Deep learning-driven quantitative spectroscopic photoacoustic imaging for segmentation and oxygen saturation estimation

• A deep neural network is developed for spectroscopic photoacoustic imaging. • Performs segmentation of blood vessels and estimation of oxygenation • Hybrid loss function improves oxygenation predictions. • Demonstrates high accuracy on Monte Carlo simulated and experimental data. Spectroscopic photoacoustic (sPA) imaging can potentially estimate blood oxygenation saturation (sO 2 ) in vivo noninvasively. However, quantitatively accurate results require accurate optical fluence estimates. Robust modeling in heterogeneous tissue, where light with different wavelengths can experience significantly different absorption and scattering, is difficult. In this work, we developed a deep neural network (Hybrid-Net) for sPA imaging to simultaneously estimate sO 2 in blood vessels and segment those vessels from surrounding background tissue. sO 2 error was minimized only in blood vessels segmented in Hybrid-Net, resulting in more accurate predictions. Hybrid-Net was first trained on simulated sPA data (at 700 nm and 850 nm) representing initial pressure distributions from three-dimensional Monte Carlo simulations of light transport in breast tissue. Then, for experimental verification, the network was retrained on experimental sPA data (at 700 nm and 850 nm) acquired from simple tissue mimicking phantoms with an embedded blood pool. Quantitative measures were used to evaluate Hybrid-Net performance with an averaged segmentation accuracy of ≥ 0.978 in simulations with varying noise levels (0 dB-35 dB) and 0.999 in the experiment, and an averaged sO 2 mean squared error of ≤ 0.048 in simulations with varying noise levels (0 dB-35 dB) and 0.002 in the experiment. Overall, these results show that Hybrid-Net can provide accurate blood oxygenation without estimating the optical fluence, and this study could lead to improvements in sO 2 estimation in vivo .