Infusion Pump to Be Used in Magnetic Resonance Imaging
Infusion Pump to be Used in Magnetic Resonance Imaging
Team Members:
Nate Gaeckle – Team Leader
Tim Eng – BWIG
Client:
George C. Newman, M.D., Ph.D.
UW Department of Neurology
Advisor:
Naomi Chesler, Ph.D.
UW Department of Biomedical Engineering
October 9, 2003
Problem Statement:
The current magnetic resonance imaging (MRI) infusion pump used in the UW-Health Hospital, Spectris Solaris, is inefficient, time consuming, and doesn’t perform to the standard that the client wishes. There are three main obstacles that need to be overcome in order to improve MRI images. Currently a technician must manually change the gadolinium and saline reservoirs before each bolus and infusion. The second hindrance is that the technician must use a different concentration of each liquid for a second infusion. On average, this changeover takes about seven minutes. In this time, the circulatory pattern of the patient could change due to movements of the patient. This has a negative effect on the results of the MRI. Finally, the gadolinium and saline come in pre-packaged amounts. Some patients do not need the whole bottle of solution, and it is tossed out after the infusion. In this way, the current set-up is inefficient.
Background Information:
MRI is a technique that is used to visualize the inside of a patient non-invasively. The physics behind it are complex, but will be summed up generally. The images that are created are due to the spinning of protons in water molecules of the body. When a patient is placed under the magnet, which is on the order of 1.5-3 tesla, the protons of water molecules align in certain orientations. A radio frequency is then emitted from the machine which changes the alignment of the protons. This change is detected by a coil which displays an image (“How MRI Works, 2003). Gadolinium is a contrast that is injected into the patient in order to control how fast the excited water molecules return to their original state. Different amounts of gadolinium perform with various results for patients. By administering the gadolinium properly for each patient, the images will form more resolutely. Gadolinium comes pre-packaged in bottles, but it is unlikely that a patient needs all of the gadolinium to perform the scans. The extra fluid is simply thrown away. A 20 mL bottle of gadolinium is priced at about $100, but UW-Health purchases this fluid in bulk at about $60 a bottle. On average, the patient only uses about ¾ of a bottle, so this is a significant amount of lost money after many infusions. The bottles of saline pose the same hindrance, but saline is not as expensive to purchase. The reason only a certain amount of contrast and saline can be used by the Spectris Solaris is due to the design of the injector. The design consists of two syringes of fixed volumes that contain the two fluids. Therefore, each time the injector needs to be used, a technician must refill the syringe with the pre-packaged bottles. A different pumping mechanism would ameliorate this problem.
There are two stages in an MRI scan. The first stage is the bolus, which is a short injection of 15 mL of contrast at a rate of 3 mL/s. This is followed by a 20 mL injection of normal saline at 3 mL/s. It is used to determine the mean transit time which then elucidates the cerebral blood flow. After the blood flow is determined, the proper rate and concentration of contrast can be administered during the second stage of the scan, which is the infusion. The infusion is also dependent on the weight of the patient. During the infusion, 1 mL of the combination of contrast and saline is given every second. The amount of each depends on the results from the bolus, and the weight of the patient. Currently, there is about a 7 minute gap between the bolus and the infusion where a technician must replace the almost empty syringes with new contents. Also, the amount of gadolinium usually needs to be diluted for different patients. This takes time. During this gap, the patient can make movements, which disrupts the blood flow that was found during the bolus. It would be optimal to have about a 6-7 second break in-between the bolus and infusion where saline could be injected to flush out the contrast. The proposed design reduces the waste of contrast and saline, decreases the amount of time in-between the bolus and infusion, and is able to change the concentration of contrast.
Device Constraints:
In order to design a proper device, one main issue needs to be factored in. That is the device must be able to be used safely in a MRI room. Because there is a magnet on the order of 1.5-3 teslas, there can be no ferrous material unless it is properly shielded. It would be potentially harmful to the technicians and patient if this was not taken into account. In addition, the pump must be able to maintain its sterility between usages with different patients. This point given, it would be most advantageous to design some sort of pump that can be easily cleaned, or that does not make any contact with the solutions passing through so that there is no possibility of deterioration after many procedures. On average, there are approximately 2-3 perfusion protocols per day, but it can get up to as many as 7. The pump will be used on a daily basis, so it must be very reliable and perform with accurate results.
