Automated segmentation by deep learning of loose connective tissue fibers to define safe dissection planes in robot-assisted gastrectomy

Abstract

The prediction of anatomical structures within the surgical field by artificial intelligence (AI) is expected to support surgeons’ experience and cognitive skills. We aimed to develop a deep-learning model to automatically segment loose connective tissue fibers (LCTFs) that define a safe dissection plane. The annotation was performed on video frames capturing a robot-assisted gastrectomy performed by trained surgeons. A deep-learning model based on U-net was developed to output segmentation results. Twenty randomly sampled frames were provided to evaluate model performance by comparing Recall and F1/Dice scores with a ground truth and with a two-item questionnaire on sensitivity and misrecognition that was completed by 20 surgeons. The model produced high Recall scores (mean 0.606, maximum 0.861). Mean F1/Dice scores reached 0.549 (range 0.335–0.691), showing acceptable spatial overlap of the objects. Surgeon evaluators gave a mean sensitivity score of 3.52 (with 88.0% assigning the highest score of 4; range 2.45–3.95). The mean misrecognition score was a low 0.14 (range 0–0.7), indicating very few acknowledged over-detection failures. Thus, AI can be trained to predict fine, difficult-to-discern anatomical structures at a level convincing to expert surgeons. This technology may help reduce adverse events by determining safe dissection planes.

Can a single-port robot be safely used for robotic total gastrectomy for advanced gastric cancer? First experience using the da Vinci SP platform

“…developed a deep-learning model to predict the presence of loose connective tissue fibers during robotic gastrectomy. This technique may help surgeons to determine the correct anastomical planes for their procedures [ 22 ].…”

Computer-aided anatomy recognition in intrathoracic and -abdominal surgery: a systematic review

“…Year of publication ranged from 2008 until 2021. Fourteen of the 23 studies (61%) were published in 2020 and 2021, and all used AI-based algorithms [10,11,[16][17][18][19][20][21][22][23][24][25][26][27]. The majority of studies (n = 20; 87%) used a retrospective study design, a prospective or mixed design was used in three studies [20,23,28].…”

“…Of the 18 AI-based algorithms, 13 studies (72%) used a deep learning algorithm [10,11,[16][17][18][20][21][22][23][24][25][26][27]. Four studies based their model on U-net and developed a variation [18,20,21,25]. A total of five model-based algorithms were found [28,29,33,34,36].…”

“…Concordance via a third annotator or mentioning of the interannotator differences was done in five studies (31%) [10,17,18,23,27]. In the four remaining studies (25%), annotation was provided by either non-medical researchers, medical students, or a medical doctor [16,21,28,34].…”

“…Eight studies (35%) performed the annotations of the reference standard with low risk of bias [10,11,17,18,[22][23][24]27]. No studies were attributed high risk of bias for the flow and timing criterion.…”