JQR-Net: joint quantitative reconstruction network for dual-modal photoacoustic and optical coherence tomography imaging

Photoacoustic (PA) and optical coherence tomography (OCT) imaging are complementary imaging modalities with distinct contrast mechanisms, penetration depths, and spatial resolutions. Integrating these two modalities into a dual-modal PA-OCT imaging system enables the simultaneous acquisition of multimodal signals within a single scan. This integration supports quantitative reconstruction of tissue characteristics, offering a more precise and comprehensive analysis than single-modal imaging. In this paper, we propose a deep learning approach for joint quantitative reconstruction in dual-modal PA-OCT imaging, potentially advancing imaging capabilities for detailed tissue examination and disease analysis. We develop a deep neural network that performs end-to-end mapping from photoacoustically induced pressure signals and backscattered OCT signals to parametric images representing the spatial distribution of optical absorption and attenuation coefficients. This network provides both morphological and functional insights. To the best of our knowledge, this is the first deep learning model designed to simultaneously reconstruct multiple tissue characteristic parameters from dual-modal imaging signals, facilitating in-depth tissue characterization.