Author : Dr. Xiao Di
Last Updated : 05 June 2007
Transperineal Prostate Needle Biopsy Robot
Background
The
prostate is a gland of the male reproductive system. It is located in front of
the rectum and just below the bladder. In modern society, prostate cancer is the
most common among men between the ages of 60 and 80.
In
clinic, prostate-specific antigen (PSA) and digital rectal examination (DRE) are
commonly used as screening tools for prostate cancer. If a patient had elevated
PSA level and/or abnormal DRE, he will undergo a transrectal ultrasound (TRUS)
guided biopsy of the prostate. Technically speaking, a biopsy of the prostate
would require the biopsy needle to pass through the cancer site (if one exists)
and retrieve a tissue sample for further histological examination. Clinically,
guided by TRUS, the urologist manually manoeuvres the probe to obtain 2D
transverse or longitudinal images of the prostate. More commonly, relying on a
series of 2D images, the urologist would perceive the size, shape and location
of the prostate before inserting the biopsy needle, transrectally or
transperineally, to try to reach an area of interest inside the prostate. These
two kinds of methods, called transrectal ultrasound-guided (TRUS) and
transperineal ultrasound-guided (TPUS) biopsies, are commonly used for screening
and diagnosing prostate cancer (Figure 1).
Due
to the manual operation of the prostate biopsy procedure and that current
protocol of TRUS and TPUS guided prostate biopsy more depends on the skills and
experience of the urologist, the limitations of the current methods include:
•
Gross spatial inaccuracies as surgeon tries to locate a point in space based
on 2-D images
•
High false-negative results
•
Multiple puncture holes on rectal or perineal wall
A
robotic system, which can perform accurate biopsy of the prostate, would indeed
eliminate many of the existing drawbacks of conventional prostate biopsy
procedures. The robot’s greatest contribution would be accuracy. It would be
capable of delivering a biopsy needle to a predefined point in the prostate with
minimal needle placement errors. This ensures a uniform standard in prostate
biopsy which is independent of the urologist’s skills and experience. In this
project, we developed a robotic system, which integrates advanced 3D graphics
technologies to implement transrectal ultrasound image acquisition of prostate
and provide transperineal prostate biopsy guidance.
Objectives
of project
•To
develop a prostate needle biopsy robot
•To
improve the accuracy of prostate biopsy
•To
reduce false-negative results with accurate biopsy of the prostate
•To
perform multiple biopsies with a single puncture point at the perineal wall
System
overview
Figure
2 is a schematic diagram of the overall system configuration and Figure 3 is a
photograph of the entire robotic system. Physically, the robotic system is made
up of 3 distinct entities; computer trolley, ultrasound machine and biopsy
robot.
Hardware
Design
A
photograph of the biopsy robot is shown in Figure 4. It is a stiff platform
structure with 9 DOFs. The TRUS probe and the biopsy gun with a needle attached
are integrated into the robot as shown. After the robot is positioned near the
patient’s perineal area, a foot pedal is depressed which causes the robot to
be raised slightly and be supported by 4 rubber-padded legs rather the wheels.
At this point, the height and tilt of the operating table may be adjusted so
that the patient’s rectum is horizontal and about the same height as the TRUS
probe. The following procedure would be to manoeuvre the base of the robot,
which allows adjustments in the horizontal plane and vertical axis, such that
the TRUS probe can be inserted into the patient’s anus with ease. Situated on
top of the base structure is a gantry structure, which supports and manipulates
the biopsy needle. Similar to the base structure, 2 pairs of linear slides and a
pair of lead screws enable the urologist to manually manoeuvre the needle to its
intended point of entry at the perineal wall. This point of entry is also the
pivot point for subsequent biopsies. The remaining 3 DOFs define the trajectory
of the biopsy needle. These 3 DOFs enable multiple biopsy cores to be retrieved
from various parts of the prostate.
