Volumetric evaluation of ^99mTc-pyrophosphate SPECT/CT in patients with transthyretin cardiac amyloidosis: optimization and correlation with cardiac functional parameters

Introduction: Volumetric evaluation of ^99mTechnetium-pyrophosphate (^99mTc-PYP) SPECT/CT is an objective and useful method for diagnosing transthyretin cardiac amyloidosis (ATTR-CA), but the method has not been standardized. The aim of this study was to optimize the method of volumetric evaluation and to assess its relationship to other quantitative PYP evaluation and cardiac functional parameters.

Methods: We retrospectively evaluated 43 patients who underwent ^99mTc-PYP SPECT/CT at our hospital between October 2018 and October 2020. All patients underwent endomyocardial biopsies (EMB; n = 37) and/or TTR gene tests (n = 20). Twenty (47%) patients were diagnosed with ATTR-CA (10 wild types and 10 hereditary types), and 23 (53%) patients with non-ATTR (control: 1 light chain cardiac amyloidosis, 1 cardiac sarcoidosis, and 21 other myocardial disorders; all control patients underwent EMB and were negative for ATTR amyloid). Patients were imaged at 1 hour (n = 18) or 3 hours (n = 25) after ^99mTc-PYP injection using both SPECT/CT and planar studies. First, we evaluated an aortic blood pool ^99mTc-PYP activity (ABPmax) using SUVmax in the ascending aorta at the level of the pulmonary artery bifurcation. Next, we objectively evaluated the total volume of the region where ^99mTc-PYP uptake > 1.0×, 1.2×, and 1.4× of ABPmax within the left and right ventricular myocardium and defined as cardiac metabolic volume (CMV_ABP1.0, CMV_ABP1.2, and CMV_ABP1.4: volume of voxels with abnormal ^99mTc-PYP uptakes). SUV and CMV were calculated using xSPECT Quant (Siemens). We compared CMV to cardiac metabolic activity (CMA: CMV×SUVmean within the abnormal ^99mTc-PYP uptake region) and a conventional quantitative parameter (planar heart to contralateral ratio of ^99mTc-PYP uptakes: H/CL). Accuracy for identifying ATTR-CA was examined by receiver operating characteristic analysis and the area under the curve (AUC) was calculated. Correlation of quantitative ^99mTc-PYP scintigraphy metrics with brain natriuretic peptide (BNP) and left ventricular ejection fraction (LVEF) were measured.

Results: CMV_ABP1.2 achieved the highest sensitivity and specificity of 90% and 100%, respectively (Figure). AUC of CMV_ABP1.0, CMV_ABP1.2, CMV_ABP1.4, CMA_ABP1.2, and H/CL were 0.96, 0.95, 0.90, 0.93, and 0.93 (p = ns), respectively. In patients with ATTR-CA (n = 20), CMV_ABP1.2 strongly correlated with CMA_ABP1.2 (r = 0.97, p < 0.0001), moderately correlated with H/CL (r = 0.58, p = 0.008) and BNP (r = 0.45, p = 0.045), and moderately negatively correlated with LVEF (r = -0.61, p = 0.005). However, H/CL did not correlate with BNP and LVEF. CMV_ABP1.2 did not differ between patients with wild type ATTR-CA and hereditary type ATTR-CA. Although ABPmax was significantly lower in 3-hour image compared with 1-hour image (p = 0.0003), in subgroup analysis the accuracy was similar for identifying ATTR-CA between patients imaged at 1 hour (AUC = 0.92) and 3 hours (AUC = 0.96). CMV_ABP1.2 showed the highest sensitivity and specificity for identifying ATTR-CA at both timings. ABPmax did not correlate with CMV_ABP1.2, BNP, and LVEF. While ABPmax moderately and negatively correlated with estimated glomerular filtration rate in patients imaged at 3 hours (r = -0.48, p = 0.01), it did not at 1 hour.

Conclusions: CMV_ABP1.2 could identify patients with ATTR-CA more accurately than other CMV and H/CL, and strongly correlated with CMA. Although CMV significantly correlated with cardiac functional parameters, H/CL did not.

• Download figure
• Open in new tab
• Download powerpoint