Catheterization and Cardiovascular Diagnosis 3: 101-106
(1977)
A SIMPLIFIED METHOD OF ECHOCARDIOGRAPHIC DATA ANALYSIS Richard M. Schieken, M.D., Horst R. Jordan, William Clarke, Ph.D., and Ronald M. Lauer, M.D. Rapid accurate analysis of echocardiographlc data is accomplished using a sonic digitizer and programmable calculator. This method allows the echocardiographer to select technically optimal areas of the recording for analysis. The resolution of the measuring device is 0.1 mm. A hardcopy printout of both measurement and calculation is provided. Instead of expensive on-line computer, an inexpensive programmable calculator Is used. Key words: ultrasound, computer, echocardiography
INTRODUCTION
The echocardiogram, an ultrasonic image of cardiac structures, has gained acceptance as a safe, reliable diagnostic tool (1). Recent advances in the use of the echocardiogram require accurate measurement of both right and left heart structures (2, 3). This paper describes a simplified approach for rapid echocardiographic analysis using a sonic digitizer and a programmable calculator. MATERIALS AND METHODS
All echocardiograms were performed using a Smith Ekoline 20 ultrasonascope with a 2 . 2 , 3 . 5 , or 5.0 mHz transducer with a 1 psec transmission time and a repetition rate of 1000hec. The echocardiograms were recorded on a Honeywell 1856 strip chart recorder with a 1 cm and 0.2 sec calibration, run at a paper speed of 2.5 cm/sec. The data were analyzed using a sonic digitizer coupled to a programmable calculator. The Graf-pen GP-3 (Science Accessories Corporation) is a digitizer, which converts graphic or planar information into numerical data. It generates a specific set of coordinates (x and y) for every point on the plane area, the size of which is defined by the length of the axis from which x and y distances are
From the University of Iowa Hospitals. Iowa City, Iowa
Reprint requests to: Richard M . Schieken, M.D.. Division of Pediatric Cardiology, University of Iowa Hospitals, Iowa City, Iowa 52242 Received August LO, 1976; revision accepted October 15, 1976
101
0 1 9 7 7 Alan
R. Liss,
Inc., 1 5 0 Fifth Avenue, New York, NY 10011
102
Schieken et al.
r
Z
G k NF k T Y LINE SEGMENT SUBROUTINE AND-STORE FOR LATER USE I
FROM
l
1
(7 ENTER
GET END POINT I (X,Y,) FROM SWRK PEN
GET END POINT 2 (&,Y2) FROM %?IRK PEN LVDd LV
SEGMENTLENGTH =
1
STOP
Fig. 1. Flow chart of echocardiographic analysis program.
I (RETURN) Fig. 2. Flow chart of line segment measure subroutine.
measured. Hypersonic impulses are generated at the point of the stylus. The times required for the sound wave to reach 2 linear sensory microphones which measure x and y distances respectively are converted into distance measurements. The resolution of the GP-3 is 0.1 mm. In this study, the sonic digitizer is coupled, using a model 1461 interface (Science Accessories Corporation), to the Monroe 1860 programmable calculator (Monroe the Calculator Company, Division of Litton Industries). Flow charts of the program and the line segment subroutine are diagrammed in Figs. 1 and 2. METHODS
An echocardiographic strip chart recording is inspected. An area of the recording that shows both anterior mitral leaflet and chordae tendiniae with rapidly moving, clear, endocardial interface, and a rapidly posterior moving, systolic, septal echo is chosen for analysis (Fig. 3). A vertical line is drawn through the shortest systolic distance between the left septal surface and the endocardial surface of the left ventricular posterior wall. This distance is identified as the left ventricular systolic dimension LVDs. A vertical line is drawn through the entire echo record beginning on the Q wave of the chosen beat. The following interfaces are marked: anterior
Echocardiographic Data Analysis
103
Fig. 3.A selected portion of an ~ h ~ a r d ~ ~ r atracing p h i cwith the transducer directed towards the body of the left ventricle. The areas measured in this portion of the tracing are marked. ARVW = anterior right ventricular wall. RVlD = right ventricular internal dimension. IVS = l n t e ~ e n t r k u lar septum. LVDd = left ventricular diastolic dimension. LVD, = left ventricular systolic dimension. LVPW = left ventricular posterior wall. The stylus is touched at 2 points to recordthe length of these dimensions. In similar fashion, a 2nd portion of the ~hocardiographic tracing with the transducer directed to record the aorta, aortic valve leaflets, and left atrium (not shown in illustration) is placed on the active tablet and measured.
