Applied Radiology

Mr. Margolin is the Vice President of Marketing for Algotec, Inc., Duluth, GA.

[Editor's note: This article is presented as a counterpoint to the article, Web technology and its relevance to PACS and teleradiology by Wayne T. DeJarnette, PhD, published in the August 2000 issue of Applied Radiology, pages 9-12. Applied Radiology welcomes articles contributing to the dialogue regarding the state of radiology practice and the use of new technology.]

Though the terms "Web" and "the Internet" are often used synonymously with their enabling technologies, I agree with Dr. DeJarnette's position that some distinction between these terms is important in understanding and analyzing their relevance to PACS and teleradiology. In my opinion, though, a primary distinction should be made between Web technologies and the Internet. Web technologies include a multitude of different technologies, all of which are in common use over "the Internet"­­a global, public network, also known as the "information superhighway."

To clarify the above distinction, here are some examples:

1) E-mail technology has been widely used over the Internet for more than a decade as a means for information exchange and communication. In recent years, there has been growing use of e-mail technology, clearly a Web-related technology, for intra-organizational communications. In this application, e-mail messages will typically move on the organization's local area network (LAN), without ever crossing the public Internet. Still Web technology based, the application can then be referred to as an "intranet application," i.e., an application of Web technology over a LAN.

2) Java technology was introduced in the mid-1990s by Sun Microsystems as a tool to write platform-independent software code. In other words, with Java, code can be written and compiled only once, and then run on any machine, irrespective of its architecture or operating system. Though Java can be used as any other object-oriented programming language (e.g., C++) to code software applications, it is easy to understand its special benefits for applications that run in the context of the Internet. Clearly, on the Internet, many different target computers coexist, and the original application developer has no means to reach all of them or even anticipate their make or model. Therefore, it is not surprising that Java has become one of the most common programming languages for Internet applications. Hence, unlike C++, we can safely consider Java a Web technology.

Web technology in PACS and teleradiology applications

In his article, Dr. DeJarnette suggests that different Web technologies have not been equally useful for PACS and teleradiology applications. Clearly, no one can dispute the usefulness of technologies such as TCP/IP communications (which provide the foundation for the DICOM-3 standard) and the Internet (with its widely available, low-cost communication infrastructure). But it is exactly at this point that Dr. DeJarnette's arguments become entangled. He goes on to say that other Web technologies have much more limited use or no use whatsoever in PACS and teleradiology. In my opinion, almost all Web technologies can and will be useful for these applications. The grounds for this general statement had best be laid by the distinction made above between Web technologies and the Internet.

As defined above, what Web technologies have in common is their usefulness for Internet applications. In other words, these are technologies that either enable applications to run over the Internet, or make use of the Internet. This definition includes one fundamental advantage of Web technologies for PACS and teleradiology. In fact, I would argue that this definition shows that the future of PACS and teleradiology is immersed in Web technologies. To understand why, let us first step back a little into the history of PACS and teleradiology.

"Traditional" PACS applications, which started in the early 1990s, were all focused on solving the departmental issue of digital clinical image management. These applications were hardware- and network-intensive, and could by no means tackle the much larger problem of enterprise-wide clinical image distribution. Accordingly, these applications, without exception, were met with commercial failure. Since films were still needed to complete the life cycle of the imaging study outside the Radiology department, cost justification of PACS became practically impossible.

"Traditional" teleradiology applications, by far more commercially successful than their contemporary PACS applications, have found their success through a much narrower definition of the problem they sought to solve. The "traditional" teleradiology problem entailed images that were acquired at point A, with a need for interpretation by a doctor at point B. If it is less expensive to transport the images from point A to point B than to transport the doctor the other way, "traditional" teleradiology is appropriate. Under this simple paradigm, "traditional" teleradiology found its application in the market. But what happens if the attending physician happens to be at point C; or if there are multiple attending physicians to review the images? With these and similar questions, the "traditional" teleradiology paradigm starts to break down. The application of teleradiology has thus remained extremely limited relative to its potential market.

