Test-Retest Repeatability of Patlak Slopes versus Standardized Uptake Values for Hypermetabolic Lesions and Normal...

Area of Science:

• Nuclear Medicine
• Oncology
• Radiomics

Background:

• Quantitative metrics in oncologic [18F]FDG-PET/CT are crucial for treatment response assessment.
• Standardized uptake value (SUV) is widely used, but Patlak slope (PS) offers potential advantages.
• Test-retest repeatability is essential for reliable longitudinal tracking of disease progression or regression.

Purpose of the Study:

• To evaluate the test-retest repeatability of quantitative metrics derived from Patlak slope (PS) compared to standardized uptake value (SUV).
• To assess these metrics in both malignant lesions and normal organs using oncologic [18F]FDG-PET/CT.
• To determine the suitability of PS-based metrics for monitoring therapeutic response in cancer patients.

Main Methods:

• Prospective, single-center study involving adults undergoing oncologic [18F]FDG-PET/CT.
• Reconstruction of early (35-50 min) and late (75-90 min) SUV and PS images from dynamic PET data.
• Repeat imaging within 7 days, with quantitative metrics extracted from lesions and organs.
• Repeatability assessed using mean test-retest percent changes, within-subject coefficients of variation (wCVs), and intra-class correlation coefficients (ICCs).

Main Results:

• Nine subjects completed the test-retest protocol; 17 [18F]FDG-avid lesions were analyzed.
• Early PS metrics showed higher lesion wCVs (worse repeatability) than early SUV metrics.
• Late PS and SUV metrics demonstrated similar lesion wCVs and high ICCs (0.93-0.98).
• A significant bias toward higher retest values was observed for late SUV and PS metrics in lesions.

Conclusions:

• Patlak slope (PS) based metrics exhibit similar repeatability to SUV metrics at late time points in [18F]FDG-PET/CT.
• These findings suggest PS metrics may be viable for tracking oncologic therapy response.
• Further validation is needed due to study limitations, including small sample size and observed bias.