Abstract & Overview
Background Confocal fluorescence microscopy (CFM) is a powerful optical biopsy technique which captures cellular resolution images of the tissue surface without the need for tissue fixation or sectioning. The evolution of CFM with miniaturization and fibre-based optics now allows rapid capture of wide field images with microscopic resolution. For in-situ diagnostics, there is growing evidence that CFM systems could rapidly and accurately identify breast cancer with clinically actionable results. Review Focus This comprehensive review discusses different technological advances in CFM systems and explores emerging trends in Artificial Intelligence (AI) and robotic integration in breast cancer imaging. The review further discusses the clinical implications of these technologies, including their potential to reduce re-excision rates following breast conserving surgery (BCS) and improve surgical workflow efficiency. Methods A comprehensive literature review using PubMed, Embase and Web of Science databases was conducted by three reviewers independently covering studies published from January 2013 to December 2024. We included studies that provided human tissue data (preclinical and clinical) relevant to breast cancer imaging, focusing on the technological features, intra-operative usability, and ease of use of different bench-top and fibre-based CFM systems. Research focusing on future trends and emerging challenges in standardizing imaging protocols for breast cancer CFM imaging and automating diagnostic workflows were also considered. Results and conclusion Of 1382 articles identified from database screening, 28 fulfilled the inclusion criteria. Only 10 clinical studies reported statistical differentiation among specimens. Bench-top CFM systems demonstrated high-resolution imaging with accuracy ranging 83%–99.6% making them effective for detailed tissue analysis. However, their size and operational complexity limit their use during live surgery. In contrast, fibre-based CFM systems offer miniaturized flexible micro-endoscopes that enable real-time, in-situ imaging with accuracy upto 94% demonstrating suitability for intra-operative diagnosis. Notably, fibre-bundle based Cellvizio® confocal laser endomicroscopy (CLE) system and line-scan CLE system can identify breast pathology but data is lacking on intra-operative diagnostic accuracy for margin assessment on wide local excision specimens. New developments like the commercial Histolog® Confocal Microscopy system (SamanTree Medical SA, Lausanne, Switzerland) has potential to identify missed tumour margins in up to 75% of cases, enhancing the accuracy of margin assessments. While these technologies are promising, several obstacles must be overcome before CFM can be widely adopted in routine surgical practice. Additionally, AI- powered automation in CFM, although promising, requires large-scale validation to ensure accurate real-time tissue classification. Integrating robotics and AI-enhanced CFM could greatly improve real-time surgical decision-making, minimizing interpretation errors and enhancing workflow efficiency.