臨床三倍頻顯微術: 表皮游離神經末梢之光學虛擬切片

Abstract

周邊神經病變是臨床上常見的神經纖維受損相關疾病，它會導致周邊神經纖維的麻木、刺痛及慢性疼痛，並隨著時間推移而惡化，若未能及早診斷病因並正確治療，病情惡化可能導致死亡。周邊神經病變在初期極易被忽略，目前已知病因有數百種之多，正確診斷方式以了解病因極為重要。 臨床上，糖尿病周邊神經病變為大宗，超過60 %糖尿病患者之神經系統會由表皮神經末梢開始退化(degeneration)而併發周邊神經病變，未及早治療將引起嚴重的足部問題，例如畸形、感染、潰瘍，更嚴重的患者須進行截肢。其中以小纖維感覺神經系統受損為主，常以皮膚切片(俗稱皮膚神經活檢)作為小神經疾病診斷及嚴重程度的準則，也是目前臨床上能觀測皮膚游離神經末梢纖維(free intraepidermal nerve endings，簡稱FINEs)結構的唯一方法。此觀測神經末梢纖維結構至關重要，目的在定量腳踝表皮神經纖維密度(intra-epidermal nerve density，簡稱IENFD)，並以此關鍵資訊從事臨床決斷。由於腳踝皮膚的神經活檢技術門檻相當高，臨床上只有大型教學醫院才有能力從事皮膚神經活檢，並無可非侵入式觀察皮膚神經末梢的臨床工具。此外，皮膚切片為侵入性方法，不建議用於糖尿病等凝血功能異常病患，且同時存在取樣錯誤及組織處理過程錯誤而造成的偽影，導致誤診等問題。目前非侵入式方式只能在人體和動物皮膚內對髓鞘大神經纖維進行成像，但無法用於小神經纖維結構影像。基於周邊神經病變之龐大病患數，在美國單一國家即超過4000萬人以上，一項能夠不需要切除組織，不造成皮膚傷害，但卻能夠獲取皮膚下末梢神經結構影像且提供高穿透度、高解析度、量化的病理級臨床神經影像是目前亟需的。 本研究將基於三倍頻顯微術開發出全球唯一能在無需取出皮膚與染色的情況下，能成功非侵入式於人體皮膚中取得關鍵無髓鞘之表皮游離神經末梢影像之技術。影像之正確性透過成功從事ex vivo與皮膚神經活檢之比對，即在同一離體皮膚組織切片中完成與PGP 9.5化學免疫病理染色得到證實。此體外研究結果表明此影像技術具有無需外加染劑在人體皮膚中直接觀測表皮神經末梢纖維結構之能力。透過在不同小鼠的腳趾中進一步注射神經染劑-亞甲基藍(methylene blue)及縱向追蹤神經受損動物模型，再次驗證此影像技術中的信號來自表皮游離神經末梢纖維，並同時確立活體應用之可行性。 在臨床試驗中，本研究同時建立了對表皮內神經纖維進行量化之操作型定義，即表皮神經纖維指數(IENF index)。經由臨床試驗分析，健康受試者及糖尿病神經病變患者的表皮神經纖維指數具有顯著差異，並與皮膚神經活檢技術才能提供之關鍵表皮神經纖維密度(IENF density)具有相關性。研究結果表明，三倍頻顯微術即為目前亟需的無創、可逆、定量和微觀神經測定臨床影像技術，以作為未來神經內外科之診斷與治療評估之有效追蹤工具。

Peripheral neuropathy (PN) represents a spectrum of neurological diseases that damage the peripheral nervous system of human body, for which hundreds of etiologies have been identified. One of the leading causes for the development of PN is diabetes. Diabetic peripheral neuropathy (DPN) also was a major complication of diabetes and a frequent cause of foot ulcers and amputations. Most of the studies indicated DPN has a predilection for small sensory nerve fibers, suggesting small fiber neuropathy (SFN) as an early stage of DPN, and terminal small nerve fibers of lower distal extremities are the early targets for screening and diagnosis. Skin biopsy was the current gold standard to provide structural information to provide free-intraepidermal-nerve-endings (FINEs) structural information for the differential diagnosis of SFN. Its invasive nature was particularly unfavorable for patients with diabetic coagulation abnormalities thus there is an unmet clinical need for a non-invasive FINEs imaging tool. However, the previous study reports that the current in vivo optical imaging tools can only provide contrast through myelin sheath, a clear indication that no techniques are available to in vivo image human fibers without myelin sheath, especially unmyelinated FINEs for human skin. In this study, we developed the world's first-ever unmyelinated FINEs visualizer for the lower distal extremities of human skin by using a tightly-focused epi-Third-harmonic-generation microscope (TFETM). Its label-free capability was confirmed by PGP9.5 immunohistochemistry staining and a longitudinal spared nerve injury model study. The proposed TFETM was shown to be able to delineate the 3D structure of unmyelinated FINEs that are confirmed by PGP 9.5 immunohistochemistry staining from the same tissue section of human skin. Moreover, through multiple modalities nonlinear microscopy which combined the third-harmonic generation (THG) and two-photon fluorescence (2PF) channels with its clinical capability proved. We take advantage of the far-red emission of nerve-specific methylene blue (MB) stain with a two-photon excitation wavelength of 1260 nm, not only capable to provide the highest optical penetration in skin and to avoid melanin absorption but also capable of simultaneously detecting the epi-THG signals (around 420nm) and 2PF signals of MB (around 685nm). With this character in optics and clinical applications, we first inject the MB solution of consistent concentrations into the third toe of the hind limb of wild-type mice and examined in vivo the correlation between the fluorescence of MB signal and epi-THG signals to double confirm whether the fiber-like signals were generated from nerves fiber. Moreover, through proposing a dot-connecting algorithm, we established the operational protocol to count three-dimensionally the intraepidermal nerve fibers (IENF) and define the quantitative IENF index. Our clinical trial showed that the label-free IENF index can differentially identify SFN (P-value=0.0102) and was well correlated with IENF density of skin biopsy (Pearson’s correlation, R-value= 0.98) in the DPN group. Our study suggested that the unstained dot-connecting TFETM imaging can noninvasively provide FINEs structure information assisting in diagnosing SFN. This unprecedented non-invasive imaging system would meet the unmet clinical need for the FINEs imaging tool not only to assist in the screening and differential diagnosis of DPN but also for surgical evaluation and efficacy assessment of radiculopathy treatment and therapy in the future.