- Email: [email protected]
Characterization of Normal and Abnormal Pulmonary Surface by Reflected Ultrasound
Characterization of Normal and Abnormal Pulmonary Surface by Reflected Ultrasound* Kiran B. Sagar, M.D.; Theodore L. Rhyne, Sc.D.; Gordon S. Myers, M.D.; and Robert S. Lees, M.D.
The use of re8ected ultrasound In the diaposis of palmonary disease ... beeD Umited due to the serolll re8eetion and high absorption of acoustic eDellY by alveoU containing air. Our study reports the u1trasoDic: characteristics of normal aDd diseased pubnonary surface. Two measurements, the coeftieient of pulmonary reftection aDd the tissue frequency signature, were studied with speciaDy calibrated instruments. The coefticient of pulmonary reflection in 23 normal subjeds was -19.6 ± 0.1 dB at 5.5 MHz. In ten patients with documented pulmonary emboli, the coefficient of pulmonary re8edion was -47.4 ± 0.1 dB, significantly less than normal (P < 0.1).
I n recent years, reflected ultrasound has found
o
The tissue frequency signature In 18 normal subjeds was specular at JdPer and IlCMISpeCUIar at low. frequendes, with a typical dip at 5.2 MHz. In paden. with duonic obstruc:tive pulmoDary disease, the tissue frequency Bignature flat, aDd the typical dip W8S alJseDt. Ia patients with pulmonary emboli, the dlaracteristic shape of the tissue frequency signature w. preserved, but there W8S a generalized loss of magnitude of the reftection. Thus, our preUminary dam indicate tbat reftec:ted uItnuJound provides a noninvasive method for cUtrerentiating normal from diseased pulmonary 8IIIfaces.
w.
Virginia, RichflWfld 23298
the problems of reverberation by modifying the transducer and the technique; however, with his method and this technique, the angular sensitivity of the ultrasonic beam was still a problem, and, in addition, at the frequency used, the lung still had specular characteristics. The present study was undertaken to evaluate the acoustic properties of the normal and abnormal pulmonary surface using an absolutely calibrated instrument over a relatively wide range of frequencies. The normal lung, including air-containing alveoli, forms an irregular surface for reflection of ultrasound. Figure 1 is a schematic representation of the peripheral portion of the lung, consisting of pleura, lung, parenchyma, and alveolar sacs. The dimensions of a normal alveolar sac is 2BO",. 9 At the routinely used echocardiographic frequency of 2.25 MHz, the depth of the interstitium between adjacent alveolar sacs is smaller than the wave length of the ultrasound, so that the surface reflects in a specular fashion. One of us!O has recently analyzed in detail the ultrasonic reflection from the normal pulmonary surface. The measurement of absolute pulmonary reflection, as compared to a perfectly reflecting plane, was defined as the coefficient of pulmonary reflection. The predicted coefficient of pulmonary reflection at varied frequencies plotted graphically is the tissue frequency signature of the lung (Fig 2 represents computation from an analytic
CHEST. 74: 1. JULY. 1978
NORMAL AIID ABNORMAL PULMONARY SURFACE 29
widespread application as a noninvasive diagnostic tool; however, its use in the diagnosis of pulmonary disease has been limited by the high absorption and widespread scatter of ultrasound by aircontaining pulmonary tissue.' The surface of the normal lung at the routinely used echocardiographic frequency of 2.25 MHz is a highly reflective mirrorlike (specular) surface, thereby producing reflections highly sensitive to the angle of the ultrasonic beam, with little penetration of ultrasound into the lung. Nevertheless, several Investigatorst" have attempted to use ultrasound in various pulmonary diseases. Miller et al2 and joyner" used a &equency of 2.25 MHz for diagnosis of pulmonary emboli. They had difficulty in identifying the echoes from the pulmonary surface and in determining whether multiple reverberations of the transducer &om the pulmonary surface contributed to the normal and abnormal echographic patterns. In addition, the lack of calibration of the instruments and the angular sensitivity of the pulmonary surface contributed to variability in their procedure. Cordons overcame ·From the Noninvasive Diagnostic Laboratory, Cardiac Unit, Massachusetts General Hospital, Boston. Supported by grants from the Ambrose Monell and the G. Unger Vitlessen Foundations. Manuscript received July 22, revision accepted January 2. Reprint requesta: D1'. Sagat', Ben 51, Medical CoUege of
Non Specular Region
Specular Region
Destructive Interference Frequency
~Alveolar
I Mean of Reflection
Sac
__.....'-----:.-/_CoeffiCient------
3
456
FREOUENCY (MHz)
FiGURE 2. Predicted tissue frequency signature of normal lung, showing mean and modulation (standard deviation) of coefficient of pulmonary reflection (computation from analytic model of type shown in Fig 1). FicURE 1. Schematic representation of normal pulmonary surface.
