Introduction Elastic scattering spectroscopy (ESS) is an emerging optical diagnostic technique which detects cellular abnormalities in vivo. In patients undergoing endoscopic surveillance for Barrett's oesophagus (BE), we have demonstrated ESS is able to detect high-grade dysplasia (HGD) and DNA ploidy, both associated with significant increase in 5-year cancer risk (59% and 28%, respectively). The spectrometer used had a wavelength range of 350–800 nm, where there is significant haemoglobin absorption. We hypothesise that a red enhanced spectrometer, more sensitive in the range 450 nm–1000 nm, would detect more scattering data and may improve detection rates. The aim of this in vivo study was to assess the influence of the spectral range of the spectrometer on accuracy of diagnosis.

Methods Spectral data for each spectrometer were collected at matched sites in vivo. A biopsy of the site was then taken from each site and histopathology reviewed. DNA ploidy was then assessed by image cytometric DNA analysis (ICDA) on these biopsies as previously described. A diagnostic algorithm of best fit was generated by principle component and linear discriminant analysis on the histology and DNA ploidy data.

Results 49 patients were enrolled in the study, with 217 sites analysed. ESS using a red enhanced spectra was able to distinguish squamous and columnar tissue with comparable accuracy to a standard spectrometer (91% sensitivity and 91% specificity). When comparing HGD vs non-dysplastic tissue, there was also little difference between the two spectrometers. Sensitivity with the red enhanced spectrometer was 84%, with specificity of 83%, respectively. The standard spectrometer sensitivity was 86% and specificity of 83%. When combining aneuploidy+HGD vs normal columnar tissue, however, the red spectrometer was more accurate with sensitivity 90% and specificity 80%. The standard spectrometer sensitivity was 90% but specificity was 74%.

Conclusion These data demonstrate ESS is highly accurate for the detection of BE and associated HGD and molecular abnormalities in vivo, whichever spectrometer is used. The red enhanced spectrometer is somewhat more accurate for the detection of DNA ploidy abnormalities than a standard spectrometer. We intend to test this spectrometer in a prospective equivalence study, comparing ESS with targeted detection of HGD and aneuploidy vs surveillance with quadrantic biopsy.