The Target Task
The SAGES CVS Challenge aims to target classification probabilities of the three criteria of the Critical View of Safety (CVS) in laparoscopic cholecystectomy (surgical gallbladder removal) (Strasberg et al. J Am Coll Surg. 2010). The CVS is a well-established surgical safety measure that targets the prevention of common bile duct injuries, a feared complication with a significant impact on patients’ recovery and survival. The CVS consists of three criteria, defined by the dissection and isolation of three anatomic landmarks, which can be visually distinguished in the intraabdominal video obtained from the laparoscope. For this challenge, these criteria were defined in a consensus-based Annotation Protocol as:
The SAGES CVS Challenge aims to develop models capable of reliably and consistently classifying each of the CVS criteria. Different metrics and evaluations have been designed to evaluate different aspects of the model performance, namely:
- Subchallenge A (CVS Classification): Confidently predicting the CVS criteria as assessed by the majority of annotators (ground truth annotation of 3 criteria, generated by majority vote)
- Subchallenge B (Uncertainty Quantification): Representing inherent uncertainties of the task (CVS criteria assessment), such as subjectivity of assessment and annotation difficulty.
- Subchallenge C (Robustness): Robustly accounting for potential data distribution shifts when moving to "unseen" data.
Participants will make one submission, which is evaluated according to all three subchallenges. However, prizes are allocated separately for each subchallenge. Participants are required to ensure that their submissions are causal at inference time (i.e., they should only make use of past and current frames). Check the evaluation criteria page (accessible to registered participants) for more information on the specific metrics used to evaluate these characteristics.
SAMPLE CODE COMING SOON!
