This study validates perfusion defect extent and severity as derived by PET polar maps in vivo against measurements derived from radiolabeled microspheres.
Methods: In seven open-chest dogs, either the left anterior descending (n = 11) or left circumflex coronary artery (n = 13) were ligated sequentially from distal to proximal. After each occlusion, gated PET images were acquired with 13N-ammonia (20 mCi) while radiolabeled microspheres were administered into the left atrium. The transaxial PET images were reoriented into left ventricular short-axis cuts, including the apex, and polar maps were generated from circumferential activity profiles. PET polar maps were then compared with polar maps derived from microspheres after normal databases for 13N-ammonia and for microspheres were established. Nitrogen-13 or microsphere activities of less than 1.5 s.d. below the mean were defined as hypoperfused.
Results: The extent (percent of left ventricular mass) and mean severity of the hypoperfused myocardium in the postmortem microsphere measurements ranged from 3% to 69% and 3% to 58%, respectively. The estimated extent by summed PET and by microspheres correlated by y = 4.95 + 0.95x (r = 0.91, s.e.e. = 0.085, p < 0.001) and mean severity by y = 5.52 + 0.87x (r = 0.85, s.e.e. = 0.101, p < 0.001). The extent and severity were similar for summed and gated PET studies.
Conclusion: The current study validated a polar map approach that provides accurate, quantitative assessment of the extent and severity of myocardial perfusion defects in vivo. Gating did not yield an improved correlation between PET and microsphere measurements. Thus, ungated PET images can be used to assess accurately the extent and severity of perfusion defects.