Key result
Roboformer deep learning accurately quantifies robotic surgeries, achieving ~0.97 AUC for gesture classification across centers.
Why the study?
Understanding how surgical activity affects patient outcomes requires reliable, objective, and scalable quantification of the core elements of surgery, which is currently lacking.
Does the Roboformer deep learning framework accurately quantify surgical phases, gestures, and skills from robotic surgery videos?
Does the Roboformer deep learning framework accurately quantify surgical phases, gestures, and skills from robotic surgery videos?
A novel deep learning framework, Roboformer, successfully quantifies surgical phases, gestures, and skills from robotic surgery videos, demonstrating good generalization and potential for providing targeted surgical feedback.
May enable objective robotic surgery feedback; leaves open prospective validation of training or patient outcomes.
Surgery is a high-stakes domain where surgeons must navigate critical anatomical structures and actively avoid potential complications while achieving the main task at hand. Such surgical activity has been shown to affect long-term patient outcomes. To better understand this relationship, whose mechanics remain unknown for the majority of surgical procedures, we hypothesize that the core elements of surgery must first be quantified in a reliable, objective, and scalable manner. We believe this is a prerequisite for the provision of surgical feedback and modulation of surgeon performance in pursuit of improved patient outcomes. To holistically quantify surgeries, we propose a unified deep learning framework, entitled Roboformer, which operates exclusively on videos recorded during surgery to independently achieve multiple tasks: surgical phase recognition (the what of surgery), gesture classification and skills assessment (the how of surgery). We validated our framework on four video-based datasets of two commonly-encountered types of steps (dissection and suturing) within minimally-invasive robotic surgeries. We demonstrated that our framework can generalize well to unseen videos, surgeons, medical centres, and surgical procedures. We also found that our framework, which naturally lends itself to explainable findings, identified relevant information when achieving a particular task. These findings are likely to instill surgeons with more confidence in our framework's behaviour, increasing the likelihood of clinical adoption, and thus paving the way for more targeted surgical feedback.
No takes yet. Share an insight, caveat, or question.
Kiyasseh et al. (2022) studied Robot-assisted radical prostatectomy (RARP) and partial nephrectomy (RAPN) (n=251). Roboformer (vision-based deep learning framework) vs. Ablation models (e.g., without test-time augmentation, without optical flow, without self-attention) was evaluated on Model performance (AUC) for surgical phase recognition, gesture classification, and skills assessment. The Roboformer deep learning framework successfully quantified robotic surgeries, achieving high discriminative performance (e.g., AUC up to 0.974 for gesture classification) and generalizing across unseen videos, surgeons, and medical centers.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: