Hsiang-Chieh Lee

Effective indicators for diagnosis of oral cancer using optical coherence tomography

A swept-source optical coherence tomography system is used to clinically scan oral precancer and cancer patients for statistically analyzing the effective indicators of diagnosis. Three indicators are considered, including the standard deviation (SD) of an A-mode scan signal profile, the exponential decay constant (alpha) of an A-mode-scan spatial-frequency spectrum, and the epithelium thickness (T) when the boundary between epithelium and lamina propria can still be identified. Generally, in abnormal mucosa, the standard deviation becomes larger, the decay constant of the spatial-frequency spectrum becomes smaller, and epithelium becomes thicker. The sensitivity and specificity of the three indicators are discussed based on universal and individual relative criteria. It is found that SD and alpha are good diagnosis indicators for moderate dysplasia and squamous cell carcinoma. On the other hand, T is a good diagnosis indicator for epithelia hyperplasia and moderate dysplasia.

Differentiating oral lesions in different carcinogenesis stages with optical coherence tomography

A swept-source optical coherence tomography (SS-OCT) system is used to clinically scan oral lesions in different oral carcinogenesis stages, including normal oral mucosa control, mild dysplasia (MiD), moderate dysplasia (MoD), early-stage squamous cell carcinoma (ES-SCC), and well-developed SCC (WD-SCC), for diagnosis purpose. On the basis of the analyses of the SS-OCT images, the stages of dysplasia (MiD and MoD), and SCC (ES-SCC and WD-SCC) can be differentiated from normal control by evaluating the depth-dependent standard deviation (SD) values of lateral variations. In the dysplasia stage, the boundary between the epithelium (EP) and lamina propria (LP) layers can still be identified and the EP layer becomes significantly thicker than that of normal control. Also, in a certain range of the EP layer above the EP/LP boundary, the SD value becomes larger than a certain percentage of the maximum level, which is observed around the EP/LP boundary. On the other hand, in the ES-SCC and WD-SCC stages, the EP/LP boundary disappears. Because of the higher density of connective tissue papillae in the ES-SCC stage, the SD values of the slowly varying lateral scan profiles in the ES-SCC samples are significantly larger than those in the WD-SCC sample. Also, ES-SCC can be differentiated from WD-SCC by comparing the exponential decay constants of averaged A-mode scan profiles. Because of the higher tissue absorption in the WD-SCC lesion, the decay constants in the WD-SCC samples are significantly higher than those in the ES-SCC samples.

Delineation of an oral cancer lesion with swept-source optical coherence tomography

We demonstrate the ex vivo imaging of an oral cancerous sample with a swept-source optical coherence tomography (SS-OCT) system. With the axial resolution of 8 microm in free space and system sensitivity of 108 dB, we can well distinguish the normal and abnormal tissue portions in a sample. In particular, we analyze the lateral variation of A-scan profiles to show two parameters of SS-OCT signal for delineating an oral cancer lesion. One of the parameters is the decay constant in the exponential fitting of the SS-OCT signal intensity along depth. This decay constant decreases as the A-scan point moves laterally across the margin of a lesion. The other parameter is the standard deviation of the SS-OCT signal intensity fluctuation in an A-scan. This parameter increases significantly when the A-scan point is moved across the transition region between the normal and abnormal portions. Such parameters are useful for determining the margins of oral cancer.

Diagnosis of oral submucous fibrosis with optical coherence tomography

The epithelium (EP) thickness and the standard deviation (SD) of A-mode scan intensity in the laminar propria (LP) layer are used as effective indicators for the diagnosis of oral submucous fibrosis (OSF) based on the noninvasive clinical scanning of a swept-source optical coherence tomography (OCT) system of approximately 6 mum in axial resolution (in tissue) and 103 dB in sensitivity. Compared with the corresponding parameters in healthy oral mucosal mucosa, in OSF mucosa, the EP thickness becomes smaller and the SD of A-mode scan intensity in the LP layer (LP SD) also becomes smaller. The LP SD can also be used for effectively differentiating OSF (small LP SD) from lesion (large LP SD). This application is particularly useful in the case of a lesion without a clear surface feature. Meanwhile, the use of the SD of A-mode scan intensity in the EP layer (EP SD) can further help in differentiating OSF (medium EP SD) from healthy oral mucosal (small EP SD) and lesion (large EP SD) conditions. Compared with the conventional method of maximum mouth opening measurement, the use of the proposed OCT scanning results can be a more effective technique for OSF diagnosis.

Study of photon migration with various source-detector separations in near-infrared spectroscopic brain imaging based on three-dimensional Monte Carlo modeling.

We have simulated photon migration with various sourcedetector separations based on a three-dimensional Monte Carlo code. Whole brain MRI structure images are introduced in the simulation, and the brain model is more accurate than in previous studies. The brain model consists of the scalp, skull, CSF layer, gray matter, and white matter. We demonstrate dynamic propagating movies under different source-detector separations. The multiple backscattered intensity from every layer of the brain model is obtained by marking the deepest layer that every photon can reach. Also, the influences of an absorption target on the brain cortex are revealed.

Effective indicators for oral cancer diagnosis based on optical coherence tomography

A swept-source optical coherence tomography system is used to clinically scan oral precancer and cancer patients for statistically analyzing the effective indicators of diagnosis including the signal standard deviation, spatial-frequency spectral shape, and epithelium thickness.

Optical coherence tomography for oral cancer diagnosis

Incorporating with a flexible-holder probe, a time-domain optical coherence tomography system is built to achieve 7 microns in axial resolution, 80-90 dB in sensitivity, and sub-sec imaging rate for clinical oral cancer diagnosis.

Near-infrared spectroscopy studies of visual evoked responses during photon stimulation

We demonstrate the utility of near-infrared spectroscopy in monitoring of the human visual cortex. The typical hemodynamic signals expected are observed in our preliminary experiments.

Photon migration in a three-dimension human brain model: Monte Carlo simulation

In this paper, we demonstrate the time-resolved photon migration phenomena in a 3-D human brain model at variable wavelengths with Monte Carlo simulations. In our simulations, the brain model consists of five layers: scalp, skull, cerebrospinal fluid, gray matter, and white matter. The model imitates the geometry of a real human brain obtained from MRI images. Our simulations can reveal the intensities of photons reaching different layers. We can optimize the source-detector separation in experiments by analyzing the partial intensity variation of the different layers in the brain model.

Correlations between temporal and spatial photon distributions in turbid media: Monte-Carlo modeling of time-, aperture-, angle-, and polarization-gating methods

For understanding the correlations between temporal and spatial photon distributions in the process of random scattering in turbid media, the Monte Carlo algorithm is used to describe various optical gating conditions. We study the distributions of transmitted photon with time-, aperture-, and angle-gating techniques. We record the optical paths of transmitted photons with different aperture sizes and angles. Time-resolved profiles and polarization states of transmitted photons are also discussed.