REMOTE AUSCULTATORYPATIENT MONITORING DURING MA6NETIC RESONANCEIMAGING S. Henneberg, MD,* B. Htk,~" L. Wiklund, MD,¢ and G. Sjtdin¢
From the *Department of Anesthesiology and Intensive Care, Odense University Hospital, Odense, Denmark; tthe Electronics Department, Institute of Technology, University ofUppsala, Uppsala, Sweden; and ~the Department of Anesthesiology and Intensive Caxe, University Hospital, Uppsala, Sweden. Received Apr 5, 1990, and in revised form Jan 25, 1991. Accepted for publication Feb 12, 199t. Address correspondence to Dr Henneberg, Department of Anesthesiology and Intensive Care, Odense University Hospital, DK-5000, Odense C, Denmark.
Henneberg S, H6k B, Wiklund L, Sj6din G, Remote auscultatory patient monitoring during magnetic resonance imaging. J Clin Monit 1992;8:37-43 ADSTRACT.A system for patient monitoring during magnetic
resonance imaging (MRI) is described. The system is based on remote auscultation of heart sounds and respiratory sounds using specially developed pickup heads that are positioned on the precordium or at the nostrils and connected to microphones via polymer tubing. The microphones operate in a differential mode outside the strong magnetic field to reduce various sources of interference from the MRt equipment. After amplification, the signal is transmitted as infrared light to a small, battery-operated receiver and a headphone set. Thus, the patient can be simultaneously auscultated both inside and outside the shielded MRI room by infrared transmission through a metal mesh window. Bench tests of the system show that common mode acoustic noise is suppressed by approximately 30 dB in the frequency region of interest (100-1,000 Hz), and that polymer tubing having a diameter of approximately 2 mm can be used for efficient sound transmission. Recordings in situ show satisfactory detection of both heart sounds and respiratory sounds, although the signal is somewhat masked by noise during imaging. A clinical test incorporating 17 sedated or anesthetized patients was also performed. In all but four cases, the quality of the breath and heart sounds was regarded as acceptable or better. KEY WORDS.Monitoring: heart, respirations. Measurement techniques: magnetic resonance imaging; infrared telemetry.
Magnetic resonance imaging (MRI) has evolved as a major diagnostic tool in radiology. A review o f the physical and radiologic aspects o f M R I has been given by Kean and Smith [1]. M R I at the present state o f development requires relatively long exposure times, up to 15 minutes. This in turn requires that the patient be either cooperative or anesthetized. The latter is always the case with small children, and sometimes with adults. The anesthesiologist is then left with a choice o f either administering a sedative drug or general anesthesia. In either situation, monitoring o f ventilation and circulation becomes mandatory. H o w e v e r , this monitoring can pose technical difficulties, due to the fact that M R I equipment is a strong source o f various types o f interference, yet is also a delicate instrument, sensitive to different kinds o f disturbances. O u r purpose is to point out remote auscultation as a possible means o f patient monitoring during MRI. The possibility has been discussed before [2,3], but without specifying and addressing the interference-related p r o b lems that are encountered. In our experience, conventional stethoscopes, both acoustic and electronic, cannot be used successfully due to p o o r signal to noise (S/N) ratio [4]. We present a new approach to this problem, based Copyright © 1992 by Little, Brown and Company 37
38 Journal of Clinical Monitoring Vot 8 No 1 January 1992
on a differential technique to improve S/N ratio. The system is described and measurements on its performance are presented, together with a clinical study incorporating both sedated and anesthetized patients. MATERIALSAND METHODS
Design Principles The design of a monitoring system for MRI should be adapted to the specific conditions set by the MRI equipment. These conditions include the following (Fig 1): 1. The patient is placed within a magnet several cubic meters in dimension and is not accessible to direct observation. 2. The magnet is placed within an electromagnetically shielded r o o m with limited possibilities for signal transmission through the shield (feed-through and a metal-mesh window are usually available). 3. The static magnetic field within the magnet is usually in the range of 0.02 to 2 T, but falls offrapidly outside the magnet. 4. Despite the relatively small magnetic field strength outside the magnet, there is concern among the personnel involved about long-term exposure to this environment. Thus, it is desirable to avoid unnecessary exposure. 5. A pulsed-gradient magnetic field is present during imaging. This field also creates mechanical vibrations and acoustic noise with an appreciable intensity. Amplitude, time sequence, and frequency spectrum may vary considerably from one setup to another and also among different exposure types.
f
MRI control equipment
\
I IR transmitter // [ ,..,Z, Micro-
k U l P°nes
~Metal mesh Polymer tubing window
Pickup heads for heart sounds and respiratory sound!
II
i
[
MRI magnet 0.02 - 2 T
\ Shielded room
Fig 1. Schematicpresentation of the setup during MRI.
