GB2333882A

GB2333882A - Apparatus for and method of assessing surgical technique

Info

Publication number: GB2333882A
Application number: GB9801618A
Authority: GB
Inventors: Ari Warkes Darzi; Nicholas Taffinder
Original assignee: Imperial College of Science Technology and Medicine
Current assignee: Imperial College of Science Technology and Medicine
Priority date: 1998-01-26
Filing date: 1998-01-26
Publication date: 1999-08-04
Anticipated expiration: 2018-01-26
Also published as: GB9801618D0; GB2333882B

Abstract

A system for assessing surgical technique has surgical instrument position measurement means (4) for detecting the position of a surgical instrument (2) whilst performing a surgical task, data storage means for recording position and time data representing the position of said surgical instrument at a spaced series of times whilst performing said surgical task, and assessment processing means (6) for determining from said position and time data at least one quantitative measurement of surgical technique. The at least one quantitative measurement of surgical technique may be one of: (i) a distance travelled by said surgical instrument whilst performing at least part of said surgical task, (ii) when said surgical task requires said surgical instrument to be moved to a predetermined location, the number of times the distance between said surgical instrument changes from decreasing to increasing, (iii) an average velocity of said surgical instrument whilst performing said surgical task, (iv) a number of velocity maximums in movement of said surgical instrument and (v) a time taken to complete said surgical task. The surgical task may be a predetermined simulated task such as performed in a virtual reality simulated environment. Alternatively the task may be a real task performed upon a patient. The position measurement means may comprise an electromagnetic transducer fixed to one or both of said surgical instrument or the hand of the surgeon.

Description

APPARATUS FOR AND METHOD OF ASSESSING SURGICAL TECHNIQUE This invention relates to an apparatus for and a method of assessing surgical technique.

Surgery, such as laparoscopic surgery, is difficult to perform and extensive training is necessary to master the skills to operate safely (see Royston c, Lansdown M, Brough W. Teaching laparoscopic surgery: The need for guidelines. British Medical Journal 1994;308:1023-5). Although technical expertise is only one of many factors which determines outcome from surgery, failure to address the issue of training in psychomotor skills has become more apparent with the advent of laparoscopic surgery. The hand-eye co-ordination required to operate within the constraints of a two-dimensional image on a video monitor involves skills that are unfamiliar to most surgeons (see McDougall-EM, Soble-JJ, Jr W-J, Nakada-SY, Alashry-OM, Clayman RVAD. Comparison of three-dimensional and two-dimensional laparoscopic video systems. J-Endourol 1996;10(4):3714). It has been estimated that some 10% of surgeons lack the visuo-spatial skills necessary to perform this type of surgery (see Cushieri A. Reflections on Surgical training. Surgical Endoscopy 1993;7:73-74).

Practice on patients is no longer acceptable and a large number of courses for surgical trainees in all disciplines ofminimal access surgery have been developed (see EAES. Training and assessment of competence. Surgical endoscopy 1994;8:721722). Tasks include simple transfer of small inanimate objects between containers up to highly complex tasks such as practising laparoscopic suturing on animal tissue.

The only assessment possible has been a subjective rating by the instructors of the trainees. This is widely regarded as an unsatisfactory method, prone to bias and variability (see Elliot D, Hickman D. Evaluation of physical examination skills.

Reliability of faculty observers and patient instructors. JAMA 1987;258(23):34053408). The only absolutely objective method of measuring performance has been to record the time taken to complete a task (see Crosthwaite-G, ChungT;, Dunkley P;,Shimi-S;, Cushieri-AAD. Comparison of direct vision and electronic two- and three-dimensional display systems on surgical task efficiency in endoscopic surgery.

Br-JOSurg 1995;82(6):849-51). Although good surgeons tend to perform procedures fairly quickly, using speed as a goal for surgical training is not ideal.