right wall; right ventricular septum; interventricular septum; left ventricular diastolic dimension; left ventricular posterior wall. A second portion (not illustrated) of the strip chart recording is chosen. In this section the aortic valve leaflets are identified. A vertical line is drawn through the largest left atrial dimension, which is the point where the aortic valve leaflets close in diastole. The aortic root dimension is measured at the Q wave. The left ventricular ejection time is marked from the opening of the aortic valve leaflet to closure. The marked strip chart recording is placed on the active area tablet of a model GP-3 BCD Graf-pen (Science Accessories Corporation) (Fig. 4). The stylus is touched down upon the anterior and posterior interfaces of the recording in the following order: 1 . Two calibra~iondots (cm depth). 2. Anterior right ventricular wall (ARVW)---chest wall interface to posterior aspect of anterior right ventricular wail interface. 3 . Right ventricular dimension (RVI~)-anterior right ventricular wall interface to right interventri~ularseptal interface. 4. Interventricular septum (1VS)-right interventricular septal interface to left interventricular septal interface.
104
Schieken e t al.
GRAF E N STYLUS
SONIC DlGlTlZER
MONROE IBW SERIES
J PRWRAMMABLE
CALCULAlW
-
J
5. Left ventricular diastolic dimension (LVDd)--left interventricular septa1 interface to ventricular posterior wall endocardial interface. 6. Left ventricular posterior wall (LVPW)-left ventricular endocardial surface to left ventricular posterior epicardial surfaces. 7. Left ventricular systolic dimension (LVD,)-left septum to left ventricular endocardium in systole. The record is then moved to allow the stylus to touch down the anterior and posterior left atrial dimension, the anterior and posterior aortic dimension, and the left ventricular ejection time. The programmed magnetic card allows the calculator to make the following calculations: I . Left ventricular diastolic volume. 2. Left ventricular systolic volume. 3. Stroke Volume. 4. Ejection fraction 1 (4). 5. Ejection fraction 2 (used for children) (5). 6. %AD (6). The calculator prints a tape with the dimensions and calculations. The following reliability study was performed. Ten physicians or technicians familar with the instrumentation each measured 10 xeroxed echocardiograms. The 1Oechocardiographic copies were made up of 5 randomly ordered pairs. The echos were first measured using the Spark pen-calculator and then by hand using a single steel millimeter ruler. There was excellent reproducibility of the paired measurements. A n analysis of variance showed no significant differences between hand-or Spark-pen-calculator measurement (Table I).
Echocardiographic Data Analysis
105
TABLE I. Reliability Study.
Hand
Spark pen
Within observer Variance (cm) Correlation (r)
0.23 0.986
0.17 0.990
Between observers Variance (cm) Correlation (r)
0.29 0.977
0.20 0.988
DISCUSSION
Other methods exist for the computer analysis of echocardiogram. Griffith and Henry described a semiautomated method of analysis using a closed circuit television scanning system and an analog recorder (8). Other investigators have utilized real time digitization transmitted to a computer (9). We have tried to approach this problem with a minimum of hardware. The elimination of the video scanning device not only reduces the cost of the operation but also allows the echocardiographer the opportunity to select optimal points for analysis. Judgments can then be made as to the clarity of the interfaces and the reliability of the interpretations. The stylus measurement ofthe points is as accurate as hand measurement with a ruler, and less fatiguing. The Monroe programmable calculator allows the data for both the measurement of the line segments and the storage for calculation to be introduced with a single side of a magnetic card. The magnetic card programming system allows for easy, low-cost modification of the program. Further work in echocardiography may be necessary to validate the usefulness of these ultrasound measurements. We believe that our method provides an inexpensive, accurate method for echocardiographic data analysis. It uses a sonic digitizer and a programmable calculator. The echocardiographer selects the points for measurement, hence he is satisfied that the interpretable interfaces are used for calculation. We will furnish our program upon request. ACKNOWLEDGMENTS
The authors wish to express their appreciation to Mr. William Schoon of the Monroe Company for his assistance in developing our program, and to Mr. Robert Forlaw of Cain Engineers, who is a representative of Science Accessories Corporation, for assisting in the use of the sonic digitizer. REFERENCES 1 . Grossman H, Felman A , Kirkpatrick JA, Shopfner CH, Swischuk LE, Taber P, Tefft M and Goldberg BB: Ultrasonography in children. Pediatrics 54:480-481, 1974. American Academy of Pediatrics on Radiology
106
Schieken et al.
