Our main research line is aimed at biomedical imaging in deep tissue using label free microscopy. We use the physics of the interaction of short laser pulses with biological tissue to generate label-free microscopic images of volumes of tissue, in particular through second and third harmonic generation processes. With this we pursue several lines of interest, aimed at clinical diagnostics and at the study of fundamental processes in live tissue. A short summary, citing a few highlights: We started imaging mouse brains to demonstrate the versatility of the technique1. We have used SHG/THG to detect healthy and diseased states in excised tissue2-6 and have developed automated image analysis tools for computer aided diagnosis based on classical algorithms7-9 and neural networks. Applications include label-free analysis of clinical tumor samples2-6, pathological states of connective tissue in inflammation and mechanisms underlying wound healing10-12 as well as scoring and detection of resection margin of cancer. In glioma samples, normal brain was discriminated from infiltrative glioma with 96.6% sensitivity and 95.5% specificity, in nearly perfect (93%) agreement with pathologic scoring2 and concurring results were reached in healthy breast tissue and in lung tumor tissue3,6.
We are currently bringing transportable SHG/THG tissue analysers developed by Femtodiagnotics BV (Disclosure: MLG has indirect interests in Femtodiagnostics BV) into the clinic6 and are developing THG/SHG fiberscopes to be used in-situ, to help determine what tissue to excise during surgery. Furthermore, we are developing machine learning algorithms for automated image analysis of the 3D THG/SHG images. Supported by the Dutch Burn Foundation and TKI Holland Health we study the relationship between the 3D microstructure of skin and its macroscopic properties elasticity and strength, in healthy and scar tissue.