Scholarlink

<p>Augmented Reality (AR) systems can enhance intraoperative anatomical understanding during Robot-Assisted Partial Nephrectomy by overlaying preoperative 3D models onto the surgical view. Their usefulness, however, depends on the spatial accuracy of these overlays, which requires objective validation.&#xa0;This preclinical study quantitatively evaluated the spatial accuracy of a new AR software prototype using a validation workflow based on 3D-printed kidney phantoms and a Structure-from-Motion (SfM) reconstruction as ground truth. CT scans were segmented to create virtual kidney models that were 3D printed and filmed with the da Vinci Si camera. The AR prototype was applied to the recordings, and manual 3D registration was performed using the TilePro interface. Keyframes from manual and tracking-based registrations were analysed to compute the Target Registration Error (TRE), defined as the mean Euclidean distance between corresponding vertices of the AR and SfM models. Three operators repeated the workflow, and manual registration time was recorded to assess variability and feasibility.&#xa0;Mean TRE across experiments was below 5&#xa0;mm, indicating clinically acceptable accuracy. TRE values for individual keyframes remained consistently under 10&#xa0;mm, with limited inter-operator variability (ICC(3,k) = 0.80). Manual and tracking TREs were comparable and strongly correlated (β = 0.89; <i>p</i> &lt; 0.001). Tracking-related precision changes varied among operators, and mean manual registration time was 173.6 ± 76.4&#xa0;s. Accuracy was unaffected by keyframe count, kidney variability, or registration duration.&#xa0;This methodology offers an objective, reproducible framework for assessing AR registration in preclinical settings. The prototype showed reliable spatial performance, supporting progression toward in vivo validation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Accuracy and reliability of an augmented reality prototype for robot-assisted partial nephrectomy: a preclinical study with 3D-printed kidney phantoms (UroCCR 168)

  • 期刊论文

  • 发表日期：

    2026年3月9日

查看全文
  • F. Rubat Baleuri,
  • M. Pattou,
  • G. Margue,
  • G. Nguyen,
  • A. Pitout,
  • A. Khaddad,
  • K. Chandelon,
  • J. Desternes,
  • N. Bourdel,
  • J.-C. Bernhard,
  • A. Bartoli

摘要

Augmented Reality (AR) systems can enhance intraoperative anatomical understanding during Robot-Assisted Partial Nephrectomy by overlaying preoperative 3D models onto the surgical view. Their usefulness, however, depends on the spatial accuracy of these overlays, which requires objective validation. This preclinical study quantitatively evaluated the spatial accuracy of a new AR software prototype using a validation workflow based on 3D-printed kidney phantoms and a Structure-from-Motion (SfM) reconstruction as ground truth. CT scans were segmented to create virtual kidney models that were 3D printed and filmed with the da Vinci Si camera. The AR prototype was applied to the recordings, and manual 3D registration was performed using the TilePro interface. Keyframes from manual and tracking-based registrations were analysed to compute the Target Registration Error (TRE), defined as the mean Euclidean distance between corresponding vertices of the AR and SfM models. Three operators repeated the workflow, and manual registration time was recorded to assess variability and feasibility. Mean TRE across experiments was below 5 mm, indicating clinically acceptable accuracy. TRE values for individual keyframes remained consistently under 10 mm, with limited inter-operator variability (ICC(3,k) = 0.80). Manual and tracking TREs were comparable and strongly correlated (β = 0.89; p < 0.001). Tracking-related precision changes varied among operators, and mean manual registration time was 173.6 ± 76.4 s. Accuracy was unaffected by keyframe count, kidney variability, or registration duration. This methodology offers an objective, reproducible framework for assessing AR registration in preclinical settings. The prototype showed reliable spatial performance, supporting progression toward in vivo validation.