Improved lesion detection in Whole-Body Parametric ¹⁸F-FDG PET/CT vs. Static ¹⁸F-FDG PET/CT

Introduction: ¹⁸F-FDG PET/CT is a cornerstone in the diagnostic of malignant and non-malignant diseases, with imaging guidelines endorsing static PET images at a single time-point, usually 60 minutes after tracer administration. The method is well-proven and robust, but hampered by the inherent omission of tracer kinetics from injection to acquisition. Recently introduced methodology allows for whole-body dynamic PET/CT scans using conventional PET cameras. The dynamic whole-body data can be used to calculate parametric images of the metabolic rate of glucose using Patlak linearization¹, thus utilizing the entire information contained in the blood input function and tissue time-activity curves. It was our goal to retrospectively evaluate whether parametric ¹⁸F-FDG PET/CT images have superior lesion target-to-background and if this translates into clinically valuable information.

Methods: Retrospective analysis of 48 patients scanned using a multiparametric PET acquisition protocol on a Siemens Biograph Vision 600 PET/CT scanner. Of the 48 patients, 25 had pathological lesions spanning a range of mostly malignant diseases (7 lung cancer, 5 lymphoma, 8 infection & inflammation, 2 colon cancer, 1 paraganglioma, 1 ovarian cancer and 1 cancer of unknown primary origin.) A 70-min dynamic whole-body PET scan was performed immediately after injection of ¹⁸F-FDG, 4 Mbq per kg. The scan included a 6-min dynamic scan of the chest region followed by multiple whole-body passes from 6-70 min. Parametric Patlak images were reconstructed using data from 40-70 min and an image-derived aorta input function [Reconstruction parameters: TrueX+TOF, 8 iterations, 5 subsets, 440 matrix, 2 mm Gaussian filter]. In addition, a standard-of-care static PET image was reconstructed using data from 60-70min [Reconstruction parameters: TrueX+TOF, 4 iterations, 5 subsets, 440 matrix, 2mm Gaussian filter]. Parametric Patlak images of metabolic rate (K) and distribution volume (V) were compared to standard 60-min static SUV images using PMOD 4.0. Target lesions were selected by an experienced nuclear medicine physician and PMOD’s isocontour tool was used (threshold set to 50% of max), to obtain the final target region. Background regions were manually drawn in adjoining tissue. Standard SUV values were compared with parameters obtained from dynamic scanning using Patlak reconstructions (K and V).

Results: Whole-body parametric Patlak images were of good visual quality and comparable to standard static SUV images. Observed lesion target to background contrast was superior in the parametric K images (p<0.001), allowing for easier detection of lesions, particularly in organs with high blood pool signal. There was no observable difference between disease groups in background normalized K/SUV ratios. In three patients with focal FDG uptake in the upper arm, Patlak imaging correctly omitted the signal from the K-image avoiding false positives and subsequent biopsies. However, although lesions were easier to detect by Patlak K-images than static images, no additional lesions were detected. Conclusion: Whole-body parametric Patlak-PET/CT imaging provides supplementary information, which may improve the evaluation of PET images. Pointedly, parametric imaging reduced the number of false positive lesions and had superior target-to-background ratios enabling easier image interpretation. Future prospective studies will clarify if parametric ¹⁸F-FDG PET/CT may also detect more lesions in diseases characterized by relatively low target-to-background ratios on conventional static PET/CT.