Comparison of Conventional Methods for Bowel Length Measurement in Laparoscopic Surgery to a Novel Computer-Assisted 3D Measurement System

Purpose

Accurate laparoscopic bowel length measurement (LBLM), which is used primarily in metabolic surgery, remains a challenge. This study aims to three conventional methods for LBLM, namely using visual judgment (VJ), instrument markings (IM), or premeasured tape (PT) to a novel computer-assisted 3D measurement system (BMS).

Materials and Methods

LBLM methods were compared using a 3D laparoscope on bowel phantoms regarding accuracy (relative error in percent, %), time in seconds (s), and number of bowel grasps. Seventy centimeters were measured seven times. As a control, the first, third, fifth, and seventh measurements were performed with VJ. The interventions IM, PT, and BMS were performed following a randomized order as the second, fourth, and sixth measurements.

Results

In total, 63 people participated. BMS showed better accuracy (2.1±3.7%) compared to VJ (8.7±13.7%, p=0.001), PT (4.3±6.8%, p=0.002), and IM (11±15.3%, p<0.001). Participants performed LBLM in a similar amount of time with BMS (175.7±59.7s) and PT (166.5±63.6s, p=0.35), but VJ (64.0±24.0s, p<0.001) and IM (144.9±55.4s, p=0.002) were faster. Number of bowel grasps as a measure for the risk of bowel lesions was similar for BMS (15.8±3.0) and PT (15.9±4.6, p=0.861), whereas VJ required less (14.1±3.4, p=0.004) and IM required more than BMS (22.2±6.9, p<0.001).

Conclusions

PT had higher accuracy than VJ and IM, and lower number of bowel grasps than IM. BMS shows great potential for more reliable LBLM. Until BMS is available in clinical routine, PT should be preferred for LBLM.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s11695-021-05620-6.

Dense Depth Estimation from Stereo Endoscopy Videos Using Unsupervised Optical Flow Methods

In the context of Minimally Invasive Surgery, estimating depth from stereo endoscopy plays a crucial role in three-dimensional (3D) reconstruction, surgical navigation, and augmentation reality (AR) visualization. However, the challenges associated with this task are three-fold: 1) feature-less surface representations, often polluted by artifacts, pose difficulty in identifying correspondence; 2) ground truth depth is difficult to estimate; and 3) an endoscopy image acquisition accompanied by accurately calibrated camera parameters is rare, as the camera is often adjusted during an intervention. To address these difficulties, we propose an unsupervised depth estimation framework (END-flow) based on an unsupervised optical flow network trained on un-rectified binocular videos without calibrated camera parameters. The proposed END-flow architecture is compared with traditional stereo matching, self-supervised depth estimation, unsupervised optical flow, and supervised methods implemented on the Stereo Correspondence and Reconstruction of Endoscopic Data (SCARED) Challenge dataset. Experimental results show that our method outperforms several state-of-the-art techniques and achieves a close performance to that of supervised methods.