6. Radio frequency waves in the order of tens of megahertz are generated during exposures. 7. Sensors preferably should be nonmetallic and definitely nonmagnetic in order not to disturb the imaging process. 8. The monitoring equipment should not generate any signals in the radio frequency band used for the imaging process. In our system, the primary signals are accoustic. C o m m o n microphone elements work well approximately 1 m outside the magnet, but not within it. Therefore, to transmit the signal from the source to outside the magnet, we adopted a simple acoustic transmission line consisting of a plastic tube. O f course, such a transmission line will redistribute the frequency spectrum o f the signal by causing both attenuation and standing waves within the tubes. In an open-ended tube, frequencies coinciding with the relation fr open = i'c/4 L
i = 1, 3, 5. . . .
(1)
will be augmented and, in a closed-ended configuration, the corresponding relation is fr dosed = j'c/2 L
j = 1, 2, 3. . . .
(2)
L is the length of tubing and c is the velocity of sound (340 m/s in air). Attenuation will result from loss due to the air viscosity. Assuming laminar flow, the flow resistance, R, determined by the ratio between the pressure drop and the flow rate, will follow Poiseuille's law: R = 128 bt,1/.'rrd4
(3)
where ~ is the viscosity and d is the diameter of the tubing. Note the strong diameter dependence of the resistance. This attenuation and redistribution of the frequency spectrum is, of course, a serious limitation if one wants to extract precise information from the character of the sound. If, however, the object is to detect the presence and timing o f certain sounds (heart and respiratory sounds), such a transmission system may be fully adequate. There is an optimization problem, however, to designing an acoustic system that is adapted to the specific requirements. The mechanical vibrations and acoustic noise from the equipment must be suppressed by orders o f magnitude, especially in MRI equipment with a horizontal magnetic field orientation, where the patient is confined within a narrow tunnel that functions as an acoustic resonator. The mechanical vibrations of the MRI equip-
Henneberg et al: Auscultatory Patient Monitoring
I
Housing
A Microphone
°0\ I
J
Voltage supply Output
Ground
Voltage supply
Output
Turbulent air flow
Ground
B Threeway
Microphone Voltage supply
;e ord,a, Tubing
//
Output
Ground
Nasal/oral head
C
Fig 2. Design principles of the microphones and the pickup heads. (A) Microphone bridge connected to a precordial pickup head, sensing skin vibrations. (B) Open tube endings (oral~nasal pickup head) sensing turbulent airflow and connected to a corresponding microphone bridge. (C) Arrangement with three-way stopcocks to connect one microphone bridge to two pickup heads.
ment are usually so intense that they can be felt by the hand against the patient bed. Acoustic noise is suppressed in our system by operating two acoustic transmission lines differentially. This is done by connecting two standard electric elements (Matsushita Electric WM-063, Yokohama, Japan) in a bridge configuration, as depicted in Figure 2A. The output voltage of this configuration is proportional to the instantaneous difference in the acoustic pressure at the microphone inputs. Individual variations in sensitivity of the microphone elements are compensated for by introducing shunt resistors and adjusting their value until the response to a common source ls zero. This adjustment is done during assembly of a microphone pair; no readjustments are needed. A further increase in the S/N ratio was achieved by designing dedicated microphone heads for precordial application (heart sounds and respiratory sounds) and oral or nasal application (respiratory sounds). The precordial microphone head (Fig 2A) actually converts mechanical vibrations into airborne acoustic signals. It consists of a plastic housing with an interior air chamber divided into two equal volumes. The divi-
39
sion is obtained by a mass element suspended by a flexible rubber membrane. When the housing is subjected to vibrations, the mass-membrane combination will cause compression and expansion of the respective parts of the chamber. If both parts are connected to the differential microphone inputs, the output signal will represent the vibration of the housing. Suppression of acoustic noise is obtained both from the differential design and from the fact that the interior (the air-filled chambers and the connecting tubes) constitutes a closed system. The oral/nasal microphone head simply consists of two open tube endings connected to a similar pair of microphone elements. The detection principle is based on the turbulence and the associated acoustic noise created by an obstructive object in a stream of air. Suppression of ambient sound is provided by the differential technique, as mentioned before. The turbulence taking place in the two tube endings are uncorrelated with each other; therefore, the signals S~ and S2 in the two microphone elements add according to the following equation: S = (812 -+ 822) 1/2
(4)