The Internet

First, the basic question: "What is the Internet?" must be addressed. In his article, Dr. DeJarnette defines the Internet as "a publicly owned, network-based communication infrastructure that grew out of DARPA (Defense Advanced Research Project Agency) research in the 1970s and 1980s." Clearly, this communication infrastructure has taken a leap in the 1990s, becoming more and more of an all-encompassing, global network, where people throughout the world find information, communicate, buy, sell, and seek entertainment. Now consumer-oriented, the Internet has become more widely available, faster, and cheaper to attain than any communication infrastructure ever before.

The "Internet Revolution" of the 1990s has paved the road for a multitude of new business models and markets, collectively referred to as "e-businesses." With a good deal of hype, and the help of some sharp copywriters, fields like e-travel, e-commerce, e-healthcare, and many others have emerged. Among these, it has become increasingly clear that the place of e-radiology will not remain absent for long. But what is e-radiology and what promise does it hold for the future of radiology?

The answers to the above questions bring us back to the breakdown of "traditional" PACS and teleradiology. These applications remained limited because they could not break out from the confines of the department (in the case of PACS) or some predefined points A and B (in the case of teleradiology). But what if a global network is available with enough bandwidth to support these data-intensive applications and at a price everyone can afford? The outcome is simple: a new type of PACS and teleradiology applications, fit for enterprise-wide clinical image sharing and community-wide or even worldwide radiology consultation and other remote services. Hence the new business of e-radiology.

It is clear that the Internet has great potential to revolutionize or reinvent the business of radiology. But this alone does not explain the usefulness of Web technologies for PACS and teleradiology applications. As Dr. DeJarnette suggests, it is not impossible to run non-Web applications over the Internet. The only question here is: at what cost?

Low-cost Web technology­­A myth or a fact?

As defined above, Web technologies either enable the Internet or make use of it. As such, they create an infrastructure for the development of applications designed to run over the Internet. The same two Web technologies mentioned have potential applications for clinical image management:

1) E-mail has recently been proposed by some vendors as a tool for the distribution of clinical results, including the transmission of radiology reports and reference (or key) images to referring physicians. Clearly a Web technology - why is e-mail suitable for this task? Here are some good reasons:

a) E-mail systems are prevalent and standard; they do not require any special application software on the referring doctor's computer. Therefore, there is one less application to purchase, install, and maintain on the doctor's computer.

b) Many doctors are very familiar with e-mail, which enables faster learning curves and lower training and education costs.

c) E-mail messages inherently consist of multimedia contents; this makes them ideal to transport multimedia clinical results.

d) E-mail systems are highly integrated with Web browser technology, which opens the way for a closer, smoother integration of clinical results with more sophisticated and specialized applications for their processing and analysis.

2) Java has conquered more and more ground as a prime software development tool in general and in PACS and teleradiology applications in particular. Dr. DeJarnette declares that Java "should be of no interest to a PACS or teleradiology user or purchaser, other than if the use of this technology results in a more reliable, more maintainable, higher performing, and/or lower cost system." I totally agree, but I maintain that Java has tremendous potential to provide exactly that: a lower cost system. Here are some of the reasons in the specific context of (on-call) teleradiology:

a) Java provides a framework for remote, platform-independent application use. With a pure Java application, or applet, no application software whatsoever is required on the end-user's station to run the application. Everything needed is downloaded automatically and transparently from the application Web server. This means that any doctor can operate a Java-based image viewing application from his/her home PC (using any make and model) as is! No need to purchase, install, or maintain any application software on the home PC, since none is required in the first place!

It can be argued that some non-Java applications may be less expensive than the Java-based ones because the image viewing application itself is very cheap, or even provided for free ("freeware"), whereas the cost for a Java-based system is in purchasing the server. I contend, however, that even in the case of freeware, these applications end up being more costly than the Java-based ones due to the associated costs of installation and maintenance. The latter must not be underestimated. General statistics in the United States show that an organization spends approximately $1500 annually on the maintenance of each of its PCs. This amount is comparable to the original purchase price of the PC itself. Though similar statistics for the consumer market is harder to obtain, many of us have witnessed first-hand the ramifications of an operating system upgrade, or even the installation of new application software, on the performance and viability of other applications. With Java, these concerns and costs are largely eliminated.

b) Typically, Java is invoked from a standard Web browser. With its widely familiar interface, the Web browser thus becomes the access mechanism for the doctor onto the teleradiology application. Again, this translates into shorter learning curves and lower training costs.

c) Java is easy to integrate with other Web-based technologies and systems. A Java-based teleradiology application, providing clinical image viewing capabilities and tools, can easily be integrated into a Web-based clinical results browser (e.g., patient charts viewer) or an electronic medical records (EMR) system.

