|
|||||||||||||||
|
Home
|
Time to Change — Managing
Image Volume Through Advanced Visualization The evolution of diagnostic imaging technology has brought with it numerous benefits, as well as new challenges. Perhaps none is more apparent than the overwhelming amount of data generated by increasingly sophisticated equipment. This particular issue was readily identified two years ago, and the Mayo Clinic’s Bradley J. Erickson, MD, PhD, tagged it with a forceful phrase that succinctly defined the problem. Heading into the 2005 meeting of the Society for Computer Applications in Radiology (now the Society for Imaging Informatics in Medicine)—an event that examined the past, present, and future of medical imaging—Erickson commented that one of the largest emerging issues in radiology involved what he called image overload. Image overload has become a reality as the ever-increasing number of images has overwhelmed radiologists. As observers have noted, image generation has surpassed workforce capabilities. So the problem is how to handle the flood of available information in such a way as to make workflow more manageable. Interpreting exams through 3D reconstructions instead of individual axial slices is one tool the radiology community can use to combat image overload. More Images, Less Time “The development of MDCT brought vast improvements in image quality that was coupled with a decrease in the time needed to acquire the datasets,” says Gordon J. Harris, PhD, director of the 3D Imaging Service and Radiology Computer Aided Diagnostics Laboratory at Massachusetts General Hospital (MGH). “As such, we now have much broader body coverage in a shorter amount of time,” Harris adds. “MDCT enables us to scan more rapidly and with thinner slices,” adds Eliot L. Siegel, MD, professor and vice chairman of the University of Maryland School of Medicine department of diagnostic radiology and chief of radiology and nuclear medicine for the VA Maryland Healthcare System. Improved image quality enhances diagnostic capabilities and enables more accurate treatment planning, as well as generates increasingly larger volumes of data. The larger number of acquired images for each case turns interpretation into a more complicated and time-consuming process. Also compounding the problem is the increased number of applications now available (eg, CT angiography [CTA] and MR angiography [MRA]) and their increased usage. In particular, CT utilization has substantially increased. “For instance, in our emergency room, it has risen from about 60 in every 1,000 patients in 1998 to nearly one in every three patients today,” Siegel explains. Further, the complexity of studies has increased, as functional information is now coupled with anatomic information. “Whether we do a head study, an abdominal CT, or multiphase studies, increased scanner speed enables us to extract significantly more information, especially for contrast-enhanced studies, and generates a much greater number of images,” says Siegel. For example, Siegel points out that a typical abdominal and pelvic CT used to produce approximately 60 to 70 images, but MDCT has increased that number. “With the advent of thinner sections, the average number of images we obtain ranges between 700 and 900,” he says. He adds that images produced by a CTA study could reach 2,000 or more, while cardiac studies routinely reach anywhere from 6,000 to 8,000 images. “For the new generation of dual-source CT scanners, cardiac imaging studies can create as many as 15,000 images,” he points out. A radiologist attempting to go through each of those slices individually would be restricted to a small number of cases per day, as it would be impossible to view all the images without a major reduction in productivity. All these elements created the pressing need for effective management of image volume. As the number of images will only increase, healthcare professionals had to find ways to make the process more efficient. Part of the solution comes from 3D renderings and advanced visualization, which can reduce the number of images to a manageable level. The realm of advanced visualization includes hardware innovations, including workstations with larger memory and graphics capabilities, and more complex software, as well as the human element entailing defined responsibilities and appropriate training. Challenges Still, there’s a distinct lack of uniformity. Adaptation is conducted in variable fashion. “Different places have different philosophies about who should do the processing,” says Harris. “At some places, radiologists are used to having the 3D work done for them. At other places, radiologists do their own 3D processing or may have some of the scanning technologists do part-time 3D processing.” Further, the challenges can be daunting. Harris notes that many facilities struggle to develop a viable 3D laboratory. One major issue is integrating all of a department’s workflow, clinical protocols, billing, and compliance. Department heads also need to determine who will do the 3D work. Once that is decided, training becomes an issue. “It’s an inefficient use of time