Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

PitRSDNet: Predicting Intra-operative Remaining Surgery Duration in Endoscopic Pituitary Surgery

Anjana Wijekoon, Adrito Das, Roxana R. Herrera, Danyal Z. Khan, John Hanrahan, Eleanor Carter, Valpuri Luoma, Danail Stoyanov, Hani J. Marcus, Sophia Bano

TL;DR

PitRSDNet is presented for predicting RSD during pituitary surgery, a spatio‐temporal neural network model that learns from historical data focusing on workflow sequences that improves RSD precision on outlier cases utilising the knowledge of prior steps.

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 improving 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 paper 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 improve RSD precision on outlier cases utilising the knowledge of prior steps.

PitRSDNet: Predicting Intra-operative Remaining Surgery Duration in Endoscopic Pituitary Surgery

TL;DR

PitRSDNet is presented for predicting RSD during pituitary surgery, a spatio‐temporal neural network model that learns from historical data focusing on workflow sequences that improves RSD precision on outlier cases utilising the knowledge of prior steps.

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 improving 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 paper 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 improve RSD precision on outlier cases utilising the knowledge of prior steps.
Paper Structure (16 sections, 1 equation, 6 figures, 3 tables)

This paper contains 16 sections, 1 equation, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Problem formulation
  4. Proposed PitRSDNet
  5. Baseline methods
  6. Dataset Description
  7. Images
  8. Annotations
  9. Duration distribution
  10. Evaluation
  11. Implementation details
  12. Evaluation metrics
  13. Results and Discussion
  14. Quantitative results
  15. Qualitative results
  16. ...and 1 more sections

Figures (6)

  • 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 minutes (inter-quartile range (IQR): 61-80 minutes), and for Pit-88, it is 64 minutes (IQR: 53-84 minutes)
  • 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.
  • ...and 1 more figures