Video networking of cardiac catheterization laboratories.

UNLABELLED
A video telecommunication network was established to transmit coronary images between a cardiac catheterization laboratory and a remote core laboratory. In 40 patients during interventional procedures, cine angiograms, live fluoroscopy, intravascular ultrasound studies, and images of the cath lab were transmitted in real time over a T1 line at 768 kbits/second at 15 frames/second.


RESULTS
Measurements of angiographic and intravascular ultrasound parameters were very close between the original studies and the transmitted images. The telecommunication link up during the interventional procedures had a significant affect in 58% of cases.

T HIS STUDY was performed to assess the feasibility and accuracy of a video communication network to transmit coronary images in real time through commercial telephone lines to permit interaction between personnel in a cardiac catheterization laboratory and a remote core laboratory.

METHQDS
A point to pomt VTEL 225 BK telecommunication system (VTEL. lnc. Austin, TX) was installed in the core laboratory at the Urfiversity of California, Irvine and conr~ecmd with another VTEL 225 BK module 40 miles away in the cardiac catheterization laboratory at Kaiser Sunset Hospital, Los Angeles, CA. The VTEL syslem uses proprietary data compression algorithms to transmit data over telephone lines. The two laboratories were connected by a TI phone line that can carry up to 24 channels. The VTEL system includes an IMUX (ASCEND multiband with TI-Primary Rate Interface) to convert the data to an accep[able format for the phone lines. This system allows video signals to be digitally converted, compressed, and sent at 768 Kbits per second using 12 of the telephone channels simuItaneously, lmages were transmitted at 15 frames per second.

CORE LABQRATORY COMPUTER SYSTEM
In the core laboratory, the video images transmitted from the remote lab were passed through a vŸ recorder (SVHS Panasonic Model 1960) and then serially senl to the digitizing board of ah Apple Power Pe 8500 computer. Selecled frames of the coronary angiograms of intravascular ultrasound images were digitized at 640 • 480 pixels using software by AVID Video Shop 3.0.2 (Avid Technology, Tewksbury, MA). While the clinical case was in progress at the remote laboratory, measurements of tumen diameter from the angiograms or p]aque and lumen cross-sectional areas from the ultrasound images were performed using public domain software (NIH Image v1.59), The images were analyzed using digitaI calipers with operator defined edge detection. The angiographJc images were scaled using a 2• zoom mode and calibrated by the diameter of the guiding catheter. Afler an image was analyzed, it was moved to a second monitor attached to the computer. The output of the second monitor's video card was connected to the video input of the VTEL telecommunication system for transmission back to the remote laboratory. In this manner, the images originally downloaded from the remote laboratory could be measured and annotated by the core laboratory with the results of the quantitative measurements wfitten on the image and transferred back to the laboratory for their review dnring the case.

DATA ANALYSIS
Following the catheterization procedure, the original videotape of the ultrasound images and the cineangiogram were m~i 91 to the core laboratory. The original cineangiograms and ultrasound images obtained in the cardiac cath lab were compared with their corresponding paired images, downloaded over the telecommunications system at the time of the initial procedure. The sets of paired images were compared by linear regression analysis and the difference in measurements in the paired images were assessed by the Bland-Altman analysis.

RESULTS
Forty patients were randomized per protocol. There were no statistical differences in the clinical characteristics of the patients assigned to the angiographic versus intravascular ultrasound (IVUS) groups. The mean age of the patients was 59.8 -2_ 8.9 years; 32 (80%) of the patients were men. The baseline angiographic mean lesion length, mean minimum lumen diameter (MLD) pre-and poststent, and the mean reference lumen angiographic diameters are shown in Table 1. There were no significant differences between the two groups for the baseline angiographic parameters. The final "punch-out" ba[loon mean diameter was 3.7 _+ 0.4

Telecommunications
The telecommunication network was successfully completed in 39 (98%) cases. The linkup was delayed in three (8%) cases due to routing problems within the phone company. The quality of the video images was impaired in one (3%) case due to "tileing" from inadequate speed of transmission when the T1 line was not used, and the quality of the audio was impaired in one (3%) case due to intermittent drop out of the voice channel from the cath Lab.
The results of the quantitative comparison between the original cineangiograms and the angiographic images digitized over the telecommunications network are shown in Table 2. There were no significant differences between the original cine and the telecommunications angiograms for the lesion length or the reference lumen diameter proximal or distal to the stent, both pre-and poststent insertion. The baseline MLD was 0.8 -+ 0.3 mm for the original cine angiograms and was 0.7 +_ 0.3 mm for the telecommunication group, P < .008. The correlation of angiographic mini- The results of the IVUS measurements compa> ing the original S-VHS tapes and the images digitized during the telecommunication study are shown in Table 3. There were no significant differences between these methods for assessment of lumen diameters or areas either within the stent or at the proximal or distal reference zones. The correlation coefficients between the original studies and the telecommunication system were between 0.93 and 0.99. The Bland-Altman analysis for stent area by IVUS showed only minor differences between the original ultrasound tape and the telecommunication images, (average difference = 0.0 • 0.3 mm2).

DISCUSSION
This is the first time, to our knowledge, that a telecommunication network has been used to transmit interventional procedures from a catheterization laboratory to a core laboratory. This study demonstrates that a vŸ telecommunication network provides excetlent image quality with accurate representation of angiographic anatomy and lesion severity. In addition, IVUS lumen and plaque geometry also correlated closely with the original images observed direct]y in the cath lab. The telecommunication images minimally underestimated the diameter of the stenoses compared with the original angiograms (0.7 _+ 0.3 mm v 0.8 _+ 0.3 mm), but this was not clir~ically significant.
The video telecommunication network had a significant impact on the performance of 23 (58%) cases in this small clinical trial. The causes of this influence were (1) the presence of another reviewer for discussions of image interpretation or recommendations of techniqª on-tine dudng the intervention, (2) the abi[ity to have quantitative coronary angiography performed by an independent observer at the time of the procedure, and (3) the presence of an experienced reviewer of IVUS images both for diagnosis and measurement at the Core Lab. The implications of this initial study for the use of telecommª in interventional procedures ate significant. Telecommunication net-works could be used to link community hospitals to universities of other centers of excellence to facilitate physician training and improve technical skills and judgement during interventional cases. This coª also assist managed care networking to a tertiary center for review of cases for clinical conferences. FinaIly, this project demonstrates a paradigm of how future multicentered ctinical trials could be operated through vŸ telecommunication networks so thal the procedures could be reviewed on-line for appropriate inclusion criteria, correct adherence to the protocol, and immediate assessment of the adequacy of the results.