Quantitative fluorescence angiography versus hyperspectral imaging to assess bowel ischemia: A comparative study in enhanced reality
Quantitative fluorescence angiography versus hyperspectral imaging to assess bowel ischemia: A comparative study in enhanced reality
复制标题
DOI：
10.1016/j.surg.2020.02.008
复制
发表时间：
2020-07-01
期刊：
影响因子：
3.8
通讯作者：
Diana, Michele
中科院分区：
文献类型：
--
作者：
Barberio, Manuel;Felli, Eric;Diana, Michele
背景资料：基于荧光增强现实是一种软件，通过计算静脉注射吲哚菁绿色后的荧光强度达峰时间，提供定量荧光血管造影。高光谱成像是一种无对比度，光学成像模式，测量组织oxygenation.Methods：在8只猪，缺血性肠段通过划分拱廊分支成像使用高光谱成像和基于荧光增强现实。通过高光谱成像系统获取组织氧合值。随后，在静脉注射0.2 mg/kg吲哚菁绿色后，使用近红外腹腔镜照相机进行荧光血管造影。通过专有软件分析达到峰值的时间荧光信号以实现灌注图。这被叠加到实时图像上，以获得基于荧光的增强现实。同时，选择9个相邻的感兴趣区域并叠加到实时视频上，从而获得基于高光谱的增强现实。叠加了基于增强现实和基于超光谱的增强现实，从而可以比较两种成像模式。在感兴趣的区域对局部毛细血管乳酸水平进行采样。两个预测模型，使用局部毛细血管乳酸水平外推的基础上，两个imaging system.Results：所有区域的利益，平均局部毛细血管乳酸水平为4.67 +/- 4.34 mmol/L，平均组织氧合为45.9 +/-18.9%，平均时间为10 +/- 9.4秒的峰值。在相应的感兴趣区域，基于荧光的增强现实时间-峰值和高光谱成像-组织氧合之间的Pearson检验给出了R = -0.66（P < .0001）。高光谱成像乳酸预测模型被证明比基于荧光的增强现实模型更准确（P < .0001）。结论：使用高光谱成像和荧光血管造影定量肠灌注。高光谱成像比荧光血管造影获得了更准确的结果。基于高光谱的增强现实可能被证明是一种有用的、无对比度的术中工具，可以量化肠道缺血。(C)2020爱思唯尔公司All rights reserved.
Background: Fluorescence-based enhanced reality is a software that provides quantitative fluorescence angiography by computing the fluorescence intensity time-to-peak after intravenous indocyanine green. Hyperspectral imaging is a contrast-free, optical imaging modality which measures tissue oxygenation.Methods: In 8 pigs, an ischemic bowel segment created by dividing the arcade branches was imaged using hyperspectral imaging and fluorescence-based enhanced reality. Tissue oxygenation values were acquired through a hyperspectral imaging system. Subsequently, fluorescence angiography was performed using a near-infrared laparoscopic camera after intravenous injection of 0.2 mg/kg of indocyanine green. The time-to-peak fluorescence signal was analyzed through a proprietary software to realize a perfusion map. This was overlaid onto real-time images to obtain fluorescence-based enhanced reality. Simultaneously, 9 adjacent regions of interest were selected and superimposed onto the real-time video, thereby obtaining hyperspectral-based enhanced reality. Fluorescence-based enhanced reality and hyperspectral-based enhanced reality were superimposed allowing a comparison of both imaging modalities. Local capillary lactate levels were sampled at the regions of interest. Two prediction models using the local capillary lactate levels were extrapolated based on both imaging systems.Results: For all regions of interest, the mean local capillary lactate levels were 4.67 +/- 4.34 mmol/L, the mean tissue oxygenation was 45.9 +/- 18.9%, and the mean time-to-peak was 10 +/- 9.4 seconds. Pearson's test between fluorescence-based enhanced reality-time-to-peak and hyperspectral imaging-tissue oxygenation at the corresponding regions of interest gave an R = -0.66 (P < .0001). The hyperspectral imaging lactate prediction model proved more accurate than the fluorescence-based enhanced reality-based model (P < .0001).Conclusion: Bowel perfusion was quantified using hyperspectral imaging and fluorescence angiography. Hyperspectral imaging yielded more accurate results than fluorescence angiography. Hyperspectral-based enhanced reality may prove to be a useful, contrast-free intraoperative tool to quantify bowel ischemia. (C) 2020 Elsevier Inc. All rights reserved.