Software Design
The
computer software is relative independent but an integral part in the robotic
biopsy system. As shown in Figure 5, the software framework adopts a 3-layer
architecture. At the bottom are the hardware control modules, including motion
control module and image acquisition module, which call low-level drivers for
hardware control. The middle-layer modules are the kernel algorithms,
independent from the GUI (Graphical User Interface) and hardware control
modules. The GUI layer undertakes the tasks of calling all functional modules
and organizing all displaying and interactive panels. Currently, the software is
programmed with C++ language under Linux Red Hat 8.0 platform. In the system,
all kernel modules can be summed up to four functional parts; ultrasound image
acquisition, prostate modeling, biopsy path planning and hardware operation
guidance.
•
Ultrasound image acquisition consists
of two main modules: motion control and
image acquisition. It is used to acquire the transverse ultrasound image frames
with a predefined distance interval during the probe’s motion (Figure 6).
•
Prostate modelling required
the interaction of an urologist with the
•
Biopsy path planning consists
of two aspects; the definition of puncture point at the perineal wall and the
biopsy path calculation. Firstly, the biopsy point needs be defined on a
selected 2D image by the urologist. Then, define the biopsy point. With
information of the biopsy and puncture points, the software can compute the
trajectory of the needle for the currently selected biopsy point and simulate
the needle’s path on 3D
scene. Figure 8 shows the relationship
of the prostate surface, needle, needle trajectory and defined biopsy point 3D.
•
Hardware operation guidance can
be considered a part of the biopsy path planning procedure. A separate panel is
used to implement this function. The urologist can read and input the scaled
markings from the robot’s passive joints. The software module will compute the
output parameters from these inputs and give instructions.
Surgical
procedure
The
main procedure to operate the robot includes:
(1)acquiring
image slices into computer;
(2)extracting
prostate information;
(3)guiding
biopsy process.
The
following describe the procedure in greater detail:
•Positioning
of robot with respect to patient
•Acquisition
of ultrasound image slices
•Delineation
of prostate boundary
•3D
modeling of prostate
•Pivot
point definition
•Definition
of biopsy point of interest
•Computing
trajectory of needle
•Adjustment
of trajectory of needle on robot
•Insertion
needle and firing biopsy gun
Experiments
The
main objective of the robotic system is to improve percutaneous needle placement
accuracy within the prostate. As such, it has undergone three phases of rigorous
experimentations (phantom,
cadaveric and
clinical trials)
to validate its actual performance.
Phantom
Test
The
phantom experiment aimed at testing the initial accuracy of the robotic system,
analyzing error and making a further fine-tuning by mechanical adjustment and
compensation from software algorithm. After that, a series of tests on gelatine
phantom and mixture phantom of pork meat and gelatine showed the placement error
is no more than 1.0mm in all three axes. Figure 9 shows one phantom test.
Cadaveric
Trial
Before
moving to clinical trials, a cadaveric trial was performed to verify the
efficacy of the robotic system. Prior to the experiment, two sterile copper
seeds (cylindrical in shape measuring 1.0mm in diameter and 2.5mm in length)
were implanted into the cadaveric prostate; one in each lobe. The distance
between the needle tip and the center of the seed was measured using a uni-planar
fluoroscope by comparison to an object of known length in the same field of
view. Figure 10 illustrates one result of cadaveric trials.
Clinical
Trials
With
favourable results from the phantom experiments and cadaveric trial, the team
proceeded to perform clinical trials. The trials (with approval from the Medical
Ethical Committee and consent from the patients) were performed on patients just
prior to them undergoing RRP (Retro Pubic Prostatectomy) or TURP (Transurethral
Resection of the Prostate). In both cases, the implanted target copper seeds
would ultimately be removed from the body. Figure 11 shows the error curve of
needle placement in a series of clinical trials.
The objective of the pilot study (Phase I)
was to treat 20 patient with follow-up examination over 1 week, in order to
monitor for any side-effect of the transperineal biopsy.
Also, to ensure proper device functionality, to monitor safety, to
evaluate the usability of the device, address any sterility issue that may arise
and to gain experience in using the device prior to commencing the primary phase
of the clinical study.
The
clinical trials were conducted with the approval of the Medical Ethical
Committee and the consent of the patients. The study was conducted in
agreement with the Declaration of Helsinki, revised version of 1983 (World
Medical Assembly, Venice 1983). The physician informs each patient about the
goals, methods, expected use as well as possible risks and side-effects. The
patient is informed that he has the right to refuse to participate in the study
and to interrupt the study at any time. The declaration of informed consent must
be submitted in writing. All patients’ confidentiality is to be maintained and
respected during the course of the studies.
