| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Clinical Observations |
1 All authors: Spatial Medicine and Physiology Unit, Department of Nuclear Medicine and Ultrasonography, CHU Trousseau, Tours, France 37044.
Received March 17, 2005;
accepted after revision November 8, 2005.
Address correspondence to P. Arbeille
(arbeille{at}med.univ-tours.fr).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CONCLUSION. Robotic telesonography can be used for reliable diagnosis without moving the patient. No false diagnoses were made in this study. A bandwidth of 250 Kbps via integrated services digital network or satellite is required for reliable diagnosis. Such a system can provide diagnostic information that is currently unavailable in isolated or inaccessible areas and on rescue vehicles.
Keywords: abdomen • echography • robot • sonography • telemedicine
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Various pathologic conditions, such as abnormal heart rate, pericardial collection, cholecystis, renal lithiasis, normal and ectopic pregnancies, ovarian cyst, acute appendicitis, and phlebitis, can affect persons whose medical history is not available. Furthermore, after trauma, a physician may look for a lesion of an abdominal organ (liver, kidney, spleen) and check if there is blood collection in the pouch of Douglas. Sonographic and Doppler examinations are noninvasive methods well suited to the diagnosis of such lesions and are currently used in hospital emergency departments. One of the major advantages of sonography in medicine is the possibility of evaluating the degree of emergency by confirming the absence of suspected pathologic conditions. In routine emergency practice, physicians face highly varied pathologic conditions. Unfortunately, many small medical centers and isolated sites do not have well-trained sonographers to perform the initial emergency evaluation. Because transfer of a patient from an isolated site (e.g., a dispensary) to an expert center can be difficult and expensive, a reliable diagnosis has to be made as soon as possible.
The objective of this study was to design and validate a method of delivery of a reliable sonographic diagnosis in an isolated site by a medical sonographer located in an expert center, that is, the nearest main hospital. It was assumed that there would be no sonographer at the patient site and that there would be a transmission system (videoconference by telephone or satellite) between the two sites for audio, video (sonographic and ambient images), and digital data transfer. The system was tested on 87 patients at our hospital.
This study was based on the use of a robotic arm specially developed for the project. The design was not based on a full robotic system but on a portable robotic arm holding a sonographic probe on the patient, mimicking the movement made by the sonographer's hand on a dummy probe at the expert center. The robotic arm was located and held on the patient's body by an operator who was not a trained sonographer. Unlike other methods designed for managing sonographic examinations of distant patients, our method requires that the expert search and control the sonographic view directly by moving a probe at the expert center. The operator at the patient site does not manipulate the probe (Figs. 1, 2, 3A, 3B and 4A, 4B, 4C, 4D).
|
|
|
|
|
|
|
|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recruitment of Patients
Over a period of 1.5 months, 87 of the patients received in the abdominal
sonography unit of our emergency department were included in the study and
underwent both conventional and robotized sonography. The patients were
informed about the entire protocol and signed an informed consent form. The
protocol was approved by the French committee for the protection of persons in
biomedical research (no. 02/07) and by our university hospital administration
(hôpital promoteur).
Robotic Design and Performance
Principles—At the patient site, the robotic arm, on which
the sonographic probe is fixed, is handled by a physician or paramedic who is
not a qualified sonographer. The robotic arm reproduces the movements of the
expert's hand. At the expert center, the expert moves a dummy probe connected
to computer 1. Via integrated services digital network (ISDN) or satellite
links, computer 1 sends the changes in coordinates representing the movements
of the dummy probe to computer 2, which is at the patient site. Computer 2
controls the robotic arm, which reproduces the movement of the dummy probe on
the sonographic probe. The expert cannot move the probe holder laterally, so
the operator at the patient site follows oral instructions given by the expert
(from the expert center) and moves the robotic arm under video surveillance
(Figs. 1,
2,
3A,
3B and
4A,
4B,
4C,
4D).
