Clinical evaluation and application of cardiac laboratory high-definition video systems

Catheterization and Cardiovascular Diagnosis 10:63-71 (1984)

Clinical Evaluation and Application of Cardiac Laboratory High-Definition Video Systems

David R. Holmes, Jr., MD, Hugh C. Smith, MD, Joel E. Gray, PhD, and Merrill A. Wondrow

This study evaluated a new high line rate (1,023-line) high band-width (10 MHz) system, compared it to our high-definition, 525-line video system, and compared both video systems to 35-mm cine angiography in 117 patients undergoing clinically indicated coronary angiography. The subjective image quality of the 1,023-line video system was not significantly better than that of the 525-line system. High line rate, high band-width systems may have theoretical advantages, but the increased noise level under clinical angiographic conditions reduces their diagnostic quality. Com- parison between video (all modes) and cine images resulted in cine being rated as the best modality, but the differences were slight. Radiation levels required for video imaging, however, were significantly less than those required for cine recording. Current clinical and computer interactive uses of video systems in the cardiac laboratory are described. The eventual role and potential replacement of cine recording by video tape systems will depend on continued developments in video recording techniques.

Key words: coronary angiography, high-definition video, computer interaction, video camera iris, radiation levels, television systems, video systems

INTRODUCTION

Since the development of coronary angiography, cine recording techniques have been used almost exclusively because they best met the temporal, spatial, and contrast requirements for cardiac imaging. Recently, however, video recording techniques have greatly improved, and systems are now available that provide stable slow motion

From the Mayo Clinic and Foundation, Rochester, Minnesota.

Address reprint requests to David R. Holmes, Jr., MD, Division of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, 200 First Street S.W., Rochester, MN 55905.

Received July 12, 1983; revision accepted October 1, 1983.

0 1984 Alan R. Liss , Inc.

64 Holmes et al

stop-action images suitable for angiographic analysis. Video systems have unique features which prompt consideration for expanded application in cardiac imaging. These include instant replay, reduced costs and radiation requirements, and computer interactive capabilities.

The eventual role and optimal configuration of these video systems is as yet undefined. It is the purpose of this paper 1) to evaluate high line rate (1,023-line), high band-width video systems under clinical coronary angiographic conditions, compare these to high-definition (525-line) video systems, and then compare both types of video systems to conventional 35-mm cine angiography; 2) to define the current and potential future applications of video systems in the cardiac laboratory.

MATERIALS AND METHODS Present High-Definition System

Our current system is a high-definition system (525-line, 5 MHz) which consists of a conventional General Electric TE-40 video camera in a Fluorocon 300 system with an Ampex VPR-2 tape recorder [ 11.

The VPR-2 video tape recorder provides images which can be reviewed on a high- quality monochrome monitor in all of the cine film viewing modes available including real-time, forward or reverse, slow motion, and stable stop-action without image breakup. This tape recorder has a signal-to-noise ratio of 46 dB at a video band width of 5 MHz achieved by the automatic scan tracking features which maintains maximum video signal during playback from the tape [2]. In addition, it incorporates a digital time-base corrector which eliminates time base misregistrations [ 11.

High Band Width and/or High Line Rate System

The high line rate system consists of a General Electric video camera and an Ampex XVR-2 tape recorder which is capable of operating at either the 525-line or 1,023- line rate at 10 MHz band width. A time-base corrector is not available for the Ampex XVR-2 recorder. This system was used in conjunction with our standard General Electric biplane cardiac catheterization laboratory [I]. The systems could be confi- gured so that two combinations could be compared with our present high-definition system: 1) 525-line, 10-MHz system, and 2) 1,023-line, IO-MHz system.

Methods

After optimizing the imaging chains, comparison images were produced using the 6-inch mode of the image intensification system in patients undergoing clinically indicated diagnostic cardiac catheterization. Catheterization including left ventricular angiography and coronary angiography was performed using either a percutaneous femoral or brachial artery cutdown approach. Multiple projections of both the right and left coronary arteries were obtained.