model of the type shown in Fig 1). The coefficient of reflection approaches a perfect reflection (0 dB) at low frequencies and exhibits a minimum reflection at higher frequencies. The minimum reflection is due to the destmctive interference which occurs when the frequency (5.5 MHz) is adjusted so that onehalf ultrasonic wave length equals the depth of tissue between the adjacent alveolar sacs ( 135Jl ). Destmctive interference is a phenomenon where sound waves reflected from various points on a surface cancel each other, resulting in diminished echoes. MATERIALS AND METHODS
Studies were performed in normal subjects and in patients with various pulmonary disorders. The normal pulmonary surface was studied in 23 healthy hospital employees of both sexes. All had normal findings on physical examination and normal chest x-ray films. Abnormal lung was studied in the following pathologic states: (1) pulmonary emboli (experimentally induced pulmonary emboli in five dogs and ten patients with pulmonary emboli documented on pulmonary angiogram) ; ( 2) other infiltrative pulmonary disease (five patients); ( 3) pleural effusion (three patients); and (4) chronic obstructive pulmonary disease documented with pulmonary function tests (six patients). A hand-held transducer was coupled to the skin with a hydrophilic coupling gel (Aquasonic 100) and was moved along the intercostal spaces in order to scan the underlying lung. The A-mode echographic pattern from the skin, soft tissue. and lung was observed on the ultrasonic scope. The pulmonary echo was identified by its distinctive temporal modulation with breathing. Patients were studied in either supine or sitting positions while breathing quietly. Data were digitized and stored in a minicomputer. Two measurements, both de6ned previously herein, were
30 SAUR ET AL
studied, fe, the coe1Bcient of pulmonary reSection and the tissue frequency signature. The former was measured at one specific frequency, 5.5 MHz. At this frequency, the angular sensitivity of reflection by the pulmonary surface is minimal The tissue frequency signature was then measured at multiple frequencies between 3 MHz and 6.9 MHz, each measurement being the coefficient of reflection at that frequency.
Inst",mentaffon The instruments used included a transducer that was onefourth inch in diameter and had a circular flat radiating disk operated by a transmitter-receiver unit and a digital computer. Ultrasonic radiation from the transducer propagated through the skin and adipose tissue to the underlying pulmonary surface with negligible interference from the ribs. The ultrasonic pulses were bursts of radio frequency sinusoids of calibrated amplitude whose frequency and duration were determined by the transmitter. Carrier bursts with a duration of 5 microseconds were utilized, which permitted selective study of echoes of the edge of the lung. The frequency of the radio frequency carrier burst was selectable from discrete frequencies spaced 0.1 MHz apart over the range of 3 MHz to 6.9 MHz. By methods reported previously,lO the echoes returned from the pulmonary surface were corrected for (1) the loss from the transducer, (2) the amplitude of the transmitted pulse, and (3) the loss to bulk tissue. A multiplefrequency scan was performed from 3 MHz to 6.9 MHz, and the coefficient of pulmonary reflection was graphicaI1y plotted to obtain the tissue frequency signature. Single-Erequency measurements of the coefficient of pulmonary reBection in decibels were obtained at 5.5 MHz. Mean and root mean square values and modulation were calculated and tabulated.
REsuLTS Coeflicient of Reflection at 5.5 MHz Normal Sub;ects. Figure 3 shows the echogram of the pulmonary surface from one of the normal volunteers. The echoes at left represent static echoes
CHEST, 74: 1, JULY, 1978
a
a
b
b
c
c I
5
a from the soft tissue. The echo at 2.2em was from the pulmonary surface and was recognized by its modulation with respiration, the echo being most prominent at the end of inspiration and disappearing from view at the end of expiration. Similar echograpbic patterns were observed from the 23 normal wlunteers. The coef6cient of pulmonary reHection was - 19.1 ± 3.2 dB. No significant differences were observed between male and female subjects or between smokers and nonsmokers. Patients with Pulmonary Diaease. Figure 4 shows the echographic pattern obtained at 5.5 MHz from a patient with pulmonary emboli. Figure 4B shows that the modulating pulmonary echo is markedly diminished. Figure 4A shows the pulmonary echo from the corresponding normal side. An echographic pattern similar to that in Figure 3 is seen. Similar patterns were obtained in aD of the patients with in6ltrations from other causes or from pleural effusions. Table 1 shows the coef6cient of pulmonary reflection in normal volunteers and in patients with various pulmonary diseases. There is a statistically significant reduction in the coefBcient of pulmonary reflection in the patients. The coefficient of pulm0nary reflection appeared to be a sensitive method of differentiating between the normal and abnormal pulmonary surfaces; however, the patterns from different pulmonary diseases were nonspeciBc.