Viewed from one aspect the invention provides an apparatus for assessing surgical technique, said apparatus comprising: surgical instrument position measurement means for detecting the position of a surgical instrument whilst performing a surgical task; data storage means for recording position and time data representing the position of said surgical instrument at a spaced series of times whilst performing said surgical task; and assessment processing means for determining from said position and time data at least one quantitative measurement of surgical technique.

Viewed from another aspect the invention provides a method of assessing surgical technique, said method comprising the steps of: detecting the position of a surgical instrument whilst performing a surgical task; recording position and time data representing the position of said surgical instrument at a spaced series of times whilst performing said surgical task; and determining from said position and time data at least one quantitative measurement of surgical technique.

At least preferred embodiments of the invention seek to provide: 1. A reliable and feasible method to assess laparoscopic psychomotor skills on a virtual reality surgery simulator (MIST VR simulator); 2. A validated scoring system by comparing surgeons of different laparoscopic experience on the MIST VR simulator; and 3. A quantified measurement of the effect of a standard laparoscopic training course on MIST VR simulator score.

An embodiment of the invention will now be described, by way of example only, with reference to the accompanying drawings in which: Figure 1 shows the effect of previous laparoscopic experience on performance in VR; Figure 2 shows the effect of experience on psychomotor skill assessed on a complex MIST VR task; Figure 3 shows the effect of a training course on performance in VR; Figure 4 shows the effect of a training course on psychomotor skill assessed on a simple MIST VR task; and Figure 5 illustrates an apparatus in accordance with one embodiment of the invention.

All the subjects were assessed on a virtual reality laparoscopic surgery simulator. The Mist VR system used was based on a PC ruining Windows 95. The PC was configured with a Pentium 133 Mhz processor, 32 MB of RAM, a 1.6 GB hard drive, a Mattrox Mystique 4MB video card and a 21-inch monitor. The laparoscopic interface was a standard Immersion Corporation unit , with the addition of a foot pedal for the diathermy tasks. The trials ran Mist VR version 1.2 which currently utilises the WorldToolKit Version 6 and Microsoft Direct 3D Version 3 graphics libraries. Frame rates averaged around 15 fps and did not drop below 10 fps during the evaluations.

Two tasks were chosen for detailed analysis: a simple task that involved picking up a virtual ball, placing it in a box and releasing it; and a complex task which required participants to hold a target still while the other hand burns off subtargets by using a foot pedal to simulate surgical electrocoagulation. The x,y,z co-ordinates for both virtual instrument tip positions were recorded approximately every 0.05 - 0.1 seconds (depending on screen refresh rates). The kinematic data were re-sampled to ensure regular time stamping of 0.05 seconds, extrapolating to recalculate tip positions. The data were then filtered using a fourth order 1.5Hz Butterworth filter.

Path length, path velocities and distance to target profiles were analysed to calculate: 1. Efficiency of movement (actual path length/ideal path length) 2. Number of submovements (number of velocity peaks (local velocity maximums)) 3. Errors (number of movements away from the target) 4. Time taken to complete the virtual task.

Thirty subjects (experienced laparoscopic surgeons with > 100 cases (n=10); trainee surgeons (n=10) and non-surgeons (n=10)) were recruited for the validation study. All subjects were given identical tuition on the MIST-VR simulator to control for the limited cognitive skills required to perform the tasks and to familiarise them with the equipment. Six tasks which demanded different co-ordination skills were given in a progressive and identical sequential order. The final complex virtual task which involved two-handed co-ordination and use of a foot pedal to simulate diathermy was used for the assessment. The scores for the repetitions were averaged and the groups compared.

To assess the effect of a training course, ten junior surgeons attending a Royal College of Surgeons of England basic surgical skills course were recruited. All subjects had assisted in laparoscopic procedures but never performed any laparoscopic surgery. On the first day of the course, all 10 subjects were assessed on the same simple virtual task. On the following day, 5 subjects were re-assessed after the skills training course. The course involved extensive hands-on training on standard closed box tasks developing simple hand-eye co-ordination manipulating laparoscopic instruments to dissect, cut, clip and transfer inanimate objects. The remaining 5 were used as a control group to correct for the learning curve of the MIST VR simulator itself, and were re-tested before any hands-on training. Both groups had identical exposure to the MIST VR simulator. The Wilcoxon Rank Sum test was used for nonparametric analysis; p < 0.05 was considered statistically significant.

