VISITS Visual Surgical Instrument Tracking System

1
VISITSVisual Surgical Instrument Tracking System

• PostPC Course Project
• Yishai Beeri
• Dudi Einey

2
Problem Retained Instruments

• Instruments forgotten inside patients
• Happens in average more than once a year for a
  large hospital

3
Problem Retained Instruments

• Estimated 1500 cases / year in the US
• Account for over 50 of malpractice claims
• Average of 60,000 compensation fees
• Total hospital loss about double
• Several past cases of death
• Hospital reputation
• implication

4
Its a Real Problem

• 2/3 of incidents occurred even though equipment
  was counted before and after procedure
• Most were sponges and needles, but also metal
  clamps, electrodes and retractors
• Most (70) patients needed additional surgery to
  remove object

Incidents are rare but have a huge impact
5
Risk Factors and Implications

• Problem worse for fat patients
• Promoted by fatigue, emergencies or complications

6
Risk Factors and Implications
Emergency cases 9 times more likely
Complications requiring change in procedure 4 times more likely
More than one surgical team 3.5 times more likely
7
Current Approaches

• Ignore the problem
• Does not make the problem go away
• Increase liability (no best effort)
• Count instruments before and after surgery
• Error prone, fails to deliver desired results
• 88 of incidents had a good count!
• Doctors do not rely on nurse counts

8
Current Approaches cont.

• Semiautomatic instrument tracking
• Requires changed or specially made instruments
• Time consuming manually scan each instrument

9
Post Operational Imaging

• Not all instruments easily detected
• May require multiple images or image manipulation
  
• Costly (from 100 / film), might be harmful
• Not systematic
• What instrument are we looking for?
• Time consuming, occurs after stitching

Recommended only as complementary measure
10
Post Operational Radiography

• Sponges etc. have radiological markers to help
  find them in post-operation images

11
Solution Use Vision Technology
Identify instruments on trays and disposal
surfaces before, during and after surgery

• Automatically detect missing instruments
• No change to current instruments
• Simply mount a non-obtrusive camera in OR ceiling
  or wall
• Same handling for all instrument materials

12
VISITS how is it used?

• Take snapshot image of instruments on trays prior
  to surgery

13
VISITS how is it used?

• Scan additional trays during surgery, as required

14
VISITS how is it used?

• Compare subsequent snapshots (e.g. before
  stitching) to base. Also cover disposal and other
  locations
• System identifies missing instruments

15
VISITS how is it used?

• Operator can manually reconcile discrepancy (e.g.
  implant, etc.), rescan or bypass
• Minimal intrusion to operation procedure

16
System Components

• Wall or ceiling mounted digital cameras
• Several cameras for several instrument areas
• Central computer for processing images
• Interface via a simple mounted 15x10 screen
  with 2-5 buttons / touch-screen
• All important information at a glance

17
VISITS in the Operating Room

• Mount cameras on ceiling or wall above tray
  tables and above operation bed
• Install UI screen

18
VISITS - Technology

• Vision technology used to match images of
  instruments in snapshot to pre-existing database
• Database may contain 2D, 3D and other models for
  instruments
• Connection with hospital networks allows system
  to anticipate specific instrument sets

19
Vision Technology

• Wide variety of currently available techniques
• Appearance-based/similarity matching
• Recognition using 2D silhouettes
• Invariant feature matching
• Regular mesh tessellation
• Neural networks algorithms
• A popular and rapidly developing field
• Better/faster future techniques are imminent

20
Vision Challenges Shape Changes

• Some instruments change shape
• HMM and other technologies allow detection of
  varying shapes

21
Vision Challenges Objects Overlap

• Instruments may lie overlapping one another
• Since original shapes are known, identification
  is possible

22
Vision Challenges Multiple Objects

• Many objects to detect in one frame
• Current vision technology can handle this - even
  in motion

23
VISITS Cost and Deployment

• Less than 300 per camera
• Less than 10K for entire system
• Hassle-free installation
• Virtually no training required
• Fits in existing administrative OR procedures
• Reduces time of manual counts
• Reduces instrument loss

24
VISITS - Summary

• Harness advanced vision technology to solve a
  real problem retained instruments
• Low TCO, minimal maintenance
• No interference with surgeons work
• Dramatically reduce liability and avoid
  image-tarnishing incidents

25
References and Links

• Dont show this slide!
• http//www.sccs.swarthmore.edu/users/01/xianglan/e
  27/objRecognition.html - 2D object recognition
• Traceability
• http//www.permanentmarking.com/etch-surgical.php
• http//www.newco.co.uk/traceability.htm
• http//www.scantracksa.com/its.html
• http//www.scantrack.net/ITS20Overview.pdf
• http//www.steribar.com/traceability_why.html
• http//www.hospitalmanagement.net/contractors/surg
  ical/newco/
• http//ordesignandconstruction.com/index.html

26
(No Transcript)
27
(No Transcript)