Endoscopic multispectral imaging sensor systems for intraoperative tissue differentiation.

Abstract

Significance: To remove the entire tumor area in cancer surgery, the margins are currently assessed by frozen sectioning or by analyzing the resected tissue afterwards. By utilizing spectral differences between healthy and malignant tissue, multispectral imaging (MSI) enables real-time tissue differentiation. Through miniaturization, the system can be used intraoperatively, thereby enabling the surgeon to define tumor margins, which reduces the need for a follow-up surgery.

Aim: By obtaining characteristic spectral differences of healthy and malignant breast tissue, a miniaturized multispectral sensor system should be developed. The system should be compatible to commercial endoscopes, so that it can be used in minimally invasive endoscopic procedures.

Approach: We analyzed hyperspectral measurements of an ex vivo breast cancer study in the visible to shortwave infrared spectral range (450 to 1390 nm) and defined characteristic spectral differences. With these, we designed two MSI sensor systems, which only use few spectral channels, gaining a potentially real-time-capable sensor system. By miniaturization, they can be used intraoperatively with a commercial endoscope.

Results: The final setups can be used with a commercial endoscope and allow for 90 deg field of view and a high spatial resolution, but showed limitations in the SWIR transmission.

Conclusion: The two demonstrated systems are promising steps towards real-time endoscopic tissue differentiation.