The assessment of left ventricular volume and function in gated small animal 18F-FDG PET/CT imaging: a comparative study of three commercially available software tools.

ABSTRACT

BACKGROUND: Several software tools have been developed for gated PET imaging that use distinct algorithms to analyze tracer uptake, myocardial perfusion, and left ventricle volumes and function. Studies suggest that different software tools cannot be used interchangeably in humans. In this study, we sought to compare the left ventricular parameters in gated 18F-FDG PET/CT imaging in mice by three commercially available software tools PMOD, MIM, and QGS. METHODS AND RESULTS: Healthy mice underwent ECG-gated 18F-FDG imaging using a small-animal nanoPET/CT (Mediso) under isoflurane narcosis. Reconstructed gates PET images were subsequently analyzed in three different software tools, and cardiac volume and function (end-diastolic (EDV), end-systolic volumes (ESV), stroke volume (SV), and ejection fraction (EF)) were evaluated. While cardiac volumes correlated well between PMOD, MIM, and QGS, the left ventricular parameters and cardiac function differed in agreement using Bland-Altman analysis. EDV in PMOD vs. QGS r = 0.85; p < 0.001, MIM vs. QGS r = 0.92; p < 0.001, and MIM vs. PMOD r = 0.88; p < 0.001, showed good correlations. Correlation was also found in ESV PMOD vs. QGS r = 0.48; p = 0.07, MIM vs QGS r = 0.79; p < 0.001, and MIM vs. PMOD r = 0.69; p < 0.01. SV showed good correlations in PMOD vs. QGS r = 0.73; p < 0.01, MIM vs. QGS r = 0.86; p < 0.001, and MIM vs. PMOD r = 0.92; p < 0.001. However, EF among correlated poorly PMOD vs. QGS r = -0.31; p = 0.26, MIM vs. QGS r = 0.48; p = 0.07, and MIM vs. PMOD r = 0.23; p = 0.41. Inter-class and intra-class correlation coefficient were > 0.9 underlining repeatability in using PMOD, MIM, and QGS for cardiac volume and function assessment. CONCLUSIONS: All three commercially available software tools are feasible in small animal cardiac volume assessment in gated 18F-FDG PET/CT imaging. However, due to software-related differences in agreement analysis for cardiac volumes and function, PMOD, MIM, and QGS cannot be used interchangeably in murine research.