BBI Contributing Writer

CHICAGO, Illinois There can be little argument that technological advances are major drivers of growth in sales of products and services to the radiology community. To be sure, end-user demand must exist, but the proven clinical value of radiological science assures the meeting of that criterion.

Technological advances again were most apparent on the McCormick Place exhibit floors at the 2003 annual meeting of the Radiological Society of North America (RSNA; Oak Brook, Illinois), despite a slight reduction in the number of exhibitor pre-registrants in comparison to 2002's gathering here. Foot traffic was brisk in the sprawling exhibit area of the McCormick Place convention center along the Lake Michigan shoreline, despite an imposing schedule of scientific and clinical sessions attended by more than 20,000 prospective radiologist customers.

The quality of the exhibits at the meeting, held in its usual time slot the week after Thanksgiving, was typically first-rate. Summarizing the offerings of the 600-plus exhibiting firms on hand is an impossible challenge, but several compelling technological and commercial advances are worthy of note, as are a few unique entries.

Computer-aided detection/diagnosis

Computer-aided detection (CAD) has come of age, as the initial turf-trespassing concerns of the clinical community evaporate in the face of proven clinical values of various computer aids. More than 40 exhibitors highlighted their CAD offerings, not counting the numerous presentations by academic and clinical groups devising home-grown CAD solutions a veritable cottage industry of software aids to image-based detection and diagnosis. The surest sign of acceptance is the adoption of various CAD capabilities as integrated features of imaging systems offered by the industry leaders. But the CAD specialty houses continue to innovate, one example being R2 Technology (Sunnyvale, California). R2 is moving from CAD's initial proving ground, mammography, to lung screening via computed tomography (CT), virtual colonoscopy and CAD integration onto image management system networks (picture archiving and communications systems, or PACS).

Having demonstrated its mettle in mammography, lung screening appears to be the next CAD frontier for R2. (See the RSNA Notebook in the January 2004 issue of The BBI Newsletter for more details.) Other CAD specialty firms including CADx Systems (Beavercreek, Ohio) and Deus Technologies (Rockville, Maryland) also are active in this space, the latter in partnership with GE Medical Systems (Waukesha, Wisconsin). Another specialty player, iCAD (Nashua, New Hampshire), announced its merger with CADx during the RSNA meeting. The two companies together claim more than 300 installed CAD systems.

3-D imaging

Another image-processing methodology, three-dimensional (3-D) imaging, also has established a strong foothold in medical imaging. Nearly 50 exhibiting firms included 3-D processing capability as a feature of their imaging systems or as standalone software solutions. 3-D first emerged when computed tomography (CT) burst on the scene in the 1970s but found little traction at that time. Radiologists were content to examine one CT "slice" at a time, since each dataset was manageable in size. Modern CT and magnetic resonance imaging (MRI) scanners have changed all that. For example, the newest multi-detector CT (MDCT) scanners produce as many as 5,000 0.5-mm slices per study; to fully exploit the entire imaging study essentially requires 3-D reconstruction and visualization.

Beyond 3-D integration into CT and MRI scanners, 3-D has begun (albeit slowly) to be integrated into PACS, which yields 3-D capabilities across all the imaging modalities supported by a PACS installation. The applications are widespread today: ultrasound, CT, virtual colonoscopy, vascular studies, orthopedics and neurology.

The earliest innovator in the field was Vital Images (Plymouth, Minnesota). Now, all scanner producers offer 3-D workstations to complement their CT and MRI systems. Other standalone workstations and software solutions are offered by specialty firms, including TeraRecon (San Mateo, California), Voxar (Edinburgh, Scotland) and Able Software (Lexington, Massachusetts). Viatronix (Stony Brook, New York) focuses on vascular imaging, while several suppliers including GE Medical, Medison America (Cypress, California) and Philips Medical Systems (North Andover, Massachusetts) have recently developed ultrasound scanners that render 3-D images in real time, an implementation that some believe will become a high-end standard of care in ultrasound.

The clear benefits of 3-D do not come without cost, however. Radiologists viewing 3-D images require training in addition to their usual skill development, a burden to be reckoned with in large teaching institutions, not to mention community hospitals and small radiology practices. But 3-D's benefits will force change as surgeons, referring physicians and patients themselves come to expect review of an intuitive volumetric image.

CT and PET

These two very different imaging modalities are considered together because both are experiencing dramatic increases in popularity, and their blend has found important clinical applications combining anatomical and metabolic imaging.

MDCT, now about five years old, is invigorating the CT market with placements of new 6-slice and 16-slice machines. In fact, most new placements are multi-slice systems. Apparently, users appreciate the ability to combine the advantages of high resolution and fast scanning in a single system, while 3-D processing and CAD enable rudimentary management of the resulting large datasets.

The leading suppliers of these machines as one would expect are the leading CT scanner makers: GE Medical, Philips Medical, Siemens Medical Solutions (Malvern, Pennsylvania) and Toshiba America Medical Systems (Tustin, California).

Three challenges confront this arena. First, image management of the resulting datasets is problematic, and needed progress by PACS vendors is slow in coming. Second, the industry would like to infringe on the territory of cardiac catheterization, but the consistency of CT results is not yet adequate to that task. Third, some are concerned about radiation exposures from multi-slice CT scans and from the large number of single-slice scans being performed.

Functional imaging, and in particular positron emission tomography (PET), is hot. Sixteen firms with PET offerings displayed their wares at RSNA. In particular, Siemens Medical Solutions, Philips Medical Systems and CTI Molecular Imaging (Knoxville, Tennessee) were exhibiting new systems offering faster scans and more rapid image reconstructions. Also, new radiotracers for bone imaging were introduced, as were systems to produce radiopharmaceuticals on site to alleviate a significant PET shortcoming.

