Coherent Raman imaging can provide rapid, label-free virtual histology, but its clinical translation is constrained by detection geometry. Standard stimulated Raman scattering (SRS) relies on forward detection, which is poorly suited to thick, highly scattering tissue and therefore limits intraoperative implementations. In our ERC-funded research, we introduced backscattering coherent Stokes Raman scattering (b-CSRS), a coherent Raman modality whose red-shifted emission enables epi detection and is compatible with laser scanning and spectrally resolved acquisition. However, our current b-CSRS implementation operates in the short-wave infrared (SWIR), where detector noise prevents shot-noise-limited, high-speed imaging and spectroscopy.
This Proof-of-Concept will remove that bottleneck by developing a compact, low-noise SFG/OPA readout module based on periodically poled lithium niobate, enabling frequency conversion and optical amplification of b-CSRS photons into detector-friendly wavelength ranges. The project will deliver (i) an alignment-robust module compatible with laser-scanning microscopes, (ii) quantitative benchmarks against state-of-the-art direct SWIR detection in terms of sensitivity, speed, and spectral fidelity, and (iii) a validated route toward parallel, multi-channel spectroscopic b-CSRS for tissue classification. Together, these advances will transform b-CSRS from a promising laboratory method into a scalable platform for real-time molecular tissue assessment. We will pursue a licensing-first exploitation strategy with Lightcore Technologies to integrate the module into nonlinear imaging platforms and accelerate translation toward intraoperative cancer diagnostics.
