Training systems for endoscopic soft-tissue surgery
Dr. M.P.S.F. Gomes-Bharath, Research Associate
Mr. A.R.W. Barrett, PhD student
Mr. A.G. Timoney, Consultant Urologist
Mr. P.V.S. Kumar, Clinical Research Fellow
Duration: 1 April 1998 to 31 March 2001
Imperial College of Science, Technology and Medicine
Non-Uniform Rational B-Splines
TransUrethral Resection of the Prostate
The original objectives, as presented in the grant proposal, are:
To identify generic criteria for in vitro and in vivo computer-based endoscopic surgery training
based in vitro prostatectomy training aid.
To provide a tracking system and ultrasound measurement facility as a computer-based in vivo
2. Background/ context
TURP has been described as the most difficult operation to learn and to teach. Training is passive
initially as the trainee observes the trainer perform the operation. After a certain period of observation,
the trainee is allowed to perform the operation under close supervision, without any transition between
observation and practice. The trainers themselves may be sometimes unaware of some of the
difficulties the trainee can encounter.
There are a few courses available to the trainee where more theoretical knowledge can be gained.
These courses also utilise practical skills stations where the trainee performs a TURP under supervision
on a resectable prostate model.
In addition, the Calman training scheme for higher surgical training, introduced in the United
Kingdom since 1993, has reduced the training period significantly. The possibility of a European
directive, which may further reduce training hours, is becoming realistic and may compound this
factor. All these facts highlight the need for a simple, realistic teaching aid for this operation that will
provide active feedback.
3. Methodological approach and achievements against objectives
Objective 1 was addressed in lengthy discussions between the researchers at Imperial College and the
clinical collaborators. Different scenarios were presented in the meetings, drawing on existing research
on surgical training, surgical simulation and computer-assisted surgical systems, on observations of
surgical procedures, as well as on previous research work at MIMLab.
Objective 2 was realised in the development of CASTS, by instrumenting surgical tools for optical
tracking and providing global positional information of the tools in relation to a prostate phantom
inserted in a mock-up abdomen. CASTS was fully tried by the surgical collaborators and the findings
Objective 3 was attained by designing a second generation of tools, fully sterilisable, and by
developing a method for building the patient’s prostate model from transrectal ultrasound scans. Ethics
Committee approval was obtained for the operating room trials. Components of CASMS were
successfully tested in the OR and the test of the integrated system is underway (see section 6 for
4. Project’s summary
The prostate gland is a male organ and surrounds the urethra at the bladder outlet (below, left). With
age, a benign enlargement of the tissue can develop, leading to obstruction of the bladder outlet and
urine retention. Transurethral Resection of the Prostate (TURP) aims to alleviate the urinary
constriction by debulking the enlarged tissue from within the urethra.
The urologist carries out a
TURP using a resectoscope
inserted through the shaft of the
penis, under endoscopic
In CASTS (left), the trainee
resects a prostate phantom,
inserted in a mock-up abdomen,
using standard real surgical
tools, instrumented for optical
tracking (right). A GUI displays
a model of the phantom, the
position of the resectoscope and
the tissue resected.
Three aspects of the surgical tools were instrumented. The first relates to the resectoscope’s position
and orientation. Three generations of IRED tools for optical tracking were developed. The first
prototype (below, left) is a 100mm-diameter aluminium ring, with 8 metallic IREDs along its
perimeter, at equally spaced intervals. The second (below, centre) is made of a lighter material and has
12 metallic IREDs for improved visibility. The third (below, right) has 14 ceramic IREDs in a non-
coplanar arrangement for improved accuracy, and it can be sterilised. It has been especially designed
for CASMS (OR).
The position of the cutting loop in relation to the resectoscope’s sheath is also known. For CASTS, a
potentiometer arrangement was attached to the resectoscope (below, left). For CASMS, an alternative
sterilisable IRED probe was designed (below, centre).
Finally, the system needs to know when the surgeon is cutting or coagulating. Two switches have
been connected to the diathermy pedals and to an A/D board in the PC (below, right). The pedal is
placed in a sealed plastic bag in the OR.
The GUI (below, left) has been kept simple and the different views (3D, 2D, thumbnails) can be turned
measuring hard casts of the Limbs & Things phantom used for conventional urological surgery training
(below, right), and using software to build NURBS surfaces from the measured points.
