Proceeding of the 2nd Asia Pacific Association of Medical Informatics Conference (Sydney), Plenenary Session, 1997.

Telemedicine and multimedia technology

Hiroshi Takeda

Department of Medical Information Science, Osaka University Hospital 2-25, Yamada-Oka, Suita 565 Japan


Abstract: Telemedicine covers wide range of medical information which includes text, numeric, image and voice data. It is natural to apply multimedia and communication technology to telemedicine. In this paper, early experiments toward establishing high quality image-oriented telemedicine and HDTV/SHD and B-ISDN/CS technology are installed and evaluated. Some obstacles for deployment of the telemedicine are also described.

1. Introduction

Telemedicine, the "application of telecommunications to the health care," has been widely spreading in the field of teleradiology, telepathology, home telecare and teleconference. In a Japanese recent survey, 192 telemedince projects were identified, of which 88 were teleradiology related [1]. Owing to successful installation of the hospital information system of Osaka University Hospital [2], our next goal is the development of PC-based PACS [3,4,5], electronic patient record system [6,7] and telemedicine [8]. For two years we have carried out basic experiments to clarify feasibility for application of multimedia technology for high quality-oriented image transfer.

2. Experiment 1: "MEDIKORE-Boderless Project" teleconference using HDTV images and B-ISDN

Both Kyoto and Osaka University Medical School have started a new teleconference experiment, called as "MEDIKORE (Medical Information Interchange between Kyoto and Osaka University for Research and Education Purposes)-Boderless Project" in order to exchange lectures and conferences with each other by using HDTV (High Definition Television) and B-ISDN (Broadband integrated services digital network) channel in cooperation with BBCC (the association of Broadband Business Chance and Culture creation) in 1995.

2.1 System Design

Diagram of the system is described in Figure 1. The HDTV original 1.2 Gbps images (1440 pixels x 1035 lines x 8 bit depth, 60 frames per sec) were compressed to 150 Mbps by means of CODEC (Mitsubishi Electric) which had both analog to digital/digital to analog conversion and encoding/decoding function. B-ISDN has the transfer speed of 156 Mbps and a pair of B-ISDN channel was able to transfer both NTSC (National TV Standard Committee, 720 pixels x 480 lines x 8 bits depth), HDTV and voice data bi-directionally. The input resources were composed of an acquisition equipment for 35 mm slides projector (Nikon) and over head projector (OHP)(Nikon), HDTV video recorder (Matsushita) for HDTV images and three sets of a NTSC camera (Cannon) were used for taking live scene of a speaker, a discusser and audiences. The display resources were main display of HDTV rear of front projector (120 to 200 inches) and two pairs of NTSC and HDTV monitors for speakers and operators. Two sets of voice system were arranged for own voice and partners voice transfer.

2.2 Operational procedure

One of two medical schools was in charge of either a teacher site or a student site. The director of the teacher site had to take care of managing the whole process and choosing the images (still images of a slide or a OHP, teacher's picture or students' picture and so on).

The management of selecting input resources and overlaying the pictures was so complicated that a director-assisting system (Mitsubishi Electric) was developed on a Macintosh computer system. The system was able to control matrix switches, NTSC cameras and other resources. A staff could easily register some screen patterns in advance, and image patterns were selected by a function key on a site. Several staffs were needed to handle the experiment at first, but recent experiment required a few operators at both sites.

2.3 Evaluation

Various kinds of teleconferences have been held over ten times since the spring of 1995. Students' and audiences' responses in the early experiments were evaluated by questionnaires. In their answers, the highest score was obtained for the significance of the Borderless Project and most audience expected to continue the interactive teleconferences. However the quality of moving pictures and still images were evaluated to be improved. The resolution of the color slides in particular was not enough and the problem was due to input devices which take slide images by HDTV cameras. Direct input method through a high resolution computer bus may solve the problem. Since pointing devices and interactive discussions were assessed rather high, improvement of input device would make the teleconference more effective. Although the system contains technological and operational problems, the potentiality of HDTV and B-ISDN technology for distance medical learning between two medical schools was demonstrated in the experiment.

3. Experiment 2: Teleconference system using SHD images and B-ISDN

In clinical process such as care for a cardiovascular disease patient, movies of high definition picture are necessary for diagnosis and the case will be the most difficult to be dealt in telemedicine. In order to certify that specification of SHD (super high definition) [9] image system and B-ISDN is enough for high quality oriented- telemedicine, we constructed a teleconference system in which all kind of medical pictures are archived and retrieved , and experiments of teleconferences were carried out.

3.1 System configuration

The SHD image format is defined as spatial resolution of 2048 x 2048 and 8 bits depth. The SHD filing system is structured as four components of data acquisition, data processing and storage, display and communication (Figure 2).

