A Digital Reference Object to Analyze Calculation Accuracy of PET Standardized Uptake Value

Area of Science:

• Medical Imaging
• Radiochemistry
• Quantitative Analysis

Background:

• Standardized uptake value (SUV) is crucial for quantitative analysis in PET/CT imaging.
• Variations in SUV computation across different software packages can impact diagnostic accuracy.
• A need exists for standardized methods to ensure consistent SUV measurements.

Purpose of the Study:

• To quantify the variability in body weight-normalized SUV (SUV(BW)) calculations across various software packages.
• To introduce a Digital Imaging and Communications in Medicine (DICOM) reference test object for standardizing SUV computation.
• To ensure consistency in SUV measurements on medical image viewing workstations.

Main Methods:

• A synthetic positron emission tomographic/computed tomographic (PET/CT) digital reference object (DRO) with known SUV was created.
• The DRO was distributed to 16 sites for evaluation on 21 different PET/CT display software packages.
• Users reported SUV metrics (max, min, mean, SD) from defined regions of interest (ROIs), and deviations from expected values were analyzed.

Main Results:

• Significant errors were observed in reported SUV metrics, with maximum SUV errors ranging from -37.8% to 0% and mean SUV errors from -1.6% to 100%.
• Variability was also noted in minimum SUV and standard deviation SUV calculations.
• The observed discrepancies highlight substantial inconsistencies in SUV(BW) computation.

Conclusions:

• The variability in computed SUV(BW) across different software packages is substantial.
• A reference standard, such as a DICOM test object, is necessary to ensure the standardization of SUV computation.
• Implementing a standardized approach will improve the reliability and comparability of quantitative PET/CT data.