; radiation_group
Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out

radiation_group

VIEWS: 6 PAGES: 23

  • pg 1
									Robot v. Surgeon:
A Comparative Study on Needle Accuracy and Radiation Exposure in Robotic versus
Manual Percutaneous Biopsy Procedures


Johns Hopkins University, Department of Computer Science
Senior Design Project: Computer Integrated Surgery II




May 14, 2003



Ayushi Ahuja
Tristen Johnson
Brandon Rochelle
Jie Zheng


{ayushi,tristen,brandon,jiezheng}@jhu.edu
Abstract

       Needle Accuracy during percutaneous biopsy procedures is important to effective

diagnosis. However, the patient and surgeon are exposed to radiation during CT guided

biopsy procedures that could possibly lead to potential health problems. An accuracy

study was done to compare the effectiveness of the Percutaneous Access to Kidney –

Remote Center of Motion robotic system versus normal biopsy. Radiation exposure was

also measured during the procedures. It was found that the robot was more accurate by

over 2 mm in all cases (more in most cases) compared to that of the needle placement in

manual procedures. Radiation exposure levels were reduced by at least a factor of 3.


1. Introduction

       Computer assisted surgeries are quickly becoming the next-generation step in the

operating room. Through computers, surgeons are offered a level of precision thought

possible by only the most skilled of surgeons. With the aid of computer tomography (CT)

images, surgeons have an extra-sensory tool for target visualization. However, along with

the advent of CT and CT-Fluoroscopy (CT-Fluoro) comes another risk; the surgeon and

patient are exposed to radiation. Surgeons are faced with the trade-off between better

visualization and possible health hazards associated with radiation exposure. Medical

robotics offers a way to have the best of both worlds. In the example of percutaneous

biopsies, surgery times can be reduced by up to 40%1 through the use of medical

robotics, specifically the Percutaneous Access to Kidney – Remote Center of Motion

system (PAKY-RCM) 2. Along with reduced surgery times, it has been hypothesized by

Solomon et al1 that a reduction in the amount of radiation seen to the patient and surgeon

is possible, as the robot can be more accurate with fewer CT or CT-Fluoro images taken.
         The purpose of our study was to quantify the reduction in radiation realized

through the use of the PAKY-RCM in percutaneous biopsy procedures (subtitled

“Radiation Study” throughout this paper). The study also quantified the increase in

accuracy seen with the robot with respect to the biopsy target (subtitled “Accuracy

Study” throughout this paper”).



2. Materials and Methods

Robot

         The robot used in this study was the PAKY-RCM (Percutaneous Access to

Kidney – Remote Center of Motion) system. It consisted of a passive robotic arm with

seven degrees of freedom attached to the robot base that was mounted on the CT table.

At the end of the passive arm was the RCM which allowed the needle three degrees of

movement to obtain the correct trajectory to target while maintaining a single entry point

on the patient’s skin. The PAKY, attached to the RCM, held an 18-guage needle close to

the insertion point on the body to avoid needle deflection distal to the entry point on the

skin. The PAKY was equipped with a needle driver that advances the needle at a constant

speed.

         During the robotic procedures, the target BB was located with sequential 2mm CT

scans of the phantom. The image with the target was registered to the robot. Next, the

pre-selected entry point was registered to the robot via a laser registration method.3 After

picking the entry point and target from CT images registered to the robot, the needle was

advanced using the automatic targeting feature of the PAKY-RCM system.
Phantoms

       The phantoms used in this study were hemi-spheres made of gelatin mixed at

three times the manufacturer’s recommended amount. After the gelatin solidified in its

hemi-spherical mold, a 2mm BB was placed in the center at the bottom of the hemi-

sphere and imbedded 5mm. A hemi-sphere phantom was used to ensure uniform depth to

the target when approached from any entry angle. Three different sized phantoms were

used with radii measuring 2.25, 3.0 and 4.25 inches respectively. The phantoms had 90,

70 and 50 degree entry angles, with respect to the horizontal plane, marked on the

surface. The smallest half-sphere was covered with cheesecloth to keep the surgeon from

seeing the BB while performing the biopsy.