Data acquisition and print out: The film scanner (Leafscan45, Leaf Systems) and the digital still camera (Leaf Digital Studio Camera, Leaf Systems) are connected to the personal computer (MAC 8100/80). The film digitizer (LD-4500, Konica) and the color printer (Fujix Pictograph 3000, Fuji Photo) are linked to the workstation (Sun SS20). The film digitizer and the digital camera which allow flexible input have spatial resolution of more than 3000 pixels x 2000 lines at more than 12 bits depth. After the input image data are processed on the computers, connected to the input device, they are then transferred to the SHD image server via LAN.

Data processing and storage: The image server (SHD-1000, Mitsubishi Electric) and an image database are installed. The server is able to contain 256 SHD images on local RAM that corresponds to 3 Gigabytes data volume. The SHD image data for each RGB signal channel are transmitted through 8 Gbps display bus, and converted to analog output with the rate of 357 mega-samples per second. Frame rate is achieved as progressive 60 frames per second. A video board for simple television meeting can also be connected to this server to digitize NTSC signals. The video signals input into the video board are used as a device to input non SHD images with low resolution.

Display: The SHD monitor (DDM2801, SONY) is used for the display which is able to demonstrate 2048 pixels x 2048 lines and temporal resolution is up to 60 frames per sec with non-interlace RGB 8 bits image.

Communication: The SHD system provides an environment for multipoint SHD conference. The three sets of the local system are connected via B-ISDN. The pictures in the system are transferred to another one beforehand and can be controlled from remote points (Figure 3). An electronic television conversation system is also installed in the control unit of the workstation to communicate among the sites.

In constructing an image filing system which archives a variety of medical images, a cardiovascular disease case was selected as an example. Since the patient underwent twice PTCA (percutaneous transluminal coronary angioplasty) because of restenosis of the artery, her clinical history can be divided into 4 phases (before and after the first and the second PTCA). In the filing system, we stored the pictures of case summary, electronic cardiogram, chest roentgenogram (CX), ultrasonic cardiogram (movie)(US), myocardial scintigram (MS), coronary angiogram (movie) (CA) and left ventriculogram (movie)(LV) in each phase. The total number of stored pictures was 243 frames. In order to improve retrieval operability, a GUI (Graphic User Interface) displaying a matrix was developed. The vertical axis indicates the types of medical modalities and the horizontal axis, the time course. By clicking a button on the matrix, users can retrieve and instantly display the chosen modality image from the SHD image frame memory.

3.2 Evaluation

At first, this system has made centralized and instantaneous retrieval was made possible for a variety of high resolution medical images by a SHD image filing system with effective GUI. In order to obtain medical comments regarding this system, twenty-one doctors from Osaka University Hospital were asked to evaluate its operability of the retrieval system, quality of the images and possibility of medical use after operating the system.

As for the operability, all images were stored in the frame memory and waiting time for the image to be displayed was extremely short. The GUI matrix which presented seven types of modality-specific images in four phases was good to help the user operate several images and to select the desired image easily.

As for resolution, CX, CA, LV received high evaluation and US and MS received low evaluation. There may be a problem in the images' input methods for US and MS. And the fact that images of CX, CA, LV that have high resolution in origin were evaluated highly indicated that display of these images was feasible by the SHD format. The above suggested that the SHD image format would be sufficient in terms of quality at least for the types of images handled in this experiment. The system was also capable of displaying animated images without losing its high quality.

Doctors also shared opinions regarding the system's effectiveness in the medical practice. Physicians' need to compare two or more images of different phases on the same screen surfaces was fulfilled in the experiment. Out of US and LV and other images which are generally poor in quality, two or more selected images with different phases were simultaneously displayed on one screen. Its convenience in observation was graded highly for effectiveness in a patient care. Use in education system and study groups were evaluated as "can be used" and use as clinical research database was evaluated as "can mostly be used."

In teleconference experiments between Osaka University and Kyoto University via B-ISDN, it took 2 hours to transfer all the pictures (about 3GB) beforehand. Owing to the prefetch procedure, we could select the picture quickly using user friendly interface tool which also controlled the remote system simultaneously. The system also showed pointer function not only in static picture but also in movie. When we stopped the movie, we could see the same frame in both side. The conference finished without any trouble as though it held in the same room. The success demonstrated that super high definition system and B-ISDN based system is enough specification to deal the high quality-oriented telemedicine.

4. Experiment 3: Medical Information Network by Communication Satellite for University Hospitals

Another experiment of the high quality-oriented telemedicine was started since 1996. The project, MINCS-UH (Medical Information Network by Communication Satellite for University Hospitals) was initiated by the support of the Ministry of Education, Science and Culture of Japan and was participated with eight national university hospitals. The objectives of the project are 1) live broadcasting for an advanced medical state of art, 2) joint teleconferences such as clinical case conferences, 3) joint lectures and tutorials for students and trainees, 4) technological training for hospital staffs, 5) contribution to regional medical care by a network, 6) telemedicine in disasters and 7) inter-network of hospital information systems. The system features 1) the digitalized HDTV broadcasting via a communication satellite (CS), 2) encryption of the image data to protect the privacy and the security, 3) interactive and multi-lateral communication for question-answer mode, and 4) computer assisted "one man operation" at a lecture hall.

