In situ detection of nuclear atypia in Barrett's esophagus by using angle-resolved low-coherence interferometry.

2007 
Background Monitoring of patients with Barrett's esophagus (BE) for dysplasia, currently done by systematic biopsy, can be improved through increasing the proportion of at-risk tissue examined. Objective Optical biopsy techniques, which do not remove the tissue but interrogate the tissue with light, offer a potential method to improve the monitoring of BE. Frequency-domain angle resolved low-coherence interferometry (fa/LCI) is an optical spectroscopic technique applied through an endoscopic fiber bundle and measures the depth-resolved nuclear morphology of tissue, a key biomarker for identifying dysplasia. The aim of the study was to assess the diagnostic capability of fa/LCI for differentiating healthy and dysplastic tissue in patients with BE. Methods Depth-resolved angular scattering data are acquired by using fa/LCI from tissue excised from 3 patients who had esophagogastrectomy. The data are processed to determine the average nuclear size and density as a function of depth beneath the tissue surface. These data are compared with the pathologic classification of the tissue. Main Outcome Measurements Average of depth-resolved nuclear diameter and nuclear density measurements in tissue samples. Results Upon comparison to pathologic diagnosis, the fa/LCI data results report the nuclear atypia characteristic of dysplasia in the epithelial tissue. Examination of the average nuclear morphology over the superficial 150 μm results in complete separation between healthy columnar and BE dysplastic tissues. Limitations Lack of in vivo data; lack of nondysplastic BE data because of limited sample size. Conclusions In complicated tissue structures, such as BE, depth-resolved nuclear morphology measurements provided an excellent means to identify dysplasia. The preliminary results demonstrate the great potential for the in vivo application of fa/LCI as a targeting mechanism for physical biopsy in patients with BE.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    10
    References
    48
    Citations
    NaN
    KQI
    []