Our target patient population is
male, below the age of 70 with at least one normal previous prostate biopsy and
at least 2 weeks since prior prostate biopsy with increasing prostate-specific
antigen (PSA) level. Patient was made
aware of this procedure by the surgeon, who also explains to the patient the
benefits and risks of the procedure. Patient consent was also obtained by the
surgeon after a pre procedural check up to ensure that patient is fit for the
procedure. Patient trials were conducted in three phases.
Phase I: Between May 2006 and
June 2006, ‘warm up’ trials on 3 patients were conducted to get familiar
with the system and understand the operating protocol. Patients, placed in
lithotomy position, underwent ultrasound scanning. The urologist modeled the
prostate on the acquired images and planned the most optimal cores to be taken.
However, no needling of patient was done.
Phase II: In August 2006, we
conducted a seeding test on 1 patient to test the system accuracy by targeting
the biopsy needle tip to pre-implanted gold seeds in the prostate and using
X-ray to verify the results. The TRUS was used to retrieve ultrasound images.
The metal seeds were identified and a core was planned at the centre of it. A
small cut is made on the perineal wall to allow manual insertion of the biopsy
needle. The needle follows the trajectory determined by the software. A C-arm
fluoroscopy unit is used to view the needle placement accuracy and observe the
absolute error.
Phase III: Between September 2006
and January 2007, 25 patients underwent transperineal biopsy with assistance
from BioXbot. Our target patient population is male, who have undergone prostate
biopsy with negative for adenocarcinoma but have increasing prostate-specific
antigen (PSA) level. The transperineal biopsy is performed by the surgeon with
assistance from BioXbot. Once the urologist specifies the scan range, the
acquired images are presented for subsequent modeling and planning. The gantry
unit is moved towards the perineal wall to mark the entry points to allow the
insertion of the biopsy needle. A non-crossing protocol is used i.e., left pivot
point for biopsy sites on the left part of prostate, and likewise right pivot
point for sites on the right. No needle trajectory crosses the middle of the
prostate. This technique avoids the urethra and reduces pain and discomfort for
the patient. The urologist needs to place the biopsy gun on the gun stopper in
order to obtain the required depth for the biopsy point. The surgeon then
manually inserts the biopsy gun to retrieve the core.
Even
in its initial learning stage, prostate cancer was diagnosed in 4 out of 25 of
these patients. This demonstrates that BioXbot is able to detect cancers which
were missed by conventional biopsy with improved patient safety profile. These
patients did not have pain or discomfort and showed no signs of haematuria.
There was no incidence of sepsis.
Based
on the pathology results, some tissue were non-prostatic or skeletal muscles,
especially in cores taken from the extreme peripheral areas. This could be due
to the needle sliding on top of the prostate towards the peripheral areas
instead of penetrating through the gland itself.
The
execution of the whole procedure takes between 30 minutes to 45 minutes on an
average to complete a 20 core biopsy. This includes, setting up of machine,
putting the patient under general anesthesia, setting up the ultrasound machine,
cleaning of the patient and draping him, aligning of the machine with patient,
performing the ultrasound, modeling the prostate and making the 3D image,
planning the biopsy and finally the execution of the actual biopsy.
ACKNOWLEDGEMENTS
The
research group wishes to acknowledge the support of National Medical Research
Council (NMRC) Grant 0537/2001, SingHealth Grant CC011/2001 and SingHealth Grant
RP002/2001.
Awards
A
Robotic Prostate Biopsy Device: The Answer to Our Current Inaccurate Manual
Biopsy System, Urology Fair 2004, Feb., 2004. (Best Presentation Award)
Ultrasound Guided Robotic System for Transperineal Biopy of the Prostate, IEEE
International Conference on Robotics and Automation April, 2005, Barcelona,
Spain. (Best Paper Award).
Publications related to
Cancer Detection.
We would be glad if you could sign our guest
book.
For more information, please contact the principal investigator:
A/P Ng Wan Sing
School of Mechanical & Aerospace Engineering
Nanyang Technological University
Nanyang Avenue, Singapore 639798
Fax:(65) 6791 1859