Probe holder—The probe holder is supported by a frame so that the entire system can easily be held on the patient's body. The bottom ring (7-cm diameter) of the frame maintained in contact with the patient's skin by the operator provides a stable contact [1]. To reach most organs, including the heart and abdominal and pelvic organs, the system was designed for use with all sonographic probes (mechanical oscillating probe, electronic curved probe, or phased-array probe). The system accepts probes for visualizing superficial (5-10 MHz) and deep (2-5 MHz) organs. The weight (2.5 kg) of the probe holder applies pressure that ensures contact between skin and probe. The expert can exert additional pressure up to 15 N on the sonographic probe by pushing on the dummy probe. The maximum displacement of the probe through the ring is 1.5 cm. This feature is particularly useful when the probe is applied to irregular surfaces, such as the rib cage and some abdominal areas. Because of the very limited pressure exerted and the small movements of the probe, there is no feedback control of the force applied to the probe.
Communication—The communication link used for validation consisted of ISDN channels of 128 Kbps for transferring the control data to the robot and of 192 Kbps for transferring the sonographic and ambient video images. Satellite tests were performed with two Inmarsat channels providing 64 Kbps each. With ISDN (terrestrial) telephone lines, there was less than 0.5 second of delay between movement of the dummy probe at the expert center and reception of the video image from the patient site. With satellite links, the delay was approximately 1 second.
Validation Procedure at the Hospital (ISDN) and Other Sites (Satellite Link)
The expert was in a room 20 m from the patient's room. The robot was held
on the patient by an operator whose only role was to locate the probe holder
on the appropriate area according to the instructions of the expert. The
position of the robotic probe holder was monitored in real time by the expert
using video images transmitted from the patient site to the expert site with a
conventional videoconference system. During this phase, the expert had a
real-time ambient view of both the patient and the position of the sonographic
probe and had an audio link with the operator (Fig.
3A,
3B).
The expert asked the patient-site operator to adjust the setting of the sonographic unit (gain, depth) depending on the patient's size and the echogenicity. Using simple anatomic references (e.g., breast, axilla, umbilicus), the expert told the operator where to position the robot. When the operator had correctly positioned the robotic arm in the required area (so that the expert could see at least part of the organ to be examined), the robotic probe holder system was held motionless in position. The expert then tried to find the sonographic view needed for the diagnosis by moving the dummy probe and analyzing the sonographic images received.
After the telesonographic diagnosis was made, another expert sonographer from the same department performed a conventional sonographic examination of the patient using the standard procedures as for any patient. A probe similar to that used for telesonography was used for conventional sonography, and the sonographic expert was not informed of the results of the telesonographic diagnosis. The patient remained in the same room for both examinations, which were performed at the same stage (e.g., distended bladder for pelvic sonography).
Evaluation of the Performance of Robotic Sonography
In both cases, each organ was visualized continuously from longitudinal to
transverse cross-section. To validate the method under conditions as close as
possible to real medical practice, the method was tested on different groups
of organs. One organ usually is not sufficient for a reliable diagnosis. One
has to find the expected abnormalities in one organ and make sure there are no
additional lesions in other organs that may be related to the symptoms. We
designated three groups of abdominal organs corresponding to the minimum
number of organs to be visualized for a reliable diagnosis in the most common
emergency situations: digestive symptoms (liver, gallbladder, biliary
ductules, pancreas, right kidney; n = 27), trauma such as car crashes
(liver, right and left kidneys, spleen, aorta, bladder; n = 30), and
pelvic and urinary symptoms (bladder, kidneys, prostate, uterus, ovaries;
n =30).
The reliability of the robotic sonographic examination was scored by comparison of the results with those of conventional sonography. The visualization score was expressed as the percentage [(Nrob/No) x 100] of patients for whom all organs were visualized with the robotic system (Nrob) with respect to the number of patients for whom all organs were visualized with conventional sonography (No). A score of 80% meant that with telesonography all organs were visualized in only 80% of the patients in whom all organs were visualized with conventional sonography. The score was obtained by comparing the reports from the telesonographic and the conventional sonographic examinations. This score indicated the ability of the robot to depict all of a group of organs needed to reach a diagnosis.
