Abstract
1432
Objectives: scintigraphic evaluation of coronary flow reserve (CFR) is usually performed by dynamic SPECT (1) or dynamic PET (2-3). We propose an alternative method, using routine 99mTc-tetrofosmin SPECT (one day stress-rest protocol), with computation of a coronary reserve index (CRI).
Methods: Myocardial tracer uptake is related to the coronary flow. Theoretically, the ratio [myocardial stress counts/myocardial rest counts] represents the coronary reserve. But several adjustments are necessary to correct this stress/rest counts ratio. So, we developed a 5 steps software (Aladdin language, GE Xeleris station) in order to compute the correction factors. The myocardial counts were measured on all myocardial slices, using an automatic myocardial ROI (50% isocontour). 1. Subtraction of the estimated stress residual activity interfering with the rest images: the residual activity is calculated with the decay of 99mTc-tetrofosmin stress counts. 2. Normalization of the time acquisition duration: for 17 seconds per projection at stress, and 15 seconds at rest, the correction factor is 0.88. 3. Normalization of the injected tracer dose activity: this correction factor is obtained by the ratio Rest MBq / Stress MBq. 4. Correction related to the myocardial tetrofosmin extraction: to compensate the underestimation of uptake, we used a curve uptake issued from experimental studies (4). We applied a fitted curve (3rd polynomial function: Y = 0.02 X3 - 0.25 X² + 1.12 X + 0.012, with Y for the uptake and X for the flow) that indicates the true uptake for a flow value (knowing the Y value, we estimate the corresponding X value by applying an iterative process). 5. Normalization of the central ventricular pixel counts at stress vs at rest: blood pool activity must be low and similar at stress and at rest, as a background, 99mTc-tetrofosmin being cleared from the blood 10 minutes after injection. The normalization coefficient has been applied to the total myocardial counts at stress and at rest. In order to validate this CFR assessment method, this software was applied to a series of 135 patients who were addressed to our institution, for a routine myocardial SPECT stress (exercise, dipyridamole or regadenoson) and rest. Among these patients, an invasive coronary angiography (ICA) was realized in 27 cases. This software is applied after routine slices reconstruction in an automatic mode: the operator indicates the injected activity at stress and rest, and the delay between the 2 injections. The computer gives the Coronary Reserve Index (CRI) immediately. The results of ICA were compared to the CRI: sensitivity (SENS), specificity (SPE), positive predictive value (PPV) and negative predictive value (NPV) were calculated. CFR is related to epicardial coronary arteries and to microvascular arteries (3,5).
Results: For the CRI, we obtain: mean CRI = 2.85 - Standard Deviation (STD) = 1.97 (135 patients), and for the patients with ICA: mean CRI = 1.85 - STD = 0.07 (27 patients). These values are in the range of the CFR evaluated by other methods (1,3,5,6). For 27 patients with ICA, we found the optimal threshold value at 2.15 . This CRI value allows a safe exclusion, or a strong motivation for asking an ICA: SENS 79% - SPE 88% - PPV 94% - NPV 88%, area under Receiver-Operator-Characteristic curve: 0.909 . ICA is considered as the gold standard, although it doesn’t assess routinely microvascular lesions (6). This CFR evaluation method has allowed to diagnose a 3 vessels disease for 2 patients and diffuse atheroma in 1 patient (out of 27) who were missed by the standard SPECT images. The main limit is sometimes the presence of digestive activity, adjacent to the myocardium.
Conclusions: Assessment of CFR is feasible with routine 99mTc-tetrofosmin, allowing to decrease the number of false negative SPECT. This method is easy to use on a routine basis, totally reproducible (automatic process ). There is no additional time or cost to the routine SPECT. Sensitivity and specificity are very good for CAD diagnosis.