The mechanical vibrations of the bed induced by the MRI equipment can be damped by placing the patient on a soft mattress or a pillow. The use of acoustic signals throughout makes it possible to connect and mix signals from different sites simply by introducing stopcocks between the tubings. Figure 2C depicts such a system, which uses three-way stopcocks to connect one precordial and one oral/nasal probe to one differential microphone. Evidently a variety of systems could be built, based on stereophonic or quadraphonic audio technology. We built and tested only one of these. In our system, the acoustic signals are converted into frequency-modulated infrared light and wirelessly transmitted [3] to battery-operated stethoscope receivers. These can be tocated either within the shielded room or in the control room. In the latter case, the signal is transmitted through the metal mesh window. A detailed description of the transmission system has been given earlier by H6k et al [3]. The system is available at a price of $4,200 from H6k Instrument AB, Flottiljgatan 55, S-721 31 V~ster~s. Sweden. Experiments
We first tested the system in a series of bench-top experiments and then in actual use on patients during normal imaging procedures. The clinical part of the study was
40 Journal of Clinical Monitoring Vol 8 No 1 January 1992
approved by the local ethical committee. Parental consent was also obtained for each patient. The signal transmission properties of the system have been tested by performing measurements of the crucial components. The infrared transmission system had been found earlier to operate linearly and with a flat frequency response in the frequency range of interest [3]. Therefore, for this study, the experiments could concentrate on the transmission properties of the differential microphone and the acoustic transmission line. We measured the characteristics of the differential microphone by applying a sinusoidal sound source with variable frequency to the microphone and calculating the ratio o f the signal output obtained with both inputs connected to the source compared with only one input connected. This experiment gives a quantitative assessment of the suppression of ambient noise over the whole frequency range of interest. The second experiment was outlined to study the effect of the acoustic transmission line. Sinusoidal sound was applied to one end o f a plastic tube, and measurements of the sound intensities were performed at both ends o f the tube. The ratio of the measured values then gives a measurement o f the transmission at a particular frequency. A clinical test was performed on 17 children and teenagers. Ages ranged from 1 month to 18 years, and body weights ranged from 4.4 to 70 kg. Fifteen of these patients were sedated with a mixture of pethidine, promethazine, and chlorpromazine; this sedation was supplemented with small intravenous doses of thiopental as required. T w o patients were given a general anesthetic. The sedated patients breathed spontaneously, whereas assisted ventilation was used in the anesthetized patients. The stethoscope was applied, and the audibility of heart and breath sounds was estimated on a scale of 0 to 3, where 0 is no audibility at all, 1 is poor, 2 is acceptable, and 3 is good. The results were noted before and during exposure. Finally, in separate studies, tape recordings were made in situ on 2 volunteers and 2 patients. The main part (15 patients) of the clinical study was performed and evaluated by one of the authors (G.S.) at the MRI laboratory at the University Hospital, Uppsala, Sweden, on Siemens Magnetom 0.5 T equipment. T w o patients were investigated by Dr. Hans Sellddn at St Gtran's Hospital, Stockholm, on a 0.02 T Instrumentarium unit. RESULTS
In Figure 3, the results of measurements of the characteristics of the differential microphone are shown. In
Line Chart for columns: X I Y 1
i i
-40
.
. 200
1 400
•
i
l 600
•
I 800
iii.
.
l 1000
.
t
I 1200
.
I 1400
•
I 1600
Frequency (Hz)
Fig 3. Measurement of the attenuation of common acoustic signals provided by the microphone bridge in comparison with a single microphone element.
10,
0
Line • CoJumn 2 OCoturnn 5
200
Chert
for columns: X 1 ¥ 1
-.-X1¥4
(}Column 3
400
600
Frequency (Hz)
ACotumn 4
800
1'000
1200
Fig 4. Sound transmission in PVC tubing of different bore and length. Solid squares--1 mm diameter, 50 cm length; open circles--I mm diameter, 150 cm length; solid triangles--1 mm diameter, 300 cm length; solid circles--2 mm diameter, 300 cm length.
the frequency range 100 to 1,000 Hz, the suppression of common mode sound is approximately 30 dB. At lower frequencies, the suppression is somewhat smaller. A possible explanation is mismatching of the frequency response of the two microphone elements in this region. It may be possible to improve this performance by adding a reactive compensation in the bridge circuit. The sound transmission in PVC (polyvinyl chloride) tubing is shown in Figure 4. In Figure 4, tubing with a bore of I m m was tested for three different lengths. In the 3-m tubing, resistive attenuation (eq. 3) dominates over the standing wave phenomenon, thereby producing a decreased transmission of sound at higher frequendes. In the shorter lengths (too short in this actual application), resonant peaks can be detected according to equations 1 and 2. Using tubing with a larger bore, 2 m m (Fig 4), gives a satisfactory transmission in the frequency range of interest, but with some attenuation of the resonant peaks. A still larger bore would be more clumsy and would also give an unwanted "volume loading effect" on the microphone head.
Henneberg et al: Auscuttatory Patient Monitoring
..............................................................................................................
41
i ............................
H i .......................... :
Fig 5. Complete system for remote auscuttatory monitoring.
