larry@kitty.UUCP (Larry Lippman) (02/14/89)
I thought I would take a stab at creating some interesting topics
of discussion (and speculation) concerning the application of engineering
to the challenges of biomedical measurement. I will pose a measurement
problem, and see if anyone can propose a solution. The problems should
be of interest to those who like to apply physics and electronics to real
world applications. After a week or so of discussion, I will post the
actual methods used and the rationale behind them. If this article
results in productive discussion, I will continue to pose these measurement
problems on a regular basis. if not, well, it seemed like an interesting
idea at the time. :-)
A brief bit of background: I manage the development of scientific
and chemical process measurement and control instrumentation. I am both
an electrical engineer and biochemist, and have worked on a number of
biomedical instrumentation projects in the past 19 years that I have been
in private industry. I have had firsthand experience with anything that
I may describe on this topic. In recent years, though, most of my work has
been in the chemical process area, but I am still manage one active
biomedical instrumentation project.
Some of you reading the Net are physicians or others who may have
intimate knowledge and experience with the subject matter; I would urge
that you sit back for a few days and see what some others have to say.
Okay, here we go...
Reduced to simplest terms, the heart is a pump. As a pump it
therefore has an output rating, which is termed "cardiac output" and
is usually measured in liters/minute. While the heart is actually a
synchronized dual pump, for the purpose of this discussion we are
dealing with the left side of the heart which pumps oxygenated blood
into the arterial system. The left side of the heart has one outlet:
the aorta.
Normal cardiac output ranges between 4 and 8 liters/min. Cardiac
output is normalized to "cardiac index" which divides cardiac output
by surface area of the skin; normal cardiac index ranges from 2.5 to 5
liters/min/meter^2. Cardiac output and cardiac index are valuable
measurements used in the diagnosis of cardiovascular disease.
Now the problem: how can we measure cardiac output without
major surgery to expose the aorta and attach a flowmeter? There
have been three major techniques used over the years, all of which
share a common principle. Hint: bear in mind that we are taking a
relatively short-term measurement, so you can assume that cardiac
output is the same as venous return to the right atrium. You may also
assume that the output of the right ventricle into the pulmonary artery
is the same as the pulmonary venous return into the left atrium; i.e.,
both sides of the heart are pumping at the same rate with the blood
volume within the lungs remaining the same.
<> Larry Lippman @ Recognition Research Corp., Clarence, New York
<> UUCP: {allegra|ames|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry
<> VOICE: 716/688-1231 {att|hplabs|mtune|utzoo|uunet}!/
<> FAX: 716/741-9635 {G1,G2,G3 modes} "Have you hugged your cat today?" oconnor@nuke.steinmetz (Dennis M. O'Connor) (02/18/89)
An article by larry@kitty.UUCP (Larry Lippman) says: ] Now the problem: how can we measure cardiac output without ] major surgery to expose the aorta and attach a flowmeter? Well, you could : Inject the subject with a technicium bound into a pyrophosphate compound and place the subject in front of a gamma-ray camera. Record the amount of gamma-rays ( = k*amount of technicium = K*amount of blood ) in the heart using a timescale much finer than a single heartbeat. Calculate the difference beteen the maximum amount of blood in the ventricle during a heartbeat and the minimum amount. Multiply by beats/time-unit to obtain a flow rate. This technique assumes all the valves in the heart are functioning correctly. Or, you could : Use doppler ultrasound on the aorta to measure the velocity, and use an NMR or CAT scan to determine the cross-section, and multiply to obtain the flow rate. Or you could : Inject a VERY small transmitter ( about the size of a red blood cell would be nice, even though that will get stuck in the capillaries ) into a major vein that broadcast a long pseudo-random number and track it using multiple ( at least three ) recievers, using the delay to each receiver to precisely locate the unit, and measure it change in position as it flows through the aorta, then use an NMR or CAT scan to determine the cross-section, and multiply to obtain the flow rate. Or you could : Use bolus injection of a radioisotope and see how quickly it moves to and through the aorta using a gamma-camera, then compute the flow as you did for the radio transmitter. Or you could : Compute the velocity of the flow out through the valve by analysis of the noise it makes as it passes through, then proceed as abover to get the volume of material going thrnough. I thought of heating a section of the aorta with microwaves or a particle accelorator and measuring the cooling rate, monitoring the temperture with a fiber-optic thermometer ( we could use the same fiber to heat it with a laser, come to think of it ) or monitoring the temperature with NMR sensors, but this seems a bit invasive to me ( although not as bad as inserting a flowmeter ). -- Dennis O'Connor oconnor%sungod@steinmetz.UUCP ARPA: OCONNORDM@ge-crd.arpa "...the bastard got away. God always fights on the side of the bad man"