This study was designed so that whenever possible angiographic images were recorded simultaneously from cine and different video system configurations. This not only greatly reduced the need for additional contrast injections and radiation exposure, but also facilitated comparisons because the two imaging systems would be displaying the same radiographic image. Prior informed consent was obtained from those patients who participated in the portion of the study that required an additional

Videoangiography-Clinical Use 65

selective coronary contrast medium injection with attendant radiation over and above the routine diagnostic study.

Two cardiologists (HCS, DRH) viewed 60-fps images at the same time from the same patient side by side on either two TV monitors or on one TV monitor and one Tagarno cine projector. Images were subjectively assessed for resolution, noise, contrast, gray scale, and overall image quality. The numerical scores by each cardiol- ogist according to the scoring system outlined in Table I were analyzed using Student’s t-test. In most cases, images could be compared directly, one against the other. However, for two of the video image configurations (Table II), it was necessary to make an indirect comparison in that both video images were directly compared against the same cine image from the same patient and thus indirectly compared to each other. In the case of these indirect comparisons (Table 11) the difference between the image quality scores against the common cine images was determined and a paired t- test performed on the resultant difference.

TABLE 1. Clinical (Diagnostic) Image Quality Differences

Score Basis of comparison

-3 B is significantly better than A -2 B is much better than A - 1 B is slightly better than A

0 B is similar to A 1 A is slightly better than B 2 A is much better than B 3 A is significantly better than B

TABLE II. Image Quality Difference Results

Direct or Best Number __ HCS DRH

Systems compared indirect system comparison” Average P Average P

I . 525-line, 5-MHz, 4:3 vs Direct Cine 29 0.57 <0.01 0.31 <0.05

2. 525-line, 10-MHz, 1 : 1 vs Direct Cine 25 0.34 <0.05 0.16 NSD

3. 1,023-Iine, 10-MHz, 1:l vs Direct Cine 23 0.91 <0.01 1.91 <0.01

4. 525-line, 5-MHz, 4:3 vs Direct 525-line, 33 0.21 <0.05 0.09 NSD

5. 52S-line, 5-MHz, 1 : l vs Direct - 30 0.10 NSD 0.06 NSD

6. 525-line, 5-MHz, 4:3 vs Indirect - 29/25 0.23 NSD 0.15 NSD

7. S25-line, 5-MHz, 4:3 vs Indirect 525-line, 29/23 0.32 NSD 1.60 <0.01

8. 525-line, 5-MHz, 4:3 high ’ Low dose 34 1.19 <0.01 0.77 <0.01

“Number of patient examinations compared. Where two numbers arc given (eg, 29/25), the values are the number of patients in the two groups compared indirectly. ’Different injection, same patient. Note: All comparisons based on t-test for direct comparisons or on paired t-tests for indirect comparisons.

cine

cine

cine

525-line, 10-MHz, 4:3 5-MHz

525-line, 10-MHz, 1: 1

525-line, 10-MHz, 1: 1

1,023-line, 10-MHz, 1: 1 S-MHZ

vs low dose rate rate

66 Holmes et al

Because criteria used by the two cardiologists viewing these images were necessarily subjective, a more extensive subjective assessment was obtained. Images from six patients were selected which provided a range of patient chest wall thicknesses and represented the spectrum of conventional angiographic views. These images were then evaluated by 80 participants in a study performed in an exhibit at the annual meeting of the American College of Cardiology (ACC) in April 1982. Neither patients nor institutions were identified and the study was carefully presented in a neutral manner to avoid any possible study bias. The following comparisons were made: 1) 525-line, 5-MHz, 4:3 aspect ratio versus cine (2 patients); 2) 525-line, 5-MHz, 1:l aspect ratio versus cine (2 patients); and 3) 1,023-line, 10-MHz, 1: 1 aspect ratio versus cine (2 patients).

RESULTS

There were 174 comparisons in 117 patients. There were 85 males and 32 females with a mean age of 58 years (range 35-81 years). The mean weight of these patients was 76.9 kg with a range of 48-127 kg.

Comparison between video and cine images by the two cardiologists resulted in cine being rated as the best modality, but the differences were perceived as slight (Table 11). In some cases the video images were rated better than cine images owing to slight overexposure of the cine images along the cardiac border. Video images may provide better results in this situation because the brightness and contrast of the video image can be adjusted while viewing, an option not available with cine. In those few instances where cine images were rated much better than video images, the video images were obtained with video levels less than 50% of the dynamic range of the video system.

