compartment-model kinetic analysis that accounts for spillover and partial volume effect via physiologic model was introduced. We aimed to determine the impact of model-based correction in determining the insulin-stimulated glucose transport capacity in a rat model of diabetes.

Methods ¹⁸F-FDG dynamic PET imaging was conducted in Zucker diabetic fatty (ZDF) rats (n=6), a model of type 2 diabetes mellitus, and Zucker lean (ZL) control rats (n=6) at age of 13 weeks. Under hyperinsulinemic-euglycemic clamp, ¹⁸F-FDG (37MBq) was administered followed by a dynamic 35min PET acquisition. Time activity curves (TAC) were generated for the myocardium (mid ventricular short axis), left ventricle blood cavity and left atrium blood cavity. A two-tissue, three-compartment model analysis was conducted including model based spillover and partial volume correction (Fang YH et al. JNM2007) or using the conventional standard method without spillover correction.

Results Myocardial ¹⁸F-FDG activity in ZDF rats (30min after tracer injection) was significantly decreased compared with ZL controls (0.92±0. 27 and 1.84 ±0.18 %ID/g, respectively, P<0.001). The input function derived from model-based correction was significantly smaller than the non-corrected input function directly calculated from PET measurements (area under the curve of input function with and without model-based correction were 5.85±1.58 and 9.11±0.84, respectively. p<0.005) and showed a good agreement with the time activity curve of the left atrium that has less spillover effects from myocardial activity (R² = 0.851, p<0.0001). Estimated Influx Ki values of myocardial ¹⁸F-FDG kinetic parameters in ZDF rats and ZL controls were 0.00179 ±0.00024 and 0.00333 ±0.00095 (P<0.05 vs controls) with the model-based correction approach, and 0.00188 ±0.00067 and 0.00668 ±0.0017 (P<0.001 vs controls) with the conventional non-corrected approach, respectively. Interestingly, Ki values were significantly changed by model correction in ZL controls (p<0.05) whereas they remained unchanged in ZDF diabetic rats (n.s.).

Conclusions In rat hearts, model-based spillover correction has a significant impact on ¹⁸F-FDG PET-derived glucose kinetic parameters estimation including influx Ki, especially in the presence of high myocardial activity.

• Download figure
• Open in new tab
• Download powerpoint