New, smaller and less-expensive PET scanners for small animals also were introduced for use by companies that are active in drug development, an important emerging application for PET that ultimately will migrate to the clinic for human clinical trials.

An especially important trend is the integration of PET and CT scanners to assess form and function simultaneously. (Several clinical applications of this new imaging modality were addressed in the January issue of BBI in the first part of this two-part series.) CPS Innovations (Knoxville, Tennessee), a supplier to both CTI Molecular Imaging and Siemens Medical Solutions, announced a new-generation PET/CT scanner with three times greater spatial resolution than existing scanners. The new technology, dubbed HI-REZ, is integrated with MDCT, including the latest generation of 16-slice and 6-slice CT scanners.

Image management

Purveyors of information technology products occupied what seemed to be a majority of exhibitor floor space, despite the numerous and expansive walk-in accommodations offered by the leading scanner makers. Consider the numbers: more than 100 PACS suppliers, almost 70 vendors of radiology information systems (RIS), and about 30 producers of hospital information systems (HIS). Some of the same names were attached to two or three of these groups. This emphasis on image management in particular and information management generally is inevitable, as there is simply too much data to collect, display, integrate, store, report and archive to even consider a non-systematic solution in any real healthcare enterprise. Consequently, PACS spending by healthcare providers is forecast to grow at about 11% per annum in the near term, while these same providers are forecast to increase spending only about 3% to 4% on information technology generally.

We are seeing RIS and HIS suppliers poaching on the PACS space and, surprisingly, new PACS vendors are entering the fray to challenge the established scanner makers and film makers that have dominated the PACS scene in recent years. As a result, a dizzying array of information products were on display, leading to some quizzical looks by attending radiologists. These vendors run the risk of stimulating user confusion in making product choices, potentially leading to market stagnation.

Another interesting saga is the continuing effort to integrate PACS with RIS capabilities and to a lesser extent with HIS offerings. Integration is an incredibly obvious need, yet it has proven to be an elusive goal over the last decade. Simplistically, the problem is that PACS uses the DICOM interface standard to communicate, while RIS uses HL7. But the problem goes beyond the technical to the inability of RIS and PACS vendors to collaborate in meaningful ways. The answer may lie in appropriate single-vendor solutions or in the use of Internet protocol (HTTP) as arbiter, or both. Several reported acquisitions are emblematic of the need to combine forces: RIS supplier VitalWorks (Ridgefield, Connecticut) announced an agreement to acquire PACS firm Amicas (Boston, Massachusetts), while Eastman Kodak (Rochester, New York) plans to acquire PACS maker Algotec Systems (Raanana, Israel).

Additional observations

While not new initiatives, several continuing trends were widely evident on the RSNA exhibit floor. Digital mammography is even more pervasive than before. More high-field (3-Tesla) MRI machines were in evidence. Both direct radiography (DR) and computed radiography (CR) methods continue their advance to digitize X-ray imaging, which remains the majority method of generating clinical images today; CR seems to be the most popular alternative by virtue of the number of systems exhibited at the meeting.

Tools for the interventional radiologist were exhibited in greater numbers than in prior years, perhaps an inevitable reaction to the widespread interest in less-invasive procedures by clinicians, payers and patients.

The acquisition of Amersham (London) by General Electric (Fairfield, Connecticut), which is moving through the numerous steps necessary before completion, is intriguing. The most obvious synergy is GE's stated interests in functional imaging and molecular medicine coupled with Amersham's offerings in radiopharmaceuticals. But one also wonders if GE has more expansive interests in in vitro clinical diagnostics generally.

RSNA experimented with a small section of the exhibit floor dedicated to a Mobile Computing Pavilion. About 20 information technology suppliers made 30-minute presentations to audiences of modest size over a five-day period on topics including wireless technologies, tablet PC applications, secure communications, personal digital assistant (PDA) applications, and point-of-care information access. The process was surely technology-push as opposed to demand-pull. Attending radiologists were curious, but few had direct experience with mobile computing within their institutions. PDA use by radiologists today appears to mimic applications exploited by the general population so far.

MaxxVision (Gainesville, Florida) exhibited a clever, hand-held device that subjects selected portions of a mammography film to digital enhancement and re-display on a compact LCD monitor. The so-called iView is essentially a replacement for the radiologist's optical eyeglass commonly employed in reading mammography films. The vendor claims image enhancement features of digital mammography at a fraction ($34,500) of its cost. According to the supplier's product literature, a small clinical study of the device was conducted at the H. Lee Moffitt Cancer Center and Research Institute (Tampa, Florida) involving 114 cases carried out over eight weeks with five radiologists. The study demonstrated 11% reader performance improvement in cancer detection and 4% reduction in recalls for negative cases.

More LCD displays were evident at the RSNA meeting than in prior years. The CRT/LCD comparisons now appear to be tipping in favor of LCDs, which offer high-quality color rendition (required by some imaging modalities), adequate resolution, low display noise, acceptable gray-scale presentation for many applications, and greatly reduced acquisition and service costs.

Digital mammography supplier Hologic (Bedford, Massachusetts) showed work in progress on a device it characterizes as a full-field digital breast tomosynthesis research system. The system can acquire 1 mm image slices and reconstruct them in a 3-D presentation. The company intends that the technology will reduce or eliminate the appearance of overlapping tissue, which can hide breast lesions. Application for FDA approval has not yet been made.

Alara (Hayward, California) demonstrated its new CRystalView desktop computed radiography (CR) system. The device employs storage phosphor technology to create high-quality digital images. Different from contemporary CR systems that are large and cost hundreds of thousands of dollars, the Alara entry is about the size of a large laser printer and is priced at about $75,000.