It was found that the Limbs & Things phantom moved and deformed whilst being resected, due to
forces exerted by the resectoscope, release of internal forces and to absorption of irrigant. This was
taken into account by: a) pre-calculating deformations and selecting appropriate control points for the
NURBS surfaces (below, left and centre), b) partially constraining the motion with a set of Perspex
profiles (below, right), and c) using ultrasound for tracking motion, with an optically tracked
ultrasound probe (below, right).
Since the Limbs & Things phantom is opaque to ultrasound, alternative phantoms, made of gelatine
and an ultrasound scatterer, were developed. The first prototype (below, left) was based on a resection
shape from anatomical diagrams. The second prototype (formers and moulds below, centre and right),
based on the Limbs & Things phantom, includes anatomical features, thus allowing a laboratory-based
assessment of the in vivo system, more closely mimicking the actual operation.
The gelatine phantoms were used to form ultrasound images with
of this quality image is feasible.
Transrectal ultrasound images of a patient’s prostate are much
more difficult to analyse automatically, as they depict artefacts due
to the shadow created by the resectoscope (below, left), and by the
diathermy current when cutting (below, centre) and coagulating
In CASMS (left), the surgeon acquires
a set of 2D pre-operative transrectal
ultrasound scans of the patient’s
prostate, using an optically tracked
ultrasound probe. The scans are then
manually segmented by the urologist
who delineates the areas to be
resected. A 3D model is automatically
built and rendered in a computer
During the TURP, the display shows the current status of the resection by superimposing a rendering of
the resectoscope onto the patient’s prostate model. The resected cavity is also rendered. Per-operative
ultrasound, with a tracked and motorised probe, is used to detect and compensate for the movement of
Following Ethics Committee approval, mock-up of the tools were tried in the OR during TURP
(below, left). The optical tracker was suspended from the ceiling of the OR, on a counterbalanced arm
(below, top right). Tests of visibility of the IRED tools in the OR were successful (below, bottom
5. Main findings/ key advances
the skills of minimal access surgery, there is still a large gap between such systems and the real surgical
procedure. The dual system CASTS/CASMS reduces this gap by providing assistance during both in
combination of computer graphics and VR, haptic feedback, and finite element modelling. However,
even the most advanced haptic systems are seen by some as not giving the same feel as cutting through
real human tissue. The use of physical phantoms in CASTS is more realistic and provides a natural
progression from traditional training.
Measurable performance CASTS/ CASMS provide objective data to enable the measurement of
performance and improvement, such as the amount of tissue resected, perforation or damage to high-
risk areas, and the time taken to complete the procedure.
Role of CASTS The system can be used as an extra cue for endoscopic navigation, as a “black box”
for logging the session, or as the basis of a VR system.
for logging the surgical procedure, or as a pre-operative planner.
display, and uses both hands to operate the resectoscope and ancillary equipment, as well as one foot to
activate the diathermy unit. In CASTS/ CASMS, extra intervention or information has therefore been
kept simple and to a minimum to avoid changing the procedure.
Real surgical equipment is used. The weight of the instrumentation used to track the resectoscope was
kept to a minimum.
Prostate phantoms The hygroscopic behaviour from the Limbs & Things prostate phantom causes
great difficulty in the modelling of the procedure. Alternative phantoms were developed which have
overcome this problem.
Deformation and movement The prostate phantom (and the prostate) moves and deforms in
unexpected modes. This was tackled by introducing partial constraints and precalculated deformations
(in CASTS), and by using per-operative ultrasound (in CASMS).
Pre-operative and per-operative ultrasound A transducer, small and powerful enough to fit around
the resectoscope’s sheath, would avoid the need for transrectal ultrasound and simplify the procedure.
strategies for this were adopted: models that can be rendered quickly, precalculated deformations,
partial constraining of motion, multi-threaded code.
Tool tracking A novel IRED probe for tracking the position and orientation of a resectoscope
(EndoTracker) has been developed in collaboration with Traxtal Technologies Ltd, Canada. Its average
error is less than 0.2mm and the probe is visible over 90% of the time, in spite of ±180° axial rotations.
6. Project plan review
The original research proposal stated that “The target [the prostate] would be assumed not to move
significantly intra-operatively so that pre-operative 3D models can be used”. This was found to be only
valid in the previous research context, where all data was in the same coordinate system. The need for
tracking motion and deformation of the prostate (and the prostate phantoms) in external (world)
coordinates was successfully tackled and accommodated within the project plan.
Two unforeseen factors slowed down the progress of the OR trials. Our main surgical collaborator
needed back surgery which forced him to be away from the operating table for several months,
impeding the trial of the integrated CASMS, scheduled for January-February 2001. Also, the
development of EndoTracker, in collaboration with Traxtal Technologies Ltd, was severely delayed
due to a shortage of ceramic IREDs.