In order to solve the problem of wasting contrast and saline, a pump needs to be developed that can provide a continuous flow of liquid. This is unlike the injectors that are currently used, where only a certain amount of solution can be inserted before having to refill. The flow rate of contrast and saline need to have a range of .25 mL/second to 5 mL/second in order to suit the research needs of the client. This should be able to be controlled through the use of computer software from the observation room. This range of flow rate would solve the gap in-between the bolus and the infusion. Because there would be a continuous flow, and the ability to change the concentration of the gadolinium, there would be no reason for a technician to have to dilute the gadolinium to a specified concentration. It would all be done through the use of computer software. Currently, the Spectris Solaris is mounted on a portable stand. This makes it convenient for the technicians to move the injector close to the patient, but it is not a mandatory requirement that the client has specified
Alternative Solutions:
Solution 1: With the constraint to only non-ferromagnetic materials allowed in the MRI room, a solution was proposed to mount the device in the control room, while running tubing to the patient in the MRI room (Figure 1). Multiple bars comprising a finger pump (Figure 2) will press against the tubing in succession to force the liquid along (Hoffmann 2002). Run by servo motors, these finger pumps are able to supply continuous flow. The main advantage of this solution is the elimination of the non-ferromagnetic limitation. There are vast possibilities in material selection, which would facilitate the design process. The use of finger pumps also simplifies sterilization, since nothing but tubing makes contact with the liquids. Disadvantages also arise, primarily in the area of efficiency. While no syringes need to be replaced as with the current device, the amount of tubing required for each patient is immense. Furthermore, after the bolus and infusion, contrast will remain in the tubing, which will be discarded. Accurate flow rates would be difficult to maintain with such lengths of tubing. Very little information was found related to the medical use of finger pumps, so locating two pumps with the required specifications would be difficult.
Figure 1: Schematic of Solution 1
Figure 2: A finger pump (Hoffman 2002)
Solution 2: In contrast with Solution 1, this design will be placed inside the MRI room. A similar setup with two pumps run by servo motors will be used. The choice of gear pumps (Figure 3) was made based on their ability to provide continuous flow as well as the ease with which flow rates can be altered. The whole device (pumps and motors) will need shielding by MuShield® to prevent both attraction to the MRI magnets and interference in the images. A plastic stand will be used to allow movement. This is desired to minimize the length of tubing required. The main problem with this design is that the gears will be exposed to gadolinium and saline, making it difficult to keep this device sterile.
Figure 3 Figure 4: A gear pump (Hoffman 2002)
Proposed Solution:
The design that currently is most feasible (Figure 5) consists of two peristaltic pumps (Figure 6) run by bell-type armature motors, which are controlled by a computer in a separate room. The possibility of using labVIEW software is currently being researched to see if it can accurately control the pumps. Peristaltic pumps consist of rollers linked to the ends of arms, which are able to rotate, moving the fluid along the flexible tubing (Hoffman 2002). Titanium, a non-ferromagnetic metal, will be used for the arms of the pumps, providing strength while eliminating magnetic interference. The principle behind this concept is that the rollers create a vacuum, which pulls the fluid along. This way, each liquid may be continuously pumped at varying flow rates as specified by the technician. The device will be mounted on a plastic stand, providing portability and durability. On the tubing just below the bags of contrast and saline, there will be a check valve followed by a micron filter. These are standard on IV units to prevent leakage from the bags as well as to ensure the filtering of exceptionally small particles. The check valve consists of a roller on a track, which pinches the flexible tubing, preventing seepage. As another standard precaution, a micron filter will be placed below the point of mixture of the two solutions.
The Bell-type armature motor that will be incorporated is made of all non-ferrous material, with the exception of the magnet that is found in all motors. Even though the magnet of the motor creates a magnetic field, the magnet is placed in a position that minimizes the flux out of the system (Lahmadi, 2003). In addition, the motor will be surrounded by magnetic shielding. This will again decrease the chance of the MRI magnet interfering with the motor’s duty. The amount of torque needed to rotate the arms of the peristaltic pump in order to transport fluid is still undetermined.
In order to solve the problem created by the immense magnetic field, a MuShield® model #062 will be formed in a 1 inch radius around the motor (Grille 2003). It is unknown whether this material will totally prevent attraction of the device to the MRI magnets, so further research needs to be performed. It was noticed that the current device used in the UW hospital has shielded cables that block the magnetic field formed by current carrying wires. Similar shielding will be utilized for all cables running to the control room.
The attributes of this solution include uninterrupted flow of gadolinium and saline along with adjustable flow rates. An online search for peristaltic pumps yielded information concerning the widespread use of peristaltic pumps in current medical devices. Thus, there are few worries about these pumps performing in a medical environment. The design of the peristaltic pump offers an answer to the sterilization dilemma that Solution 2 encountered. Since no parts of the pump touch the fluids, only the tubing will need to be replaced.