lharris@gpu.utcs.toronto.edu (Leonard Harris) (02/20/89)
In article <13175@steinmetz.ge.com> oconnor%sungod@steinmetz.UUCP writes: >An article by larry@kitty.UUCP (Larry Lippman) says: >] Now the problem: how can we measure cardiac output without >] major surgery to expose the aorta and attach a flowmeter? > >Well, you could : >Inject the subject with a technicium bound into a pyrophosphate compound >and place the subject in front of a gamma-ray camera. Record the amount >of gamma-rays ( = k*amount of technicium = K*amount of blood ) in the >heart using a timescale much finer than a single heartbeat. Calculate >the difference beteen the maximum amount of blood in the ventricle >during a heartbeat and the minimum amount. Multiply by beats/time-unit >to obtain a flow rate. This technique assumes all the valves in the >heart are functioning correctly. > >Or, you could : >Use doppler ultrasound on the aorta to measure the velocity, and use >an NMR or CAT scan to determine the cross-section, and multiply to >obtain the flow rate. > >Or you could : >Inject a VERY small transmitter ( about the size of a red blood cell >would be nice, even though that will get stuck in the capillaries ) into >a major vein that broadcast a long pseudo-random number and track it >using multiple ( at least three ) recievers, using the delay to each >receiver to precisely locate the unit, and measure it change in position >as it flows through the aorta, then use an NMR or CAT scan to determine >the cross-section, and multiply to obtain the flow rate. > >Or you could : >Use bolus injection of a radioisotope and see how quickly it moves to >and through the aorta using a gamma-camera, then compute the flow >as you did for the radio transmitter. > >Or you could : >Compute the velocity of the flow out through the valve by analysis of >the noise it makes as it passes through, then proceed as abover to get >the volume of material going thrnough. > >I thought of heating a section of the aorta with microwaves or >a particle accelorator and measuring the cooling rate, monitoring >the temperture with a fiber-optic thermometer ( we could use >the same fiber to heat it with a laser, come to think of it ) or >monitoring the temperature with NMR sensors, but this seems a bit >invasive to me ( although not as bad as inserting a flowmeter ). > > >-- > Dennis O'Connor oconnor%sungod@steinmetz.UUCP ARPA: OCONNORDM@ge-crd.arpa > "...the bastard got away. God always fights on the side of the bad man" Get yourself a Swan-Ganz catheter - put it into the internal jugular or subclavian and thread it into the left vatrium. Inject saline or water of a known temperature and measure the temp of the saline at the tip of the swan ganz using the temperature probe incorporated. The decay curve of temperature gives the cardiac output.
dietz@cs.rochester.edu (Paul Dietz) (02/21/89)
How about using the Mossbauer effect to measure the Doppler shift of gamma rays from small tracer particles? This extraodinarily sensitive phenomenon can measure speeds as low as a few centimeters per second. One might use it to get a measure of the amount of blood near the heart moving at a certain velocity relative to the detector. I read somewhere the Mossbauer effect has been used to measure the velocity distribution in ant colonies (given them sugar-coated tracer particles). Also, I think someone has used it to measure the vibration of eardrums. Paul F. Dietz dietz@cs.rochester.edu
sac@conrad.UUCP (Steven A. Conrad) (02/23/89)
In article <13175@steinmetz.ge.com> oconnor%sungod@steinmetz.UUCP writes: >An article by larry@kitty.UUCP (Larry Lippman) says: >] Now the problem: how can we measure cardiac output without >] major surgery to expose the aorta and attach a flowmeter? > >Well, you could : >Inject the subject with a technicium bound into a pyrophosphate compound >and place the subject in front of a gamma-ray camera. Record the amount >of gamma-rays ( = k*amount of technicium = K*amount of blood ) in the >heart using a timescale much finer than a single heartbeat. Calculate >the difference beteen the maximum amount of blood in the ventricle >during a heartbeat and the minimum amount. Multiply by beats/time-unit >to obtain a flow rate. This technique assumes all the valves in the >heart are functioning correctly. This method is used quite successfully for determining the ejection fraction as you describe, but is rather lousy for actual flow calculations. Because of attenuations, inability to record all emitted radiation, etc. it doesn't cut it. >Use doppler ultrasound on the aorta to measure the velocity, and use >an NMR or CAT scan to determine the cross-section, and multiply to >obtain the flow