7. Research impact and benefits to society
Patients, Surgeons, Health Providers, and the NHS benefit from improved training and performance.
Surgeon Training Establishments benefit from improved training, assessment and computer-based
Medical Instrumentation Industry benefits from improved trainers and trackable surgical tools.
The Research Community benefits from computer-assisted training, generic tracking of tools,
construction and tracking of ultrasound phantoms, modelling and segmentation of ultrasound images.
11 papers (all refereed, 3 journal publications) were written and presented to scientific and medical
international audiences. A poster was presented at EPSRC’s Theme Day in Machine Vision and Image
Processing, on 7
June 2000. Summaries of the papers and other information on the project are
available on the web (http://www.me.ic.ac.uk/case/mim/). Links to and from our main surgical
collaborators at the Bristol Urological Institute have been established (http://www.bui.ac.uk). Project
posters are also displayed in the Mechatronics in Medicine Laboratory of Imperial College and are
used to present the project to visitors. A4 copies of the posters have been sent to all collaborators. All
collaborators were sent progress reports throughout the project.
9. Further research
Industrial collaboration to develop commercial products is under investigation.
The test of the integrated CASMS will be finished when the surgeon is fully recovered.
Results on phantoms and models, soft tissue modelling, and motion tracking with ultrasound have
direct relevance to the project at Imperial College on Robotic Systems for a Range of Urological
Disorders (EPSRC grant GR/M53394/01), and also to the project on Innovative Haptics for Risk
Mediation in VR Arthroscopic Training, jointly with Sheffield and Warwick Universities (EPSRC
A PhD project, carried out by Mr. A.R.W. Barrett, entitled An Investigation into Monitoring
Minimally Invasive Surgery has been underway since September 1998.
Know-how in optical tracking is of general importance in the activities of MIMLab, and it has a
special synergy with the Acrobot project for Robotic Knee Surgery.
10. List of publications
M. P. S. F. Gomes, A. R. W. Barrett, and B. L. Davies, “Computer-assisted soft-tissue surgery training
and monitoring”, MICCAI 2001, Utrecht, The Netherlands, Springer, 2001. Submitted
M. P. S. F. Gomes, P. V. S. Kumar, A. G. Timoney, and B. L. Davies, “Tool tracking for endoscopic
surgery”, Measurement+Control, 2001. To appear.
P. V. S. Kumar, M. P. S. F. Gomes, B. L. Davies, and A. G. Timoney, “A computer-assisted surgical
trainer for transurethral resection of the prostate”, Journal of Urology, 2001. Submitted
M. P. S. F. Gomes and B. L. Davies, “Computer-assisted surgical training for TURP”, First IFAC
Conference on Mechatronic Systems, Darmstadt, Germany, VDI/ VDE - GMA, vol.2, pp. 531-535,
M. P. S. F. Gomes and B. L. Davies, “Computer-assisted TURP training and monitoring”, MICCAI
P. V. S. Kumar, M. P. S. F. Gomes, B. L. Davies, and A. G. Timoney, “Computer-assisted surgical
training system for transurethral resection of the prostate”, American Urology Association 95
training system for TURP”, British Association of Urological Surgeons (BAUS) Annual Meeting,
Birmingham, BJU Intl, vol.85: Suppl 5 63A, pp. 27, 2000.
B. Davies, M. P. S. F. Gomes, A. G. Timoney, and P. V. S. Kumar, “A computer-assisted surgical
trainer for transurethral resection of the prostate (TURP)”, International Workshop on Mechatronic
Tools for Surgery (MTS), Rome, pp. 24, 1999.
M. P. S. F. Gomes, A. R. W. Barrett, A. G. Timoney, and B. L. Davies, “A computer-assisted training/
monitoring system for TURP - Structure and design”, IEEE Transactions on Information Technology
in Biomedicine, vol. 3:4, pp. 242-251, 1999.
M. P. S. F. Gomes, P. V. S. Kumar, A. G. Timoney, and B. L. Davies, “A computer-assisted surgical
trainer for TURP - Instrumenting the tools”, Second EUREL (European Advanced Robotic Systems
Development) Workshop, Pisa, Italy, pp. 124-126, 1999.
P. V. S. Kumar, M. P. S. F. Gomes, B. L. Davies, and A. G. Timoney, “Transurethral resection of the
prostate - A computer-assisted training system”, 17
World Congress on Endourology, Rhodes, J.