A New Method for Hot-Spot Quantification of Hybrid SPECT/CT Cardiac Images: Methodology and Preliminary Phantom Validation

摘要

We present a new method for quantification of "hot-spot" cardiac single-photon-emission computerized tomography (SPECT) and preliminary phantom validation. Our quantification method was developed based on the circumferential count profiles of reference perfusion and focal hot-spot SPECT images. Normal hot-spot limit was estimated from the circumferential count profiles using the Bayes estimator and joint entropy maximization. Focal hot-spot uptake in the myocardium was calculated based on the normal hot-spot limit. The quantification method was validated using a cardiac phantom filled with 201Tl solution simulating normal myocardial perfusion. A hot-spot insert was filled with a series of different concentrations (5 muCi/mL-25 muCi/mL) of 99mTc solution to simulate a wide range of myocardial focal hot-spot uptake. A spherical point source filled with a known dose of 99mTc solution was used as an external reference. No extra cardiac and background activities were simulated. Dual-isotope images were acquired simultaneously using a hybrid SPECT/CT imaging system. Acquired images were reconstructed with and without CT-based attenuation correction and were quantified using the new quantification method. Correlation between SPECT quantified activities and actual hot-spot activities measured in a dose calibrator was fair without CT attenuation correction (r=0.78) and was improved with CT attenuation correction (r=0.84 and 0.997). Hot-spot activity was considerably underestimated in the absence of attenuation correction (Mean error=-146 muCi). Accuracy of the quantified hot-spot activity was markedly improved when the attenuation correction was employed (Mean error=-50 muCi and -10 muCi). In conclusion, our method for assessment of absolute hot-spot activity with incorporation of the CT-based attenuation correction may allow for accurate quantification of myocardial hot-spot uptake