Optical biopsy of breast tissue using differential path-length spectroscopy

Abstract

Differential path-length spectroscopy (DPS) was used to determine the local optical properties of breast tissue in vivo. DPS measurements were made on healthy and malignant breast tissue using a fibre-optic needle probe, and were correlated to the histological outcome of core-needle biopsies taken from the same location as the measurements. DPS yields information on the local tissue blood content, the local blood oxygenation, the average micro-vessel diameter, the beta-carotene concentration and the scatter slope. Our data show that malignant breast tissue is characterized by a significant decrease in tissue oxygenation and a higher blood content compared to normal breast tissue.

Confidence intervals on fit parameters derived from optical reflectance spectroscopy measurements

“…Thus for large noise amplitudes ͑0.1 and 0.5, corresponding to a signal-tonoise ratio at = 800 nm of 10 and 2, respectively͒, the fitted parameters are also correctly within 2 CIs in not significantly less than 95% of the simulations, because the CIs are dramatically increased compared to the low noise amplitude ͑0.01, or signal to noise of 100͒. Note that for the low noise amplitude, which is typical for our in vivo data, [11][12][13][14] the CI of the fitted blood saturation a 6 is smaller than 1% for blood volume fractions a 5 Ͼ 1%. Even for a blood volume fraction as low as 0.2%, the CI of a 6 is only 5% when the correct model is used in the fit.…”

“…Therefore, to be safe, we have in previous publications restricted the calculations of average blood saturation to spectra with blood volume fractions larger than 1%. [11][12][13][14] It is important to note that the saturation CI was calculated for our specific DPS measurement geometry with 400-m fibers, featuring a path length of 0.4 mm. Since the saturation CI will depend on signal attenuation rather than on blood volume fraction, the saturation CI ͑and potential bias͒ will be smaller for similar blood volume fractions in the case of DPS with larger fiber diameters, or any other measurement geometry with a longer path length.…”

“…In this work we use the empirical model used to analyze a particular type of reflectance spectroscopy that we have been investigating over the past 5 years, differential path-length spectroscopy ͑DPS͒, [11][12][13][14] as an example. The model to which the differential path-length spectra are fitted can be written as:…”

“…Input spectrum a Bcar ͑͒ is the specific absorption coefficient of betacarotene, 17 a HbO2 ͑͒ is the absorption coefficient of fully oxygenated whole blood, and a Hb ͑͒ is the absorption coefficient of fully deoxygenated whole blood. 18 With our experimental setup, [11][12][13][14][15][16] the data acquired in each reflectance spectrum consist of N = 2048 pixel values R data ͑j͒ corresponding to the differential reflectance at wavelengths ranging from 350 to 1000 nm. Since the transmission efficiency of the setup is different for different wavelengths, the signal-to-noise ratio ͑SNR͒ is wavelength dependent.…”

Spectral discrimination of breast pathologies in situusing spatial frequency domain imaging

“…The scattering spectrum can distinguish tissue pathologies when the optical signal is sampled locally [14-18] or filtered by using polarization techniques [19-21] to minimize the collection of multiply scattered light. Localized methods, such as optical coherence tomography (OCT) [22], Raman spectroscopy [23], and confocal sampling [24,25], have been applied to surgical-margin assessment, but they are fundamentally limited in depth sampling by scattering attenuation in tissue.…”