These two examples demonstrate that the low-cost nature of Web technologies is not only an "article of faith," as suggested by Dr. DeJarnette. It is based on actual facts.

DICOM and Web technologies: A contradiction in terms?

Dr. DeJarnette makes a distinction between DICOM-based teleradiology systems and Web-based teleradiology systems. In fact, he even provides a comparison chart between the two. But, are these two technologies really mutually exclusive? The answer is definitely no. There is no reason why these two technologies cannot be mixed. In fact, some applications in the market today offer exactly this mix. The concept is straightforward: enjoy the benefits of both technologies.

How can DICOM and the Web work in concert to provide the best of all worlds for teleradiology? Here is a simple recipe:

1) The end-user (e.g., a radiologist on call using his home desktop PC) opens a Web browser exactly as if surfing the Web.

2) The user then selects an IP address (or jumps to a predefined bookmark) of a teleradiology Web server.

3) The user's Web browser and the server communicate via the HTTP protocol to initiate the communications.

4) Once the user is authenticated, control is transferred to a Java applet, which goes from the server to the browser.

5) Through the Java applet, the user can query any DICOM device on the network and retrieve images from it to the browser. The DICOM commands are encapsulated in the Web protocol for the communications between the client and the Web server. Direct DICOM communications prevail between the Web server (which can also act a DICOM device) and the other DICOM devices on the LAN.

Pulling, pushing, and compressing on the Web

Another disadvantage of Web technology, according to Dr. DeJarnette, is its lack of support for the push model: the traditional teleradiology model. But does the Web really preclude the use of the push model? The answer is clearly no. First, we already mentioned e-mail as an obvious example of Web technologies. E-mail is also a push-type of technology. As mentioned above, I see the main application for e-mail in the distribution of clinical results to referring physicians. E-mail remains limited in its application for teleradiology, i.e., remote interpretation of diagnostic images.

But e-mail is not the only push-type Web technology. There are other technologies available that allow pushing data to the user as opposed to having them pull this data. In fact, some commercial Web-based systems make use of such technologies in the context of PACS and teleradiology applications today. The main benefit of the push model is in cutting short the waiting time for the data to download. When used over slow communications (e.g., plain old telephone system), this model can make a significant difference, especially for the radiologist on call. But, since it is available on both Web and non-Web systems, this cannot be taken as a differentiation between the two.

When dealing with slow communications, the push model is only one option. Another important factor is compression. Dr. DeJarnette claims that Web systems generally offer only JPEG compression, which has been known to be inferior to Wavelet compression techniques in handling clinical images. While this may be the case, since JPEG is so commonly used on Web pages, it must be noted that nothing on the Web prevents the use of Wavelet compression. And again, some commercial Web systems already offer Wavelet compression.

Moreover, by making use of the concept of progressive compression prevalent on Web pages, some Web systems actually reduce the user's waiting times even more. The concept is simple and powerful: images are first loaded and displayed in crude, low-resolution versions, which come up very quickly. The images then build up their resolution to show more and more of the fine details, while the user already has access to most of the data.

Web without Internet?

Highlighting the benefits of the Internet, Dr. DeJarnette mentions that non-Web systems can make use of this global network. He also mentions that the Internet still has its drawbacks, including bandwidth and security questions. I agree with him on both points. I believe, however, that the complete picture must be examined in context.

As explained above, I believe a primary distinction must be made between the Internet and the Web, i.e., Web technologies. We must understand that there are four generally possible combinations for using the Internet and/or Web technologies (Table 1). In other words, some non-Web systems can run over the Internet. Most Web-based systems can also run over any communications that support TCP/IP. In particular, the application of Web-based systems over LAN is common, and is referred to as an "intranet application."