Tissue Frequency Signature
Nonnal Subfects. Figure 5A shows the tissue frequency signature from one of the normal individuals. It closely resembles the tissue frequency
CHEST, 74: 1, JULY, 1978
b
c
~1
3 4 5 RANGE (em)
I
6
F1coBB 4. Filtered (a> aod UDflltered (b) A-mode echoes (c is range marker). A (lop), Normal region of lung in patient with pulmonary emboli Echo from pulmonary sudace is at 2.2 em. B (boIIom), Echo from region of pulmonary embolus in same patient. Note marbdly dhnlnfshed echo at 2.2 em.
signature curve of Figure 2 predicted by the analytic model, the lung being specular at lower frequencies of 3 MHz to 4 MHz but nonspecular at higher frequencies of 5 MHz to 6.9 MHz. A characteristic dip due to destructive interference" was seen at 5.2 Table l~.fidenIa01 ~ R.efl-ell-" 5.511&.
Root Mean Square ±8E,dB
Mean, dB
Modulation, dB
Normal 8ubjects
-19.6±0.1
-20.5
-26.8
Chronic obstructive pulmonary disease
-21.2±0.1I
-22.1
-28.3
Pulmonary emboli
-47.4±0.1I
-46.6
-54.7
Other pulmonary diseases
-31.3±0.5
-32.2
-39.0
Group
IIORIIAI. AlIt AIIIOIIIAL PUl.IIOlAIY SUlfACE 31
Cii °A ~
---
'~ -10
-_--_
~ ~
~
""'J
~
~
~
~
~
-~-
- --
-----------.----
- 30L.----L.-------JL----....L..-----L-~ 0----------------.
B
-10-- Mean Reflection Coefficient
(::5
- --
~
G -30
----~---..--_-.....-
-- -
:-...;. -20- __
J.( 1;"-
~
------
---~
_ Modulation
-20
-.....J
"'~"
Mean Reflection Coefficient
-
-
-- -
~-~
Modulation I
3.5
-
_
- __-
I
4.5
~
-
___
--
I
5.5
--
--
--_ I
6.5
FREOUENCY (MHz)
FIGURE 5. Tissue frequency signature of lung, showing mean and modulation of coefficient of pulmonary reflection. A (top), Normal subject. Note dip (due to destructive interference) of mean echo at about 5 MHz. B (bottom), Patient with severe chronic obstroctive pulmonary disease. Note absence of dip (due to destructive interference) of mean echo,
MHz. This pattern was seen in the 18 normal volunteers in whom the multiple-frequency scans were obtained. Patients with Pulmonary Disease. In patients with chronic obstructive pulmonary disease, the tissue frequency signature (Fig 5B) was diHerent from the predicted and observed normal tissue frequency signature. The curve became essentially Hat at about -18 dB and rose by about 5 dB at the lower frequencies. The single-frequency coefBcient of pulmonary reflection at 5.5 MHz was essentially normal. In pulmonary embolic disease the distinctive feature of the tissue frequency signature was a significant loss of magnitude. The frequency signature became of constant magnitude at approximately -45 dB over all of the frequencies studied. Drscossrox
We have de6ned the patterns of reflected ultrasound from normal and abnormal peripheral portions of the lung. The normal pulmonary surface has a well-defined coefficient of pulmonary reflection and tissue frequency signature. In the studies reported herein, any departure from these patterns has been associated with the presence of disease. The absolute reflection of ultrasonic energy de-
32 SAUl ET AL
pends on several variables, te, the amount of energy transmitted to the lung, the distance between the source of energy and the lung, the loss due to diffraction, and the absolute reHectance of the lung, which, in tum, depends upon the distribution of air in the alveolar sacs. When corrected for the loss of energy due to its passage through the tissues, diHraction, and loss from the transducer, the measured reHection depends only on the absolute reflectance of the lung. lo.n We anticipated that normal individuals would show relatively uniform absolute reflectance of the lung. This was confirmed in the present study using carefully calibrated instruments, with appropriate corrections for loss of acoustic energy. In addition, the coefBcient of pulmonary reHection and the tissue frequency signature obtained from normal individuals showed little variation. Thus, two criteria for the normal pulmonary reHectance were obtained, and any departure from these indicated abnormality. This was confirmed by the study of patients with different pulmonary diseases, where it could be assumed that air in the alveoli was displaced secondary either to pulmonary embolus or to other infiltrative pulmonary disease. The coefBcient of pulmonary reflection was diminished significantly in these patients. In Patients with chronic obstructive pulmonary disease in which there was no indication of displacement of air from the alveoli, the single-frequency signature followed a pattern diHerent from the normal. The dip due to destructive interference was not present, and the curve was Hat. In the past, otheIT.8 have used the routine echocardiographic frequency of 2.25 MHz to diagnose various pulmonary disorders; however, such studies do not clearly define the characteristics of the normal lung. Other factors affecting absolute pulmonary reflection were not taken into consideration. The present study defines the criteria for normality and corrects for other factors affecting reflection of acoustic energy from the pulmonary surface. Our preliminary data lack SPecificity in differentiating various pulmonary disorders from one another, as did the data of Miller and associates," With the coefficient of pulmonary reflection, we could not differentiate among pulmonary embolus, atelectasis, and infiltrative disorders. Coexistence of chronic obstructive pulmonary disease did not interfere with the diagnosis of infiltration due to any cause. ModiBcations in the instrumentation which would provide the ability to measure the How of blood in an affected region of the lung may allow us to distinguish pulmonary emboli from infiltrative disease and may increase the SPecificity of noninvasive diagnosis.
