TU-G-204-03: Dynamic CT Myocardial Perfusion Measurement Using First Pass Analysis and Maximum Slope Models




To evaluate the accuracy of dynamic CT myocardial perfusion measurement using first pass analysis (FPA) and maximum slope models.


A swine animal model was prepared by percutaneous advancement of an angioplasty balloon into the proximal left anterior descending (LAD) coronary artery to induce varying degrees of stenosis. Maximal hyperaemia was achieved in the LAD with an intracoronary adenosine drip (240 µg/min). Serial microsphere and contrast (370 mg/mL iodine, 30 mL, 5mL/s) injections were made over a range of induced stenoses, and dynamic imaging was performed using a 320-row CT scanner at 100 kVp and 200 mA. The FPA CT perfusion technique was used to make vessel-specific myocardial perfusion measurements. CT perfusion measurements using the FPA and maximum slope models were validated using colored microspheres as the reference gold standard.


Perfusion measurements using the FPA technique (P_FPA) showed good correlation with minimal offset when compared to perfusion measurements using microspheres (P_ Micro) as the reference standard (P _FPA = 0.96 P_Micro + 0.05, R2 = 0.97, RMSE = 0.19 mL/min/g). In contrast, the maximum slope model technique (P_MS) was shown to underestimate perfusion when compared to microsphere perfusion measurements (P_MS = 0.42 P _Micro −0.48, R2 = 0.94, RMSE = 3.3 mL/min/g).


The results indicate the potential for significant improvements in accuracy of dynamic CT myocardial perfusion measurement using the first pass analysis technique as compared with the standard maximum slope model.