Date: Thu, 29 Jan 1998 10:23:24 -0800 (PST)
From: Lee Hirsch <firstname.lastname@example.org>
Subject: Project Report
Measurement of Neonatal Residual Volume
Our group enthusiastically accepted this project because it was different than most of the other posted projects. When students enter the biomedical engineering curriculum as freshmen they think of designing prosthetics or imaging equipment. But as we progress through the curriculum we see that BME's do much more; they strive to design or reengineer any facet of the medical field which may need fixing or improving. Maybe the protocol for hospital patient admissions is inefficient or the orthopedics department needs a user-friendly database for examining case studies-Biomedical Engineers do all of these things. And it is typically the latter sorts of projects that are requested of BME's for senior projects.
But the neonatal project was unique, and it caught our eyes. The sponsoring doctor wanted a couple of BME's to find a technique of measuring the dynamic change residual volume of neonatal lungs in order to properly adjust their respirator settings. He proposed a technique called Electrical Impedance Tomography (EIT). This made the project even more exciting since our group mates were very interested in imaging. But this is where we screwed up. We were so fascinated with the idea of coming up with either a simulation or working model of this project that we did not bother to perform a sufficient amount
of preliminary research. The only thing we knew was that we wanted this project! However, if we would have shopped around, talked with professors, and done some reading we might have saved ourselves a lot of time (a whole semester in fact).
So here were the blunders of our project:
If our group would have bothered to do some simple early research on
the neonate's lung physiology, then we would have seen the pitfalls of this project. The
average adult lung has a volume of about 7 liters. A neonate's lung, however, is about 64
times less; this is roughly 100 ml's. The ratio of total lung capacity/residual volume
does not change much throughout development; as a result, the average residual volume of a
neonate would be about 12 ml's. This is a very small amount of
volume to detect-even with a spirometer. Moreover, this volume has to be distinguished from the other present volumes such as anatomical dead space. If we had this knowledge before accepting the project, then we would have seen that this volume would be extremely difficult to detect in vivo no matter what approach we took.
This technique was recommended because it was known that the impedance across the chest
of a patient is proportional to the total lung volume. Therefore, it would make sense that
you could possibly extrapolate the residual volume from the impedance remaining after
expiration. But we later found out that although the lung volume is proportional to
impedance, the change in lung volume was not proportional to the change in impedance. In
other words, the EIT may give a static representation of the lung volume, but it could not
give us the dynamic measurements that we wanted.
So after blindly accepting the project, we began our research and quickly discovered a couple of other complications. First of all, there are two types of EIT's: 1) Single and 2) Multifrequency. The latter allows the image to have a full impedance spectrum at each pixel. The first type of EIT was appropriate for our purposes because we weren't as interested in an image as we were in a correlation between impedance and volume-the second type would have given us a lot of useless data. The single frequency EIT consists of a 16 electrode system that surrounds the circumference of the infant's chest. Each electrode takes a turn sending out a 20 kHz signal while the other fifteen measure the impedance relative to the signal source. After each electrode has been a signal source, background projection may be used to calculate the impedances throughout the infant's chest.
A fundamental problem with this design is that the electrodes are placed onto the chest. Therefore, there is no constant reference frame. For example, a MR scanner has a coil around the patient that emits/detects the radio frequencies. But with the electrodes placed on the chest, this reference frame moves with the patient. This is a problem due to the fact that the patient "flattens" when they lie down; as a result, the EIT would pick up less resistance between electrodes and mistakenly detect this phenomena as a decrease in volume.
The EIT uses electricity to measure volume. Unfortunately, the body happens to be an
excellent conductor of electricity, but in variable amounts. Electricity will follow the
path of least resistance, which, as Dr. Galloway noted, will cause the EIT current to flow
in non linear paths through the chest cavity. This causes mathematical problems when
calculating impedances that are not encountered with the highly linear x-rays of the CAT
scanner. Blood volume is largely responsible for changing the impedance measurements of
the EIT. One final bungler of the EIT measurements is the heart. The heart is, of course,
a large source of electrical activity that often gives rise to considerable noise in the
left lung measurements.
After coming to the resignation that the EIT would not work, we tried to come up with
another solution to the problem. We consulted with various Dr.'s, Professors, and
engineers about solutions, but none of them fared fruit. Regardless, here are a couple of
other solutions that we brainstormed.
Deuterated Water Vapor
This technique would have worked in a similar fashion to the He dilution method. The
helium dilution method is the widely used
method for measuring residual volumes. It uses simple mass balance equations. The patient inhales a known concentration of He gas after forced expiration. The volume and concentration of He in the exhaled gas is then measured. What ever He gas was not exhaled will be residual volume (subtracting anatomical and instrument dead space).
We proposed a deuterated water vapor method to perform this same operation in the hopes
that the deuterated water could be measured with greater precision than the He gas
concentration. Deuterated water has a strong absorbance in the IR region of the spectrum.
Therefore a laser/spectrophotometer setup could closely monitor the concentrations
of inhaled/exhaled deuterated water vapor. But there were two problems with this solution.
1) Even if it were more precise than its predecessor (He dilution), it would still be
nondynamic. 2) After talking with Dr. Harris we found that the water vapor would adhere to
the lung walls (unlike a gas), which would therefore include surface area measurements as
Next, we thought that ultrasound might provide an image of the neonatal lung. But
problems arise when the acoustic wave travels past the skin and into the ribs. These bony
structures have a higher acoustical impedance than the flesh, and as a result will reflect
a large portion of the sound waves. Therefore, it is difficult to receive a decent image
of the lungs behind the ribs. And even if we could, we are left with the hard time of
discerning between different volumes within the image, such as anatomical dead space,
alveolar space, and residual volume.
One other idea that we did not really pursue was the acoustic resonance approach.
Imagine that the lungs were like hollow pipes; then hypothetically, if you could emit the
proper frequency the sound would resonate within this volume. Of course, the lungs are in
no way shaped like a cylindrical pipe. Moreover, the perimeter of the lungs is surrounded
by a rough surface of surfactants, branches, and alveoli which will deflect the sound
waves in an erratic matter. Nevertheless, Dr. Galloway thought that this was a fun
So here we are at the end of a whole semester with no progress on our project. Left
with no other alternative, we had to abandon it. But with the benefit of hindsight, we see
how easy it would have been to avoid the errors of this project. The first thing anybody
should do before accepting a project is perform a little preliminary research on the
subject; make sure that the project is suitable for undergraduate work instead of a Ph.D.
thesis. After doing much research on the EIT, we talked with Dr. Galloway about the idea.
He told us that Ph.D's have been working with the EIT for many years and have yet to
a reliable instrument. If we had this knowledge from the start, we would have saved ourselves a lot of time.
Another thing to avoid is narrowing your scope too soon. We painted ourselves into a
corner very quickly by hastily choosing the EIT as our measurement device. We came up with
the other measurement ideas after we spent considerable time on the EIT. Even though none
of the other ideas worked either, the point is the same--do a lot of brainstorming at the
beginning of the project. Then select the best approach; do not bet on the first horse
that you find.