Evaluation of biexponential relaxation processes by magnetic resonance imaging. A phantom study.

Despite the complexity of biologic tissues, a monoexponential behaviour is usually assumed when estimating relaxation processes in vivo by magnetic resonance imaging (MRI). This study was designed to evaluate the potential of biexponential decomposition of T1 and T2 relaxation curves obtained at 1.5 tesla (T). Measurements were performed on a phantom of bicompartmental perspex boxes with combinations of different CuSO4 concentrations. T1 determination was based on a 12-points partial saturation inversion recovery pulse sequence. T2 determination was provided by a multiple spin echo sequence with 32 echoes. Applying biexponential curve analysis, a significant deviation from a monoexponential behaviour was recognized at a ratio of corresponding relaxation rates of about 3 and 2, estimating T1 and T2 relaxation, respectively (p less than 0.01, F-test). Requiring an SD less than or equal to 10 per cent for each set of parameters, the biexponential model was accepted when this ratio exceeded a factor of 5 and 8, respectively. Referring to 'expected' T1 and T2 values, however, an accuracy within 20 per cent only was observed at ratios of at least 6 and 15. It is concluded that quantitative estimation of individual and specific relaxation components in complex biologic tissues by MRI may prove very difficult.