Flow analysis to help identify false positives in N13-ammonia cardiac perfusion imaging

Objectives Flow analysis improves diagnostics of multi-vessel, balanced or microvascular diseases. This work studies if flow analysis can also help identify false positives in late perfusion imaging.

Methods 35 patients referred for rest/stress N13-ammonia myocardial perfusion scans were acquired on a Siemens Hi-Rez (BMI=29.3±6.3). The average injected dose was 15.2±3.2 mCi. When a CT-PET misalignment was identified, CT-PET registration was performed manually. Static and 16-frame images were reconstructed. Each study was semi-quantitatively scored by an experienced physician based on static images only. Patients underwent interventional coronary angiography (ICA) when stenosis was diagnosed. Flow analysis was performed retrospectively with Siemens syngoMBF and results were consolidated with perfusion and ICA findings.

Results 15 patients had normal perfusion and normal flow (average flow: 1.00±0.25 ml/mg/min at rest, 2.63±0.34 ml/mg/min at stress) while 12 patients had both abnormal perfusion and abnormal flow (average flow: 0.78±0.26 ml/mg/min at rest, 1.43±0.56 ml/mg/min at stress). The abnormal diagnostics were confirmed by ICA. 5 patients were called normal by perfusion reading but abnormal by flow analysis (average flow: 0.94±0.37 ml/mg/min at rest, 1.56±0.50 ml/mg/min at stress). 3 of those 5 patients underwent ICA because of lasting symptoms and all 3 had 50% stenosis in one of the coronary. The last 3 patients had abnormal perfusion but the flow analysis returned normal results (average flow: 0.93±0.30 ml/mg/min at rest, 2.46±0.27 ml/mg/min at stress). ICA found no stenosis in any of the coronaries and diagnostics were normal for all 3 patients.

Conclusions With this work, we show that flow analysis is more accurate in the diagnostics of stenosis. Flow analysis is also less sensitive to CT-PET improper registration and it can help identify false positive perfusion readings