RESULTS 1. Effect of previous laparoscopic experience on performance in virtual reality Experienced surgeons performed significantly better in virtual reality than trainee surgeons on efficiency ratio (2.3 vs 3.3 (p < 0.04)), number of submovements (10.4 vs 15.9 (P < 0.02)) and time to complete task (13.7 vs 19.3 seconds (p < 0.02)). there was no significant difference in the number of errors made (4.8 vs 6.3). As shown in Figure 1 experienced surgeons performed better than non-surgeons on all 4 criteria (p < 0.01), but there was no significant difference between trainees surgeons and non-surgeons. There was a large variation in ability in the inexperienced groups with some trainees scoring as well as the best experienced surgeons for efficiency and errors (Figure 2).

2. Effect of the training course As shown in Figure 4, both groups were well matched on the first assessment.

On the second assessment the trained group were more efficient than the untrained group (2.8 vs 3.5 (p < 0.02)) and made less errors 2.3 vs 3.6 (p < 0.02)). There was a small reduction in the number of submovements (8.4 vs 10.3) and time taken (9.3 vs 10.4 seconds) but this did not reach significance (Figure 3).

CONCLUSION The MIST VR simulator offers an opportunity to assess a surgeon's technical expertise in laparoscopic surgery. By tracking the movement of the instruments and correlating that with the virtual tasks. a detailed analysis of how well the task has been completed, rather than simply how quickly it has been done, can be produced.

Use of the invention has demonstrated that experienced surgeons outperform trainee surgeons by measurable, objective criteria on a virtual reality simulator.

Moreover, a randomised controlled study has shown that a Royal College of Surgeons approved course has taught juniors to make less errors and to operate more efficiently than their untrained contemporaries, without significant decrease in time to complete the tasks.

An unbiased method of measuring psychomotor skill in laparoscopic surgery according to the invention could guide training by providing detailed and constructive feedback and could identify those who require additional training before being allowed to operate on patients.

Examples of the algorithms that are required in these examples as part of or may be used to gain a quantitative measure of surgical technique are: 1. Resampling The record is initially presented as a series of timestamped rows of data, where tl, t2, t4... are the timestamps and dl, d2, d3, d4... are the rows of position indicating data. The timestamping does not necessarily take place at regular intervals so that ti+l- ti.^tj+l-tj in some cases. Resampling is therefore required. For resampling, a sampling interval t is chosen. The kth sample is then given by first finding i such that ti < k.t and ti+l > k.t Then using linear interpolation: dk = di + [(k.t.-ti)(di+1 - di)/(ti+l - ti)j 2. Velocity calculation The data for each timestamp consists of 3D co-ordinates for the tip of both tools. If r is such a co-ordinate then we can estimate instantaneous speed by: vk+l =Irk+l -rkI /(tk+l-tk) 3. Closing rate calculation For each timestamp, a co-ordinate corresponding to the current target is also given. If this is given by q then the speed with which the tool tip is closing on the target can be found by: uk+I =(Irk+l -qk+l -Irk-q)/(tk+I tk) 4. Noise gating In order to eliminate small scale tremors and vibrations associated with periods of intended non-movement, a moving average for instantaneous speed was calculated.

The following rule was then used to perform the noise gating vavk > vNoiseThresh: vk = vk otherwise: vk = 0 VNoiseThresh was chosen empirically. The same process was applied to uk.

5. Smoothing The velocity data uk and vk were smoothed using a Digital 4th Order Low Pass Butterworth Filter, using an empirically determined cut-off frequency.

6. Trimming Periods of inactivity at the beginning and ends of datasets were discarded. Working in from the end of the record, all end rows of data where vk < vTrimThresh were discarded. Again, vTrimThresh was empirically determined.

