利用全域式光學同調斷層掃描術與自編軟體進行角膜神經的量化分析

Abstract

在眼角膜的檢測當中，建構高解析度的三維影像，可以提供臨床醫師更多的診斷資訊，對疾病做精準的判斷。例如：角膜各層結構之形貌與厚度。除此之外，角膜富含感覺神經，許多的系統性與眼表面疾病，皆與角膜亞基底神經之形貌變化有關。目前文獻上也報導了許多與角膜神經相關的疾病，例如：帕金森氏症、糖尿病與乾眼症。因此精確定量角膜亞基底神經形貌，就有機會對疾病進行早期的診斷。 在本研究當中，使用高空間解析度之全域式光學同調斷層掃描，其具有等向性及微米等級解析度之特性，能夠展示角膜各橫平面、縱平面與三維立體影像。除此之外，本系統還具有大面積掃描的能力，能夠藉由拼接的方式組合多個視域之大面積影像。 藉由檢體小鼠與兔子實驗，證實本系統具有細胞等級的成像能力，且能夠對角膜亞基底神經做定量的分析。在兔子的實驗當中，我們發現了許多影響神經量化的複雜形貌，例如：K結構與基底上皮細胞邊界。而小鼠角膜神經形貌定量結果為：主幹神經密度136.93 ± 19.18 mm-2；分枝神經密度731.00 ± 186.19 mm-2；分枝神經連接點密度608.01 ± 116.99 mm-2；角膜亞基底神經密度61.46 ± 1.78 mm/mm2；神經寬度1.08 ± 0.32 µm；扭曲度0.0681 ± 0.0132；平均分枝點4.57 ± 1.41；分枝主幹比5.53 ± 2.04；K line平均值為 Grade 1；短神經15 ± 3.5根；平行性0.7500 ± 0.0475；單一影像中之平行性標準差0.30 ± 0.02。 利用全域式光學同調斷層掃描，可清楚地拍攝出小鼠角膜亞基底神經細微的結構，並同時得到角膜之縱平面影像。接著透過自行編寫的神經分析軟體TCCMetrics，精確的將神經形貌定量為十二項量化參數。結合以上兩者技術，在臨床上有望促進疾病早期之診斷。

In cornea examination, high-resolution three-dimensional images can provide clinicians with more diagnostic information, such as corneal thickness and cellular morphology. In addition, the cornea is a densely innervated tissue with sensory nerve fibers. Both ocular surface and systemic diseases are associated with corneal nerves. Therefore, accurate quantification can help in disease diagnosis at earlier stages. In this study, we use high spatial-resolution full-field optical coherence tomography (FF-OCT) with isotropic cellular resolution as an examination instrument. This technique can provide images from reconstructed three-dimensional information and a large field of view by stitching tomograms side by side. We validated the imaging ability with ex vivo mice and rabbit, and quantified the corneal sub-basal nerve morphology by twelve parameters. The quantitative results of ex vivo mice are as follows: NFD, 136.93 ± 19.18 mm-2; BND, 731.00 ± 186.19 mm-2; BNCD, 608.01 ± 116.99 mm-2; NFL, 61.46 ± 1.78 mm/mm2; NFW, 1.08 ± 0.32 µm; TC, 0.068 ± 0.013; BNCM, 4.57 ± 1.41; BMR, 5.53 ± 2.04; average K line grading, Grade 1; numbers of short nerve fiber, 15 ± 3.5; parallelism, 0.75 ± 0.05; and standard deviation of single-image parallelism, 0.3 ± 0.02. The use of FF-OCT can clearly capture the sub-basal nerve morphology in mice. By means of the self-written neural analysis software TCCMetrics, the nerve's morphology is accurately quantified through twelve parameters. The combination of these two technologies is expected to facilitate the early clinical diagnosis of a wide range of diseases.