Total-Body PET Multiparametric Imaging of Cancer Using a Voxelwise Strategy of Compartmental Modeling

Abstract

Quantitative dynamic PET with compartmental modeling has the potential to enable multiparametric imaging and more accurate quantification as compared to static PET imaging. Conventional methods for parametric imaging commonly use a single kinetic model for all image voxels and neglect the heterogeneity of physiological models, which can work well for single-organ parametric imaging but may significantly compromise total-body parametric imaging on long axial field-of-view scanners. In this paper, we evaluate the necessity of voxel-wise compartmental modeling strategies, including time delay correction and model selection, for total-body multiparametric imaging. Methods: Ten subjects (5 patients with metastatic cancer and 5 healthy volunteers) were scanned on the uEXPLORER total-body PET/CT system following injection of 370 MBq 18 F-fluorodeoxyglucose (FDG). Dynamic data were acquired for 60 minutes. Total-body parametric imaging was performed using two approaches. One is the conventional method that uses a single irreversible two-tissue compartmental model with and without time delay correction. The second approach selects the best kinetic model from three candidate models for individual voxels. The differences between the two approaches were evaluated for parametric imaging of micro kinetic parameters and FDG net influx rate Ki. Results: Time delay correction had a non-negligible effect on kinetic quantification of various organs and lesions. The effect was larger in lesions with higher blood volume. Parametric imaging of Ki with the standard two-tissue model introduced artifacts in vascular regions, which was overcome by the voxel-wise model selection strategy. Conclusions: The time delay and appropriate kinetic model vary in different organs and lesions. Modeling of the time delay of the blood input function and model selection improved total-body multiparametric imaging.

18F-FDG gallbladder uptake: observation from a total-body PET/CT scanner

“…This assumption might be supported by the higher frequency of early tracer detection in the wall (69%, 26%, 17%, 5% and 0% at 40-, 60-, 90-, 120-, and 180-min timepoints, respectively). However, this hypothesis needs further validation, for example through kinetic modeling and parametric imaging [ 21 ].…”

Neural KEM: A Kernel Method With Deep Coefficient Prior for PET Image Reconstruction

“…11), here we focused on parametric imaging of early-dynamic data. A two-tissue compartment model with voxel-wise time delay estimation [41] was used to generate parametric images from the earlydynamic data. For each reconstruction method, the blood input function was derived from the descending aorta ROI.…”

Total Body PET – Will it Change Science and Practice?

“…The acquisition of large volumes of 4D data (3D plus time) lends itself to, and even necessitates, innovative approaches to image analysis. The inclusion of both diseased and normal tissue, and differences in radiopharmaceutical transport and metabolism between different tissues, may require adjusting the kinetic model to different tissues (11) or considering dynamic PET data as a spatially varying mixture of characteristic kinetic curves. approaches to estimate key components such as tracer delivery and retention may offer computational efficiency and linear scaling that can be especially helpful for large 4D datasets.…”