Accuracy Study Procedure

       The robot and the surgeons who participated in this study used a standard eight-

inch needle for all biopsies. Accuracy was calculated based on the error distance which

was defined as the distance from the needle tip to the BB. The surgeon was evaluated on

the error distance after the first needle placement as well as the number of tries to hit the

BB. The robot was only allowed one needle placement, and then the error distance was

measured.

       The error distance was obtained from a CT image of the final needle position and

the target BB using the built in evaluation function of the Siemens Plus-4 CT Scanner.

When the insertion point was not in the CT plane, the final needle tip position and the

target BB were not always visible on the same scan. Using a 2mm thick scan,
consecutive images were taken through the phantom to find the target BB and the tip of

the needle. The height difference and length difference (measured as the difference in

table position) could then be appreciated, and the true distance from the BB was

calculated using simple geometry.



Radiation Study Procedure

       The radiation study was performed using dosimeters from Landauer Inc. with

sensitivities of 5 mGys. Using the CT scanner’s built in section-selection laser; a

dosimeter was placed in the plane of the imbedded BB to measure the radiation a patient

would be receiving for the procedure. In manual trials, a dosimeter was placed on the

surgeon’s dominant hand middle finger to measure the radiation exposure to the surgeon

on each biopsy. While in robotic trials the dosimeters were placed on the arm of the

PAKY. After the procedure, the dosimeters were sent to Landauer Inc. for analysis.



Statistical Analysis

       The mean, standard deviation, and the 95% confidence interval of the error

distance measurements and radiation exposure were calculated. Based on the standard

deviation, a student t-test was performed to analyze the significance of the mean accuracy

and radiation exposure between robotic and manual surgeries. The t-test measures the

probability that the two means (robotic and manual) came from the same population.
3. Results

Accuracy Study

       The relationship between accuracy of the procedure and the insertion angles at

various depths in manual and robotic procedures was examined. Using the 4.25 inch

phantom, the mean error distance during manual procedure at 90 degree insertion angle

was 5±3.9mm (n=4), at 70 degree insertion angle was 5.37±8.3mm (n=4), and at 50

degree insertion angle was 17±3.0mm (n=4); during robotic surgery the mean error

distance at 90 degree insertion angle was 2±0.72mm (n=4), at 70 degree insertion angle

was 0.5±0.88mm (n=4), and at 50 degree insertion angle was 0.33±0.65mm (n=4).

       Using the 3.00 inch phantom, the mean error distance during manual procedure at

90 degree insertion angle was 3.3±1.7mm (n=4), at 70 degree insertion angle was

7.5±4.9mm (n=4), and at 50 degree insertion angle was 5.3±4.6mm (n=4); during robotic

surgery the mean error distance at 90 degree insertion angle was 2.25±0.49mm (n=4), at

70 degree insertion angle was 0.5±0.57mm (n=4), and at 50 degree insertion angle was

0.5±0.65mm (n=4).

       Using the 2.25 inch phantom, the mean error distance during manual procedure at

90 degree insertion angle was 7.0±6.8mm (n=4), at 70 degree insertion angle was

10.3±8.5mm (n=4), and at 50 degree insertion angle was 9.66±5.8mm (n=4); during

robotic surgery the mean error distance at 90 degree insertion angle was 1.0±1.13mm

(n=4), at 70 degree insertion angle was 0.75±1.13mm (n=4), and at 50 degree insertion

angle was 0.7±0.84mm (n=4).
Radiation Study

       During manual surgery the radiation exposure to surgeons was measured as

follows: at 90 degree insertion angle the radiation exposure was 460 mrads (n=1), at 70

degree insertion angle the radiation exposure was 30 mrads (n=1), and at 50 degree

insertion angle was also 30 mrads (n=1). However, during robotic procedure, the

radiation exposure to surgeon was 0 mrads (n=3) at all three insertion angles.