CHEST, 74: 1, JULY, 1978
In conclusion, we have defined the ultrasonic features of normal lung and the criteria that allow differentiations between normal and abnormal peripheral pulmonary surfaces. Further refinement of ultrasonic techniques may increase the specificity and clinical usefulness of such studies in the future.
REF'ERENCD 1 Dunn F: Attenuation and speed of ultrasound in lung. J Acoust Soc Am 56:1638-1639.1974 2 Miller LD, Joyner CEo Dudrock S, et &1: Clinical use of ultrasound in the early diagnosis of pulmonary embolism. Ann Surg 166:381-393. 1967 3 Joyner CE: Ultrasound in the diagnosis of venous thromboembolism. Ira Moser KM, Stein M (eds): Pulmonary Thromboembolism. Chicago. Year Book Medical Publishers, 1973, pp 252-268 4 Buddee FW, Johnson DC, JeDinus J, et &1: Experimental and clinical experiences in the use of ultrasound for the
early detection of pulmonary emboli: A preliminary reJ Aust 1:195-197. 1969 Hendrin A: Ultrasonic pulmonary densitometry: Preliminary studies. Invest RadiollO:258-262, 1975 Doust BD. Bamn JI{, Maklad MF. et al: Ultrasonic evaluation of pleural opacities. Radiology 114: 135-140. 1975 Ross AM, Crentin E, Holmes JM: Ultrasonic examination of the lung. J Lab Clio Med 72:556-564. 1968 Gordon D: A new ultrasonic technique for lung diagnosis Ira Ultrasonics in Medicine (de Vliager M. White DN. Mc Cready VR, eds). Rotterdam, International Congress Series 390. 1933, 206-211 Weibel ER: Morphometry of the Human Lung. New York, Academic Press. Inc, 1963 Rhyne TL: Acoustic Instrumentation and Characteristics of Lung Tissue. Forest Grove, Ore. Research Studies Press. 1977 Bauld TI. Schwann lIP: Attenuation and reflection of ultrasound in canine lung tissue. J Acoust Soc Am 56: port. Med
5 6
7
8
9
10
11
1630-1637. 1974
American Lung Association Fellowships Tralfaing F ellowahips directed toward a career in the prevention and control of lung disease are offered to: physicians entering the second or later year of residency in internal medicine, pediabics, thoracic surgery, or other specialties, including basic sciences, relevant to lung disease; individuals holding the degree of M.D., Ph.D., or Sc.D. for further training as scientific investigators in this 6eld; graduate students holding a Ph.D. degree or others with comparable qualifications for research related to the lung. Priority will be given to applicants interested in academic careers. Limited to U.S. citizens to train in the U.s. or Canada and to Canadian citizens or holders of bona Hde permanent u.s. visas for training in u.s. institutions. Awards are up to $12,000. Renewal is possible for a total of two years of support.
Edward Livingston Trudeau Fellowships are awarded
to physicians who have completed graduate training in the Held of lung disease and who have appointments in schools of medicine. The Fellowships are intended to
CHEST, 74: 1, JULY, 1978
give promising instructors an opportunity to stay in an academic setting and to prove themselves as teachers and investigators. The award requires supplement by the school. Awards are up to $15,000. Renewals are possible for three years of Trudeau Fellowship support. H preceded by a Training Fellowship, the maximum Fellowship support allowable is four years. Limited to U.S. citizens in U.S. 01' Canadian schools and to Canadian citizens in u.s. schools. Completed appllctJtionB must be receioed by October 1. Payments are made directly to the Fellow on a quarterly basis. Fellowships are granted for one year and may begin on any date approved by the Review Committees. The usual beginning date is July 1.
Fellowship grants wiD be given only to individuals in institutions identified as Equal Opportunity Employers. Further infonnation and application fonns may be obtained from: American Lung Association, Director of Medical Affairs. 1740 Broadway, New York, N. Y.10019.
