Elastic Scattering Spectroscopy for the Detection of Pre-Cancer and Early Cancer of the Breast

摘要

Elastic scattering spectroscopy (ESS) interrogates tissue with short pulses of white light, obtaining diagnostic information from immediate spectral analysis of light scattered back. From data pairs correlating spectra with conventional histology at individual points, we developed diagnostic algorithms to detect cancer by spectral analysis. Pixels containing less than 20% cancer were detected in mastectomy specimens. In initial testing on axillary nodes, some cancers were missed due to inadequate sampling. This was addressed by scanning the cut surface of excised sentinel nodes (400 pixels/cm2) to create an optical image which improved sensitivity to 77.8% with 97.5% specificity. This is comparable to touch imprint cytology, an established technique for rapid assessment of nodes, but with no need for tissue processing or for a pathologist's opinion. Greater sensitivity should be possible by scanning more layers of nodes (ESS interrogates 0.3-0.5mm into tissue). Preliminary studies suggest that ESS may also be able to diagnose lesions seen at ductoscopy (where biopsy is not feasible) and detect aneuploidy, an important prognostic factor. These results justify further prospective studies. The technique is simple in concept, uses low cost equipment and has considerable potential for the immediate detection of cancer in vivo or ex vivo in many tissues.