       The radiation exposure to phantom during manual surgery was measured as

follows: at 90 degree insertion angle the radiation exposure was 11000 mrads (n=1), at 50

degree insertion angle the radiation exposure was 5470 mrads (n=1), and at 50 degree

insertion angle radiation exposure was 14190 mrads (n=1). During robotic procedures,

the mean radiation exposure at 90 degree insertion angle was 2214±733 mrads (n=3), at

70 degree insertion angle was 1966±962 mrads (n=3), and at 50 degree insertion angle

was 2959.6±2059 mrads (n=3).



4. Discussion

Accuracy Study

       Manual procedures showed a decrease in procedure accuracy as the insertion

angle increased from 90 degrees to 50 degrees with respect to the horizontal plane. This

was especially prominent at the insertion depth of 4.25 inch (figure 1). However, the data

showed the accuracy of robotic procedures were not affected by the insertion angles as

the 95% confidence interval for the mean accuracy across all insertion depth and

insertion angle overlapped (figures 1, 2, and 3). It is interesting to note that in Figure 1

and 2, the accuracy of the robot increase as the insertion angle increased. This seemingly
odd trend could be explained by the robot operator’s learning curve of the laser

registration method. However, as the number of procedures increased, the operators’

skills increased and the trend is not visible at the insertion depth of 2.25inch (figure 3).

When comparing the mean error of the robotic procedure to that of the manual procedure

using the t-test, significant differences were observed. For insertion angles of 70 and 90

degrees, the t-test value was less than 0.1 which confirms that the mean error of the

robotic procedure and manual procedure came from two different populations (figures 1,

2, and 3). However, at the 90 degree insertion angle, the confidence interval of manual

and robotic procedures overlapped, indicating that when the insertion angle is in the plane

of the CT scanning field, the surgeon may be as accurate as the PAKY-RCM. Further

study was carried out to examine accuracy versus insertion depth. The data depicted no

conclusive trends as the t-test values were greater than 0.1 and there was large overlaps in

the 95% confidence intervals (figure 4, 5, and 6). This indicates that it is the insertion

angle but not the insertion depth that is the major contributor to error in both the robotic

and manual procedures.



Radiation Study

       The data showed a consistent trend which was inline with previously published

trends on radiation exposure to patients and surgeons.4 Radiation exposure to the surgeon

in manual surgeries greatly exceeded the radiation exposure in robotic surgeries (figures

8 and 9). During manual surgery at the 90 degree insertion angle, the radiation exposure

was significantly higher than that of 70 and 50 degree insertion angles which indicate that

when the insertion angle is in the plane of the CT scanning field, the surgeon must place
his hand in the field to advance the needle and thus greatly increases his radiation

exposure (figure 8).

       The radiation exposure to phantom was significantly higher in manual surgery

than in robotic surgery across insertion angles (figure 9). However, when examining the

data for robotic procedures only, there was no difference in radiation exposure across

insertion angle. This was as expected, since the radiation exposure is dependent on the

number of CT slices, but during robotic procedures the number of CT slices is

independent of the insertion angle.



Future Studies

       In this study, we were able to quantify the widely accepted hypothesis that robotic

surgeries increase accuracy while reducing radiation exposure to surgeons and patients.

However, due to limited time and funding, further radiation studies were not conducted.

Future radiation studies should be conducted to confirm the existing radiation exposure

data at a statistically significant level. Furthermore, to validate the use of robotic system

in surgeries, the protocols described in this paper must be carried into patient studies as

the phantoms used in the study cannot accurate simulate humans with respect to

breathing and tissue-density variation.
Appendix A: Data Tables
DATA TABLES

Accuracy Study

Robotic: (all measurements in mm)
depth = 4.25 in
 Entry Angle trial1 trial2 trial3 trial4
      90          1      2      2           3
      70          0      0      2           0
      50          0      0      1

depth = 3.00 in
 Entry Angle trial1 trial2 trial3 trial4
      90          3      2      2           2
      70          0      1      1           0
      50          1      0      0

depth = 2.25 in
 Entry Angle trial1 trial2 trial3 trial4
      90          2      0      0           2
      70          1      2      0           0
      50          1      2      0           0



Manual: (all measurements in mm)
90 degree entry point
  Depth(cm)        Trial 1    Trial 2           Trial 3
      4.25                9          3                    3
      3.00                2          5                    3
      2.25               11         10                    0