2. McDonald IG: Echocardiographic assessment of left ventricular function in aortic valve disease. Circulation 53:860-864, 1976. 3. Schieken R and Kerber T: Echocardiographic abnormalities in acute rheumatic fever. Am J Card 38:458462, 1976. 4. Feigenbaurn H. Popp R , Wolfe SB, Troy BL, Pombo JF, Haine CL and Dodge HT: Ultrasound measurements of the left ventricle. Arch Intern Med 1?9:461467, 1972. 5 . Myer RA, Stockert J and Kaplan S: Echocardiographic determination of left ventricular volumes in pediatric patients. Circulation 51:297-303. 1975. 6. Gutgessell HP, Paquet M, Duff D F and McNamara D: Left ventricular function in normal children: Effects of age and heart rate. Circulation 52:II-p, 1975. 7. Karliner JS, Gault JH, Eckberg D, Mullins CB and Ross J. Jr: Mean velocity of fiber shortening. Circulation 44:3?3-333, 1971. 8. Griffith JM and Henry WL: Videoscaner-Analog computer system for semiautomatic analysis of routine echocardiograms. Am J Card 32:961-964, 1973. 9. Hirsch M, Sanders WJ, Popp RL and Harrison DC: Computer processing of ultrasonic data from the cardiovascular system. Computers and Biomedical Research 6:336-346, 1973.
(1977)
A SIMPLIFIED METHOD OF ECHOCARDIOGRAPHIC DATA ANALYSIS Richard M. Schieken, M.D., Horst R. Jordan, William Clarke, Ph.D., and Ronald M. Lauer, M.D. Rapid accurate analysis of echocardiographlc data is accomplished using a sonic digitizer and programmable calculator. This method allows the echocardiographer to select technically optimal areas of the recording for analysis. The resolution of the measuring device is 0.1 mm. A hardcopy printout of both measurement and calculation is provided. Instead of expensive on-line computer, an inexpensive programmable calculator Is used. Key words: ultrasound, computer, echocardiography
INTRODUCTION
The echocardiogram, an ultrasonic image of cardiac structures, has gained acceptance as a safe, reliable diagnostic tool (1). Recent advances in the use of the echocardiogram require accurate measurement of both right and left heart structures (2, 3). This paper describes a simplified approach for rapid echocardiographic analysis using a sonic digitizer and a programmable calculator. MATERIALS AND METHODS
All echocardiograms were performed using a Smith Ekoline 20 ultrasonascope with a 2 . 2 , 3 . 5 , or 5.0 mHz transducer with a 1 psec transmission time and a repetition rate of 1000hec. The echocardiograms were recorded on a Honeywell 1856 strip chart recorder with a 1 cm and 0.2 sec calibration, run at a paper speed of 2.5 cm/sec. The data were analyzed using a sonic digitizer coupled to a programmable calculator. The Graf-pen GP-3 (Science Accessories Corporation) is a digitizer, which converts graphic or planar information into numerical data. It generates a specific set of coordinates (x and y) for every point on the plane area, the size of which is defined by the length of the axis from which x and y distances are
From the University of Iowa Hospitals. Iowa City, Iowa
Reprint requests to: Richard M . Schieken, M.D.. Division of Pediatric Cardiology, University of Iowa Hospitals, Iowa City, Iowa 52242 Received August LO, 1976; revision accepted October 15, 1976
101
0 1 9 7 7 Alan
R. Liss,
Inc., 1 5 0 Fifth Avenue, New York, NY 10011
102
Schieken et al.
r
Z
G k NF k T Y LINE SEGMENT SUBROUTINE AND-STORE FOR LATER USE I
FROM
l
1
(7 ENTER
GET END POINT I (X,Y,) FROM SWRK PEN
GET END POINT 2 (&,Y2) FROM %?IRK PEN LVDd LV
SEGMENTLENGTH =
1
STOP
Fig. 1. Flow chart of echocardiographic analysis program.
