Multicenter evaluation of single-photon emission computed tomography quantification with third-party reconstruction...

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Quantitative Analysis

Background:

• Single-photon emission computed tomography (SPECT) was historically not considered quantitative.
• Recent advancements in reconstruction algorithms have enabled quantitative SPECT imaging.
• Reproducibility is critical for the clinical adoption of quantitative SPECT.

Purpose of the Study:

• To evaluate the reproducibility of quantitative SPECT.
• To assess the performance of novel third-party reconstruction software.
• To investigate multicenter SPECT quantification using phantoms.

Main Methods:

• Phantom studies were conducted across five hospitals using eight SPECT scanners (GE and Siemens).
• A Jaszczak phantom determined conversion factors; a NEMA-IEC phantom assessed quantitative accuracy with spherical inserts.
• Reconstruction was centralized using HERMES HybridRecon with corrections for attenuation, scatter, and collimator-detector response.

Main Results:

• The coefficient of variation (CoV) for recovery coefficients in the NEMA-IEC phantom ranged from approximately 19% to 5%, improving with larger sphere diameters.
• Intersite CoV was comparable to intrasite CoV, indicating consistent performance across different centers and scanners.
• Quantitative SPECT demonstrated high reproducibility in a multicenter setting.

Conclusions:

• Quantitative SPECT is a reproducible imaging modality in a multicenter environment.
• Third-party reconstruction software can achieve reliable SPECT quantification.
• These findings support the clinical utility of quantitative SPECT.