72 degree entry point
   Depth(cm)       Trial 1    Trial 2           Trial 3
      4.25               14             0                 3
      3.00               10             5
      2.25                6             6             19

51 degree entry point
   Depth(cm)       Trial 1    Trial 2           Trial 3
      4.25               12         17                16
      3.00                3          3                10
      2.25               14         11                 4
Radiation Study

Radiation exposure to surgeons: (all measurements in mrads)
Insertion Angle 90 Degree 70 Degree 50 Degree
    Robotic              0      0           0
    Manual              460    30          30



Radiation exposures to phantom: (all measurements in mrads)
Insertion Depth / Insertion Angle 90 Degree 70 Degree 50 Degree
              4.25 in               2132        1180    1200
              3.00 in               1470        1680    1599
              2.25 in               3040        3040    6080
Appendix B: Figures
FIGURES

                                                                                               Accuracy vs. Insertion Angle At 4.25 inch Insertion Depth


                                                          20

                                                          18

                                                          16
                                                                                                                                                                        t-test value = 0.0007
 Accuracy - Distance from needle tip to bb (mm)
 (error bars represent 95% confidence interval)




                                                          14

                                                          12

                                                          10
                                                                                                                                                                                                Manual
                                                                       8
                                                                                                                                 t-test value = 0.21                                            Robotic
                                                                                t-test value = 0.14
                                                                       6

                                                                       4

                                                                       2

                                                                       0
                                                                                90 Degree                                      70 Degree                          50 Degree
                                                              -2

                                                              -4
                                                                                                            Insertion Angle (with respect to horizontal plane)




Figure 1




                                                                                                Accuracy Vs. Insertion Angle At 3.00 inch Insertion Depth


                                                                       14



                                                                       12
                      Accurary - Distance from Needle Tip to bb (mm)




                                                                       10
                      (error bars represent 95% confidence interval)




                                                                           8                                                           t-test value = 0.01



                                                                                                                                                                                                Manual
                                                                           6
                                                                                                                                                                        t-test value = 0.1      Robotic

                                                                                      t-test value = 0.23
                                                                           4



                                                                           2



                                                                           0
                                                                                 90 Degree                                      70 Degree                         50 Degree

                                                                           -2
                                                                                                             Insertion Angle (with respect to horizontal plane)




Figure 2
Figure 3




Figure 4
                                                                    Accuracy vs. Insertion Depth At 70 Degree Insertion Angle


                                                   20




                                                   15
 Accuracy - Distance from N eedle Tip to bb (mm)
 (error bars represents 95% confidence interval)




                                                   10                                                                                 t-test value = 0.04


                                                                                                                                                             Manual
                                                                                                   t-test value = 0.01
                                                                                                                                                             Robotic
                                                              t-test value = 0.21
                                                    5




                                                    0
                                                        4.25 inch                           3.00 inch                           2.25 inch




                                                   -5
                                                                                          Insertion Depth




Figure 5




                                                                    Accuracy vs. Insertion Depth at 50 Degree Insertion Angle


                                                   20


                                                   18
                                                              t-test value = 0.0007
                                                   16
 Accuracy - Distance from N eedle Tip to bb (mm)
  (error bars represent 95% confidence interval)




                                                   14


                                                   12


                                                   10
                                                                                                                                      t-test value = 0.017   Manual
                                                                                                                                                             Robotic
                                                    8


                                                    6
                                                                                                   t-test value = 0.1

                                                    4


                                                    2


                                                    0
                                                        4.25 inch                           3.00 inch                           2.25 inch
                                                   -2
                                                                                          Insertion Depth




Figure 6
Figure 7




Figure 8
                                                                                                      Radiation Exposure to Phantom

                                                                                  16000

   R adiation Exp osure ( mr ads) ( err or bars r epresent sta ndard deviation)

                                                                                  14000



                                                                                  12000



                                                                                  10000

                                                                                                                                                  Manual
                                                                                  8000
                                                                                                                                                  Robotic
                                                                                  6000



                                                                                  4000



                                                                                  2000



                                                                                     0
                                                                                          90 Degree               70 Degree           50 Degree
                                                                                                                Insertion Angle




Figure 9
Appendix C: Management Summary
MANAGEMENT SUMMARY

       The main management oversight in this project was needed not for the team

members themselves, but for the many clinicians, technicians and qualified personnel that

were needed to perform the trials. On numerous occasions, the trials were postponed due

to one member of the support staff backing out of the project. Eventually, the robot

technician taught the group how to use the robot instead of waiting for the robot operator

to be available. The CT technicians that Dr. Solomon had enlisted to help with our

project kept falling through. The group was forced to rely on help from CT technicians

who were working in the adjacent CT machine on the weekends until the group learned

enough to operate the CT machine independently for the final robotic trials.

       Although there were great difficulties in arranging common times for the study, in

most cases it was due to the clinicians and staff’s over exuberance to help. In most cases,

they would volunteer for times that they were not available, only to cancel at a later date.

       The core four members of this study worked well together to organize procedures

and adjust to last minute changes in the schedule. Brandon Rochelle had the daunting

task of coordinating the outside help for the project. Scheduling a robotic tech to set up

the machine, a CT tech to run the machine, a robotic operator, and CT time all at the

same time proved to be a very large task, much larger than any of us had anticipated.

       Ayushi Ahuja was the main radiation badge technician. She met with radiation

experts from ImageGuide and referrals from Dr. Solomon to find the right path at the

beginning of the study. Ayushi was able to find badges with the correct sensitivity, small

enough size, and short enough turn around time for use in this study. She also ordered,

tracked, and analyzed the radiation study data from the badges.
       Jie Zheng was the main statistician for the project. Besides keeping detailed

records of each trial, she and Brandon designed the phantoms for the study. Jie was also

responsible for getting our procedure and results approved by Dr. Solomon throughout

the project. Jie and Ayushi learned how to safely and effectively operate the CT machine

for the robotic trials, enabling the trials to continue even when a CT tech was unavailable.

       Tristen Johnson was in charge of updates to the binder. Throughout the project,

she updated the binder with reading materials, presentations, major communications and

meetings within the group, and kept track of the timeline proposed at the beginning of the

semester. Tristen and Brandon were able to operate the robot during the study, something

that rose out of the necessity for a robotic operator when the volunteer was unavailable.

       Overall, this was a group project. Even though each team member had their

specific duties, every member was present at group meetings, clinical observations, and

the robotic and manual trials. The expected deliverables for this project were met.

Radiation study and needle accuracy study results are given in the project technical

summary. Continuations for this study include more radiation trials and extending the

study to include human patients. Dr. Solomon is already using the robot in biopsies and

ablations under FDA approval. By using radiation badges in his surgical procedures, and

even better understanding of the radiation exposure to patients and surgeons could be

appreciated.
Appendix D: References
REFERENCE

For more information on this study, please visit http://cis.chemicalsuperfreak.com

1
  Stephen B. Solomon, MD, Alexandru Patriciu, Mark E. Bohlman, MD, Louis R. Kavoussi, MD and Dan
Stoianovici, PhD. Robotically Driven Interventions: A Method of Using CT Fluoroscopy without Radiation
Exposure to the Physician. Radiology 2002;225:277-282.
2
  Su LM, Stoianovici D, Jarrett TW, Patriciu A, Roberts WW, Cadeddu JA, Ramakumar S, Solomon SB,
Kavoussi LR. Robotic percutaneous access to the kidney: comparison with standard manual access.
J Endourol. 2002 Sep;16(7):471-5.

3
 Alexandru Patriciu, Stephen Solomon MD, Louis Kavoussi MD, and Dan Stoianovici PhD. Robotic
Kidney and Spine Percutaneous Procedures Using a New Laser-Based CT Registration Method. LNCS
2208, pp. 249-257, 2001.


4
 Kato R, Katada K, Anno H, Suzuki S, Ida Y, Kogo S. Radiation dosimetry at CT fluoroscopy: physician’s
hand dose and development of needle holders. Radiology 1996; 201:576-578

								
To top