Improving mesoscopic fluorescence molecular tomography through data reduction.

Mesoscopic fluorescence molecular tomography (MFMT) is a novel imaging technique that aims at obtaining the 3-D distribution of molecular probes inside biological tissues at depths of a few millimeters. To achieve high resolution, around 100-150μm scale in turbid samples, dense spatial sampling strategies are required. However, a large number of optodes leads to sizable forward and inverse problems that can be challenging to compute efficiently. In this work, we propose a two-step data reduction strategy to accelerate the inverse problem and improve robustness. First, data selection is performed via signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) criteria. Then principal component analysis (PCA) is applied to further reduce the size of the sensitivity matrix. We perform numerical simulations and phantom experiments to validate the effectiveness of the proposed strategy. In both in silico and in vitro cases, we are able to significantly improve the quality of MFMT reconstructions while reducing the computation times by close to a factor of two.