Abstract
PURPOSE: We aim to verify predictions showing T(2) relaxation rate of nanoparticle clusters and its dependence on spacing, size, geometry, and pulse sequence. METHODS: We performed a laboratory validation study using nanopatterned arrays of iron oxide nanoparticles to precisely control cluster geometry and image diverse samples using a 4.7T MRI scanner with a T(2) -weighted fast spin-echo multislice sequence. We applied denoising and normalization to regions of interest and estimated relative R(2) for each relevant nanoparticle array or nanocluster array. We determined significance using an unpaired two-tailed t-test or one-way analysis of variance and performed curve fitting. RESULTS: We measured a density-dependent T(2) effect (p = 8.9976 × 10(-20) , one-way analysis of variance) and insignificant effect of cluster anisotropy (p = 0.5924, unpaired t-test) on T(2) relaxation. We found negative quadratic relationships (-0.0045[log τ(D) ](2) -0.0655[log τ(D) ]-2.7800) for single nanoparticles of varying sizes and for clusters (-0.0045[log τ(D) ](2) -0.0827[log τ(D) ]-2.3249) for diffusional correlation time τ(D) = r(p) (2) /D. Clusters show positive quadratic relationships for large (3.8615 × 10(-6) [d(pp) /r(p) ](2) -9.3853 × 10(-5) [d(pp) /r(p) ]-2.0393) and exponential relationships for small (-2.0050[d(pp) /r(p) ](0.0010) ) clusters. Calculated R(2) peak values also align well with in silico predictions (7.85 × 10(-4) ms compared with 1.47 × 10(-4) , 4.23 × 10(-4) , and 5.02 × 10(-4) ms for single iron oxide nanoparticles, 7.88 × 10(-4) ms compared with 5.24 × 10(-4) ms for nanoparticle clusters). CONCLUSION: Our verification affirms longstanding in silico predictions and demonstrates aggregation-dependent behavior in agreement with previous Monte Carlo simulation studies.