Total-body parametric imaging using the Patlak model: Feasibility of reduced scan time

Area of Science:

• Nuclear Medicine
• Radiopharmacology
• Medical Imaging Analysis

Background:

• Patlak parametric imaging is crucial for quantitative analysis in Positron Emission Tomography (PET).
• Optimizing scan times is essential for improving patient throughput and reducing motion artifacts in total-body PET imaging.

Purpose of the Study:

• To investigate the feasibility of reducing scan duration for Patlak parametric imaging on the uEXPLORER total-body PET scanner.
• To evaluate the impact of reduced acquisition times and different input functions on image quality, noise, and lesion detectability.

Main Methods:

• 65 patients underwent 60-minute dynamic 18F-FDG PET scans on the uEXPLORER scanner.
• Patlak analysis was performed using both image-derived input function (IDIF) and population-based input function (PBIF) with varying acquisition times (20-60 min).
• Non-local means (NLM) denoising was applied, and image quality, noise, lesion conspicuity, and quantitative accuracy (K_i values) were assessed by radiologists and compared to a 40-minute IDIF gold standard.

Main Results:

• Shorter scan times (e.g., 30 min) yielded diagnostic value equivalent to longer scans (40 min) for both IDIF and PBIF.
• NLM filtering significantly reduced image noise and improved lesion detection, enabling acceptable image quality with 20-minute scans (G_IDIF+NLM, G_PBIF+NLM).
• No significant differences in K_i values were observed between IDIF and PBIF, or between reduced-scan-time groups and the gold standard.

Conclusions:

• A 20-minute dynamic scan using the PBIF-based Patlak model with NLM filtering achieves comparable diagnostic efficiency to a 40-minute IDIF-based scan.
• The study demonstrates the feasibility of significantly reducing scan times for Patlak parametric imaging on the uEXPLORER, enhancing clinical utility.
• There is no significant difference in diagnostic outcomes when using IDIF versus PBIF for K_i image generation.