According to Research and Markets, the global market for medical imaging devices is estimated to be worth $30.2 billion this year (2013) and to increase to $32.3 billion next year. By 2020, this market will be worth a massive $49 billion, growing at an average annual rate of 7%.
Well-reported demographic pressures on healthcare systems – notably, aging populations in many Western economies, rising levels of ill health and lifestyle issues such as obesity – are set to be key drivers of this growth. As demand for healthcare grows (not least in the US, with the introduction of the Affordable Care Act), so too will demand for diagnosis and imaging.
Other generic factors contributing to growth in this market include demand for, and increased interest in, telemedicine-based approaches to treatment and diagnosis. This covers actual telemedicine for patients in the home, as well as, in the radiology and imaging arena, the ability to interpret images anywhere and any time, and the ways in which information and images are shared and received.
Key to this growth is imaging software. According to research by MarketsandMarkets, the growth of the medical image analysis market is contributing to (and being fed by) strong growth in the stand-alone software market. In fact, this market was found to be accelerating at a faster pace than the integrated software market.
The market researcher calculated the medical imaging analysis market to be worth $1.7 billion in 2012, and growing at a rate of 7.2% (in line with the Research and Markets’ findings), meaning that it will be worth $2.4 billion by 2017.
A lot has changed in the past 20 years, according to Dr Eliot Siegel, professor in the Department of Diagnostic Imaging at Baltimore’s University of Maryland School of Medicine, and chief of radiology and nuclear medicine at the US city’s VA Maryland Healthcare System. And there is a lot of change to come.
"One thing that has been interesting to me is that, 20 years ago, the Baltimore VA Medical Center was the first in the US, and we think in the world, to move to a filmless radiology department and hospital," Siegel says. "While we were at the cutting edge of specifying the primary parameters 20 years ago, now we are not only the first filmless department but, as a result, one of the most antiquated. So, we have been going through the process of trying to create a new specification and infrastructure.
"So much has changed," he continues. "When we went filmless, we were ahead of the idea of a patient electronic record system or electronic medical record; now, it is more integrated. We went digital using a proprietary network since we were in the pre-ethernet era in 1993, and the internet was in its infancy. I recall the cost of purchasing a 1TB archive was huge; now you can buy one for $40, but ours cost around $800,000. Network software has become much less proprietary and much less expensive as well.
"The first-generation picture archiving and communication system (PACS) image interpretation workflow was based on that PACS system itself," adds Siegel. "Our current workflow is based on our reporting system that uses speech recognition, and, consequently, we are no longer constrained to using the PACS database to mark studies as in the process of being dictated or having been dictated, but instead can use the reporting system for that function. This frees us up to use different types of specialty or ‘best-of-breed’ workstations for our reporting. I, for example, use a PET/CT workstation for PET studies, an advanced visualisation (3D) workstation for CT reporting and a separate nuclear medicine workstation for those studies. We also use a single reporting system to integrate across multiple electronic medical record/radiology information systems from different facilities."
File sharing
Innovations coming over the horizon, Siegel predicts, include radiology software that not only records when studies are completed, reported and shared with referring physicians, but also provides feedback on whether any of the referring physicians subsequently acted on the recommendations.
"Image sharing is becoming a big issue," says Siegel. "In 1993, image sharing with our PACS involved printing out the images and physically sharing them on film, then moved on to burning CDs and DVDs that, at the time, were seen as major advances because of the amount of data and number of images they could store.
"But the problem with CDs and DVDs is that they will have been written by the manufacturer, often in a different format or with a different viewer. Consequently, you may have to install that viewer or software, but the computer you are using in the hospital won’t give you administrative access, so the physician cannot install the software, which is frustrating.
"Another challenge has been addressing how to use the internet and modern web technology within radiology. People are often surprised that they cannot send images digitally from one hospital to another; but it has been difficult to accomplish this, not least because of the issue of patient permission."
Patient health records
Siegel cites the Radiological Society of North America’s Image Share network pilots (see ‘The Image Share network’) as potentially breaking significant new ground.
"The University of Maryland and the other pilot academic centres have been developing software that will allow images to be transferred electronically to a patient health record (PHR)," explains Siegel. "So, a patient signing on and creating an account with a PHR can get their MRI scan securely transferred to their record via the cloud service and then, using a password, their physician is also able to access their account. So far, it has worked out well.
"But a lot of the time and effort that has gone into it has been focused on developing a system that is as secure and private as possible. There was a lot of discussion about how many digits patients’ passwords needed to be. The irony is that a third of the patients who have taken part in the pilots have almost immediately gone and shared their images with friends and family anyway.
"But, what the Image Share network has shown is that the next generation of PACS will be able to transfer images digitally and securely from one faculty to another. We have managed to solve one of the key challenges: access and security."
Search parameters
There is a high likelihood of seeing software come to market that can produce better analytics and actively assist the radiologist in interpretation, including by offering better summarising of information and reports. Software is also being used proactively to mine current and previous radiology reports, to allow radiologists to better evaluate patient trends, as are better search tools to allow reports to be searched for by specific words and concepts and correlate them back to said reports.
"One of the biggest challenges we want to address with the next generation of PACS is the communication of important or unexpected findings," points out Siegel. "The American College of Radiology (ACR) has protocols in place for the communication of critical or serious sightings – by phone or in person – but for lesser (but nevertheless important) findings, there can sometimes be a danger that something does not get followed up. For example, a lung nodule on a patient being evaluated pre-operative by a hospital can be subtle and not necessarily life-threatening, and will go into the radiology report, but it can end up being lost or ignored.
"The next-generation system is going to take that relatively manual and time-consuming process of following up, and ensure it happens digitally. So, I want my next-generation speech-recognition system to act as if it understands my reporting, be able to highlight important findings within my report, and then, if that information needs to be sent to the referring physician, provide a one-click method to order the recommended follow-up study, and, if it is not, generate automatic reminders. This currently happens manually at our facility, but it requires a good deal of time and personnel resources.
"The next-generation systems will be more intelligent, and pass on information to the radiologist," Siegel continues. "For example, I’ve been privileged to work with IBM on one of its artificial intelligence (AI) projects, which aims to create computers that are able to answer open-domain questions.
"Future software will utilise AI to help diagnose medical conditions, provide a concise summary of the electronic medical reports, and ensure better integration from multiple sources within the medical enterprise. Smart algorithms, too, can change how radiological data is mined and discovered, and in the process help radiologists make better connections and interpretations."