The objective of robotic sonography was not only to visualize groups of organs but also to find lesions and make a reliable diagnosis. Thus, in addition to the visualization score, a diagnostic score was defined to take into account both visualization of all organs and identification of lesions. The lesions not found with robotic sonography were classified into two groups. The first group included those lesions not present on the sonographic images obtained with the robot (limited robotic movement) and taken into account in both the visualization and diagnostic scores. The second group included those lesions not seen because of insufficient gray-scale resolution in the final image, and these lesions were taken into account in the diagnostic score but not in the visualization score. Other parameters of the robotic sonographic examination measured were duration of the examination for each group of organs and number of times the probe was repositioned during both robotic and conventional sonography.
Other Tests with Satellite Links
The system was tested successfully by our group with Inmarsat satellite
links between a hospital 50 km from the nearest main hospital and the main
hospital and between a military hospital and a naval rescue ship 30 km at sea.
Only six subjects were examined in this test.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The audio and video channels and the robotic control data were transmitted between the patient site and the expert center with terrestrial ISDN telephone lines. One 128-Kbps channel had sufficient bandwidth for controlling the movements of the robot. One 192-Kbps channel was allocated, as requested by the expert, either to the sonographic images or to the ambient video of the patient.
During the 87 robotic abdominal examinations, each organ (e.g., liver, gallbladder, pancreas, kidney) was correctly visualized in 91-100% of cases in comparison with conventional sonography. In 12.6% of the robotic examinations, the expert could not visualize all organs correctly; in 11 of 87 cases at least one organ of the groups of organs was not visualized. The visualization score was therefore 87.4%. The visualization score was 85% for digestive examinations, 87% for urinary examinations, and 90% for trauma evaluations.
The probe was repositioned more often for robotic examinations (6.4 ± 2 changes in position) than for conventional sonography (5.1 ± 2 changes); that is, 24% more changes in position were made with the robot. Examinations took longer with the robot (16 ± 10 minutes) than with conventional sonography (11 ± 4 minutes), an increase of 43%.
Robotic sonography depicted 26 of the 35 lesions detected with conventional sonography; that is, 26% of the lesions were missed. Two renal cysts (2-4 mm) and one kidney stone (< 1 mm) were not found because they were relatively small and the organ could not be seen clearly. Six other lesions were not identified because of low contrast of the sonographic image received at the expert center. Three (9%) of 35 lesions were not found owing to the robot's limited movements. The other six (17%) of the lesions were not identified because they were small or diffuse and low-contrast parenchymal anomalies (e.g., cirrhosis, steatosis) and the image was degraded (compression and decompression for transmission over low-bandwidth telephone lines) during transfer between the patient and the expert center.
In a separate test, four patients in a small hospital 50 km away and two
healthy subjects cruising on a navy ship 30 km at sea were examined with
abdominal telesonography via a satellite link by our group. In both cases, one
satellite antenna was installed close to the patient and the other close to
the expert (point-topoint transmission) to minimize the delay (
1 second)
between dummy probe movement and reception of the resulting image. The two
Inmarsat channels provided 64 Kbps for transferring the robotic commands and
64 Kbps for image transfer. This system made it possible to examine the
patients with abdominal sonography in close to real time (delay, 1
second), as for the ISDN links (delay, 0.5 second), and the organs were
correctly visualized. However, the quality of the image received at the expert
center was not as good as with ISDN links because of the low image data rate
(64 Kbps).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In the second proposed method, the patient site has a 3D sonographic unit that can capture within a few seconds all of the echoes of a volume containing the organ suspected of having a lesion. The sonographic information is sent to the expert center and processed later by an expert [3-7]. The probe has to be positioned in many anatomic areas to make sure that the organ is inside one of the volumes examined. Unfortunately, 3D sonographic systems consist of high-technology equipment that is rather heavy and expensive. They therefore are rarely available at isolated sites.
In the third method, a robotic arm holding a conventional 2D echo probe at the patient site is teleoperated by an expert sonographer via telephone or satellite links (Figs. 1 and 2). The robotic arm acts as an extension of the expert's hand on the patient. The movement of the probe on the patient's skin is entirely controlled by the expert's hand. The contribution of the nonsonographer or paramedic at the patient's side is restricted to maintaining the robotic arm motionless on the patient at the anatomic site indicated by the expert. The operator at the patient site does not manipulate the sonographic probe.
Robotic surgery is probably the most popular application of robotics in medicine [8-10]. It is difficult to compare telesonography with telesurgery because the issues are different. A surgical robotic arm must reproduce very accurately the complex movements of the surgeon's hands. The surgical robot must work in three dimensions (6 degrees of freedom), whereas the sonographic robotic arm only needs to orient the probe head on the surface of the skin (the probe head stays in the same plane). Finally, there is additional human interface between the sonographic robot and the patient because the operator who handles the robotic arm can relocate it or remove it from the patient's skin at any moment. Consequently, the patient cannot be hurt by the sonographic robotic arm. Satellite links can be safely used because, unlike the situation with surgical robots, interruption of data transfer is not dangerous to the patient.
The robotic arm is a light (2.5 kg) handheld device with a 7-cm ring that is applied to the patient's skin. No discomfort was reported by the patients in this study. None of the patients refused to be examined with the robotic system, and none had complaints. Patients and physicians communicated in real time through videoconference links.
The images received from the patient site made it possible to evaluate gallbladder content and wall thickness, the quality of the hepatic parenchyma and ducts and bile ductules, the morphologic features and content of the pancreas, the integrity of the abdominal aorta, the morphologic features and content of the kidneys and urinary tract, the integrity of the spleen, the absence of perihepatic and perirenal collections, the absence of fluid in the pouch of Douglas, and the size and content of the uterus, ovaries, and prostate. Such observations should be sufficient to answer most of the questions that arise in an emergency that necessitates sonography.
The lack of abdominal images in a few cases was related to the technical limitations of the robotic arm compared with a sonographer's hand. The pressure to be applied on the skin was sometimes not sufficient, which made examination of the spleen difficult. Absence of translation of the probe increased the number of times the probe holder was repositioned, but this was not a problem.
The duration (16 ± 10 minutes) of telesonographic examination of each group of organs was approximately 40% longer that of conventional sonography. The most time-consuming maneuver was positioning of the probe over the organ, because placement was guided verbally and monitored by video by the expert. On the other hand, the probe did not have to be repositioned many times (6.4 ± 2 times per group of organs) before activation of the robotic arm. This number, however, represents a 24% greater rate of repositioning with the robotic arm than for conventional sonography and increased the duration of the examination. By the end of the study, we found that the easiest way of positioning the robotic arm on the patient was for the expert to show the desired location on his own body through the videoconference system. He then asked the operator to move the robotic arm slowly 1 or 2 cm to the right, left, up, or down.
For validation, examination of each organ was limited to 5 minutes to simulate emergency examinations. This time constraint was a strong one and was one of the causes of failure to visualize all organs completely in 11 of 87 cases.
The mean visualization score for the robot was 87.4% for the abdomen. There were no false diagnoses. Incomplete visualization of the organs in the other 12.6% cases made it impossible to make a diagnosis. The diagnostic score was 74% for all lesions and 81% excluding the low-contrast lesions not visible because of degradation of the image during transfer between the patient site and the expert center. These lesions were excluded because they are not of great importance to emergency diagnosis. At the end of the study, we used 128 Kbps for the robotic information and 384 Kbps for the sonographic image, which maintained the gray scale in the sonographic image and made it possible to identify diffuse lesions not visible at only 192 Kbps. We found that transfer of sonographic images in real time requires a minimum data rate of approximately 256 Kbps and a minimum frame rate of about 15 fps. These rates made real-time examination possible, and the images were not degraded.
