Blood-mimicking dye phantoms for evaluating photoacoustic oximetry accuracy

Many proposed clinical applications of photoacoustic imaging (PAI) rely on relative or absolute measurements of blood oxygen saturation (sO2), and evaluation of oximetry measurement accuracy is crucial for assessing device performance. Available bench test methods use phantoms connected to blood flow circuits with tunable oxygenation, but these methods are complex, costly, and pose biohazard safety risks. To address these issues, we have developed stable and tunable blood-mimicking solutions using binary mixtures of commercially available near-infrared organic dyes (NIR746A and IRA980) to enable non-biological phantom-based PAI oximetry test methods. We used spectrophotometry and a custom PA spectroscopy system to characterize dye extinction and PA response at 750 nm and 850 nm, then formulated various dye recipes mimicking sO2 levels from 40 % to 100 %. We then used a custom PAI system to image breast-mimicking polyacrylamide hydrogel phantoms with embedded tubes injected with static volumes of either dye solutions or bovine blood deoxygenated using sodium dithionite. Phantom testing with dyes produced similar performance metrics to blood, with root-mean-squared difference (RMSD) values between photoacoustic sO2 and reference sO2 of 6-17 % for blood and 4-18 % for dyes, sensitivity (slope of the regression line) ranged from 0.4 to 0.7 for blood and 0.4-0.9 for dyes, and depth-averaged bias ranged from 4 % to 17 % for blood and 3-10 % for dyes. These blood-mimicking dyes may offer a simpler, cheaper, safer, and more stable approach to evaluate PAI oximetry accuracy compared to traditional blood flow phantoms. This tool could facilitate establishment of less burdensome and more reproducible phantom-based PAI test methods, ultimately expediting clinical adoption of PAI technology.