Non-linear temporal scaling of surgical processes

Abstract

Objective. Surgery is one of the riskiest and most important medical acts that is performed today. Understanding the ways in which surgeries are similar or different from each other is of major interest. Desires to improve patient outcomes and surgeon training, and to reduce the costs of surgery, all motivate a better understanding of surgical practices. To facilitate this, surgeons have started recording the activities that are performed during surgery. New methods have to be developed to be able to make the most of this extremely rich and complex data. The objective of this work is to enable the simultaneous comparison of a set of surgeries, in order to be able to extract high-level information about surgical practices. Material and Method.We introduce Non-Linear Temporal Scaling (NLTS): a method to realign a set of surgeries along their intrinsic timeline. Experiments are carried out on a set of lumbar disc neurosurgeries. We assess our method both on a highly standardised phase of the surgery (closure) and on the whole surgery.Results. Experiments show that NLTS makes it possible to consistently derive standards of surgical practice and to understand differences between groups of surgeries. We take the training of * Corresponding author -ENSISA -Universite de Haute Alsace, 12 rue des freres Lumiere, 68093 Mulhouse, France, Tel.:+33 3 8933 6963, Fax.:+33 3 8942 3282Email addresses: germain.forestier@uha.fr (Germain Forestier), francois.petitjean@monash.edu (François Petitjean), laurent.riffaud@chu-rennes.fr (Laurent Riffaud), pierre.jannin@univ-rennes1.fr (Pierre Jannin) Preprint submitted to Artificial Intelligence in MedicineSeptember 16, 2013 surgeons as the common theme for the evaluation of the results and highlight, for example, the main differences between the practices of junior and senior surgeons in the removal of a lumbar disc herniation.Conclusion. NLTS realigns a set of sequences along their intrinsic timeline, which makes it possible to extract standards of surgical practices.Supplementary material. The computer code implementing the proposed methods.

Automatic phase prediction from low-level surgical activities

“…To create the clusters, we used the methodology proposed by Forestier et al [5,6] which relies on Dynamic Time Warping (DTW) and ascendant hierarchical clustering. This methodology has proven its efficiency in creating clusters of similar surgeries [7]. The Figure 6 presents the dendrogram obtained from the hierarchical clustering process.…”

A Data-driven Process Recommender Framework

“…The same activities may have different temporal characteristics in different traces and traces may have different duration. Although several similarity metrics exist for temporal sequences [7,8] [11,12] [14-16], none satisfies the above requirements.…”

“…Both metrics accept as input only process traces represented as sequences, which requires that concurrent activities are sequenced (e.g., by activity start time) and that temporal information on activity duration and idle times is ignored. Forestier et al [11,12] proposed dynamic time warping (DTW) as a similarity metric for process traces. The DTW, however, cannot handle concurrent activities, does not consider idle time intervals, and has other issues when used for process traces [13].…”

Optimal Sub-Sequence Matching for the Automatic Prediction of Surgical Tasks

“…Extracting useful high-level knowledge from this low-level data has been one of the research themes targeted by the field of Surgical Process Modeling (SPM) [5,6], which aims at understanding surgeries to improve the quality of care. The above-mentioned sensors capture the surgical tasks performed in real-time, which opens the door to using artificial intelligence methods to provide real-time information to the surgical team.…”