Non-invasive virtual biopsy using optical coherence tomography

We introduce the concept of a virtual biopsy, by using generative neural networks to synthesize virtual H&E sections from OCT images. To do this, we have developed a novel technique, “optical barcoding”, which has allowed us to construct a large dataset of OCT-H&E image pairs of human skin collected during Mohs surgery. Our results demonstrate the ability to use OCT images to generate high-fidelity virtual H&E sections and entire 3D H&E volumes. We hope that our novel method will allow clinicians to circumvent the time- and resource-consuming step of an invasive biopsy procedure.