OpenQPAIData: Measurements of test objects with reference labels for quantitative photoacoustic imaging research

Abstract

The goal of quantitative photoacoustic imaging (qPAI) is to determine absolute chromophore concentrations from multispectral photoacoustic images. Achieving this goal would enable bias-free molecular photoacoustic imaging and the derivation of functional tissue biomarkers, such as local blood oxygenation, with numerous clinical implications. To deliver qPAI, two inverse problems must be solved: the acoustic inverse problem of reconstructing the initial pressure distribution from measured signals, and the optical inverse problem of reconstructing spatial images of optical absorption from the initial pressure. Unfortunately, these inverse problems are difficult to solve in practice. The validation of proposed qPAI algorithms is particularly challenging due to the lack of experimentally measured ground truth of the optical and acoustic tissue properties. With OpenQPAIData, we present a dataset of 30 tissue-mimicking phantoms imaged using three different photoacoustic systems: the Multispectral Optoacoustic Tomography (MSOT) InVision commercially produced system from iThera Medical GmbH, and the custom-built Transmission Reflection Optoacoustic and Ultrasound (TROPUS) and Spiral Volumetric Optoacoustic Tomography (SVOT) imaging systems developed by Prof. Daniel Razansky’s group. The dataset includes raw measurements, reconstructed images, manually segmented reference annotations, and matched optical property maps derived from double-integrating-sphere (DIS) measurements. By providing these multi-device imaging data with reference optical properties, we believe that OpenQPAIData enables the benchmarking of both acoustic and optical inverse solvers and supports the development of image quality assessment metrics for photoacoustic imaging.