cCeLL: 2nd Generation Confocal Laser Endomicroscopy Platform for for Molecular-Guided Surgery

초록

Accurate, real-time margin assessment is a critical challenge in molecular-guided surgery (MGS), where complete tumor resection directly impacts patient outcomes. Conventional methods like frozen section analysis introduce workflow inefficiencies and sampling errors, emphasizing the need for more reliable intraoperative tools. This study introduces cCeLL, a second-generation Confocal Laser Endomicroscopy (CLE) system designed for MGS applications. The system provides immediate, subcellular-resolution images by using the exogenous fluorescent agent Indocyanine Green (ICG) and Lissajous fiber scanning technology. Its one-minute staining protocol enables seamless integration into the surgical workflow, positioning it as a powerful device for fluorescence imaging-guided resection. In our multi-center trial comparing cCeLL to frozen section analysis in brain tumors (a key clinical application), cCeLL demonstrated superior accuracy (93.9% vs. 91.5%) across 461 specimens while reducing diagnostic time by 65% (8 vs. 24 minutes). Our deep learning model, trained on approximately 18,000 annotated images, achieved 94.8% accuracy, matching expert performance and showing strong potential for AI-driven surgical decision-making. These results establish cCeLL as a promising platform for transforming intraoperative decision-making. Future work will focus on clinical translation by validating its versatility with a broader spectrum of molecular probes. This includes imaging endogenous fluorophores induced by agents like 5-ALA and collaborating with targeted agents such as antibody-drug conjugates (ADCs) utilizing dyes like IRDye800. Integrating this multi-agent capability, along with expansion to other cancers (prostate, ENT, lung), will significantly enhance tumor-specific detection and advance the frontier of MGS to improve oncologic outcomes.

키워드