A photograph of the complete system is shown in Figure 5; in Figure 6, the precordial and oral/nasal heads applied to a patient are shown. Figure 6 shows recordings of signals in situ from a 12-year-old volunteer subject. The signals from a precordial and an oral/nasal head were transmitted separately and recorded on tape before being displayed on a digital oscilloscope. The recordings in Figures 6A and B were taken before and during exposure, respectively. The upper tracings of each figure clearly show" the heart sounds (precordial head), whereas the respiratory sounds can be seen most clearly on the lower tracings (oral/nasal head). It is also evident from Figure 6B that additional noise is generated during exposure. In the clinical study, the quality of heart and breath sounds was estimated to be acceptable or better before exposure in all but one patient, who had a severely distorted anatomy, making heart sounds completely inaudible. During exposure, the sound quality was disturbed by noise from the MRI equipment and became unacceptably low in 4 of the 17 patients. The detailed results are shown in Figure 7. The stethoscope proved to be easy to apply without causing any discomfort to the patients during or after the procedure. After the application is done, guided by the intensities of the heart and breath sounds, readjustments are rarely needed. Occasionally, however, the nasal/oral head became malpositioned due to involuntary movements. Its design could possibly be improved in this respect. DISCUSSION
The problem of patient monitoring during MRI has been addressed by several workers. Roth et al recorded
A
i........................................................................................................................................... i
B Fig 6. Recording of the signals from a 12-year-old girl (A) before and (B) during exposure. The upper tracings of both A and B originate from the precordial pickup head, whereas the lower tracings come from the oral~nasal head. Horizontal scale is 0.2 s/div; vertical scale is 0.2 V/div.
blood pressure and electrocardiogram [4], while Selldfin et al [5] used photoplethysmography to record heart rate. Shellock [6] evaluated several techniques, including automatic blood pressure monitoring, electrocardiography, capnography, and laser Doppler anemometry. Goudsouzian and Vacanti [2] reported problems with using sophisticated monitoring equipment such as pulse oximeters, and claimed that simple stethoscopes and blood pressure cuffs are more operable. Recently, during a workshop that included 10 Scandinavian clinics, the monitoring problem during MRI was emphasized.
42 Journal of CtinicaI Monitoring Vol 8 No 1 January 1992
18.
18,
lSl
16, o
o
141 1 21 1 01
81 o
61 41
I cr
2
o
........I. I o
;
2
0
3
Audibility
Audibility
18.
18.
_~
1 61 8 'E
o3 I
161
141
141 121
"10 U~
2
"E
1 ol
101
81
61
o
4;
41 0
I i 1
O' Audibility
Fig 7. Quality estimation of respiratory and heart sounds before and during M R I exposure in 17 patients. 3, Good audibility; 2, acceptable; 1, poor; and O, no audibility. The vertical axes represent the numbers of patients in the respective category.
Collaboration between different clinics was encouraged [7], although the problems may differ considerably from one clinic to another. Basically, pulse oximetry should be a suitable technique since it relies on optical signals that by themselves do not interact with the MRI equipment. However, in existing instruments, the driving-current pulses to the light sources frequently cause disturbances on the MRI images, and the photodetector has interference caused by the gradients and radiowaves of the MRI equipment. Recently, an accidental, severe burn injury was reported [8], apparently caused by a looped cable that induced a high current through a pulse oximeter probe. This accident indicates that patient-connected metallic leads should be avoided completely in the MRI environment.
8
I
;
t
I
Audibility
In our experience, auscultation with precordial stethoscopes can rarely be used due to the high mechanical and acoustic noise of the MRI equipment. Similarly, blood pressure recording using the auscultatory principle is hardly feasible. Side-stream capnography can be applied to intubated patients but not to sedated, unintubated patients. Our technique of remote auscultation is useful in MRI units having a low to medium-high magnetic field. It is now being routinely used in our hospital for all MRI procedures requiring sedation or general anesthesia. In MRI units with a field o f 1.5 T or higher, the mechanical and acoustic interference during exposure is still unacceptably high. We are now working on improvements that we believe will solve these problems. The approach involves more precise balancing of the microphone pair and more sophisticated signal-filtering techniques.
Henneberg et al: Auscultatory Patient Monitoring
The authors are indebted to Dr Hans Selld6n, St. G6ran's Hospital, Stockholm, for his contributions to the clinical part of the study. This study was supported by the Swedish Board for Technical Development.
REFERENCES 1. Kean D, Smith M. Magnetic resonance imaging principles and applications. London: Heinemann Medical Books, 1986 2. Goudsouzian N, Vacanti FX. Current questions in patient safety: monitoring for MRI. APSF Newsletter 1987:33-35 3. H6k B, Bythell V, Bengtsson M. Development of a wireless stethoscope for auscultatory monitoring during anaesthesia. Med Biol Eng Comput 1988;26:317-320 4. Roth JL, Nugent M, Gray JE, Julsrud RR. Patient monitoring during magnetic resonance imaging. Anesthesiology 1985;62:80-83 5. Selld6n H, De Chateau P, Ekman G, et al. Circulatory monitoring of children during anaesthesia in low-field magnetic resonance imaging. Acta Anaesthesiol Scand t990;34:41-43 6. Shellock FG. Monitoring during MRI. Med Electron 1986;100:93-97 7. LindoffB. Problem med overvakning vid MRT. Med Tek 1989:38 8. Shellock FG, Slimp GL. Severe burn of the finger caused by using a pulse oximeter during MR imaging. Am J Roentgenol 1989;153:1105
43
From the *Department of Anesthesiology and Intensive Care, Odense University Hospital, Odense, Denmark; tthe Electronics Department, Institute of Technology, University ofUppsala, Uppsala, Sweden; and ~the Department of Anesthesiology and Intensive Caxe, University Hospital, Uppsala, Sweden. Received Apr 5, 1990, and in revised form Jan 25, 1991. Accepted for publication Feb 12, 199t. Address correspondence to Dr Henneberg, Department of Anesthesiology and Intensive Care, Odense University Hospital, DK-5000, Odense C, Denmark.
