文献速递主动脉疾病

深度学习全自动分割肾下腹主动脉瘤CT影像与医师手动分割的对比研究

Fully automatic volume segmentation of infrarenal abdominal aortic aneurysm computed tomography images with deep learning approaches versus physician controlled manual segmentation.
PubMed 文献 · J Vasc Surg · 2020-12-09

本研究评估了一种新型全自动软件(PRAEVAorta)对肾下腹主动脉瘤CT血管造影的分割质量,并与资深和初级外科医师手动校正的半自动分割进行对比。在100例术前CTA(13465层)中,全自动与手动校正分割的体积、表面积和直径高度相关(Pearson相关系数>0.90,P<.0001),资深医师手动校正分割的平均Dice相似系数为0.95±0.01。全自动分割时间27秒至4分钟/例,显著短于手动校正的5至80分钟(P<.0001)。

要点

  • 全自动软件与手动校正分割的体积、直径测量高度相关(Pearson>0.90)。
  • 全自动分割平均Dice相似系数0.95±0.01,分割质量优良。
  • 全自动分割时间27秒至4分钟/例,远快于手动校正的5至80分钟。

核心结论

  • 全自动软件可快速、可靠地完成肾下腹主动脉瘤CT的详细解剖分析,在临床实践和研究中应用前景强。

研究设计与样本

  • 研究纳入100例肾下腹主动脉瘤患者的术前CT血管造影(共13465层),将全自动分割与资深和初级外科医师手动校正的半自动分割进行定量比较。

主要发现

  • 全自动与资深医师手动校正分割的平均Dice相似系数为0.95±0.01,Jaccard指数0.91±0.02,敏感度0.94±0.02,特异度0.97±0.01,体积相似性0.98±0.01,每层平均Hausdorff距离4.61±7.26 mm。
  • 管腔的平均体积相似性达0.95±0.04,血栓的平均体积相似性达0.91±0.07。
  • 全自动与手动校正分割所测体积、表面积和直径的Pearson相关系数均大于0.90(P<.0001)。
  • 全自动分割时间27秒至4分钟/例,手动校正方法为5至80分钟(P<.0001)。

临床怎么用

  • 该全自动软件可快速提供肾下腹主动脉瘤的管腔、血栓及直径等详细解剖参数,辅助术前规划和临床研究。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

OBJECTIVE: Imaging software has become critical tools in the diagnosis and decision making for the treatment of abdominal aortic aneurysms (AAA). However, the interobserver reproducibility of the maximum cross-section diameter is poor. This study aimed to present and assess the quality of a new fully automated software (PRAEVAorta) that enables fast and robust detection of the aortic lumen and the infrarenal AAA characteristics including the presence of thrombus. METHODS: To evaluate the segmentation obtained with this new software, we performed a quantitative comparison with the results obtained from a semiautomatic segmentation manually corrected by a senior and a junior surgeon on a dataset of 100 preoperative computed tomography angiographies from patients with infrarenal AAAs (13,465 slices). The Dice similarity coefficient (DSC), Jaccard index, sensitivity, specificity, volumetric similarity (VS), Hausdorff distance, maximum aortic transverse diameter, and the duration of segmentation were calculated between the two methods and, for the semiautomatic software, also between the two observers. RESULTS: The analyses demonstrated an excellent correlation of the volumes, surfaces, and diameters measured with the fully automatic and manually corrected segmentation methods, with a Pearson's coefficient correlation of greater than 0.90 (P < .0001). Overall, a comparison between the fully automatic and manually corrected segmentation method by the senior surgeon revealed a mean Dice similarity coefficient of 0.95 ± 0.01, a Jaccard index of 0.91 ± 0.02, sensitivity of 0.94 ± 0.02, specificity of 0.97 ± 0.01, VS of 0.98 ± 0.01, and mean Hausdorff distance per slice of 4.61 ± 7.26 mm. The mean VS reached 0.95 ± 0.04 for the lumen and 0.91 ± 0.07 for the thrombus. For the fully automatic method, the segmentation time varied from 27 seconds to 4 minutes per patient vs 5 minutes to 80 minutes for the manually corrected methods (P < .0001). CONCLUSIONS: By enabling a fast and fully automated detailed analysis of the anatomic characteristics of infrarenal AAAs, this software could have strong applications in daily clinical practice and clinical research.

来源:J Vasc Surg,查看原文。版权归原作者所有。

引用

  1. Caradu C, Spampinato B, Vrancianu AM, et al. Fully automatic volume segmentation of infrarenal abdominal aortic aneurysm computed tomography images with deep learning approaches versus physician controlled manual segmentation. J Vasc Surg. 2020. PMID: 33309556. doi:10.1016/j.jvs.2020.11.036

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。