Diagnostic Value of (11)C-PIB PET/MR in Cardiac Amyloidosis

BACKGROUND: The thioflavin T derivative, (11)C-Pittsburgh-B (PIB), is used for Alzheimer's disease imaging because it specifically binds to β-amyloid protein deposits in the brain. The aim of this study was to estimate the diagnostic value of combined (11)C-PIB positron emission tomography/magnetic resonance (PET/MR) in cardiac amyloidosis (CA). METHODS: We enrolled 23 heart failure patients with suspected CA based on echocardiographic and electrocardiograph findings. All patients underwent cardiac (11)C-PIB PET/MR and non-cardiac biopsy within one week. We also enrolled eight healthy volunteers that underwent cardiac (11)C-PIB PET/MR as a control group. The cardiac magnetic resonance (CMR) protocol included cine imaging, late gadolinium enhancement (LGE), and native and post-contrast T1 mapping. Extracellular volume (ECV) was measured using pre- and post-contrast T1 mapping images. LVEF, IVSD, LVPW, LVmass, LVESV, LVEDV, native T1 value, ECV, and maximum uptake of myocardial tissue-to-blood background ratio (TBR) values were obtained from PET/MR images in all patients and healthy subjects. RESULTS: Thirteen out of twenty-three heart failure patients were clinically diagnosed with CA. The remaining 10 patients were CA-negative (non-CA patient group). Twelve of the thirteen CA patients showed diffuse transmural LGE patterns, whereas LGE was either absent or patchy in the non-CA patients. The diagnostic sensitivity and specificity of TBRmax were 92.3 and 100%, respectively, at a cut-off value of 1.09. Several CMR imaging parameters (LVEF, IVSD, LVmass, LVEDV, LVESV, LVPW, native T1 value and ECV) and TBR showed significant differences between CA patients, non-CA patients, and healthy controls (P < 0.05). Native T1 mapping values positively correlated with TBRmax values in CA and non-CA patients (r = 0.38, P = 0.0004). CONCLUSIONS: (11)C-PIB PET/MRI is a valuable tool for the accurate and non-invasive diagnosis of CA because it distinguishes CA patients from non-CA patients and healthy subjects with high specificity and sensitivity. Moreover, native T1 mapping values positively correlated with TBRmax values in CA and non-CA patients. In the future, larger cohort studies are necessary to confirm our findings.