光譜式光學同調斷層掃描術對大鼠角膜光學性質之分析

Abstract

本論文建立了以Mirau 全域式光學同調斷層掃瞄系統(Optical Coherence Tomography;OCT)架構對應的光譜式OCT 分析(Spectroscopic OCT; S-OCT)程序，並且開發使用該系統產生的光譜資訊來計算樣本光學性質參數的演算法。整套技術建於原本的OCT 系統架構上，不需額外硬體改造，且掃描的生物樣本可以保持其原始完整組織結構，不需做切片等任何侵入式的破壞。 本論文的 Mirau 全域式OCT 系統光源為實驗室自行生長的摻鈰釔鋁石榴石(Ce3+:YAG)晶體光纖的自發輻射(Spontaneous Emission; SE)，能產生中心波長560 nm，頻寬98 nm的光源。系統具備縱向1.07 μm，橫向1.35 μm 的高解析度，於掃描大鼠角膜檢體樣本時，能夠清楚解析角膜各個重要的分層與細胞結構。為了確定OCT 掃描機構的壓電掃描器，於掃描取樣點的位移能保持正確，使用OCT 掃描光學仿體所得的干涉訊號來量測掃描過程的位移曲線，對壓電掃描器非線性特性進行校正，達成控制每單位位移與理想值9.8%之誤差。 針對使用的 Mirau 全域式OCT 系統建立的光譜式OCT 分析程序，產生針對樣本中一段8 μm 厚度的局部區域，產生該位置對應的光譜資訊，並擁有約14 nm 之頻域解析度。以光譜式OCT 分析產生的系統光源光譜，與光譜儀量測結果吻合，僅受到其光譜解析度限制而使量測頻寬略高於光譜儀之光譜。 針對 Mirau 全域式OCT 系統量測樣本中的一層介質，依據Fresnel 方程式建立該介質前、後界面干涉訊號的兩個強度與兩訊號相位差與波長關係之數值模型，建立透過量測數值與模型擬合的演算法來計算出該樣本介質的折射率、衰減係數隨波長變化，以及該介質層的物理厚度。利用模擬數據進測試，解決光譜式OCT 分析產生的相位具有相位未定的問題。使用已知性質的Fused silica 樣本來做整套技術的驗證。與文獻參考數據比較，整體折射率計算值約僅2%，厚度5.5%的誤差。最後，針對大鼠完整眼球檢體中角膜的部分，進行OCT 掃描以及計算其樣本性質參數，展現整套技術用於分析生物組織的可能性，量測結果統計上，折射率的標準差為計算平均值之0.8%，衰減係數標準差為2.3 × 10-4，物理厚度標準差比上平均值的百分比為4.8%，呈現量測的可重複性。

Spectroscopic optical coherence tomography (S-OCT) analysis technique was implemented for Mirau-based full-field OCT (FF-OCT). An algorithm for calculating a layer of medium in the scanned sample, using spectral information generated with S-OCT, was also developed. The whole technique used the original OCT setup, without a need for further hardware modification, and the scanned bio-samples were not sectioned samples, still preserving its original structural integrity during measurement. The Mirau-based FF-OCT system used a lab-grown Ce3+:YAG crystal fiber as its light source. The generated 98 nm broadband spontaneous emission and 560 nm center wavelength allowed the OCT system to have 1.07 μm and 1.35 μm resolution in axial and lateral direction, respectively. Using the OCT system to scan and ex vivo rat cornea, important corneal structures and cell arrangement were clearly visible, demonstrating the high resolution advantage of the OCT system. The OCT system utilized a piezo actuator as its scanning mechanics. For piezo scanning device having non-linearity effects, a calibration was performed to ensuring that during OCT scanning the scanning displacements were equally sampled. A novel method for displacement measurement were implemented using OCT scan signal of an optical phantom. After the calibration, the unit scan displacement differed from ideal value by 9.8%. S-OCT analysis procedure for the Mirau-based FF-OCT system was implemented. The technique allowed generation of a spectrum with 8 nm spectral resolution, for an 8 μm localized depth of the scanned sample signal. S-OCT-generated spectrum of the OCT light source matched the one measured with an optical spectrum analyzer, barring a slightly extended measured FWHM bandwidth. The difference was caused by spectral resolution limitation, which was due to inherent trade-off of time-spectral resolution of the S-OCT. Theoretical models for Mirau-based FF-OCT interference signals of a medium within the scanned sample were built, based on Fresnel equations. An algorithm for characterizing the refractive index and extinction coefficient dispersion, and the physical thickness of that sample medium was developed. Verifications of the algorithm were performed by simulated interference signals and OCT signals of a known-characteristics sample. Comparing with reference value, the overall measured refractive index had error of 2.0%, and for thickness 5.5%. Lastly, the whole technique was applied to a biological sample. The demonstration was done on the corneal layer of an intact ex vivo rat eye. The measured results from different lateral position and the same OCT field of view were compared. Defined a minimum absorption threshold of 5.2 × 10-3 μm-1 for positive κ measured results. Also, statistically, overall measured refractive index had a standard deviation of 0.8%, and measured thickness standard deviation 4.8%. The results demonstrated the possibility of S-OCT analysis technique with FF-OCT system on biological samples.