Résumé
Influence of CT-based attenuation correction (CT-AC) in assessment of left and right ventricular functions with count-based gated blood-pool SPECT (GBPS) was evaluated in a mixed population.
Thirty-two patients (81% male; mean age 56 +/- A 12) referred for various symptoms or heart diseases were prospectively included. Data from 32 GBPS acquisitions were reconstructed using an iterative algorithm with (IRAC) and without (IRNC) CT-AC and analyzed using previously described segmentation software based on the watershed algorithm. LV and RV EF and volumes were assessed with and without CT-AC and compared.
EF and volumes were correlated (P < .001 for all parameters with r = 0.97 for LV and RV EF; r = 0.96 for LV EDV; r = 0.98 for LV ESV; r = 0.96 for RV EDV and ESV). The mean values using IRAC and IRNC were different for all parameters with lower EF (respectively, 49% +/- A 19% vs 51 %+/- A 18%; P = .002 for LV EF and 50% +/- A 14% vs 54%+/- 15%; P < .001 for RV EF) and higher volumes (respectively, 142 +/- A 41 mL vs 133 +/- A 40 mL; P < .001 and 79 +/- A 45 mL vs 71 +/- A 42 mL; P < .001 for LV EDV and ESV; 91 +/- A 32 mL vs 86 +/- A 31 mL; P = .003 and 48 +/- A 28 mL vs 43 +/- A 26 mL; P < .001 for RV EDV and ESV). Limits of agreement were -11% to 6% and -11% to 4% for LV and RV EF. We found wider limits of agreement for LV volumes (-13 to 32 mL for EDV and -10 to 27 mL for ESV) than for RV volumes (-13 to 23 mL for EDV and -9 to 20 mL for ESV). Taking into account all volumes, we found a trend with a significant positive correlation between means and differences in volumes assessed with and without CT-AC.
Assessment of both left and right ventricular functions by count-based GBPS with CT-AC showed higher volumes and lower EF. Differences were slight, especially for the range of normal to subnormal ventricular volumes.