PitRSDNet: Predicting intra-operative remaining surgery duration in endoscopic pituitary surgery

Abstract

Accurate intra-operative Remaining Surgery Duration (RSD) predictions allow for anaesthetists to more accurately decide when to administer anaesthetic agents and drugs, as well as to notify hospital staff to send in the next patient. Therefore, RSD plays an important role in improved patient care and minimising surgical theatre costs via efficient scheduling. In endoscopic pituitary surgery, it is uniquely challenging due to variable workflow sequences with a selection of optional steps contributing to high variability in surgery duration. This article presents PitRSDNet for predicting RSD during pituitary surgery, a spatio-temporal neural network model that learns from historical data focusing on workflow sequences. PitRSDNet integrates workflow knowledge into RSD prediction in two forms: (1) multi-task learning for concurrently predicting step and RSD; and (2) incorporating prior steps as context in temporal learning and inference. PitRSDNet is trained and evaluated on a new endoscopic pituitary surgery dataset with 88 videos to show competitive performance improvements over previous statistical and machine learning methods. The findings also highlight how PitRSDNet improves RSD precision on outlier cases utilising the knowledge of prior steps.

Keywords: computer vision; decision support systems.

FIGURE 1

Prototype with step progress indicators and estimated time remaining in the surgery.

FIGURE 2

PitRSDNet architecture and training stages.

FIGURE 3

Surgery duration distributions in (a) Pit‐33 and (b) Pit‐88 annotated with 25th (Q1) and the 75th (Q3) percentiles. For Pit‐33, the median duration is 72 min (inter‐quartile range [IQR]: 61–80 min), and for Pit‐88, it is 64 min (IQR: 53–84 min).

FIGURE 4

Distribution of time in seconds associated with each step in the Pit‐88 dataset. Points refer to the median of training, validation and testing splits.

FIGURE 5

RSD predictions over the surgery duration of two videos in Pit‐33 test set. Video 1 duration closely resembles training data while video 2 is significantly shorter in duration—see ground truth RSD indicated in Black. Other lines refer to RSD prediction methods—see legend.

FIGURE 6

RSD predictions over the surgery duration of two videos in Pit‐88. Video 1 duration closely resembles training data while video 2 is significantly longer in duration—see ground truth RSD indicated in Black. Other lines refer to RSD prediction methods—see legend.