Background: A significant number of percutaneous intrathoracic biopsy procedures result in indeterminate cytologic or histologic diagnosis in clinical practice. Diffuse reflectance spectroscopy (DRS) is an optical technique that can distinguish different tissue types on a microscopic level. DRS may improve needle localization accuracy during biopsy procedures. The objective of this study was to assess the ability of DRS to enhance diagnosis of malignant disease in human lung tissue.
Methods: Ex vivo analysis with a DRS system was performed on lung tissue from 10 patients after pulmonary resection for malignant disease. Tissue spectra measured from 500 to 1600 nm were analyzed using 2 analysis methods; a model-based analysis that derives clinical and optical properties from the measurements and a partial least-squares discriminant analysis (PLS-DA) that classifies measured spectra with respect to the histologic nature of the measured tissue.
Results: Sensitivity and specificity for discrimination of tumor from normal lung tissue were 89% and 79%, respectively, based on the model-based analysis. Overall accuracy was 84%. The PLS-DA analysis yielded a sensitivity of 78%, a specificity of 86%, and an overall accuracy of 81%.
Conclusions: The presented results demonstrate that DRS has the potential to enhance diagnostic accuracy in minimally invasive biopsy procedures in the lungs in combination with conventional imaging techniques.
Copyright © 2012 Elsevier Inc. All rights reserved.