4.1 System design

The system is composed of three parts:1) CS and ground stations, 2) intrahospital audio-visual (AV) commu-nication system, 3) information network for the system management. The framework is described in Figure 4. The system is able to transfer a series of HDTV (32 Mbps) and NTSC (5.6 Mbps) image data simultaneously and multi- laterally after original image data are digitalized, compressed with MPEG2 method and scrambled in a teaching site, and the dynamic image are descrambled, decompressed and converted from digital to analog data. For controlling the sending/ receiving site and giving authentication and a key for decryption, the client-server system is implemented among eight hospitals' stations with an exclusive digital line (DDX-P, NTT).

In a case of the Osaka university hospital, the studio functions as the interface between the CS ground station (Mitsubishi Electric, Tokyo) and halls where AV input and output resources are allocated. The surgical operation AV control center has an HTDV camera (Ikegami, Tokyo) for sending live scene of a surgical operation. The dynamic images are sent to the studio via optical fiber lines. The lecture hall and the conference hall are connected with optical fibers for image and voice data transfer, and ether net for communicating control data. The control processor (Hitachi, Tokyo) is linked to matrix switchers (Mitsubishi Electric, Tokyo) of HDTV, NTSC image and voice data. Due to the control network, an operator in a hall is able to receive the information from a distant teaching site and choose a necessary media with a "one man operation" method.

4.2 Evaluation

Several experiments has been carried out since December in 1996. Lectures, clinical conferences and radiological and pathological diagnostic conference were managed to be put into practice. Moving images such as HDTV video images from a surgical operation were highly evaluated by audiences. NTSC image based SCS (Space collaboration system) project has been operated among national universities. For medical images, MINCS-UH is much better than the SCS. There exists same problems in the experiment 1 and multiple interaction is further difficult to manage. For example, pointing function from a discussant site to a teaching site is not a easy work for the modern technology. Since the experiment includes the world-first digital HDTV broadcasting and encryption, the hardware and software are so complicated that the technology is not matured yet. The reliability and operability of the system must be improved for practical use. The real time communication for operators is also necessary by telephone or internet. Both technological and medical directors are necessary and should cooperate with each other for a successful multi-lateral teleconference.

5. Discussions

In order to respond for emerging needs for telecommunication in the medical and health care field, it is natural that multimedia would become a key technology for telemedicine. Although the definition of multimedia is still dynamically changing, a tentative definition includes the compound of several media (text, voice, image so on) and interactive tool(bidirectional, conversational) and digitalized media. Such kind of digitalized media integration with interactive communication will constitute multimedia environment.

To make telemedicine in practice, the quality of image, response time of communication ,cost and doctor's attitude are major problems. In order to solve the problem within boundary condition, commercially available HDTV systems have been chosen at first and SHD is further tried with communication technology of B-ISDN or CS. From the standpoint of the information processing and communication, the quality is the most high hurdle to clear. Conventional X-ray film has a 2K x 2K x 12 bit (6MB) and if the still image is converted to color dynamic image (30 frames per sec), the data quantity is calculated 540 MB per sec. In our experiments, more than 1K x 1K of spatial resolution and 30 to 60 frames per sec of temporal resolution were targeted and combination of multimedia and high speed communication tool of B-ISDN and CS were applied. Although qualitative evaluation has not been made, medical staffs were accepted our design of high quality oriented telemedicine.

As to the cost, high quality does not enjoy the scale merit of lowering cost in the present. The minimum initial investment will be estimated as follows: input device such as a HDTV camera, a video recorder and so on, A$0.5M, display such as a HDTV rear projector, a CRT monitor, A$0.3M, interface device such as a matrix switcher, A$0.2M. If a SHD system will be introduced, the cost will be calculated as A$1.5M and as for CS ground station it will cost about A$1.2M. Running cost for B-ISDN will be about two thousand times more than conventional telephone line use and for rental fee of a transponder of a CS is A$3K per hour. It will be natural to expect that the total cost will be decreased sharply if HDTV, SHD systems are commercially available and national infrastructure for B-ISDN is ready.

High quality telemedicine will be achieved in the combination of multimedia and communication technology. However, there are many engineering specialists for either side but a few for integrating both technology. Lack of the integrator will interfere the development of the telemedicine. It is emphasized to cultivate integrators and to generate a new field of integrated technology.