Intranet applications offer the benefits of Web technologies without the concerns associated with the public Internet. Since the application runs on a dedicated, private network, bandwidth and security are no longer serious hurdles. Here are some benefits to Web technologies running without the Internet:

* Open, standards-based architecture

* Common user interface (e.g., through the Web browser)

* Platform independence (e.g., through Java)

* Excellent scalability (consider the basic model of the Web)

One additional major benefit of Web technologies is their easy expandability onto the Internet. Even in cases in which the Internet cannot be used today, e.g., due to bandwidth limitations or security concerns, it would be unwise to preclude the use of the Internet in the future: when bandwidth becomes more readily available, or when sufficient security measures are in place.

In fact, on both the bandwidth and security questions, it would be extremely shortsighted to ignore the development of the market in recent years. High-speed Internet connections, through cable modems and digital subscriber lines (DSL) are reaching more and more of the consumer market, while encryption techniques and their derivative virtual private networks offer excellent privacy and data protection over the Internet. Far from having PACS or teleradiology in mind, these prominent market trends pave the road for more use of the Internet for these applications. Since they were originally designed to enable applications running on the Internet, Web technologies provide the best guarantee for smoothly and simply migrating new applications to the Internet. An application can begin by running on an intranet and then be expanded to the Internet when the need arises.

PACS, teleradiology, EMR and the promise of Web integration

For many years, PACS and teleradiology have been considered two separate applications. I contend that this distinction becomes unnecessary from a technology standpoint with the introduction of Web technologies. My argument is simple: no one can deny the need for PACS to expand its reach toward the larger community of referring physicians. As I argued above, this is exactly where "traditional" PACS failed. Web-based PACS, on the other hand, offer the first economically viable solution for this need. While most referring physicians will settle for what Dr. DeJarnette calls "the simple Web-server-based teleradiology model" (i.e., sending reports and reference-quality images using HTML pages), in up to 30% of cases (e.g., for neurosurgery and orthopedic applications), the image viewing capabilities required are not any different than for teleradiology. Luckily, Web technologies can address well the needs in both cases. If the physician at the end of the line happens to be a radiologist, we're back to the teleradiology application. In other words, with a Web-based PACS, which addresses the needs of the referring physician community, teleradiology becomes almost an inherent side effect. There is no need to invest in a separate application!

This consolidation of PACS and teleradiology has more far-reaching implications. In my opinion, it is only one facet of the much larger potential of the Web for integrating different applications. Perhaps the most interesting of these in the context of PACS and teleradiology is the potential for creating comprehensive EMRs. Long sought for, and still almost non-existent, the EMR can be thought of as a global access engine, available everywhere, and providing a uniform interface to comprehensive, multimedia clinical information on a patient. It is not hard to see the potential of the Web browser in this context. In fact, some precursors of this integration are already available. Interestingly enough, it is via Web-based applications that clinical images can be accessed from some radiology information systems (RIS) and other clinical information systems (CIS) today. Simply put, the Web provides common grounds for the unification of these, previously disparate applications.

Choosing between Web- and non-Web-based systems

Dr. DeJarnette says that "in order to decide whether or not to choose a Web-based system, it is important to understand the details of the application. In the end, the choice should come down to: which system addresses identified needs, as well as anticipated future needs, at the lowest cost." I could not agree more. I contend, however, that in most cases the Web system will in fact offer exactly that: better handling of present and future needs at a lower cost.

Conclusion

Dr. DeJarnette concludes: "Al-though interesting and well hyped, the use of Web technology should not be the sole basis for deciding on the purchase of PACS and teleradiology systems. The technology employed in a medical imaging product should be one of many factors considered as part of the purchasing decision." I agree completely. He then continues: "In PACS and teleradiology applications, Web technology has no de-facto advantage." As I demonstrated above, this is definitely not true. Web technology offers some very clear and prominent advantages. With these in mind, I would recommend to anyone interested in PACS or teleradiology, at a minimum, to consider these technologies seriously and without prejudice. Of course, no one can contest Dr. DeJarnette's bottom line: "As with any system purchase, no decision should be made without thoughtful consideration of the application itself and how well it meets the needs of the radiology department." But nothing in this statement can diminish the huge promise of Web technologies for clinical image management. AR