Hirose Tomoaki, Igami Tsuyoshi, Koga Kusuto, Hayashi Yuichiro, Ebata Tomoki, Yokoyama Yukihiro, Sugawara Gen, Mizuno Takashi, Yamaguchi Junpei, Mori Kensaku, Nagino Masato
Division of Surgical Oncology, Department of Surgery, Nagoya University Graduate School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya, 466-8550, Japan.
Graduate School of Information Science, Nagoya University, Nagoya, Japan.
Surg Today. 2017 Mar;47(3):365-374. doi: 10.1007/s00595-016-1394-5. Epub 2016 Aug 8.
Fusion angiography using reconstructed multidetector-row computed tomography (MDCT) images, and cholangiography using reconstructed images from MDCT with a cholangiographic agent include an anatomical gap due to the different periods of MDCT scanning. To conquer such gaps, we attempted to develop a cholangiography procedure that automatically reconstructs a cholangiogram from portal-phase MDCT images.
The automatically produced cholangiography procedure utilized an original software program that was developed by the Graduate School of Information Science, Nagoya University. This program structured 5 candidate biliary tracts, and automatically selected one as the candidate for cholangiography. The clinical value of the automatically produced cholangiography procedure was estimated based on a comparison with manually produced cholangiography.
Automatically produced cholangiograms were reconstructed for 20 patients who underwent MDCT scanning before biliary drainage for distal biliary obstruction. The procedure showed the ability to extract the 5 main biliary branches and the 21 subsegmental biliary branches in 55 and 25 % of the cases, respectively. The extent of aberrant connections and aberrant extractions outside the biliary tract was acceptable. Among all of the cholangiograms, 5 were clinically applied with no correction, 8 were applied with modest improvements, and 3 produced a correct cholangiography before automatic selection.
Although our procedure requires further improvement based on the analysis of additional patient data, it may represent an alternative to direct cholangiography in the future.
利用重建的多排螺旋计算机断层扫描(MDCT)图像进行融合血管造影,以及使用注射胆管造影剂的MDCT重建图像进行胆管造影,由于MDCT扫描时期不同,存在解剖间隙。为克服此类间隙,我们尝试开发一种能从门静脉期MDCT图像自动重建胆管造影的程序。
自动生成胆管造影的程序利用了名古屋大学信息科学研究生院开发的原创软件程序。该程序构建了5条候选胆管,并自动选择其中一条作为胆管造影的候选对象。通过与手动生成的胆管造影进行比较,评估自动生成胆管造影程序的临床价值。
对20例因远端胆管梗阻在胆管引流术前接受MDCT扫描的患者重建了自动生成的胆管造影。该程序分别在55%和25%的病例中显示出提取5条主要胆管分支和21条亚段胆管分支的能力。胆管外异常连接和异常提取的程度是可接受的。在所有胆管造影中,5例未经校正即临床应用,8例经适度改进后应用,3例在自动选择前产生了正确的胆管造影。
尽管我们的程序需要根据更多患者数据的分析进一步改进,但它可能在未来成为直接胆管造影的一种替代方法。