Input devices and their standardization for image data exchange must be also emphasized. DICOM (Digital Imaging and Communications in Medicine) standard will be effective for radiological images but another images such as movies are not targeted so far in the medical informatics community. Inter-AV equipment and AV equipment to computers and other peripherals are hardly connected. Standardization for image data exchange between AV equipment and information processing equipment would be a very urgent and important issue.

Doctors generally assume a X-ray film image as a golden standard. If the paradigm is shifted to the acceptance of digital image of 2K x 2K x 10 bit depth as a golden standard, the technological barrier will be much lowered and the cost for development of high quality image- oriented telemedicine will be reduced.

6. Conclusion

Telemedicine has been proposed as an multimedia and high-bandwidth telecommunication technology-bases system for the effective and efficient delivery of medical care. In order for the promise of telemedicine to be achieved, some pilot models must be constructed and assessed. In this paper, the developmental efforts of high quality- oriented telemedicine projects are described. At a minimum, these projects demonstrated the feasibility of and the potential benefits to be deprived from the application of telecommunication like B-ISDN and multimedia (HDTV, SHD) technology in medical care. But, equally some questions remained unanswered. These include assessment of cost effectiveness, significant limitations in the technology capability,standardization and legal issues.

At the time when national level of information infrastructure is arranged, a part of the questions will be solved. However, the others will be especially critical to the further development of telemedicine. International medical informatics community should take initiative to fulfil the goal of telemedicine for the ultimate viability, economical sustainability and efficacy in health care delivery system.


References

1) Kaihara S (1997) Why is telemedicine not yet used in Japan. In Proceedings of the 3rd International Conference of the medical aspects of Telemedicine, Kobe. p.49. Tokyo:Digital Medical Communications Co.

2) Matsumura Y, Takeda H & Inoue M (1995) Implementation of the totally integrated hospital information system (Humane) in Osaka University Hospital. In Greenes RA, Peterson HA & Protti DJ (eds) Medinfo'95 Proceedings of the Eighth World Congress on Medical Informatics, Vancouver. pp.590-593. Amsterdam: Elsevier Science Publishers

3) Kuroda C, Kondo H, Takeda H, Inoue M, Inamura K, Mori Y & Kozuka Y. (1991) Planning for PACS at Osaka University Hospital. Comp. Meth. Prog. Biomed. 36, p.147

4) Takeda H, Matsumura Y, Kondo H, Inoue M, Kondo H, Takeda I & Miyabe˙ S.(1995) In Greenes RA, Peterson HA & Protti DJ (eds) Medinfo95 Proceedings of the Eight World Congress on Medical Informatics, Vancouver. pp.430-433. Amsterdam: Elsevier Science Publishers

5) Inamura K, Kondoh H & Takeda H (199&). Development and operation of˙ PACS/Teleradiology in Japan. IEEE communication Magazine (July). pp.46-51.

6) Takeda H, Matsumura Y, Okada T, Kuwata S, Inoue M, Hazumi N & Aoki J.(1996) Development of a cardiovascular disease-oriented electronic patient record model in a Japanese university hospital. In Proceeding Toward An Electronic Patient Record'96, San Diego. Pp. 520-523. Broadview: Kelvyn Press.

7) Takeda H, Matsumura Y, Okada T, Kuwata S, Hazumi N & Nakazawa H. (1997) Dynamic template driven data entry system for an EPR system. In Proceedings of Toward An Electronic Patient Record'97, Nashville. pp139-143. Broadview: Kelvyn Press.

8) Takeda H, Matsumura Y, Okada T, Kuwata S, Wada M & Hashimoto T. (1997) Development of a medical image filing system based on super high definition image and its functional evaluation. In Proceedings of SPIE Vol.3013, San Jose, pp.149-156. Bellingham:the Society of Photo- Optical Instrumentation Engineering

9) Ono S (1993) The super high definition image", TV Society Magazine, 47(9), pp.1204-1215.

Acknowledgement

The experiments were made in collaboration with Department of Medical Informatics, Kyoto University Hospital (Prof. Takashi Takahashi, former Assoc. Prof. Kotaro Minato, Assoc. Prof. Masaru Komori and the staff), Department of Medical Information Science, Osaka University Hospital (former Prof. Michitoshi Inoue, Dr.Yasushi Matsumura and the staff), BBCC(Mr.Yoshio Fujio) and TAO (Mr. Tsutomu Hashimoto and Minoru Wada)

The study was partially supported by a grant from the Science and Technology Agency of Japan.

Figure 1 Outline of the Experiment 1: HDTV transfer with B-ISDN

Figure 2 Local system configuration of the Experiment 2: SHD Image transfer with B-ISDN

Figure 3 Tele-clinical conference design of the Experiment 2

Figure 4 System design of the Experiment 3; digital HDTV and NTSC image transfer with CS system