Feasibility of time reduction in myocardial perfusion imaging using dedicated cardiac camera.

Introduction: Myocardial Perfusion Imaging (MPI) is pivotal in coronary artery disease assessment, but prolonged procedures often cause discomfort and motion artifacts. Dedicated Cardiac Cameras (DCC) aim to mitigate these issues, ensuring shorter acquisition times while preserving imaging quality. This research investigates the impact of further time reduction on MPI diagnostic accuracy and polar plot interpretation.

Methods: Individuals aged 18-85, referred for MPI, underwent standard MPI protocol (8 minutes imaging time for stress, and 10 minutes imaging time for rest imaging) with list-mode gamma camera data acquisition. Polar plots and computer-generated scores for both rest and stress images were created using the QPS/QGS Cedar Sinai software. Time reductions from 100% to 10% were simulated by removing counts in 10% increments. The modified polar plots and scores were compared against the original data. Parameters include: total perfusion defects (TPD), summed rest score (SRS), summed stress score (SSS), summed difference score (SDS), transient ischemic dilation (TID), left ventricular ejection fraction (LVEF), end-diastolic volume (EDV), end-systolic volume (ESV), and wall motion. Image quality and interpretability were qualitatively analyzed by two nuclear cardiology experts. The sensitivity, specificity, and accuracy of time-reduced polar plot findings and the impact of time reduction on parametric scores were evaluated using statistical methods of Cohen's Kappa, t-tests, and McNemar tests.

Results: 26 patients (average age: 65 years; 54% male) were included. Quantitative analysis showed robust perfusion parameter performance. SRS sensitivity, specificity, and accuracy exceeded 0.9 until imaging time was reduced to 30%, while SSS specificity and accuracy remained &ge; 0.8 until a 50% time reduction. SDS accuracy &ge; 0.8 was sustained until a 50% time reduction. TPD maintained high specificity and accuracy until imaging time was reduced to 30% at rest and 50% at stress. Functional parameters EF, ESV, and EDV showed changes below 30% imaging time but remained the same categorization as normal/abnormal. Significant mean score changes occurred at 30% and 10% imaging time for SDS, SRS, and TPD at rest. Qualitative analysis showed substantial agreement between two readers (physicians), (K=0.74). Image interpretability remained at 98% until the imaging time was reduced to 60% and stayed high at 88% with a time reduction of 50%. Time reduction effects on polar map homogeneity varied. Two-reader reports achieved 90% and 85.5% correctness when imaging time was reduced to 80% and 50% respectively. Interpretation certainty dropped below 70% with reduced imaging time. Sensitivity >0.8 persisted until a 40% imaging time for rest plots, with specificity and accuracy near 0.9. Stress plots maintained sensitivity (0.84-1.0) but exhibited reduced specificity and accuracy (<0.8) at 50% time reduction.

Conclusions: The study explored the diagnostic implications of time-reduced CZT SPECT images in MPI. Despite lower count statistics, time-reduced CZT SPECT images exhibited acceptable diagnostic accuracy. Quantitative score changes between time-reduced and non-time-reduced images were generally insignificant. Perfusion parameters like SRS and TPD at rest maintained acceptable specificity and accuracy for certain imaging time reductions, while others required a higher threshold.

This study also uncovered significant variations in the sensitivity and specificity of polar maps during time reduction. Rest polar map interpretation maintains sensitivity and specificity down to 10% and 20% imaging time, respectively. Stress polar plot sensitivity remained robust even at an imaging time of 10% compared to non-time-reduced imaging indicating the detectability of true stress-induced perfusion defects at shorter durations, while specificity declined at 50% time reduction, posing challenges in distinguishing true defects from artifacts. Reduced imaging time introduces noise-induced defects on stress images, impacting specificity, a higher probability event can happen at a higher count rate. However, for sensitivity loss, random defects on both rest and stress images are needed to obscure true reversible defects, a low probability event at a high count rate. Our study emphasizes the feasibility of further time reduction for diagnostic purposes without compromising sensitivity, specificity, or accuracy. An imaging time reduction of 50% preserves diagnostic accuracy in the cohort evaluated in this study.