I (RETURN) Fig. 2. Flow chart of line segment measure subroutine.
measured. Hypersonic impulses are generated at the point of the stylus. The times required for the sound wave to reach 2 linear sensory microphones which measure x and y distances respectively are converted into distance measurements. The resolution of the GP-3 is 0.1 mm. In this study, the sonic digitizer is coupled, using a model 1461 interface (Science Accessories Corporation), to the Monroe 1860 programmable calculator (Monroe the Calculator Company, Division of Litton Industries). Flow charts of the program and the line segment subroutine are diagrammed in Figs. 1 and 2. METHODS
An echocardiographic strip chart recording is inspected. An area of the recording that shows both anterior mitral leaflet and chordae tendiniae with rapidly moving, clear, endocardial interface, and a rapidly posterior moving, systolic, septal echo is chosen for analysis (Fig. 3). A vertical line is drawn through the shortest systolic distance between the left septal surface and the endocardial surface of the left ventricular posterior wall. This distance is identified as the left ventricular systolic dimension LVDs. A vertical line is drawn through the entire echo record beginning on the Q wave of the chosen beat. The following interfaces are marked: anterior
Echocardiographic Data Analysis
103
Fig. 3.A selected portion of an ~ h ~ a r d ~ ~ r atracing p h i cwith the transducer directed towards the body of the left ventricle. The areas measured in this portion of the tracing are marked. ARVW = anterior right ventricular wall. RVlD = right ventricular internal dimension. IVS = l n t e ~ e n t r k u lar septum. LVDd = left ventricular diastolic dimension. LVD, = left ventricular systolic dimension. LVPW = left ventricular posterior wall. The stylus is touched at 2 points to recordthe length of these dimensions. In similar fashion, a 2nd portion of the ~hocardiographic tracing with the transducer directed to record the aorta, aortic valve leaflets, and left atrium (not shown in illustration) is placed on the active tablet and measured.
right wall; right ventricular septum; interventricular septum; left ventricular diastolic dimension; left ventricular posterior wall. A second portion (not illustrated) of the strip chart recording is chosen. In this section the aortic valve leaflets are identified. A vertical line is drawn through the largest left atrial dimension, which is the point where the aortic valve leaflets close in diastole. The aortic root dimension is measured at the Q wave. The left ventricular ejection time is marked from the opening of the aortic valve leaflet to closure. The marked strip chart recording is placed on the active area tablet of a model GP-3 BCD Graf-pen (Science Accessories Corporation) (Fig. 4). The stylus is touched down upon the anterior and posterior interfaces of the recording in the following order: 1 . Two calibra~iondots (cm depth). 2. Anterior right ventricular wall (ARVW)---chest wall interface to posterior aspect of anterior right ventricular wail interface. 3 . Right ventricular dimension (RVI~)-anterior right ventricular wall interface to right interventri~ularseptal interface. 4. Interventricular septum (1VS)-right interventricular septal interface to left interventricular septal interface.
104
Schieken e t al.
