On Point: Cardiac Toolbox
By Ernest Garcia
I were to rewrite the script to The Graduate today, it would read
something like this:
Mr. McGuire: I just want to say one word to you—just one word.
Ben: Yes sir.
Mr. McGuire: Are you listening?
Ben: Yes, I am.
Mr. McGuire: Software.
Ben: Exactly how do you mean?
Mr. McGuire: There’s a great future in software. Think about it. Will you think about it?
Ben: Yes, I will.
Mr. McGuire: Shh! Enough said. That’s a deal.
I started developing clinical application software some four years after The Graduate was filmed. While I’ve maintained this interest in the past 20 years, my journey has expanded far beyond the boundaries of software into a future where the university and the private sector need each other to make patients healthier.
In my early days of software development, the cost of computer hardware was astronomical, and software was something that was mostly included for free. Today, for approximately $1,500, one can buy a computer with a 3-GHz processor and a 250-gigabyte hard drive. Technology developer Ray Kurzweil predicts that, by the year 2020, a $1,000 computer will have the processing speed of the human brain, and by 2060, the speed of all the brains of the human race combined. Whether these phenomenal machines will perform any useful function, in medicine or otherwise, will depend on the software that drives them.
In the 1970s, my field of nuclear medicine developed the first digital diagnostic imaging modality. At the time, the scientists using the imaging system either wrote their own software to process and interpret the digital images, or they borrowed unlicensed software. Many scientists contributed these software programs voluntarily to the system manufacturers, who included them as part of a user library accessible to clinicians everywhere to use in evaluating their patients. That process was a precursor to today’s “open-source” programs, if you will, but in medicine, open-source had and still has some drawbacks.
This software was usually without documentation or validation and was cumbersome to use. Those of us who contributed software to these user libraries found ourselves spending long hours on the telephone providing support for other users. It became clear to me that this approach was actually hurting the growth of our field and that clinicians should be willing to pay for software that was well designed, particularly if it increased diagnostic accuracy and increased efficiency. Easier said than done.
Humans, it turns out, are reluctant to pay for things they have previously obtained for free. In the early 1980s, Emory researchers in nuclear medicine partnered with a nuclear pharmaceutical company to develop our first commercial program for myocardial perfusion, a procedure that evaluates the heart’s function and blood flow. New England Nuclear funded the development of the software to quantify myocardial perfusion using their drug as the imaging agent as well as a clinical trial to validate it. At the time, clinicians were having a difficult time interpreting these studies. Marketing research showed that, if properly interpreted, the use of this test could grow significantly.
So the drug company provided the imaging agent and the funding, we provided the software. So far, so good. But a complication soon arose in our relationship, which seems to often occur with this type of arrangement. The company argued that since they funded the software development, they should own the software, or, at a minimum, participate in the distribution of royalties. We successfully argued that we would be developing the intellectual property and thus that our institution should own it. We further pointed out that our software would significantly enhance the sale of their drug. If they wanted to participate in our royalties, we should participate in their increased revenues due to our software, we said. Finally we took a “deal or no deal” attitude, and the company agreed to our terms.
Their decision proved to be wise. Sales of these perfusion agents have grown tremendously over the past 20 years. They were used in approximately 12 million patient procedures last year. We did recognize that the companies that fund our work deserve some competitive advantage over their competitors who do not. To address this issue, we offered the pharmaceutical company exclusivity on the intellectual property rights for 12 to 18 months.
Over the past 20 years, our Emory R&D team has developed a vast set of tools for evaluating cardiac images known commercially as the Emory Cardiac Toolbox™. We have developed many of these tools in collaboration with other institutions, including Georgia Tech. A source of tremendous personal satisfaction is that more than 10,000 licenses exist all over the world for some of the tools included in the toolbox. As the business side grew and regulatory issues matured, we found it necessary to spin off a start-up medical device company called Syntermed, Inc. (www.Syntermed.com), a company co-owned in part by Emory and Georgia Tech. In addition to sublicensing the toolbox to large companies like General Electric, Siemens, and Philips, Syntermed provides a conduit for selling directly to the end-user who buys a standard PC and loads it with our software, thus opening up many new applications.
Without help from private investment and without the ability to set up a private company to market the software, we would have failed to have given many people the cardiac care they needed. These 20 years of commercial development at Emory have been the direct result of 20 years of funded scientific research. Money from the commercialization of our product has immediately gone to further new research. In our department, our general approach has been to (1) use NIH grants to fund basic scientific research, (2) use industry funding to translate our basic research to commercial software products, (3) use royalty funds from the software products to supplement salaries to retain faculty and help fund our lab, and (4) use royalty funding of the lab as seed funds to generate preliminary data for NIH grants to perform basic research. Thus back to step 1, and the funding cycle is complete.
I believe Cardiac Toolbox is a classic example of a best-case partnership between academic researchers and private industry. Although our path was sometimes rocky, we were able to develop a product, keeping patients and their needs as the focus and successfully marketing the product and making it profitable. There are reasons for that success. As time-consuming as the conflict-of-interest process can be, it is important that we examine our motives before we begin the process of turning a research endeavor into a profit-producing one. Our research, development, and validations should grow out of our integrity as scientists. Our first responsibility is to the patients who receive evaluations with these tools.
In Mr. McGuire’s words, “Enough said. That’s a deal.”
Ernest Garcia is a professor of radiology at Emory School of Medicine.