以三倍頻顯微術觀察並辨別白血球細胞

Abstract

血液是醫療檢查中重要的評鑑指標，紅血球、白血球和血小板的組成比例與數量都有其健康意涵，醫院對於血液的分析，多以抽血方式，再經由流式細胞儀等分析儀器進行樣本分析，此屬侵入式醫療行為，不只對受測者帶來心理負擔，也易讓抽離體外的檢體因外在因素而變質，進而影響判讀之正確性，未來，我們希望利用非線性光學顯微術，以非侵入式檢測對各類血球統計與分析，解決現有問題。 為達此弘遠目標，本研究扮演開山始祖的角色，著重血球在非線性光學下的特徵差異與辨識條件，因非線性光學顯微術具有次微米等級三維空間解析能力，可做深層光學成像，不具生物傷害性，且不需經由染劑標定，省去染劑標定可使血球更近於真實樣貌，本研究利用摻鉻貴橄欖石晶體雷射做為激發光源，擷取受測物所發出的倍頻訊號，經由FPGA(ALTERA DE4)將類比訊號轉為數位訊號且擷取影像，最後由Visual Studio撰寫出的影像軟體，成像與影像優化，並利用此系統觀察血球的各項特徵。 在三倍頻的成像下，紅血球與血小板有很強的三倍頻訊號，因尺寸差異很容易將兩者辨別，白血球三倍頻訊號較弱，主要又可分成中性粒細胞(Neutrophil)、淋巴球(Lymphocyte)、單核球(Monocyte)，在影像特徵下，中性粒細胞比淋巴球、單核球三倍頻訊號更強，細胞內有多個小核體，也常出現明顯細胞膜結構，淋巴球細胞最小，細胞內有一大圓核，且細胞內常出現一圓形亮點，單核球細胞最大，細胞核不規則狀，影像特徵可以輔助辨識，但為達更快速的辨別，我們以細胞大小、三倍頻強度和自體紅螢光強度做為辨別條件，三倍頻強度以中性粒細胞最強，單核球、淋巴球無顯著差異，自體紅螢光強度依然為中性粒細胞最強，單核球居次，淋巴球最弱，透過這些差異可有效將三類白血球分成三群。 雖目前成果離非侵入式活體血球辨識仍有段差距，但本研究結果可輔助流式細胞儀對血球的分選，流式細胞儀常為達更高的辨別正確性，使用染劑的添加，但染劑的添加不只經費負擔，對細胞的變異也存在風險，利用本研究結果，不僅可增加流式細胞儀對血球細胞的篩選條件，並可提供影像的特徵辨別，讓分選正確率再次提升。

Blood is an important index of medical examination. The percent composition of red blood cells, white blood cells and platelets in blood is related to health. In the hospital, drawing blood has been a dominant way of health exam. The method is an invasive medical practice, and vitro specimen tends to be affected by the external factors, which is possible to affect the accuracy of the estimation. To overcome the problems mentioned, we use nonlinear optical microscopy to gather statistics and analysis of all types of blood cells with non-invasive detection in the future. My research focuses on the differences of characteristics of blood cells observed by nonlinear optical microscopy and broadening the conditions of identification. Equipped with sub-micron grade three-dimensional analysis, non-linear optical microscopy can be applied to get deep optical images and causes no biological damages. In addition, no dye is included in the research, so the blood cells remain almost intact. In the research, we use Cr:forsterite laser as the excitation source and further to collect THG signals emitted by target objects. Then, we capture the images generated by the transition from analogical signals to digital signals processed by the FPGA (ALTERA DE4). Finally, the optimized images, generated by an image software (programmed by visual studio) are adopted to observe the characteristics of blood cells. In the third harmonic generation images, the red blood cells and platelets show a strong third harmonic generation signal. It is easy to identify them because of the size difference between the two. As for Leukocyte, it can be further divided into three types: neutrophil, lymphocyte and monocyte. Unlike red blood cells and platelets, the THG signal of Leukocyte is weaker. Because of the characteristics observed from optimized images, we can easily identify the specific type of Leukocyte. It is observed that neutrophil THG signal is stronger than lymphocyte and monocyte. Neutrophil has multiple small nucleus, and the cell membrane structure is obvious. In terms of the size of lymphocyte, it is the smallest and its circle nuclear is bigger, and oftentimes, it can be identified because of a circular bright spot. Monocyte, being the largest in size, has an irregular-shaped nuclear. Because of the optimized images, identification of cells become easy. In order to achieve more rapid identification. The size of cells, and the intensity of THG and the intensity of fluorescence condition identification of specific types of cells. In terms of the intensity of THG, Neutrophil exhibits the strongest, while monocyte and lymphocyte shows no significant difference. Concerning the intensity of fluorescence, neutrophil exhibits the strongest and lymphocyte the weakest. With the distinctive features of cells mentioned above, the three types of white blood cells can be effectively identified and further divided into three groups. Although much remains to be done, but this study is still valuable in that the empirical results enable the flow cytometry to sort blood cells effectively. Sorting blood cells through flow cytometry is usually conducted with dye; the use of dye is consuming and the most importantly, it is possible to cause cell mutation. With the result of this study, the broadened conditions on blood cells sorting conducted by flow cytometry and the features observed by the optimized images surely improve the accuracy of blood cells sorting.