10RIAL AND ABIORIIAI. PUI.IIOIWtY SURfACE 33
The use of re8ected ultrasound In the diaposis of palmonary disease ... beeD Umited due to the serolll re8eetion and high absorption of acoustic eDellY by alveoU containing air. Our study reports the u1trasoDic: characteristics of normal aDd diseased pubnonary surface. Two measurements, the coeftieient of pulmonary reftection aDd the tissue frequency signature, were studied with speciaDy calibrated instruments. The coefticient of pulmonary reflection in 23 normal subjeds was -19.6 ± 0.1 dB at 5.5 MHz. In ten patients with documented pulmonary emboli, the coefficient of pulmonary re8edion was -47.4 ± 0.1 dB, significantly less than normal (P < 0.1).
I n recent years, reflected ultrasound has found
o
The tissue frequency signature In 18 normal subjeds was specular at JdPer and IlCMISpeCUIar at low. frequendes, with a typical dip at 5.2 MHz. In paden. with duonic obstruc:tive pulmoDary disease, the tissue frequency Bignature flat, aDd the typical dip W8S alJseDt. Ia patients with pulmonary emboli, the dlaracteristic shape of the tissue frequency signature w. preserved, but there W8S a generalized loss of magnitude of the reftection. Thus, our preUminary dam indicate tbat reftec:ted uItnuJound provides a noninvasive method for cUtrerentiating normal from diseased pulmonary 8IIIfaces.
w.
Virginia, RichflWfld 23298
the problems of reverberation by modifying the transducer and the technique; however, with his method and this technique, the angular sensitivity of the ultrasonic beam was still a problem, and, in addition, at the frequency used, the lung still had specular characteristics. The present study was undertaken to evaluate the acoustic properties of the normal and abnormal pulmonary surface using an absolutely calibrated instrument over a relatively wide range of frequencies. The normal lung, including air-containing alveoli, forms an irregular surface for reflection of ultrasound. Figure 1 is a schematic representation of the peripheral portion of the lung, consisting of pleura, lung, parenchyma, and alveolar sacs. The dimensions of a normal alveolar sac is 2BO",. 9 At the routinely used echocardiographic frequency of 2.25 MHz, the depth of the interstitium between adjacent alveolar sacs is smaller than the wave length of the ultrasound, so that the surface reflects in a specular fashion. One of us!O has recently analyzed in detail the ultrasonic reflection from the normal pulmonary surface. The measurement of absolute pulmonary reflection, as compared to a perfectly reflecting plane, was defined as the coefficient of pulmonary reflection. The predicted coefficient of pulmonary reflection at varied frequencies plotted graphically is the tissue frequency signature of the lung (Fig 2 represents computation from an analytic
CHEST. 74: 1. JULY. 1978
NORMAL AIID ABNORMAL PULMONARY SURFACE 29
widespread application as a noninvasive diagnostic tool; however, its use in the diagnosis of pulmonary disease has been limited by the high absorption and widespread scatter of ultrasound by aircontaining pulmonary tissue.' The surface of the normal lung at the routinely used echocardiographic frequency of 2.25 MHz is a highly reflective mirrorlike (specular) surface, thereby producing reflections highly sensitive to the angle of the ultrasonic beam, with little penetration of ultrasound into the lung. Nevertheless, several Investigatorst" have attempted to use ultrasound in various pulmonary diseases. Miller et al2 and joyner" used a &equency of 2.25 MHz for diagnosis of pulmonary emboli. They had difficulty in identifying the echoes from the pulmonary surface and in determining whether multiple reverberations of the transducer &om the pulmonary surface contributed to the normal and abnormal echographic patterns. In addition, the lack of calibration of the instruments and the angular sensitivity of the pulmonary surface contributed to variability in their procedure. Cordons overcame ·From the Noninvasive Diagnostic Laboratory, Cardiac Unit, Massachusetts General Hospital, Boston. Supported by grants from the Ambrose Monell and the G. Unger Vitlessen Foundations. Manuscript received July 22, revision accepted January 2. Reprint requesta: D1'. Sagat', Ben 51, Medical CoUege of
Non Specular Region
Specular Region
Destructive Interference Frequency
~Alveolar
I Mean of Reflection
Sac
__.....'-----:.-/_CoeffiCient------
3
456
FREOUENCY (MHz)
FiGURE 2. Predicted tissue frequency signature of normal lung, showing mean and modulation (standard deviation) of coefficient of pulmonary reflection (computation from analytic model of type shown in Fig 1). FicURE 1. Schematic representation of normal pulmonary surface.