Another limitation of the system was the quality of the sonograms. With
low-performance sonographic systems, the images that reached the expert center
were of somewhat poor quality because they had been degraded (e.g.,
compressed, decompressed). With medium-performance sonographic systems, the
degradation caused by the processing required for image transfer was not
visible on the images that reached the expert center. Nevertheless, despite
their small display screens, some small portable sonographic units provided
very good quality images to the expert center, even after having been
processed for telephone or satellite transfer. Results of preliminary tests
have suggested that use of medium-performance sonographic systems costing more
than $50,000 (
40,000) will yield images of sufficient quality for
compression and decompression without major degradation. Using 384 Kbps and a
medium-performance sonographic system, we were able to detect abnormal diffuse
lesions that were not visible at a 192-Kbps image data rate or with
low-performance sonographic systems.
Although the ergonomics of the robotic system were good, use of a dummy probe introduces restrictions, whereas during conventional sonography the movements of the sonographer's hand are totally free. During the examination, the robotic system was held mainly vertically on the skin, so the patients were asked to change position (e.g., lying supine, left side, right side) depending on the organ to be examined. The weight of the robotic unit exerted enough pressure to keep the probe in contact with the skin and maintain the system stable. If the patient could not be moved, as in trauma cases, the operator held the robot perpendicular to the skin. In addition, the probe could be moved along its long axis ± 1.5 cm if the expert pushed on the dummy probe. The experts had to move the dummy probe slowly to reconstruct in their minds a 3D image of what they saw in the successive sonographic images. The sonographers (radiologists, cardiologists, obstetricians) who participated in validation of the method needed no more than 2 hours to get used to the system. They reported that the robotic arm was totally transparent for them. The system (dummy probe, communication link, and robotic arm) acted as an extension of their hands.
In many emergency situations, the need for a quick and reliable diagnosis of disorders of various organs (abdominal, cardiac, pelvic) requires several sonographers (e.g., radiologist, cardiologist, gynecologist). The possibility of teleoperation of the sonographic examination from an expert center using a robotic system will provide patients living in isolated sites the same diagnostic performance as that available in main hospitals where all types of experts are available.
Secondary hospitals in developed countries also need robotic equipment. Every day general medicine physicians in such hospitals send three to 10 patients to main hospitals for sonographic examinations for problems such as abdominal, pelvic, and leg pain. In more than 50% of cases, sonography does not reveal significant damage, and the patient can return home. Installing a robotic system in most secondary hospitals would result in at least 30% reduction in ambulance transfers to main hospitals, saving money and ensuring that patients with true diagnoses would be transferred faster.
In December 2003, six telesonographic examinations were performed via Inmarsat satellite between a naval rescue ship 30 km at sea and the military hospital ashore and between a small hospital and a hospital center 50 km away. The results of these tests confirmed that despite movement of the ship and delay caused by the satellite transmission, the sonographic examinations were teleoperated correctly. In both cases, the delay was approximately 1 second. Consequently, the examinations were performed in close to real time while the probe was moved slowly. The modest quality of the images was related to the low image data rate (64 Kbps) of the channel used.
Four robotic arms now are installed in four secondary hospitals 40 to 60 km from our university hospital. They are linked to us by four ISDN telephone lines of 128 Kbps each (386 Kbps for images, 128 Kbps for robot orders) to maintain the integrity of the sonographic images and are used for routine examinations. In the near future, the system will be modified to work via Internet, Wi-Fi, and universal mobile telecommunication system networks. A portable version of the system will be tested on rescue vehicles and major air and sea passenger carriers. The system has also been selected for use on the Antarctica Concordia base and the International Space Station.
Telerobotic sonography can be used for reliable diagnoses without moving the patient, and no false diagnoses have been reported. A bandwidth of 250 Kbps (ISDN or satellite) is required for reliable diagnosis. Such a system can provide diagnostic information that is currently unavailable in isolated or inaccessible areas or on rescue vehicles.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
