Robert N. Graf

Determining nuclear morphology using an improved angle-resolved low coherence interferometry system

We outline the process for determining the morphology of subsurface epithelial cell nuclei using depth-resolved light scattering measurements. The measurements are accomplished using a second generation angle-resolved low coherence interferometry system. The new system greatly improves data acquisition and analysis times compared to the initial prototype system. The calibration of the new system is demonstrated in scattering studies to determine the size distribution of polystyrene microspheres in a turbid sample. The process for determining the size of cell nuclei is discussed by analyzing measurements of basal cells in a sub-surface layer of intact, unstained epithelial tissue.

Dual window method for processing spectroscopic optical coherence tomography signals with simultaneously high spectral and temporal resolution.

Current methods for analysis of spectroscopic optical coherence tomography (SOCT) signals suffer from an inherent tradeoff between time (depth) and frequency (wavelength) resolution. Here, we present a dual window (DW) method for reconstructing time frequency distributions (TFDs) that applies two orthogonal Gaussian windows that independently determine the spectral and temporal resolution. The effectiveness of the method is demonstrated in simulations and in processing of measured OCT signals that contain fields which vary in time and frequency. The DW method yields TFDs that maintain high spectral and temporal resolution and are free from the artifacts and limitations commonly observed with other processing methods.

Prospective grading of neoplastic change in rat esophagus epithelium using angle-resolved low-coherence interferometry

Angle-resolved low-coherence interferometry (a/LCI) is used to obtain quantitative, depth-resolved nuclear morphology measurements. We compare the average diameter and texture of cell nuclei in rat esophagus epithelial tissue to grading criteria established in a previous a/LCI study to prospectively grade neoplastic progression. We exploit the depth resolution of a/LCI to exclusively examine the basal layer of the epithelium, approximately 50 to 100 microm beneath the tissue surface, without the need for exogenous contrast agents, tissue sectioning, or fixation. The results of two studies are presented that compare the performance of two a/LCI modalities. Overall, the combined studies show 91% sensitivity and 97% specificity for detecting dysplasia, using histopathology as the standard. In addition, the studies enable the effects of dietary chemopreventive agents, difluoromethylornithine (DFMO) and curcumin, to be assessed by observing modulation in the incidence of neoplastic change. We demonstrate that a/LCI is highly effective for monitoring neoplastic change and can be applied to assessing the efficacy of chemopreventive agents in the rat esophagus.

Nuclear morphology measurements using Fourier domain low coherence interferometry

We present a new common path configuration Fourier domain low coherence interferometry (fLCI) optical system and demonstrate its capabilities by presenting results which determine the size of cell nuclei in a monolayer of T84 epithelial cells. The optical system uses a white light source in a modified Michelson interferometer and a spectrograph for detection of the mixed signal and reference fields. Depth resolution is obtained from the Fourier transform of the measured spectrum which provides the axial spatial cross-correlation between the signal and reference fields. The spectral dependence of scattering by the samples is determined by windowing the spectrum to measure the scattering amplitude as a function of wavenumber. We present evidence that fLCI accurately measures the longitudinal profile of cell nuclei rather than the transverse profile.

Parallel frequency-domain optical coherence tomography scatter-mode imaging of the hamster cheek pouch using a thermal light source.

We use a parallel frequency-domain optical coherence tomography (FDOCT) system to generate a scatter-mode image of the hamster cheek pouch epithelium. To our knowledge, this is the first optical coherence tomography (OCT) image of a biological sample obtained using a thermal light source in the frequency domain. The system employs an imaging spectrometer to acquire depth-resolved profiles from adjacent spatial points without the need for any scanning. To enable this imaging modality, we have considered that signals originating from multiple depths combine in a different manner in FDOCT compared to time-domain optical coherence tomography (TDOCT). Because a multicomponent FDOCT signal is a coherent sum, it is necessary to limit the number of modes that contribute to the detected signal. Conversely, multicomponent TDOCT signals can be represented as incoherent sums, where increasing the number of modes improves the signal.

Temporal coherence and time-frequency distributions in spectroscopic optical coherence tomography

Traditional analysis of spectroscopic optical coherence tomography (SOCT) signals is limited by an uncertainty relationship between time (depth) and frequency (wavelength). The use of a bilinear time-frequency distribution for analysis, such as those that compose Cohens class of functions, may provide a way to avoid this limitation. Here we present the relationship between traditional SOCT analysis and the relevant Cohen class functions: the Wigner and Choi-Williams distributions. While cross terms that arise in these bilinear time-frequency distributions have been viewed as an artifact, here we identify these terms with temporal coherence, which contains significant information about the signal through phase relationships. The utility of time-frequency distributions is illustrated through analysis of calculated signals.