Figure 5 Figure 6: A peristaltic pump (Hoffman 2002)
Potential Problems:
There are a few problems that have arisen since contemplating the design. The first is finding a way to shield the magnet of the servo motor from the giant magnetic field of the MRI. There are a few materials, such as Gaussian surfaces made by various companies such as MuShield, that are made to shield magnetic fields, but unfortunately they consist of a nickel iron alloy. The very small amount of iron makes this is a semi-ferrous substance that absorbs the magnetic field and redirects it. It is not known how much of an effect the magnetic field will have on this alloy. There may be no effect at all, but if there is, a simple modification can be made. The design on a stand would change to a fixed position. It could possibly be fastened to the wall, or even the floor. This would get rid of the threat in the scanning room. By fastening the device, it hinders the ability for the mechanism to be portable, but the client has specified that it is not pertinent that the device is movable. However, by attaching it somewhere else, the amount of tubing has to be increased so that it can still reach the magnet. It is still unknown how increasing the amount of tubing will affect the flow rate. Another problem is all of the calculations that are going to be involved. In order to determine the flow rate, there are certain variables that need to be known such as, the correct diameter of tubing, the rate at which the motor must rotate, and the dynamics of the two fluids. These are all factors that haven’t been considered as of yet, but it is imperative that they are known.
References:
Grille, David. Personal Interview. 29 September, 2003.
Hoffman, Russell D. “The Internet Glossary of Pumps.” Accessed: 22 September, 2003.
“How MRI Works” http://www.physics.buffalo.edu/grad/MedPhys/mri.html. Accessed: 25 September, 2003.
Lahmadi, Wahid. Personal Interview. 3 October, 2003.
Appendix
MRI Infusion Pump
Team Members:
Nate Gaeckle
Tim Eng
9-18-03
Function: The current MRI infusion pump is inefficient, time consuming, and doesn’t perform to the standard that the client wishes. Currently a technician must manually change the gadolinium and saline reservoirs before each bolus infusion. After the bolus, the technician must use a different concentration of each liquid for a second infusion. On average, this changeover takes about seven minutes. In this time, the circulatory pattern of the patient could change due to movements of the patient. This has a negative effect on the results of the MRI. Also, the gadolinium and saline come in pre-packaged amounts. Some patients do not need the whole bottle of solution, and it is tossed out after the infusion. In this way, the current set-up is inefficient.
Client Requirements:
- The pump must be made of a non-ferrous material.
- The pump should be design so that no gadolinium or saline is wasted.
- It must be able to accurately control the flow rates of gadolinium and saline throughout the scan by the use of computer sequencing.
- Flow rate needs to have a range from .1 mL/s to 5 mL/s.
- Needs to be a way to sterilize each accessory of the pump
- Must be able to perform the contrast infusion right after the bolus.
- The pump must be able to infuse saline alone, gadolinium alone or a combination of the two.
Physical and Operational Characteristics:
a. Performance Requirements: The device will be used on a daily basis, changing syringes for each patient for sterility. The components of the pump in contact with the saline or Gadolinium must be able to be sterilized. It must be able to survive differing amounts of standby and usage times as well as be durable to perform equally well for all patients throughout the day. On average, the hospital sees 2-3 patients a day, but it can get up to about 7 patients a day. The average usage time of the pump is around several seconds for the bolus injection of contrast followed by infusion lasting 1-2 minutes. The pump must be readily available for use at all times. Air bubbles cannot be present in the pump as they will cause artifacts on the image. The amount of time allotted between patients is approximately 10 minutes for set up. Although it is not of high priority, it may be possible to reduce the set up time.
b. Safety: The device must be MRI compatible, so as to not pose any complications with the scan or endanger the patient’s life. The device must be able to operate in a magnetic field of 0.5 to 2.0T. Therefore it must not contain any ferromagnetic metal. Additionally, the tubing connecting the pump to the patient’s IV must be long and pliable to allow for comfort and ease of movement of the patient in and out of the MRI scan area (~5-6 feet). Coiled tubing is now being used and is satisfactory for the client’s needs. Tubing that is not sufficient in length can cause danger and discomfort to the patient as well as affecting scan quality with slight instability of the patient. The syringes cannot be put under such pressure that they or the tubing will burst.
c. Accuracy and Reliability: The pump must be reliable and accurate to deliver fluid in 1 (or 0.5) mL increments. No leakage or coagulation is allowed, as the infusion must be controlled in a constant fashion by the operator. The bolus must be injected in the range of 10-25 mL with +/- 0.5 mL accuracy at a flow rate of 3 mL/sec +/- 0.2 mL/sec. Gadolinium and saline are then optimally infused simultaneously at diluted contrast with a flow rate of 0.25 to 0.35 mL/sec for Gadolinium and 0.65 to 0.75 mL/sec for saline +/- 0.02 mL/sec or better.