rate. This method is used clinically. Actually the cross section is obtained with 2D/M mode echocardiography, usually with the same machine used for the Doppler. However, it is subject to quite a bit of error, most commonly due to errors in calculation of aortic area and in the assumptions about the flow. It assumes a flat velocity profile, which truly occurs only in the very base of the aorta. It is much better used for following changes in cardiac output than for measuring the actual value. >Inject a VERY small transmitter ( about the size of a red blood cell >would be nice, even though that will get stuck in the capillaries ) into >a major vein that broadcast a long pseudo-random number and track it >using multiple ( at least three ) recievers, using the delay to each >receiver to precisely locate the unit, and measure it change in position >as it flows through the aorta, then use an NMR or CAT scan to determine >the cross-section, and multiply to obtain the flow rate. A little wild, maybe? Nonetheless, it is well known that a single red blood cell may travel at a variety of velocities, depending on its proximity to the aortic wall, the diameter of the vessel, eddy currents at the valves, etc. Wouldn't be practical. >Use bolus injection of a radioisotope and see how quickly it moves to >and through the aorta using a gamma-camera, then compute the flow >as you did for the radio transmitter. Again, the major problem is calculating aortic diameter. >Compute the velocity of the flow out through the valve by analysis of >the noise it makes as it passes through, then proceed as above to get >the volume of material going thrnough. No good relationship between noise and velocity. >I thought of heating a section of the aorta with microwaves or >a particle accelorator and measuring the cooling rate, monitoring >the temperture with a fiber-optic thermometer ( we could use >the same fiber to heat it with a laser, come to think of it ) or >monitoring the temperature with NMR sensors, but this seems a bit >invasive to me ( although not as bad as inserting a flowmeter ). This is probably the closest to the most common way in which we do measure cardiac output. There are two major ways of measuring C.O. The oldest is the Fick method, in which the oxygen content difference across the pulmonary capillaries is related to the amount of oxygen taken up by the lungs: C.O. = [oxygen consumption] / [arterial-venous oxygen content difference] A more recent introduction are the indicator dilution methods. Cardiogreen dye was first used (by Wood from the Mayo Clinic, I believe). It is injected into the right atrium, and its concentration curve is monitored downstream, in the arterial system. With the introduction of the bedside pulmonary artery catheter in the early 70's with thermistor probes, however, the indicator now used is heat (actually cold, or negative heat). The injection of cold fluid is made into the right atrium, and the temperature is monitored in the pulmonary artery, with the right ventricle mixing the bolus well. It is based on the following simple principle: _ [mass] = _/ F(t)C(t) Mass is replaced by heat quantity, and concentration C by temperature. If we assume flow F to be constant, then we can take it out of the integral and rearrange: _ [cardiac output] = [heat quantity] / _/ T(t) There are some correction constants and others such as specific heat, density of fluids and blood, etc. that don't affect the overall meaning of the above equation. Notice that a number of assumptions are made. However, the method has less than about 15% biological variation, and has been accepted clinically in the critical care unit and cardiac cath lab. Steve. -- Steven A. Conrad, Department of Medicine (Critical Care) Louisiana State University Medical Center, Shreveport, LA UUCP: sac@conrad.UUCP, Internet: conrad@manta.pha.pa.us "Silence is the only successful substitute for brains"
sac@conrad.UUCP (Steven A. Conrad) (02/23/89)
In my previously submitted article, I neglected to mention a method which has recently hit the market, thoracic bioimpedance. The method involves running a constant current through the chest by means of electrodes placed at the neck and base of the thorax, and measuring voltage changes. A sufficiently high frequency is used so as not to intefere with biological function. Equations have been worked out for relating the rate of change of impedence with each stroke of the heart to the volume ejected by the heart (stroke volume). Multiply this by heart rate and you have cardiac output. We have evaluated this instrument and found it to be of sufficient accuracy for clinical uses. It perhaps best supplements, not replaces the pulmonary artery catheter, since the catheter has other important functions. Several investigators have reported reasonable correlations with thermodilution. Steve -- Steven A. Conrad, Department of Medicine (Critical Care) Louisiana State University Medical Center, Shreveport, LA UUCP: sac@conrad.UUCP, Internet: conrad@manta.pha.pa.us "Silence is the only successful substitute for brains"