Dependence of two-phase 19F NMR oxygenation measurements on dispersed-phase droplet size distribution

¹⁹F magnetic resonance has been explored in the context of medicine as a technique for quantifying oxygenation in blood, tissues, and tumors [1]. The technique is based on the oxygen affinity of fluorocarbons, such as hexafluorobenzene (HFB), that exhibit linear dependence of R₁ on local oxygen concentration. As HFB is immiscible in water, it partitions in tissues when injected and the impact of susceptibility artifacts at interfaces on T₁ measurements from dispersed HFB in a continuous aqueous phase may exhibit dependences on droplet size [2], and thus not directly correlate with measurements from a single HFB phase, leading to error in oxygenation determination. To evaluate this effect, the impact of HFB droplet size on spatially-resolved T₁ measurements was analyzed using agarose gel as an aqueous phase surrogate for tissue. Gel samples of various HFB concentrations were made and the two phases were homogenized with a shear mixer at 5,000 rpm (low-dispersion) or 15,000 rpm (high-dispersion) to generate different droplet size distributions. T₁ inversion recovery imaging revealed a strong dependence of spatially-resolved T₁ distribution on mean droplet radius, with smaller droplets displaying higher T₁ values. Our results suggest that applying a standard curve generated from single-phase T₁ measurements to measurements from a dispersed HFB phase may result in systematic overestimation of local oxygenation. This research is a first step toward developing a technique to measure the spatiotemporal distribution of oxygen in clinically relevant biofilms such as those of Staphylococcus aureus.

Figure 1. Left: 19F sagittal image of HFB in agarose for low-dispersion sample (A) and corresponding T1 map (C). Right: 19F sagittal image of HFB in agarose for high-dispersion sample (B) and corresponding T1 map (D)