Cine was rated more consistently better than the 1,023-line video images. The perceived image quality difference between the 525-line, 5-MHz and 1,023-line video images were quite variable owing chiefly to differences in individual observer per- sonal preference, particularly with respect to noise levels. The subjective image quality of the 1,023-line video system was not significantly better than that of the 525-line system. A higher noise level in the high-line rate, high band-width video system was the major reason for subjectively less satisfactory 1,023-line images. There was no statistical difference in perceived image quality between the two configurations of the 525-line system (5 MHz, 4:3 aspect ratio versus 10 MHz, 1:l aspect ratio). Each was compared directly to cine, however, and they could only be indirectly compared to each other.

Up to this point all video images were simultaneously recording during cine filming at cine exposure levels using a beam splitter in the optical distributor. This results in the cine film receiving 90% and the video camera receiving 10% of the light from the image intensifier.

The final comparison related to the radiation levels required for video versus cine imaging. For cine recording the entrance exposure to the patient is 65 R/min for an abdominal equivalent phantom. This results in an entrance exposure rate to the image tensifier of 34 microRoentgen/frame (6-inch mode, at 70 kVp and 60 fps). The high- definition video images were recorded at a patient entrance exposure rate of 22 R/ min (20 microRoentgen/frame at 70 kVp) at the high dose level and 18 R/min (16 microRoentgen/frame at 64 kVp) at the low dose level. This required additional


radiation exposure to the patient and one additional contrast injection. Thus, these video images were not recorded simultaneously although the same patient and projections were used. Comparison indicates that the low dose images were preferred by the cardiologists (Table 11). Contrast of the low dose images was increased slightly owing to the lower kVp's used in recording.

Comparison at ACC Meeting The SO responses from the ACC comparison are tabulated in Figures 1-4. Overall,

there was no clear-cut preference for cine or for video images obtained in this study (Figs. 1-3). However, 88% of the participants indicated that the quality of these video images was equal to or better than the cine film from their own institutions.

525 LINE, 5 MHz, 4:3 ASPECT RATIO vs CINE 80 , 50

40

Percent 3o

20

10

0 This video IS This video is This video is

1 better than equal to worse than

cine cine cine .j.

1023 LINE, 10 MHz, 1:l ASPECT RATIO vs CINE

6o 7 40

Percent 3o

20

10

Thls video IS This video 15 This video is 3 better than equal to worse than

cine cine Cine .),.. .

525 LINE, 10 MHz, 1 : 1 ASPECT RATIO vs CINE

50

40

Percent 3o

20

10

This video IS This video is This video IS

2 better than equal to worse than cine cine cine ., .,,

"HOW DO THESE VIDEO IMAGES COMPARE TO CINE IN YOUR INSTITUTION?"

7 0

51% ,,I 50

40

30

20

10

Percent

" - This video is This video is This video IS

4 better than equal t o my worse than my cine cine my cine .,..

Fig. 1. Comparison of image quality of 525-line, 5-MHz, 4:3 aspect ratio video with 35-mm cine film recording.

Fig. 2. Comparison of image quality of 525-line, 5-MH2, 1 : l aspect ratio video with 35-mm cine film recording.

Fig. 3. Comparison of image quality of 1,023-line, 1:l aspect ratio video with 35-mm cine film recording.

Fig. 4. Results of comparison between present study video images and cine images at the participant institution.

68 Holmes et al

DISCUSSION

Video systems have unique features which prompt consideration for expanded application in cardiac imaging. These include reduced costs and radiation levels, lack of film processing, and the potential for computer interfacing.

