Automated non-invasive identification of pelvic autonomic nerves with a handheld Raman spectrometer and potential application to nerve-sparing colorectal surgery: a preliminary study in surgical specimens.

Abstract

BackgroundAlthough minimally invasive surgery for colorectal cancer, whether performed as standard laparoscopic or robotic surgery, has been established as an oncologically safe procedure, postoperative urinary dysfunction and sexual dysfunction remain matters of concern, even when so-called nerve-sparing surgery is performed. We have hypothesized that Raman spectroscopy can be used intraoperatively as a non-invasive label-free means of objective identification of the pelvic nerves, and we conducted a preliminary study by applying a newly developed handheld Raman spectrometer to surgical specimens.MethodsSamples of nervous tissue, colon cancer tissue, and tissues from surrounding pelvic organs were obtained from 25 patients undergoing colectomy. Raman spectra were obtained by irradiation with the Progeny™ Raman spectrometer. We looked for characteristic Raman shifts to distinguish nervous tissue from cancer tissue. To improve discrimination between nervous tissue and other tissues, the spectral data were subjected to principal component analysis.ResultsWe detected characteristic differences in the spectra at 1,309 cm−1, 1,442 cm−1, and 1,658 cm−1. A significant difference was detected at 1,442 cm−1, and accuracy of the modality for identification of nervous tissue was 75%. The addition of principle component analysis (4 components) yielded 100% sensitivity, 85% specificity, and 90%, notably increasing accuracy from 75% to 90% in discriminating between nervous tissue and cancer tissue.ConclusionsRaman spectroscopy holds promise for non-invasive intraoperative recognition of nervous tissue. We expect the modality to become a powerful clinical tool, compensating for the lack of tactile feedback intrinsic to minimally invasive colectomy and thus thwarting the risk of postoperative urinary and/or sexual dysfunction.