model of the type shown in Fig 1). The coefficient of reflection approaches a perfect reflection (0 dB) at low frequencies and exhibits a minimum reflection at higher frequencies. The minimum reflection is due to the destmctive interference which occurs when the frequency (5.5 MHz) is adjusted so that onehalf ultrasonic wave length equals the depth of tissue between the adjacent alveolar sacs ( 135Jl ). Destmctive interference is a phenomenon where sound waves reflected from various points on a surface cancel each other, resulting in diminished echoes. MATERIALS AND METHODS
Studies were performed in normal subjects and in patients with various pulmonary disorders. The normal pulmonary surface was studied in 23 healthy hospital employees of both sexes. All had normal findings on physical examination and normal chest x-ray films. Abnormal lung was studied in the following pathologic states: (1) pulmonary emboli (experimentally induced pulmonary emboli in five dogs and ten patients with pulmonary emboli documented on pulmonary angiogram) ; ( 2) other infiltrative pulmonary disease (five patients); ( 3) pleural effusion (three patients); and (4) chronic obstructive pulmonary disease documented with pulmonary function tests (six patients). A hand-held transducer was coupled to the skin with a hydrophilic coupling gel (Aquasonic 100) and was moved along the intercostal spaces in order to scan the underlying lung. The A-mode echographic pattern from the skin, soft tissue. and lung was observed on the ultrasonic scope. The pulmonary echo was identified by its distinctive temporal modulation with breathing. Patients were studied in either supine or sitting positions while breathing quietly. Data were digitized and stored in a minicomputer. Two measurements, both de6ned previously herein, were
30 SAUR ET AL
studied, fe, the coe1Bcient of pulmonary reSection and the tissue frequency signature. The former was measured at one specific frequency, 5.5 MHz. At this frequency, the angular sensitivity of reflection by the pulmonary surface is minimal The tissue frequency signature was then measured at multiple frequencies between 3 MHz and 6.9 MHz, each measurement being the coefficient of reflection at that frequency.
Inst",mentaffon The instruments used included a transducer that was onefourth inch in diameter and had a circular flat radiating disk operated by a transmitter-receiver unit and a digital computer. Ultrasonic radiation from the transducer propagated through the skin and adipose tissue to the underlying pulmonary surface with negligible interference from the ribs. The ultrasonic pulses were bursts of radio frequency sinusoids of calibrated amplitude whose frequency and duration were determined by the transmitter. Carrier bursts with a duration of 5 microseconds were utilized, which permitted selective study of echoes of the edge of the lung. The frequency of the radio frequency carrier burst was selectable from discrete frequencies spaced 0.1 MHz apart over the range of 3 MHz to 6.9 MHz. By methods reported previously,lO the echoes returned from the pulmonary surface were corrected for (1) the loss from the transducer, (2) the amplitude of the transmitted pulse, and (3) the loss to bulk tissue. A multiplefrequency scan was performed from 3 MHz to 6.9 MHz, and the coefficient of pulmonary reflection was graphicaI1y plotted to obtain the tissue frequency signature. Single-Erequency measurements of the coefficient of pulmonary reBection in decibels were obtained at 5.5 MHz. Mean and root mean square values and modulation were calculated and tabulated.
REsuLTS Coeflicient of Reflection at 5.5 MHz Normal Sub;ects. Figure 3 shows the echogram of the pulmonary surface from one of the normal volunteers. The echoes at left represent static echoes
CHEST, 74: 1, JULY, 1978
a
a
b
b
c
c I
5
a from the soft tissue. The echo at 2.2em was from the pulmonary surface and was recognized by its modulation with respiration, the echo being most prominent at the end of inspiration and disappearing from view at the end of expiration. Similar echograpbic patterns were observed from the 23 normal wlunteers. The coef6cient of pulmonary reHection was - 19.1 ± 3.2 dB. No significant differences were observed between male and female subjects or between smokers and nonsmokers. Patients with Pulmonary Diaease. Figure 4 shows the echographic pattern obtained at 5.5 MHz from a patient with pulmonary emboli. Figure 4B shows that the modulating pulmonary echo is markedly diminished. Figure 4A shows the pulmonary echo from the corresponding normal side. An echographic pattern similar to that in Figure 3 is seen. Similar patterns were obtained in aD of the patients with in6ltrations from other causes or from pleural effusions. Table 1 shows the coef6cient of pulmonary reflection in normal volunteers and in patients with various pulmonary diseases. There is a statistically significant reduction in the coefBcient of pulmonary reflection in the patients. The coefficient of pulm0nary reflection appeared to be a sensitive method of differentiating between the normal and abnormal pulmonary surfaces; however, the patterns from different pulmonary diseases were nonspeciBc.
