Quantitative spectroscopy has been proposed as a means of improving the specificity of autofluorescence bronchoscopy by discriminating between true malignancy and suspicious but benign lesions before biopsy. This study investigated the potential discrimination ability of microvascular tissue-related parameters and relative intensity of autofluorescence, as estimated by noncontact spectroscopy.
Patients undergoing bronchoscopy for suspicion of lung cancer were enrolled from 4 sites (Canada, Russia, Slovenia, and the United Kingdom). In lesions selected for biopsy, light from a special fiberoptic bronchoscope was diverted to a spectrophotometer to obtain spectral measurements. The mucosa blood volume fraction and oxygen saturation were estimated from the computer model and, along with the autofluorescence intensity, were analyzed for discrimination potential for severe dysplasia or worse against lower histology grades.
A total of 485 patients were enrolled, from whom 352 suspicious, adequate biopsy specimens were collected. Of these, 8 specimens were severe dysplasia or carcinoma in situ, and 66 were cancer. All measures were found to be significantly altered (receiver operating characteristic curve area: 0.83, 0.74, and 0.80 for autofluorescence intensity, the blood volume fraction, and oxygen saturation, respectively) in lesions found by biopsy to have severe dysplasia or higher grade present. In addition, the estimated volume fraction of desaturated blood was found to be significantly more discriminatory than that of oxygen-saturated blood (receiver operating characteristic curve area: 0.83 vs. 0.63). Study center differences were evident and suggest that results may depend on study population or bronchoscopist experience.
Noncontact measurement and estimation of the above microvascular-related parameters, obtained during regular bronchoscope examination, may have potential for improving discrimination of severe dysplasia and cancer in lesions suspicious under white light bronchoscopy/autofluorescence bronchoscopy.
*Perceptronix Medical Inc., Vancouver, BC, Canada
†University Hospitals Coventry and Warwickshire NHS Trust, University Hospital, Coventry, UK
Supported by Perceptronix Medical Inc.
Reprints: Daniel Reinders, MSc, Perceptronix Medical Inc., Suite 400–555 West 8th Avenue, Vancouver, BC, Canada, V5Z1C6 (e-mail: email@example.com; firstname.lastname@example.org).
Received for publication January 27, 2009
accepted April 20, 2009
There is no conflict of interest.