請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86462
標題: | 以非線性顯微術擷取染色之人腦組織影像:評估膠質瘤手術邊界 Nonlinear Microscopy Imaging of Hematoxylin-Eosin Stained Whole-Mount brain Tissues: Assessment of surgical Margins in glioma removal Surgery |
作者: | 曾耀賝 Yao-Chen Tseng |
指導教授: | 孫啟光 Chi-Kuang Sun |
關鍵字: | 術中腫瘤評估,快速新鮮樣品染色法,人體腦部檢體,具十億畫素之非線性光學介觀顯微鏡,高解析度病理影像, intraoperative tumor assessment,robust fresh whole-mount tissue staining method,human brain tissue,mesoscale nonlinear optical gigascope,ultra-high-resolution pathological image, |
出版年 : | 2022 |
學位: | 碩士 |
摘要: | 在癌症的眾多治療方法中,手術為最能完整切除癌化組織的方式,但在有功能保留的組織器官,像是人體中樞神經系統中的腦部,術中能精確並有效率的分辨腫瘤及正常組織,以達到腫瘤完整切除及盡可能保留正常組織的雙重目的,便顯得相當重要。當今腫瘤切除手術的術中黃金標準流程為冷凍病理,其中樣品的製備過程需要花費相當長的處理時間,除有可能將真實的手術邊界於樣品處理流程中切除,也可能因冷凍過程造成假影。術後將會將檢體進行脫水、石蠟包埋等,以製作成術後判讀的黃金標準-石蠟切片,供病理科醫師於術後進行評估,並記錄下詳細的病理記錄。目前國際的眾多研究團隊倡議使用光學新鮮組織切片術來取代冷凍病理,但目前發表之技術除了未能良好地符合與黃金標準-石蠟切片的結果,及國際上對於數位病理影像保存與取樣之大面積高解析度規範,且眾多技術或耗時比目前冷凍病理久,或未能真實使用H&E而造成準確度偏低。
為解決此項迫切需要被改善的問題,本項研究首先透過測試老鼠腦的檢體,並模擬臨床的流程,以達到染色、實驗條件上的優化。而利用其結果,最後開發之the-RFP (true-H&E rapid fresh pathology)技術,透過改良實驗室過去開發之傳統H&E整體組織染色法的快速新鮮樣品染色流程,使非常軟的生物樣品也能進行整體組織的染色,在無需冷凍切片、能保存真實的手術邊緣區域的情況下,直接把術中取下極軟的的腦部檢體進行標準病理染劑的快速染色,再將檢體的最外緣平放於具十億畫素的非線性光學介觀顯微鏡之載物台上進行觀察,並結合以無須任何後處理的光學虛擬切片即時大面積超高解析度成像於螢幕,其檢體染色加上光學切片成像時間小於10分鐘,相當於傳統冷凍切片不到三分之一的時間,即可產生具各種組織病理資訊的大面積高解析度數位影像,以供病理師進行影像判讀。此外該影像在具極高畫素的狀態下,保持了國際數位病理影像所規範之亞微米等級的解析度,能把檢體上的病理細節鉅細靡遺地記錄,也可以提供未來的反覆讀取與後續留存備查。 為證實本技術是否能提供真實H&E、超高解析度、無任何假影且記錄各種組織上的病理特徵之新鮮病理影像,在台大醫院的腦癌手術邊緣臨床試驗中,此研究測試了50個人腦檢體,其中包含了25個經4個患者的腫瘤移除手術而切下的膠質瘤檢體,以及25個經冷凍儲存的正常人腦檢體,並將測試之影像與對應標準病理切片影像提供給台大醫院的病理科醫師進行無訓練的盲測判讀。而病理科醫師不僅能在本項新鮮病理影像技術的影像上快速且準確解讀各種組織病理學上的典型特徵,更在50個檢體的病理判讀上達到了100%的準確率。此項結果驗證了本技術能夠於現今腫瘤手術的術中鑑定過程能提供所急需之快速且準確的病理診斷,若未來使用於臨床,不僅能大幅降低手術時間,提高手術之精準度,更能有效提高醫院手術術中診斷過程之效率,以盡可能地挽救患者們的生命。 Among the multiple treatments for cancer, surgery is the most effective way to remove cancerous tissue completely. However, in some vital organs with preserved functions, such as the brain in the human central nervous system, it is very important to achieve the dual purpose of complete tumor resection and preservation of normal tissue as much as possible. The standard procedure for intraoperative tumor assessment is cryopathology, in which the sample preparation process takes a considerable amount of processing time and manpower. In addition to the possibility of excising the real surgical border in the sample preparation process, it may also cause artifacts due to the freezing process. After surgery, the removed specimen will be dehydrated, embedded in paraffin, etc., to make the gold standard for postoperative specimen assessment – formalin-fixed paraffin-embedded (FFPE) section, for the pathologist to evaluate and record pathological details and interpretion. At present, many international research teams advocate the use of optical microscopy which can do fresh tissue virtual-sectioning to replace cryopathology, but the published technologies are unable to meet the FFPE section results, do sub-micron resolution imaging under the whole-slide imaging guideline to preserve the digital pathological image details. and may take longer than current cryopathology, or fail to reach same contrast as H&E images, resulting in low accuracy. In order to solve this problem that is urgent to be improved, this study first tested mouse brain samples and tried to simulate the clinical process to achieve the optimization of staining and experimental conditions. Based on those results, we finally develop the-RFP (true-H&E rapid fresh pathology), which include the improvement with the robust fresh sample staining process of traditional H&E whole-mount tissue staining method developed in our laboratory in the past, to enable soft biological samples to be processed. For the preparation of the tissues, it is no need to do microtoming, so the real surgical margin area can be preserved. The extremely soft brain specimens were directly removed from the operation, transferred to laboratory and did true-H&E rapid soft tissue staining with standard clinical dyes, and the superficial surface of the specimens were stained. Later on, the edge was placed on the sample holder of a mesoscale nonlinear optical gigascope for observation, combining with the optical virtual sectioning without any post-processing, so the real-time large-area ultra-high-resolution images were performed on the screen. Besides, the total processing time of staining and imaging was less than 10 minutes, which was equivalent to less than one-third of the time of traditional frozen section, and a large-area high-resolution digital image with various detailed histopathological information can be generated for pathologists to perform pathological interpretation, and can also be repeated reading in the future with digital storage in the computer. To confirm whether this technology can provide true H&E, ultra-high resolution, fresh-tissue-preparation pathological images without any artifacts, and record pathological features on various tissues, this study tested 50 specimens in a clinical trial with National Taiwan University Hospital. The brain specimens included 25 glioma specimens excised from 4 subjects acquired from patients’ tumor removal surgeries, and 25 normal human brain specimens stored in the -80°C refrigerator. The-RFP images and the corresponding FFPE bright field images were provided to pathologists at National Taiwan University Hospital for pathological interpretation as a blind-test. The pathologist can not only quickly and accurately interpret the typical features of various histopathology on the the-RFP images of this new pathological imaging technology, but also achieve 100% accuracy in the pathological assessment of 50 specimens. This result verifies that this technology can provide the fast and accurate pathological procedures in the current intraoperative tumor assessment in tumor removal surgery. If it is conducted in clinical use in the future, it can not only greatly reduce the operation time, improve the accuracy of the operation with more effective preparation to save the lives of patients as much as possible. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/86462 |
DOI: | 10.6342/NTU202204215 |
全文授權: | 同意授權(全球公開) |
電子全文公開日期: | 2027-09-29 |
顯示於系所單位: | 光電工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-110-2.pdf 此日期後於網路公開 2027-09-29 | 15.77 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。