High-quality stop-action slow-motion video tape recorders (VTRs), which replay without jitter or noise bars are suitable for angiographic review, require digital time- base correction (TBC). Systems with these features are expensive ($50,000). This is a one-time cost, however, and is offset by the continued savings due to the lower cost of video tape in comparison to 35-mm cine film and its development (processor, chemicals, and personnel). Moreover, because fluoroscopy and video record times average only 4-8 min of a typical 40 to 60-min left heart catheterization, a single VTR with appropriate cabling and switching can record images from several angiographic suites. This further reduces the cost per room and per procedure. In our laboratory a single recorder in a central recording facility with a two-way intercom and video switching supports three angiographic suites with minimum queuing problems. It is anticipated that progressive reduction in electronic component size and cost should facilitate the development of less expensive VTRs with TBC features. A single video tape can store approximately 60 patients’ complete angiographic records at a cost of less than $2 per patient. This compares with cine film costs of approximately $40 per patient.

Radiation exposure in the catheterization laboratory remains a significant problem. High-definition video images of diagnostic quality can be obtained at much lower radiation levels than those required for cine angiography . In general, the exposure ranges for high-resolution video recording are twice fluoro levels and only one-third to one-fourth the cine angiographic radiation levels. These lower radiation levels become increasingly important as patients undergo repeat angiographic studies to evaluate progression of disease and results of therapy such as surgery or percutaneous transluminal coronary angioplasty (PTCA). Reduced radiation levels are important not only for the patient, but for the laboratory staff as well, particularly as therapeutic procedures such as PTCA and intracoronary streptokinase infusions often involve increased procedure duration and radiation exposures.

Video systems have a clear-cut advantage over cine systems in that processing equipment and time delays inherent in cine film processing are not necessary. This allows for decreased costs and the instant replay of any recorded images. In patients with critical coronary stenoses, unstable angina, or hemodynamics, the instant replay of any single injection in slow motion or stop-action allows the operator to fully assess anatomy and the diagnostic adequacy of angiograms during the procedure. This allows for a more data-directed study with less redundancy or omissions, and less contrast media, radiation exposure, and morbidity. We routinely replay and review the video images of all studies in the angiographic suite just prior to procedure completion. This has resulted in fewer than 1 in 2,000 cases where a repeat angiography study was required owing to diagnostic inadequacy.

Instant video replay also facilitates reaching immediate clinical decisions in emer- gency situations regarding the need for cardiac surgery or other therapy such as PTCA without the critical delay of waiting for cine film development. In our laboratory the stop-action display of the contrast-filled coronary artery on a replay monitor has provided side-by-side comparison with a real-time video monitor during therapeu-


tic interventions such as PTCA or intracoronary streptokinase infusion. This has allowed precise localization of the intracoronary catheter within the coronary artery without the need to place another marker catheter. This reduces procedure time, complexity, and morbidity.

Finally, computer interfacing plays an increasingly important role in the cardiac laboratory. Densitometric analysis of cine angiograms is difficult because of the variability of film development which necessitates careful densitometric techniques and calibrations. Measurements of ventricular functions such as left ventricular stroke work require simultaneous image and hemodynamic data. This necessitates either precise synchronization of these data from two separate recording media or some method of providing the calibrated hemodynamic measurement directly on the cine film. Neither has worked well, both are awkward and time-consuming, and they are not routinely employed in clinical cardiac laboratories.

Because video images are recorded as electronic data on magnetic tape, and no adjustments are needed for subsequent processing, they lend themselves more readily to automated analysis. Videodensitometric evaluation of cardiac valve function [3], vein graft flows [4,5], and routine biplane ventricular volume measurements from computer analysis of video angiograms [6,7] have been reported from our laboratory where all clinical angiograms are recorded dually on cine film and video tape. The angiogram is replayed from the VTR during nonrecording times and by employing an x, y plotting tablet, the recording technician can provide biplane ventricular volume and ejection fraction measurements using a Simpson’s rule ventricular volume analysis program previously described [6,7]. This analysis can be performed in 5-10 min and the data are frequently available before the patient leaves the angiographic suite. Other video-computer applications include measurements of regional myocardial blood flow [8] and analysis of regional left ventricular wall motion by a technique that is independent of spatial reference points with their inherent problems. This technique [9] tracks the epicardial and endocardia1 surface, determines left ventricular wall thickness 60 times per second, plots the wall thickness curve throughout the cardiac cycle, and determines the peak rates of systolic wall thickening and diastolic wall thinning. These measurements, which are highly sensitive indices of regional wall function and their modification by therapy [ 10-141, are not practically attainable by cine angiographic techniques. The simultaneous recording of up to seven analogue signals (ie, EKG, LV, aortic pressure) on the left margin of each video frame with a gray scale encoder system has greatly simplified the problems of analysis of composite functions such as pressure volume curves. Finally, a new and promising imaging modality, video subtraction angiography, is being evaluated in our laboratory. Initial experience has yielded promising results with the high-definition video tape recorded images in a postprocessing mode with Hybrid Video Subtraction Angiography [ 151.