Henneberg S, H6k B, Wiklund L, Sj6din G, Remote auscultatory patient monitoring during magnetic resonance imaging. J Clin Monit 1992;8:37-43 ADSTRACT.A system for patient monitoring during magnetic
resonance imaging (MRI) is described. The system is based on remote auscultation of heart sounds and respiratory sounds using specially developed pickup heads that are positioned on the precordium or at the nostrils and connected to microphones via polymer tubing. The microphones operate in a differential mode outside the strong magnetic field to reduce various sources of interference from the MRt equipment. After amplification, the signal is transmitted as infrared light to a small, battery-operated receiver and a headphone set. Thus, the patient can be simultaneously auscultated both inside and outside the shielded MRI room by infrared transmission through a metal mesh window. Bench tests of the system show that common mode acoustic noise is suppressed by approximately 30 dB in the frequency region of interest (100-1,000 Hz), and that polymer tubing having a diameter of approximately 2 mm can be used for efficient sound transmission. Recordings in situ show satisfactory detection of both heart sounds and respiratory sounds, although the signal is somewhat masked by noise during imaging. A clinical test incorporating 17 sedated or anesthetized patients was also performed. In all but four cases, the quality of the breath and heart sounds was regarded as acceptable or better. KEY WORDS.Monitoring: heart, respirations. Measurement techniques: magnetic resonance imaging; infrared telemetry.
Magnetic resonance imaging (MRI) has evolved as a major diagnostic tool in radiology. A review o f the physical and radiologic aspects o f M R I has been given by Kean and Smith [1]. M R I at the present state o f development requires relatively long exposure times, up to 15 minutes. This in turn requires that the patient be either cooperative or anesthetized. The latter is always the case with small children, and sometimes with adults. The anesthesiologist is then left with a choice o f either administering a sedative drug or general anesthesia. In either situation, monitoring o f ventilation and circulation becomes mandatory. H o w e v e r , this monitoring can pose technical difficulties, due to the fact that M R I equipment is a strong source o f various types o f interference, yet is also a delicate instrument, sensitive to different kinds o f disturbances. O u r purpose is to point out remote auscultation as a possible means o f patient monitoring during MRI. The possibility has been discussed before [2,3], but without specifying and addressing the interference-related p r o b lems that are encountered. In our experience, conventional stethoscopes, both acoustic and electronic, cannot be used successfully due to p o o r signal to noise (S/N) ratio [4]. We present a new approach to this problem, based Copyright © 1992 by Little, Brown and Company 37
38 Journal of Clinical Monitoring Vot 8 No 1 January 1992
on a differential technique to improve S/N ratio. The system is described and measurements on its performance are presented, together with a clinical study incorporating both sedated and anesthetized patients. MATERIALSAND METHODS
Design Principles The design of a monitoring system for MRI should be adapted to the specific conditions set by the MRI equipment. These conditions include the following (Fig 1): 1. The patient is placed within a magnet several cubic meters in dimension and is not accessible to direct observation. 2. The magnet is placed within an electromagnetically shielded r o o m with limited possibilities for signal transmission through the shield (feed-through and a metal-mesh window are usually available). 3. The static magnetic field within the magnet is usually in the range of 0.02 to 2 T, but falls offrapidly outside the magnet. 4. Despite the relatively small magnetic field strength outside the magnet, there is concern among the personnel involved about long-term exposure to this environment. Thus, it is desirable to avoid unnecessary exposure. 5. A pulsed-gradient magnetic field is present during imaging. This field also creates mechanical vibrations and acoustic noise with an appreciable intensity. Amplitude, time sequence, and frequency spectrum may vary considerably from one setup to another and also among different exposure types.
f
MRI control equipment
\
I IR transmitter // [ ,..,Z, Micro-
k U l P°nes
~Metal mesh Polymer tubing window
Pickup heads for heart sounds and respiratory sound!
II
i
[
MRI magnet 0.02 - 2 T
\ Shielded room
Fig 1. Schematicpresentation of the setup during MRI.
