Reversible, irreversible and effective transverse relaxation rates in normal aging brain at 3T.

Quantitative transverse relaxation rates in normal aging brain are essential to investigate pathologies associated with iron accumulation and tissue degeneration. Since absolute values depend on imaging methods and magnetic field strengths, continuous evaluation of specific reference values remains requisite. Multi-echo turbo spin echo and multi-echo gradient recalled echo imaging sequences were applied to 66 healthy subjects (18-84years) at 3T to quantify the irreversible (R2), effective (R2*) and reversible (R2'=R2*-R2) transverse relaxation rates. Representative regions-of-interest (ROIs) were determined automatically in gray matter (GM) and white matter (WM) on T1-weighted scans. Phantom experiments of different sized iron-oxide particles were conducted to explore the correlation of R2' related to R2 for the evaluation of the size of iron deposits. R2 decreased with age for the majority of ROIs, but increased for putamen, head of caudate nucleus and nucleus accumbens. R2* and R2' increased with age in deep GM structures except for the thalamus. R2* and R2' showed a distinct dependency on fiber orientation in exemplary WM regions. R2', R2 and R2* were strongly linear proportional to age-related iron content in deep GM with slopes of 0.88, 0.18 and 1.08 in [1/s/mg Fe per 100g wet tissue] and intercepts of 1.69, 9.25 and 10.69 in [1/s], respectively. Linear and non-linear curve fitting of R2' vs. R2 in phantoms revealed increased slopes with increasing particle size. In vivo, averaged R2' vs. R2 data points of patients with Parkinson's disease and progressive supranuclear palsy were above the fitted curves of healthy subjects suggesting larger sized iron deposits in these neurodegenerative diseases. Decreased R2 with age may reflect physiological tissue degeneration, whereas increased R2* and R2' with age most likely denote physiological iron accumulation. The low intercept of R2' vs. iron content suggests a nearly sole sensitivity of R2' to iron in deep GM, potentially allowing a more specific estimation of the iron content than R2 or R2*. Since R2* and R2' depend on the fiber orientation, their feasibility to estimate iron content in WM is challenging. The analysis of R2' related to R2 may provide valuable information about the size of iron deposits.