These advantages and expanded applications of video recording systems should facilitate more widespread use of these systems in the cardiac laboratory. The optimal configuration for these systems is dependent on the structure to be imaged. With technological improvements there has been interest in moving to high-resolution systems.

Although there are studies that indicate that high-line rate, high band-width video systems are superior to 525-line, 5-MHz systems, these studies have evaluated video image quality using high-contrast objects with high light levels [ 161. These low-gain, no-quantum mottle situations are not realistic or representative of clinical cardiac

70 Holmes et al

angiography where low-contrast images are obtained along with significant noise owing to both quantum mottle and electronic noise from the various components in the image chain. High-resolution systems, which may have theoretical advantages, appear to have, under clinical angiographic conditions, significant practical disadvan- tages. To accomodate the increase in the number of vertical scan lines one must increase the band width of the imaging chain. We have found that this increased band width is accompanied by a significant increase in noise which reduces the quality of the diagnostic image. Using the image chain in our laboratory, comparison images have documented that a 1,023-line video system is not significantly superior to our optimized 525-line, 5-MHz system. Clearly these results apply only to the systems under evaluation. Other commercial systems may have different performance char- acteristics. Nevertheless, it is evident that increased line rates by themselves do not necessarily produce better images. The video images produced with our 525-line system compared favorably to our 35-mm cine images.

In most commercially available cardiac angiography systems, the video camera is locked into the iris settings which are set at optimum for the cine camera, and indepedent iris settings for the video camera are not available. This is unfortunate, as there is a much higher light requirement for the cine camera. It has been our experience that under these conditions, the video cameras are usually operating only in the lower half of their dynamic range. By the simple modification of installing an independent, manually controlled video camera iris, we have utilized the full capabilities of the video camera. The improvement in video image quality from the 525-line, 5-MHz cameras following this modification was significant, and we believe this feature alone is chiefly responsible for the favorable assessment these images received at the ACC evaluation. Further attempts at image enhancement by use of a 1,023-line high bandwidth video image did not provide any further subjective improvement in image quality. It is our impression that far greater video image enhancement is obtained at far less cost and at far less system limitation by employing a manually controlled video camera iris than by switching over to a 1,023-line high bandwidth system with its inherent increased noise.

Despite the significant improvement of video images, their computer-interactive analysis capabilities, and our increasing reliance on instant replay and stop-action techniques during angiographic procedures, cine angiography remains our primary method of image recording. This continued reliance on cine recording is necessitated only by logistic problems. We are satisfied with the diagnostic quality of the video images obtained with optimum iris adjustment and replay on slow motion/stop-action VTRs. These VTRs, however, are not widely available in the medical field. This is partly because of their initial high cost and partly because they are a new product whose utility is not widely known and most cardiac laboratories already have pur- chased standard VTR equipment. For these angiograms to be optimally reviewed elsewhere, a compatible VTR should be available; otherwise the video tape must be duplicated on a 3/4 inch video cassette, or hard-copy still pictures of selected video frames must be obtained with a format camera similar to those employed with computerized axial tomography (CAT) scanners. With the latter two options there is some image degradation due to transcription, or loss of stable stop-action viewing (3/4 inch video cassette) or loss of dynamic aspects (format cameras).

While these factors currently limit the ease with which these video angiograms can be viewed by referring physicians in other centers, the rapidity with which improved


video systems are being widely applied is impressive. By comparison, cine film technology has been virtually static. Thus, the eventual role and potential replacement of cine angiograms by video tape systems will depend on not only continued developments in video recording techniques but also the widespread dissemination of standardized high-quality VTRs in the medical field.