6. Radio frequency waves in the order of tens of megahertz are generated during exposures. 7. Sensors preferably should be nonmetallic and definitely nonmagnetic in order not to disturb the imaging process. 8. The monitoring equipment should not generate any signals in the radio frequency band used for the imaging process. In our system, the primary signals are accoustic. C o m m o n microphone elements work well approximately 1 m outside the magnet, but not within it. Therefore, to transmit the signal from the source to outside the magnet, we adopted a simple acoustic transmission line consisting of a plastic tube. O f course, such a transmission line will redistribute the frequency spectrum o f the signal by causing both attenuation and standing waves within the tubes. In an open-ended tube, frequencies coinciding with the relation fr open = i'c/4 L
i = 1, 3, 5. . . .
(1)
will be augmented and, in a closed-ended configuration, the corresponding relation is fr dosed = j'c/2 L
j = 1, 2, 3. . . .
(2)
L is the length of tubing and c is the velocity of sound (340 m/s in air). Attenuation will result from loss due to the air viscosity. Assuming laminar flow, the flow resistance, R, determined by the ratio between the pressure drop and the flow rate, will follow Poiseuille's law: R = 128 bt,1/.'rrd4
(3)
where ~ is the viscosity and d is the diameter of the tubing. Note the strong diameter dependence of the resistance. This attenuation and redistribution of the frequency spectrum is, of course, a serious limitation if one wants to extract precise information from the character of the sound. If, however, the object is to detect the presence and timing o f certain sounds (heart and respiratory sounds), such a transmission system may be fully adequate. There is an optimization problem, however, to designing an acoustic system that is adapted to the specific requirements. The mechanical vibrations and acoustic noise from the equipment must be suppressed by orders o f magnitude, especially in MRI equipment with a horizontal magnetic field orientation, where the patient is confined within a narrow tunnel that functions as an acoustic resonator. The mechanical vibrations of the MRI equip-
Henneberg et al: Auscultatory Patient Monitoring
I
Housing
A Microphone
°0\ I
J
Voltage supply Output
Ground
Voltage supply
Output
Turbulent air flow
Ground
B Threeway
Microphone Voltage supply
;e ord,a, Tubing
//
Output
Ground
Nasal/oral head
C
Fig 2. Design principles of the microphones and the pickup heads. (A) Microphone bridge connected to a precordial pickup head, sensing skin vibrations. (B) Open tube endings (oral~nasal pickup head) sensing turbulent airflow and connected to a corresponding microphone bridge. (C) Arrangement with three-way stopcocks to connect one microphone bridge to two pickup heads.
ment are usually so intense that they can be felt by the hand against the patient bed. Acoustic noise is suppressed in our system by operating two acoustic transmission lines differentially. This is done by connecting two standard electric elements (Matsushita Electric WM-063, Yokohama, Japan) in a bridge configuration, as depicted in Figure 2A. The output voltage of this configuration is proportional to the instantaneous difference in the acoustic pressure at the microphone inputs. Individual variations in sensitivity of the microphone elements are compensated for by introducing shunt resistors and adjusting their value until the response to a common source ls zero. This adjustment is done during assembly of a microphone pair; no readjustments are needed. A further increase in the S/N ratio was achieved by designing dedicated microphone heads for precordial application (heart sounds and respiratory sounds) and oral or nasal application (respiratory sounds). The precordial microphone head (Fig 2A) actually converts mechanical vibrations into airborne acoustic signals. It consists of a plastic housing with an interior air chamber divided into two equal volumes. The divi-
39
sion is obtained by a mass element suspended by a flexible rubber membrane. When the housing is subjected to vibrations, the mass-membrane combination will cause compression and expansion of the respective parts of the chamber. If both parts are connected to the differential microphone inputs, the output signal will represent the vibration of the housing. Suppression of acoustic noise is obtained both from the differential design and from the fact that the interior (the air-filled chambers and the connecting tubes) constitutes a closed system. The oral/nasal microphone head simply consists of two open tube endings connected to a similar pair of microphone elements. The detection principle is based on the turbulence and the associated acoustic noise created by an obstructive object in a stream of air. Suppression of ambient sound is provided by the differential technique, as mentioned before. The turbulence taking place in the two tube endings are uncorrelated with each other; therefore, the signals S~ and S2 in the two microphone elements add according to the following equation: S = (812 -+ 822) 1/2
(4)
