Quiescent phase respiratory gating for PET/CT

Objectives Previous studies have shown that patients spend more time breathing near end expiration where less motion occurs and tend to return to the same end expiration location. Images extracted from this end-expiration quiescent interval are expected to yield an optimal SNR in PET/CT studies with respiratory motion. We report on the quiescent phase respiratory gating method and evaluate its performance compared to ungated and conventional phase gating images.

Methods We simulated noise-free PET acquisitions of the NCAT phantom with respiratory patterns derived from 1295 patient studies. Data with a corresponding respiratory displacement signal below a certain threshold of the amplitude of each cycle were extracted to form quiescent phase images. Images generated with 4 thresholds at 10%, 20%, 30%, and 50% were compared to images gated over the entire respiratory cycle using 6 equal phases (‘fully-gated’) and ungated images. Image quality was evaluated in terms of the tumor SUV_max and fraction of counts as a surrogate for image noise.

Results Tumor SUV_max decreased and the fraction of counts increased with larger thresholds. For the 20% threshold, the patients spent 40% time on average in the quiescent phase and the mean SUV_max increased 22% compared to that of the ungated images. Only 1 of the 6 fully-gated images was comparable to the quiescent phase image in terms of SUV_max but with increased noise. Patient studies are ongoing.

Conclusions Compared to the ungated and fully-gated images, quiescent phase gating yields an improved SNR. There is potential for quiescent phase gating to effectively improve tumor quantification with minimal noise increase in PET/CT.