^99mTc-PYP SPECT/CT cardiac transthyretin amyloidosis studies: cardiac-to-blood pool ratios versus visual interpretation

Objectives: ^99mTc-PYP is both sensitive & specific for cardiac ATTR amyloidosis (AT). Imaging typically consists of planar images &/or SPECT 1 hr after radiopharmaceutical injection, with delayed imaging at 3 hr in equivocal cases, i.e. persistent blood pool activity. SPECT is interpreted by visual comparison of myocardial uptake to rib uptake. Visual analysis is subjective & can vary among readers, & rib uptake depends on both systemic & local osseous metabolism, which also can affect visual analysis. The objective of this investigation was to determine if quantitative analysis could overcome the limitations inherent in visual interpretation of ^99mTc-PYP AT SPECT.

Methods: Data were reviewed retrospectively for 166 pts (103 male pts, 63 female pts, mean age = 72±12 yrs) who underwent SPECT/CT 1 & 3 hrs after I.V. injection of 555 MBq ^99mTc-PYP. Using the axial CT section of the SPECT/CT that best depicted the aorta, one individual manually drew a ROI within the aorta to sample mean blood pool counts. The aortic ROI was copied to the brightest lateral & septal myocardial wall locations to obtain 1hr & 3hr maximum cardiac-to-blood pool count ratios (CBPRs). One experienced physician without knowledge of CBPRs reviewed 1 & 3 hr SPECT/CT’s, & compared myocardial uptake to rib uptake, grading uptake as 0: no uptake (negative), 1: uptake < rib (equivocal), 2: uptake ≥ rib (positive). The 1 hr & 3 hr readings were performed blinded, without reference to one another.

Results: There was “very good agreement” (κ= 0.90) between 1 hr & 3 hr visual readings, with the same 43 pts definitively positive on both readings. CBPRs for all 166 pts were the same at 1 & 3 hr (1.34±0.75 versus 1.35±0.77, p = 0.75), & correlated significantly (r = 0.91, p < 0.001), with no trends (0.03±0.03, p = 0.38), or biases (-0.03±0.05, p = 0.51) by Bland-Altman analysis. 1 hr & 3 hr CBPRs were the same for visually negative & equivocal cases & were significantly lower than for positive cases (Table). Visually negative & equivocal cases had similar 1 hr & 3 hr CBPR values (p = 0.43), & both were significantly lower than for positive cases (p < 0.0001). ROC analysis indicated that 1 hr CBPR > 1.33 corresponded to visually positive cases at 1 hr (ROC AUC = 99±1%, sensitivity = 98%, specificity = 97%), statistically similar (p = 0.46) to 3 hr CBPR > 1.47 corresponding to visually positive cases at 3 hr (ROC AUC = 97±1%, sensitivity = 93%, specificity = 97%). The timing of image acquisition following ^99mTc-PYP injection did not affect the results, as visual & quantitative results were essentially the same for 1 hr & 3 hr SPECT/CT studies.

Conclusions: Myocardium to blood pool ratios agreed well with visual assessment & have the potential to overcome the limitations associated with visual analysis alone. Since the 1 & 3 hr results were essentially the same, a single imaging session at either time point is sufficient.

• Download figure
• Open in new tab
• Download powerpoint