The mechanical vibrations of the bed induced by the MRI equipment can be damped by placing the patient on a soft mattress or a pillow. The use of acoustic signals throughout makes it possible to connect and mix signals from different sites simply by introducing stopcocks between the tubings. Figure 2C depicts such a system, which uses three-way stopcocks to connect one precordial and one oral/nasal probe to one differential microphone. Evidently a variety of systems could be built, based on stereophonic or quadraphonic audio technology. We built and tested only one of these. In our system, the acoustic signals are converted into frequency-modulated infrared light and wirelessly transmitted [3] to battery-operated stethoscope receivers. These can be tocated either within the shielded room or in the control room. In the latter case, the signal is transmitted through the metal mesh window. A detailed description of the transmission system has been given earlier by H6k et al [3]. The system is available at a price of $4,200 from H6k Instrument AB, Flottiljgatan 55, S-721 31 V~ster~s. Sweden. Experiments
We first tested the system in a series of bench-top experiments and then in actual use on patients during normal imaging procedures. The clinical part of the study was
40 Journal of Clinical Monitoring Vol 8 No 1 January 1992
approved by the local ethical committee. Parental consent was also obtained for each patient. The signal transmission properties of the system have been tested by performing measurements of the crucial components. The infrared transmission system had been found earlier to operate linearly and with a flat frequency response in the frequency range of interest [3]. Therefore, for this study, the experiments could concentrate on the transmission properties of the differential microphone and the acoustic transmission line. We measured the characteristics of the differential microphone by applying a sinusoidal sound source with variable frequency to the microphone and calculating the ratio o f the signal output obtained with both inputs connected to the source compared with only one input connected. This experiment gives a quantitative assessment of the suppression of ambient noise over the whole frequency range of interest. The second experiment was outlined to study the effect of the acoustic transmission line. Sinusoidal sound was applied to one end o f a plastic tube, and measurements of the sound intensities were performed at both ends o f the tube. The ratio of the measured values then gives a measurement o f the transmission at a particular frequency. A clinical test was performed on 17 children and teenagers. Ages ranged from 1 month to 18 years, and body weights ranged from 4.4 to 70 kg. Fifteen of these patients were sedated with a mixture of pethidine, promethazine, and chlorpromazine; this sedation was supplemented with small intravenous doses of thiopental as required. T w o patients were given a general anesthetic. The sedated patients breathed spontaneously, whereas assisted ventilation was used in the anesthetized patients. The stethoscope was applied, and the audibility of heart and breath sounds was estimated on a scale of 0 to 3, where 0 is no audibility at all, 1 is poor, 2 is acceptable, and 3 is good. The results were noted before and during exposure. Finally, in separate studies, tape recordings were made in situ on 2 volunteers and 2 patients. The main part (15 patients) of the clinical study was performed and evaluated by one of the authors (G.S.) at the MRI laboratory at the University Hospital, Uppsala, Sweden, on Siemens Magnetom 0.5 T equipment. T w o patients were investigated by Dr. Hans Sellddn at St Gtran's Hospital, Stockholm, on a 0.02 T Instrumentarium unit. RESULTS
In Figure 3, the results of measurements of the characteristics of the differential microphone are shown. In
Line Chart for columns: X I Y 1
i i
-40
.
. 200
1 400
•
i
l 600
•
I 800
iii.
.
l 1000
.
t
I 1200
.
I 1400
•
I 1600
Frequency (Hz)
Fig 3. Measurement of the attenuation of common acoustic signals provided by the microphone bridge in comparison with a single microphone element.
10,
0
Line • CoJumn 2 OCoturnn 5
200
Chert
for columns: X 1 ¥ 1
-.-X1¥4
(}Column 3
400
600
Frequency (Hz)
ACotumn 4
800
1'000
1200
Fig 4. Sound transmission in PVC tubing of different bore and length. Solid squares--1 mm diameter, 50 cm length; open circles--I mm diameter, 150 cm length; solid triangles--1 mm diameter, 300 cm length; solid circles--2 mm diameter, 300 cm length.
the frequency range 100 to 1,000 Hz, the suppression of common mode sound is approximately 30 dB. At lower frequencies, the suppression is somewhat smaller. A possible explanation is mismatching of the frequency response of the two microphone elements in this region. It may be possible to improve this performance by adding a reactive compensation in the bridge circuit. The sound transmission in PVC (polyvinyl chloride) tubing is shown in Figure 4. In Figure 4, tubing with a bore of I m m was tested for three different lengths. In the 3-m tubing, resistive attenuation (eq. 3) dominates over the standing wave phenomenon, thereby producing a decreased transmission of sound at higher frequendes. In the shorter lengths (too short in this actual application), resonant peaks can be detected according to equations 1 and 2. Using tubing with a larger bore, 2 m m (Fig 4), gives a satisfactory transmission in the frequency range of interest, but with some attenuation of the resonant peaks. A still larger bore would be more clumsy and would also give an unwanted "volume loading effect" on the microphone head.
Henneberg et al: Auscuttatory Patient Monitoring
..............................................................................................................
41
i ............................
H i .......................... :
Fig 5. Complete system for remote auscuttatory monitoring.