Tissue Frequency Signature
Nonnal Subfects. Figure 5A shows the tissue frequency signature from one of the normal individuals. It closely resembles the tissue frequency
CHEST, 74: 1, JULY, 1978
b
c
~1
3 4 5 RANGE (em)
I
6
F1coBB 4. Filtered (a> aod UDflltered (b) A-mode echoes (c is range marker). A (lop), Normal region of lung in patient with pulmonary emboli Echo from pulmonary sudace is at 2.2 em. B (boIIom), Echo from region of pulmonary embolus in same patient. Note marbdly dhnlnfshed echo at 2.2 em.
signature curve of Figure 2 predicted by the analytic model, the lung being specular at lower frequencies of 3 MHz to 4 MHz but nonspecular at higher frequencies of 5 MHz to 6.9 MHz. A characteristic dip due to destructive interference" was seen at 5.2 Table l~.fidenIa01 ~ R.efl-ell-" 5.511&.
Root Mean Square ±8E,dB
Mean, dB
Modulation, dB
Normal 8ubjects
-19.6±0.1
-20.5
-26.8
Chronic obstructive pulmonary disease
-21.2±0.1I
-22.1
-28.3
Pulmonary emboli
-47.4±0.1I
-46.6
-54.7
Other pulmonary diseases
-31.3±0.5
-32.2
-39.0
Group
IIORIIAI. AlIt AIIIOIIIAL PUl.IIOlAIY SUlfACE 31
Cii °A ~
---
'~ -10
-_--_
~ ~
~
""'J
~
~
~
~
~
-~-
- --
-----------.----
- 30L.----L.-------JL----....L..-----L-~ 0----------------.
B
-10-- Mean Reflection Coefficient
(::5
- --
~
G -30
----~---..--_-.....-
-- -
:-...;. -20- __
J.( 1;"-
~
------
---~
_ Modulation
-20
-.....J
"'~"
Mean Reflection Coefficient
-
-
-- -
~-~
Modulation I
3.5
-
_
- __-
I
4.5
~
-
___
--
I
5.5
--
--
--_ I
6.5
FREOUENCY (MHz)
FIGURE 5. Tissue frequency signature of lung, showing mean and modulation of coefficient of pulmonary reflection. A (top), Normal subject. Note dip (due to destructive interference) of mean echo at about 5 MHz. B (bottom), Patient with severe chronic obstroctive pulmonary disease. Note absence of dip (due to destructive interference) of mean echo,
MHz. This pattern was seen in the 18 normal volunteers in whom the multiple-frequency scans were obtained. Patients with Pulmonary Disease. In patients with chronic obstructive pulmonary disease, the tissue frequency signature (Fig 5B) was diHerent from the predicted and observed normal tissue frequency signature. The curve became essentially Hat at about -18 dB and rose by about 5 dB at the lower frequencies. The single-frequency coefBcient of pulmonary reflection at 5.5 MHz was essentially normal. In pulmonary embolic disease the distinctive feature of the tissue frequency signature was a significant loss of magnitude. The frequency signature became of constant magnitude at approximately -45 dB over all of the frequencies studied. Drscossrox
We have de6ned the patterns of reflected ultrasound from normal and abnormal peripheral portions of the lung. The normal pulmonary surface has a well-defined coefficient of pulmonary reflection and tissue frequency signature. In the studies reported herein, any departure from these patterns has been associated with the presence of disease. The absolute reflection of ultrasonic energy de-
32 SAUl ET AL
pends on several variables, te, the amount of energy transmitted to the lung, the distance between the source of energy and the lung, the loss due to diffraction, and the absolute reHectance of the lung, which, in tum, depends upon the distribution of air in the alveolar sacs. When corrected for the loss of energy due to its passage through the tissues, diHraction, and loss from the transducer, the measured reHection depends only on the absolute reflectance of the lung. lo.n We anticipated that normal individuals would show relatively uniform absolute reflectance of the lung. This was confirmed in the present study using carefully calibrated instruments, with appropriate corrections for loss of acoustic energy. In addition, the coefBcient of pulmonary reHection and the tissue frequency signature obtained from normal individuals showed little variation. Thus, two criteria for the normal pulmonary reHectance were obtained, and any departure from these indicated abnormality. This was confirmed by the study of patients with different pulmonary diseases, where it could be assumed that air in the alveoli was displaced secondary either to pulmonary embolus or to other infiltrative pulmonary disease. The coefBcient of pulmonary reflection was diminished significantly in these patients. In Patients with chronic obstructive pulmonary disease in which there was no indication of displacement of air from the alveoli, the single-frequency signature followed a pattern diHerent from the normal. The dip due to destructive interference was not present, and the curve was Hat. In the past, otheIT.8 have used the routine echocardiographic frequency of 2.25 MHz to diagnose various pulmonary disorders; however, such studies do not clearly define the characteristics of the normal lung. Other factors affecting absolute pulmonary reflection were not taken into consideration. The present study defines the criteria for normality and corrects for other factors affecting reflection of acoustic energy from the pulmonary surface. Our preliminary data lack SPecificity in differentiating various pulmonary disorders from one another, as did the data of Miller and associates," With the coefficient of pulmonary reflection, we could not differentiate among pulmonary embolus, atelectasis, and infiltrative disorders. Coexistence of chronic obstructive pulmonary disease did not interfere with the diagnosis of infiltration due to any cause. ModiBcations in the instrumentation which would provide the ability to measure the How of blood in an affected region of the lung may allow us to distinguish pulmonary emboli from infiltrative disease and may increase the SPecificity of noninvasive diagnosis.
