頸部多通道近紅外光譜在活體測量及其與腦組織血氧飽和度變化之相關性

Abstract

本研究包含移動式多通道漫反射光譜硬體系統建置、光譜系統整合控制介面製作，由人體漫反射光譜提取血氧飽和度相關特徵，以及比較血氧飽和度相關特徵與腦組織血氧飽和度相關性。移動式多通道漫反射光譜系統以自製的貼片，包含自製探頭，以光源與偵測器間距(source-detector-separation，SDS)=4.5, 7.5, 10.5, 20 mm通道，加上LED光源照射仿體或人體組織，再連續測量漫反射光，以獲取仿體與人體組織漫反射光譜，由人體組織的漫反射光譜可了解如血氧飽和度變化相關的資訊。光譜系統整合控制介面，是用於控制移動式多通道漫反射光譜系統拍攝光譜的介面，控制了拍攝SDS=4.5, 7.5, 10.5 mm通道光譜的EMCCD相機，也控制了拍攝SDS=20 mm通道光譜的QE-Pro光譜儀，可調整如相機曝光時間、拍攝範圍、拍攝總時間與感測器溫度等參數。搭配設定適當的曝光時間，此介面能控制漫反射光譜系統以足夠高的取樣率拍攝SDS=4.5, 7.5, 10.5, 20 mm通道的漫反射光譜，便能觀察人體實驗測量區域—右側脖子內頸靜脈周圍區域，隨血氧飽和度變化的漫反射光譜，以及隨心跳脈搏變化的血壓波形。用整合控制介面收取各SDS通道的人體漫反射光譜後，便能處理人體漫反射光譜，再提取血氧飽和度特徵。漫反射光譜的處理，包含沿波長維度前處理去除雜訊，以及沿時間維度提取心跳脈搏成份。由伐氏操作實驗恢復期漫反射光譜，相對基線階段漫反射光譜計算∆OD光譜，便可進而提取如total hemoglobin與differential hemoglobin等與血氧飽和度相關的特徵。由分析total hemoglobin與differential hemoglobin等特徵，與商用腦血氧儀測量得到的腦組織血氧飽和度相關性，在生理上認知這些特徵與腦組織血氧飽和度具有關聯性的前提下，便可進一步了解這些特徵以什麼方式與腦組織血氧飽和度呈現不同程度的相關。由本研究建立的移動式多通道漫反射光譜系統，搭配整合控制介面測量仿體與人體漫反射光譜訊號，再對人體伐氏操作實驗的漫反射光譜進行處理，與腦組織血氧飽和度分析相關性，大幅推進了非侵入式內頸靜脈血氧儀的開發。

This research proposes a movable multi-channel diffuse reflectance spectra measurement system, an integrated interface to control this hardware system, a tailored workflow to extract the features that can represent the blood oxygen saturation, and analyzing the correlation of these features with the brain tissue’s blood oxygen saturation. The multi-channel diffuse reflectance spectra measurement system is used to measure the reflectance spectra at source-detector-separation(SDS), which represents the distance between the light source and the detector, of 4.5, 7.5, 10.5, 20 mm by a self-designed sensor module and probes. By analyzing the reflectance spectra, we can get insights into the blood oxygen saturation of the measured tissue. The integrated interface controls the EMCCD camera to take pictures of SDS=4.5, 7.5, 10.5 mm and the QE-Pro spectrometer to take photos of SDS=20 mm. One can adjust the parameters, such as exposure time, CCD regions to take photos, how long the reflectance spectra measurement system should take reflectance spectra photos, CCD temperature, etc. The interface can use sampling rates that are fast enough to acquire reflectance spectra that are affected by the blood oxygen saturation of the measured tissue, tissue around the internal jugular vein(IJV) at the neck’s right side, at SDS=4.5, 7.5, 10.5, 20 mm and blood pressure waveforms of IJV that change periodically according to heartbeats. After acquiring reflectance spectra by the developed interface, signal processing is performed and features that represent blood oxygen saturation are extracted. Reflectance spectra can be pre-processed by removing the noise along the wavelength dimension and pulsations caused by heartbeats can be extracted along the time dimension. After pre-processing the in-vivo reflectance spectra and extracting the pulsations of the recovery and baseline stage of the Valsalva maneuver experiment, ∆OD spectra, which are proportional to absorption spectra’s change of the recovery stage with respect to the baseline stage, can be calculated. After ∆OD spectra are calculated, features that represent the blood oxygen saturation, e.g. total hemoglobin and differential hemoglobin can be calculated. By knowing that these features might be correlated to the brain’s blood oxygen saturation physiologically, further analyses can be conducted to realize how these features are related to the brain’s blood oxygen saturation. By building the movable multi-channel diffuse reflectance spectra measurement system; developing the integrated interface which can control different components of this system to acquire reflectance spectra in various settings; designing the workflow to pre-process the in-vivo reflectance spectra to remove noise, extract the pulsations caused by heartbeats, calculate ∆OD spectra, and extract the blood oxygen saturation correlated features; and analyzing the correlation between these features and the brain’s blood oxygen saturation, development of the non-invasive IJV oximeter is greatly progressed.