Detecting precancerous lesions in the hamster cheek pouch using spectroscopic white-light optical coherence tomography to assess nuclear morphology via spectral oscillations.

We have developed a novel dual-window approach for spectroscopic optical coherence tomography (OCT) measurements and applied it to probe nuclear morphology in tissue samples drawn from the hamster cheek pouch carcinogenesis model. The dual-window approach enables high spectral and depth resolution simultaneously, allowing detection of spectral oscillations, which we isolate to determine the structure of cell nuclei in the basal layer of the epithelium. The measurements were executed with our parallel frequency domain OCT system, which uses light from a thermal source, providing high bandwidth and access to the visible portion of the spectrum. The structural measurements show a highly statistically significant difference between untreated (normal) and treated (hyperplastic/dysplastic) tissues, indicating the potential utility of this approach as a diagnostic method.

Nuclear morphology measurements using Fourier domain low-coherence interferometry

We have developed Fourier domain low coherence interferometry (fLCI), a novel optical interferometry method for obtaining depth-resolved spectral information, specifically for the purpose of determining the size of scatterers by measuring their elastic scattering properties. The optical system achieves depth resolution by using coherence gating, enabled by the use of a white light source in a Michelson interferometer and detection of the mixed signal and reference fields with a spectrograph. The measured spectrum is Fourier transformed to obtain the axial spatial cross-correlation between the signal and reference fields providing depth-resolution. The spectral dependence of scattering by the sample is determined by windowing the spectrum to measure the scattering amplitude as a function of wavenumber (k = 2 Pi / lambda, where lambda is the wavelength). We present a new common path confgiuration fLCI optical system and demonstrate its capabilities by presenting results which determine the size of cell nuclei in a monolayer of T84 epithelial cells.

Development of a pre-clinical Fourier domain low coherence interferometry system

Fourier domain low coherence interferometry (fLCI) is an optical technique which combines the depth resolution of low coherence interferometry with the sensitivity of light scattering spectroscopy. The fLCI system uses a white light source in a modified Michelson interferometer with a spectrograph for detection of the mixed signal and reference fields. Depth-resolved structural information is recovered by performing a short-time Fourier transform on the detected spectrum, similar to spectroscopic optical coherence tomography, and analyzing the wavelength dependent variations in scattered light as a function of depth. fLCI has been demonstrated as an excellent technique for probing the nuclear morphology of a monolayer of in vitro cancer cells. We have built a new fLCI optical system which implements an imaging spectrograph for detection and a 4- F interferometer which uses a 4-F imaging system to re-image light scattered from the experimental sample onto the slit of the imaging spectrograph. The new system has allowed us to measure light scattered from the deepest layers of thick scattering samples, such as tissue phantoms and thick animal tissues, for the first time. We now take the first steps to quantitatively determine the diameter of scatterers within a thick experimental sample using the new fLCI system along with the fLCI data processing technique.

Dual window method for processing spectroscopic optical coherence tomography signals with high spectral and spatial resolution

The generation of spectroscopic optical coherence tomography (SOCT) signals suffers from an inherent trade off between spatial and spectral resolution. Here, we present a dual window (DW) method that uses two Gaussian windows to simultaneously obtain high spectral and spatial resolution. We show that the DW method probes the Winger time-frequency distribution (TFD) with two orthogonal windows set by the standard deviation of the Gaussian windows used for processing. We also show that in the limit of an infinitesimally narrow window, combined with a large window, this method is equivalent to the Kirkwood & Richaczek TFD and, if the real part is taken, it is equivalent to the Margenau & Hill (MH) TFD. We demonstrate the effectiveness of the method by simulating a signal with four components separated in depth or center frequency. Six TFD are compared: the ideal, the Wigner, the MH, narrow window short time Fourier transform (STFT), wide window STFT, and the DW. The results show that the DW method contains features of the Wigner TFD, and that it contains the highest spatial and spectral resolution that is free of artifacts. This method can enable powerful applications, including accurate acquisition of the spectral information for cancer diagnosis.