7. Velocity Peak Counting In order to detect distinct movements, a count was made of peaks in vk. First, the rate of change of velocity at successive samples was found: ak =(vk-vk-l)/(tk-tk-l) ak+l =(vk+l -vk)/(tk+l -tk) If ak > 0 and ak+l < 0 then a peak had been found. Next, the second derivative of velocity was examined to check that the peak was sufficiently large to be counted: If(ak+l - ak)/(tk+l - tk) > daThresh then count the peak. daThresh was determined empirically.

8. Error Counting In order to detect errors, a count was made of instances where uk made the transition from negative to positive (or from moving towards the target to moving away from the target). i.e. If uk < 0 and uk+l > 0 then an error has been found.

Figure 5 illustrates an apparatus in accordance with one embodiment with the invention. A laparoscopic probe 2 is mounted on a sensing gimbal 4 that provides electrical signals indicating the position of the laparoscopic probe 2. The electrical signals generated by the gimbal 4 are passed to a computer 6 that converts these signals into position data for the laparoscopic probe 2. This position data is timestamped with its acquisition time and stored within the computer 6. The computer 6 also shows on its monitor a virtual reality model of the position of the tip of the laparoscopic probe 2 in relation to the objects that are part of the surgical task to be performed.

The subject under test manipulates the laparoscopic probe 2 whilst watching the monitor of the computer 6 to complete the simulated virtual reality task. The computer records the position data of the laparoscopic probe 2 and then analyses this in accordance with the above techniques to produce an assessment of surgical technique.

Claims

CLAIMS 1. Apparatus for assessing surgical technique, said apparatus comprising: surgical instrument position measurement means for detecting the position of a surgical instrument whilst performing a surgical task; data storage means for recording position and time data representing the position of said surgical instrument at a spaced series of times whilst performing said surgical task; and assessment processing means for determining from said position and time data at least one quantitative measurement of surgical technique.
2 Apparatus as claimed in claim 1, wherein said at least one quantitative measurement of surgical technique is one of: (i) a distance travelled by said surgical instrument whilst performing at least part of said surgical task; (ii) when said surgical task requires said surgical instrument to be moved to a predetermined location, the number of times the distance between said surgical instrument changes from decreasing to increasing; (iii) an average velocity of said surgical instrument whilst performing said surgical task; (iv) a number of velocity maximums in movement of said surgical instrument; and (v) a time taken to complete said surgical task.
3. Apparatus as claimed in any one of claims 1 and 2, wherein said surgical task is a predetermined simulated surgical task.
4. Apparatus as claimed in claim 3, wherein said predetermined simulated surgical task is defined and performed within a virtual reality simulated environment.
5. Apparatus as claimed in any one of claims 1 and 2, wherein said surgical task is a real task being performed upon a patient.
6. Apparatus as claimed in any one of the preceding claims, wherein said surgical task is a laparoscopic surgical task.
7. Apparatus as claimed in any one of the preceding claims, wherein said surgical instrument position measurement means comprises an electromagnetic transducer fixed to one or both of said surgical instrument or the hand of a surgeon.
8. A method of assessing surgical technique, said method comprising the steps of: detecting the position of a surgical instrument whilst performing a surgical task; recording position and time data representing the position of said surgical instrument at a spaced series of times whilst performing said surgical task; and determining from said position and time data at least one quantitative measurement of surgical technique.
9. A method as claimed in claim 8, wherein said at least one quantitative measurement of surgical technique is one of: (i) a distance travelled by said surgical instrument whilst performing at least part of said surgical task; (ii) when said surgical task requires said surgical instrument to be moved to a predetermined location, the number of times the distance between said surgical instrument changes from decreasing to increasing; (iii) an average velocity of said surgical instrument whilst performing said surgical task; (iv) a number of velocity maximums in movement of said surgical instrument; and (v) a time taken to complete said surgical task.
10. An apparatus for assessing surgical technique substantially as hereinbefore described with reference to the accompanying Figures.
11. A method of assessing surgical technique substantially as hereinbefore described with reference to the accompanying Figures.