ACKNOWLEDGMENTS

We would like to acknowledge the assistance of General Electric Medical Systems Division in the evaluation which was carried out at the American College of Cardiol- ogy Conference (April 1982). General Electric Medical Systems Division also provided the data and data analysis used to develop Figures 1 through 4.

REFERENCES 1. Gray JE, Wondrow MA, Smith HC, Holmes DR, Jr: Technical considerations for cardiac laboratory

high-definition video systems. Cathet Cardiovasc Diagn 10:73-86, 1984. 2. Hathaway RA, Ravizza R: Development and design of the Ampex auto scan tracking (AST) system.

J SOC Motion Picture Television Eng 89:931-934, 1980. 3. Tsakiris AG, Sturm RE, Wood EH: Experimental studies on the mechanisms of closure of cardiac

valves with the use of roentgen videodensitometry. Am J Cardiol 32: 136-143, 1973. 4. Smith HC, Frye RL, Davis GD, Pluth JR, Sturm RE, Wood EH: Roentgen videodensitometric

measure of coronary blood flow: Determination from simultaneous indicator-dilution curves at selected sites in the coronary circulation and in coronary artery saphenous vein grafts. Mayo Clin Proc 46, December 1971.

5 . Smith HC, Sturm RE, Wood EH: A videodensitometric system for measurement of vessel blood flow, particularly the coronary arteries in man. Am J Cardiol 32:144-150, 1973.

6. Ritman EL, Sturm RE, Wood EH: Biplane roentgen videometric system for dynamic (60isecond) studies of the shape and size of circulatory structures, particularly the left ventricle. Am J Cardiol

7. Fellows JL, Cravath DL, Wondrow MA, Ritman EL, Smith HC: Clinical laboratory techniques for computer-based analysis of cardiac roentgen angiograms. Anal SOC Cardiopulmonary Techno1 7(4), Oct/Dec 1977.

8. Smith HC, Robb RA, Ritman EL: Roentgen videodensitometric assessment of myocardial blood flow: Clinical applications. Second International Workshop Conference, April 1976.

9. Dumesnil JG, Ritman EL, Frye RL, Gau GT, Rutherford BD, Davis GD: Quantitative determination of regional left ventricular wall dynamics by Roentgen videometry. Circulation 50:700-708, 1974.

10. Dumesnil JG, Ritman EL, Davis GD, Gau GT, Rutherford BD, Frye RL: Regional left ventricular wall dynamics before and after the administration of sublingual nitroglycerin. Am J Cardiol 36:419- 425, 1975.

11. Chesebro JH, Ritman EL, Frye RL, Smith HC, Connolly DC, Rutherford BD, Davis GD, Danielson GK, Pluth JR, Barnhorst DA, Wallace RB: Videometric analysis of regional left ventricular function before and after aortocoronary artery bypass surgery. J Clin Invest 58: 1339-1347, 1976.

12. Chesebro JH, Ritman EL, Frye RL, Smith HC, Rutherford BD, Fulton RE, Pluth JR, Barnhorst DA: Regional myocardial wall thickening response to nitroglycerin. A predictor of myocardial response to aortocoronary bypass surgery. Circulation 57:952-957, 1978.

13. St John Sutton MG, Frye RL, Smith HC, Chesebro JH, Ritman EL: Relation between left coronary artery stenosis and regional left ventricular function. Circulation 58:491-497, 1978.

14. St John Sutton MG, Tajik AJ, Smith HC, Ritman EL: Angina in idiopathic hypertrophic subaortic stenosis-a clinical correlate of regional left ventricular dysfunction: A videometric and echocardi- ographic study. Circulation 61:561-568, 1980.

15. Wondrow MA, Gray JE, Forbes GS, Earnest F 11, Smith HC, Stears JG: Hybrid video subtraction angiography (HVSA) in neuroradiology. Submitted for publication.

16. Haendle J, Horbaschek H, Alexandresce M: High-resolution x-ray television and the high-resolution video recorder. Electromedica 3:83-91. 1977.

32:180-187, 1973.

Documents

Clinical evaluation and application of cardiac laboratory high-definition video systems