Diagnostic Efficacy of Quantitative Indexes on [99mTc]DPD SPECT/CT for Cardiac Amyloidosis: Inter-Observer Agreement and Association with Conventional Scoring

2021 
1670 Introduction: [99mTc]DPD imaging is effective for diagnosing cardiac amyloidosis (CA), and quantitative SPECT/CT is increasingly used recently. We aimed to evaluate the efficacy of visual scoring systems and various quantitative indexes of [99mTc]DPD imaging, in terms of inter-observer agreement and association with conventional scoring. Methods: Patients who underwent [99mTc]DPD imaging with suspicion of CA were enrolled. On planar scan, myocardial uptake was visually scored using two systems (“Perugini” and “Dorbala”), respectively. Also, quantitative uptake indexes were measured; heart-to-whole body ratio (H/WB) and heart-to-contralateral lung ratio (H/CL) on planar scan; SUVmax, SUVmean, total myocardial uptake (SUVmean × volume), and C-index (SUVmax normalized by spine uptake) on SPECT/CT. The association of quantitative indexes with visual grading and their inter-observer agreements were analyzed. Results: A total of 152 patients, of which 18 were positive, were included. Inter-observer agreements of both Perugini and Dorbala systems were high (kappa values, 0.96 and 0.95). However, quantitative indexes were more significantly matched with Dorbala system. In positive cases, SPECT/CT indexes showed higher intraclass correlation coefficients (ICC; > 0.95) than H/WB and H/CL (0.92 and 0.90). SUVmax showed the highest ICC (1.00) and excellent correlation with H/WB and H/CL (r = 0.90 and 0.91). The optimal cutoff values of SUVmax for Dorbala scoring system were 1.90, 3.71, and 7.77. Conclusions: Dorbala system is more significantly matched with quantitative indexes than Perugini system. SPECT/CT indexes, particularly SUVmax, are superior to planar image indexes in terms of inter-observer agreement. SUVmax can be used as a reliable and reproducible index, with excellent correlations with conventional visual scoring and quantitative indexes.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []