A new imaging platform that pairs high-resolution structural imaging with detailed molecular profiling could change the way clinicians detect and classify skin cancer. In a recent clinical study, French researchers demonstrated that combining line-field confocal optical coherence tomography (LC-OCT) with confocal Raman microspectroscopy allows for precise, label-free identification of skin cancer subtypes – without the need for a biopsy.
The system, developed by scientists at Saint-Étienne University Hospital and Paris-Saclay University in collaboration with Damae Medical, integrates two complementary technologies. LC-OCT offers real-time, cellular-level visualization of skin morphology, while Raman microspectroscopy probes the biochemical composition of selected regions. When linked via AI, the system can distinguish tumor types based on both shape and molecular makeup.
Over a year-long clinical trial involving more than 330 nonmelanoma skin cancer samples – including basal cell carcinoma (BCC) and squamous cell carcinoma (SCC) – the team used LC-OCT to flag suspicious structures, then collected over 1,300 Raman spectra for chemical analysis. These datasets trained an artificial intelligence model that could classify BCC with 95 percent accuracy and correctly identify all cancer types included in the study with 92 percent accuracy.
Notably, the dual-imaging approach revealed chemical differences between subtypes of nonmelanoma skin cancer – insights that go beyond what histology alone can offer. This structural-chemical fingerprinting could support not just diagnosis, but deeper understanding of disease progression.
The team envisions that, with further validation, the dual-mode system could guide treatment decisions, monitor therapy response, and reduce the need for invasive procedures.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