A photograph of the complete system is shown in Figure 5; in Figure 6, the precordial and oral/nasal heads applied to a patient are shown. Figure 6 shows recordings of signals in situ from a 12-year-old volunteer subject. The signals from a precordial and an oral/nasal head were transmitted separately and recorded on tape before being displayed on a digital oscilloscope. The recordings in Figures 6A and B were taken before and during exposure, respectively. The upper tracings of each figure clearly show" the heart sounds (precordial head), whereas the respiratory sounds can be seen most clearly on the lower tracings (oral/nasal head). It is also evident from Figure 6B that additional noise is generated during exposure. In the clinical study, the quality of heart and breath sounds was estimated to be acceptable or better before exposure in all but one patient, who had a severely distorted anatomy, making heart sounds completely inaudible. During exposure, the sound quality was disturbed by noise from the MRI equipment and became unacceptably low in 4 of the 17 patients. The detailed results are shown in Figure 7. The stethoscope proved to be easy to apply without causing any discomfort to the patients during or after the procedure. After the application is done, guided by the intensities of the heart and breath sounds, readjustments are rarely needed. Occasionally, however, the nasal/oral head became malpositioned due to involuntary movements. Its design could possibly be improved in this respect. DISCUSSION
The problem of patient monitoring during MRI has been addressed by several workers. Roth et al recorded
A
i........................................................................................................................................... i
B Fig 6. Recording of the signals from a 12-year-old girl (A) before and (B) during exposure. The upper tracings of both A and B originate from the precordial pickup head, whereas the lower tracings come from the oral~nasal head. Horizontal scale is 0.2 s/div; vertical scale is 0.2 V/div.
blood pressure and electrocardiogram [4], while Selldfin et al [5] used photoplethysmography to record heart rate. Shellock [6] evaluated several techniques, including automatic blood pressure monitoring, electrocardiography, capnography, and laser Doppler anemometry. Goudsouzian and Vacanti [2] reported problems with using sophisticated monitoring equipment such as pulse oximeters, and claimed that simple stethoscopes and blood pressure cuffs are more operable. Recently, during a workshop that included 10 Scandinavian clinics, the monitoring problem during MRI was emphasized.
42 Journal of CtinicaI Monitoring Vol 8 No 1 January 1992
18.
18,
lSl
16, o
o
141 1 21 1 01
81 o
61 41
I cr
2
o
........I. I o
;
2
0
3
Audibility
Audibility
18.
18.
_~
1 61 8 'E
o3 I
161
141
141 121
"10 U~
2
"E
1 ol
101
81
61
o
4;
41 0
I i 1
O' Audibility
Fig 7. Quality estimation of respiratory and heart sounds before and during M R I exposure in 17 patients. 3, Good audibility; 2, acceptable; 1, poor; and O, no audibility. The vertical axes represent the numbers of patients in the respective category.
Collaboration between different clinics was encouraged [7], although the problems may differ considerably from one clinic to another. Basically, pulse oximetry should be a suitable technique since it relies on optical signals that by themselves do not interact with the MRI equipment. However, in existing instruments, the driving-current pulses to the light sources frequently cause disturbances on the MRI images, and the photodetector has interference caused by the gradients and radiowaves of the MRI equipment. Recently, an accidental, severe burn injury was reported [8], apparently caused by a looped cable that induced a high current through a pulse oximeter probe. This accident indicates that patient-connected metallic leads should be avoided completely in the MRI environment.
8
I
;
t
I
Audibility
In our experience, auscultation with precordial stethoscopes can rarely be used due to the high mechanical and acoustic noise of the MRI equipment. Similarly, blood pressure recording using the auscultatory principle is hardly feasible. Side-stream capnography can be applied to intubated patients but not to sedated, unintubated patients. Our technique of remote auscultation is useful in MRI units having a low to medium-high magnetic field. It is now being routinely used in our hospital for all MRI procedures requiring sedation or general anesthesia. In MRI units with a field o f 1.5 T or higher, the mechanical and acoustic interference during exposure is still unacceptably high. We are now working on improvements that we believe will solve these problems. The approach involves more precise balancing of the microphone pair and more sophisticated signal-filtering techniques.
Henneberg et al: Auscultatory Patient Monitoring
The authors are indebted to Dr Hans Selld6n, St. G6ran's Hospital, Stockholm, for his contributions to the clinical part of the study. This study was supported by the Swedish Board for Technical Development.
REFERENCES 1. Kean D, Smith M. Magnetic resonance imaging principles and applications. London: Heinemann Medical Books, 1986 2. Goudsouzian N, Vacanti FX. Current questions in patient safety: monitoring for MRI. APSF Newsletter 1987:33-35 3. H6k B, Bythell V, Bengtsson M. Development of a wireless stethoscope for auscultatory monitoring during anaesthesia. Med Biol Eng Comput 1988;26:317-320 4. Roth JL, Nugent M, Gray JE, Julsrud RR. Patient monitoring during magnetic resonance imaging. Anesthesiology 1985;62:80-83 5. Selld6n H, De Chateau P, Ekman G, et al. Circulatory monitoring of children during anaesthesia in low-field magnetic resonance imaging. Acta Anaesthesiol Scand t990;34:41-43 6. Shellock FG. Monitoring during MRI. Med Electron 1986;100:93-97 7. LindoffB. Problem med overvakning vid MRT. Med Tek 1989:38 8. Shellock FG, Slimp GL. Severe burn of the finger caused by using a pulse oximeter during MR imaging. Am J Roentgenol 1989;153:1105
43