CHEST, 74: 1, JULY, 1978
In conclusion, we have defined the ultrasonic features of normal lung and the criteria that allow differentiations between normal and abnormal peripheral pulmonary surfaces. Further refinement of ultrasonic techniques may increase the specificity and clinical usefulness of such studies in the future.
REF'ERENCD 1 Dunn F: Attenuation and speed of ultrasound in lung. J Acoust Soc Am 56:1638-1639.1974 2 Miller LD, Joyner CEo Dudrock S, et &1: Clinical use of ultrasound in the early diagnosis of pulmonary embolism. Ann Surg 166:381-393. 1967 3 Joyner CE: Ultrasound in the diagnosis of venous thromboembolism. Ira Moser KM, Stein M (eds): Pulmonary Thromboembolism. Chicago. Year Book Medical Publishers, 1973, pp 252-268 4 Buddee FW, Johnson DC, JeDinus J, et &1: Experimental and clinical experiences in the use of ultrasound for the
early detection of pulmonary emboli: A preliminary reJ Aust 1:195-197. 1969 Hendrin A: Ultrasonic pulmonary densitometry: Preliminary studies. Invest RadiollO:258-262, 1975 Doust BD. Bamn JI{, Maklad MF. et al: Ultrasonic evaluation of pleural opacities. Radiology 114: 135-140. 1975 Ross AM, Crentin E, Holmes JM: Ultrasonic examination of the lung. J Lab Clio Med 72:556-564. 1968 Gordon D: A new ultrasonic technique for lung diagnosis Ira Ultrasonics in Medicine (de Vliager M. White DN. Mc Cready VR, eds). Rotterdam, International Congress Series 390. 1933, 206-211 Weibel ER: Morphometry of the Human Lung. New York, Academic Press. Inc, 1963 Rhyne TL: Acoustic Instrumentation and Characteristics of Lung Tissue. Forest Grove, Ore. Research Studies Press. 1977 Bauld TI. Schwann lIP: Attenuation and reflection of ultrasound in canine lung tissue. J Acoust Soc Am 56: port. Med
5 6
7
8
9
10
11
1630-1637. 1974
American Lung Association Fellowships Tralfaing F ellowahips directed toward a career in the prevention and control of lung disease are offered to: physicians entering the second or later year of residency in internal medicine, pediabics, thoracic surgery, or other specialties, including basic sciences, relevant to lung disease; individuals holding the degree of M.D., Ph.D., or Sc.D. for further training as scientific investigators in this 6eld; graduate students holding a Ph.D. degree or others with comparable qualifications for research related to the lung. Priority will be given to applicants interested in academic careers. Limited to U.S. citizens to train in the U.s. or Canada and to Canadian citizens or holders of bona Hde permanent u.s. visas for training in u.s. institutions. Awards are up to $12,000. Renewal is possible for a total of two years of support.
Edward Livingston Trudeau Fellowships are awarded
to physicians who have completed graduate training in the Held of lung disease and who have appointments in schools of medicine. The Fellowships are intended to
CHEST, 74: 1, JULY, 1978
give promising instructors an opportunity to stay in an academic setting and to prove themselves as teachers and investigators. The award requires supplement by the school. Awards are up to $15,000. Renewals are possible for three years of Trudeau Fellowship support. H preceded by a Training Fellowship, the maximum Fellowship support allowable is four years. Limited to U.S. citizens in U.S. 01' Canadian schools and to Canadian citizens in u.s. schools. Completed appllctJtionB must be receioed by October 1. Payments are made directly to the Fellow on a quarterly basis. Fellowships are granted for one year and may begin on any date approved by the Review Committees. The usual beginning date is July 1.
Fellowship grants wiD be given only to individuals in institutions identified as Equal Opportunity Employers. Further infonnation and application fonns may be obtained from: American Lung Association, Director of Medical Affairs. 1740 Broadway, New York, N. Y.10019.
10RIAL AND ABIORIIAI. PUI.IIOIWtY SURfACE 33