GRAF E N STYLUS
SONIC DlGlTlZER
MONROE IBW SERIES
J PRWRAMMABLE
CALCULAlW
-
J
5. Left ventricular diastolic dimension (LVDd)--left interventricular septa1 interface to ventricular posterior wall endocardial interface. 6. Left ventricular posterior wall (LVPW)-left ventricular endocardial surface to left ventricular posterior epicardial surfaces. 7. Left ventricular systolic dimension (LVD,)-left septum to left ventricular endocardium in systole. The record is then moved to allow the stylus to touch down the anterior and posterior left atrial dimension, the anterior and posterior aortic dimension, and the left ventricular ejection time. The programmed magnetic card allows the calculator to make the following calculations: I . Left ventricular diastolic volume. 2. Left ventricular systolic volume. 3. Stroke Volume. 4. Ejection fraction 1 (4). 5. Ejection fraction 2 (used for children) (5). 6. %AD (6). The calculator prints a tape with the dimensions and calculations. The following reliability study was performed. Ten physicians or technicians familar with the instrumentation each measured 10 xeroxed echocardiograms. The 1Oechocardiographic copies were made up of 5 randomly ordered pairs. The echos were first measured using the Spark pen-calculator and then by hand using a single steel millimeter ruler. There was excellent reproducibility of the paired measurements. A n analysis of variance showed no significant differences between hand-or Spark-pen-calculator measurement (Table I).
Echocardiographic Data Analysis
105
TABLE I. Reliability Study.
Hand
Spark pen
Within observer Variance (cm) Correlation (r)
0.23 0.986
0.17 0.990
Between observers Variance (cm) Correlation (r)
0.29 0.977
0.20 0.988
DISCUSSION
Other methods exist for the computer analysis of echocardiogram. Griffith and Henry described a semiautomated method of analysis using a closed circuit television scanning system and an analog recorder (8). Other investigators have utilized real time digitization transmitted to a computer (9). We have tried to approach this problem with a minimum of hardware. The elimination of the video scanning device not only reduces the cost of the operation but also allows the echocardiographer the opportunity to select optimal points for analysis. Judgments can then be made as to the clarity of the interfaces and the reliability of the interpretations. The stylus measurement ofthe points is as accurate as hand measurement with a ruler, and less fatiguing. The Monroe programmable calculator allows the data for both the measurement of the line segments and the storage for calculation to be introduced with a single side of a magnetic card. The magnetic card programming system allows for easy, low-cost modification of the program. Further work in echocardiography may be necessary to validate the usefulness of these ultrasound measurements. We believe that our method provides an inexpensive, accurate method for echocardiographic data analysis. It uses a sonic digitizer and a programmable calculator. The echocardiographer selects the points for measurement, hence he is satisfied that the interpretable interfaces are used for calculation. We will furnish our program upon request. ACKNOWLEDGMENTS
The authors wish to express their appreciation to Mr. William Schoon of the Monroe Company for his assistance in developing our program, and to Mr. Robert Forlaw of Cain Engineers, who is a representative of Science Accessories Corporation, for assisting in the use of the sonic digitizer. REFERENCES 1 . Grossman H, Felman A , Kirkpatrick JA, Shopfner CH, Swischuk LE, Taber P, Tefft M and Goldberg BB: Ultrasonography in children. Pediatrics 54:480-481, 1974. American Academy of Pediatrics on Radiology
106
Schieken et al.
2. McDonald IG: Echocardiographic assessment of left ventricular function in aortic valve disease. Circulation 53:860-864, 1976. 3. Schieken R and Kerber T: Echocardiographic abnormalities in acute rheumatic fever. Am J Card 38:458462, 1976. 4. Feigenbaurn H. Popp R , Wolfe SB, Troy BL, Pombo JF, Haine CL and Dodge HT: Ultrasound measurements of the left ventricle. Arch Intern Med 1?9:461467, 1972. 5 . Myer RA, Stockert J and Kaplan S: Echocardiographic determination of left ventricular volumes in pediatric patients. Circulation 51:297-303. 1975. 6. Gutgessell HP, Paquet M, Duff D F and McNamara D: Left ventricular function in normal children: Effects of age and heart rate. Circulation 52:II-p, 1975. 7. Karliner JS, Gault JH, Eckberg D, Mullins CB and Ross J. Jr: Mean velocity of fiber shortening. Circulation 44:3?3-333, 1971. 8. Griffith JM and Henry WL: Videoscaner-Analog computer system for semiautomatic analysis of routine echocardiograms. Am J Card 32:961-964, 1973. 9. Hirsch M, Sanders WJ, Popp RL and Harrison DC: Computer processing of ultrasonic data from the cardiovascular system. Computers and Biomedical Research 6:336-346, 1973.