for radiologists to do postprocessing, but if you’re going to have someone else do it, then you need to train them on how best to produce the most useful images,” says Harris. Massachusetts Model The lab has implemented hardware and software technology including five GE Advantage Workstations, five Vital Images Vitrea workstations, a TeraRecon Aquarius Workstation and AquariusNET, two Voxar workstations, two Linux computers with MedX/VolumePro for functional MRI, Mirada Fusion7D server workstation for image registration, three GE LOGIQworks for 3D reconstruction of ultrasound, Agfa PACS service station, and RadWorks PACS. Along with Harris, the staff includes an operations manager, administrator/billing coordinator, two technical staff members, five 3D technologists, two image analysis specialists, and a 3D ultrasound technologist. The lab currently processes more than 2,500 cases each month, with approximately 120 exams processed each day (which includes 30 MRA, 30 CTA, 30 other CT and MR exams, and 30 ultrasounds). One half of the cases fall in the realm of neuroradiology, while 20% are in vascular radiology. The remaining 30% are comprised of gastrointestinal/genitourinary, chest, breast, bone, and pediatric radiology. The facts and figures associated with MGH’s 3D lab can seem intimidating to an organization getting its toes wet in the brave new world of advanced visualization. The questions and problems about how to set up a 3D lab, how to manage the workflow, who will do the work, and how to appropriately train staff members may seem difficult to answer, even insurmountable. To help, MGH is working to provide 3D lab services for other hospitals and imaging centers, according to Harris. These facilities would benefit from the skills and protocols that MGH has already developed. “With the help of some vendors, we’re setting up a network that will allow us to connect with multiple outside sites,” he explains. “They can send us images to process, and we’ll produce the 3D images and send them back through the gateway to their PACS. It’s more cost-effective and less complex than trying to create a whole system.” The lab can provide clinical computer visualization for routine clinical use, as well as offer 3D on request, faster turnaround, full-time technologists, and full integration with hospital PACS and information and billing systems. Thinning the Field “Vendors have started to realize that the best way to make images immediately accessible to those who need access is through the thin client environment, where images are rendered on a server rather than locally through a workstation,” says Siegel. This means someone can use a PC to open a window into the memory of the vendor’s visualization server, explains Siegel. “You wouldn’t have to wait for all 800 to 1,000 images of a CT of the abdomen or pelvis, for example, to be transferred to your computer before you can view images reconstructed in different planes. Instead, by opening a window into the visualization server’s memory, you can almost immediately view images in any plane or perspective that you want because the server will already have the images in its memory. The server typically has an extremely fast connection to imaging modalities and the hospital network,” he says. As such, the PC essentially becomes a controller for the manipulation of pixels existing in a visualization server’s memory. “All you need is a user authentification mechanism to be able to access, or time share, a central visualization server,” he says. Siegel envisions that this will replace the need for multiple special purpose workstations in a radiology department. “For instance, if I am reading PET/CT scans, I have to go to the PACS workstation to find out which study I need to read next, then to the CT workstation to review multiplanar CT images and compare them to a prior study, then to the nuclear medicine workstation to obtain the SUV [standardized uptake value] and thus each workstation has a subspecialized visualization, processing, or workflow function. But, in the future, there will be a thin-client environment that will have a single PC accessing multiple software programs operating on one or more remote servers performing the workflow, visualization, and image processing.” This will minimize the need for software and hardware upgrades at each client PC, which become the responsibility of the vendor. “Thanks to innovative work being performed by companies such as Vital Images, TeraRecon, Visage Imaging, and others on server side rendering for advanced visualization, we’re rapidly moving in a direction that would eliminate the need for special purpose workstations,” says Siegel. The advantages are significant, he believes. Not only would software only have to be upgraded at the central server, but multiple users could also share server resources. “The industry is starting to come back to the idea of a personal computer as a window into servers, or supercomputers, which will enable advanced visualization and a new generation of very sophisticated image processing to be done in a centralized fashion,” says Siegel. As such, a tremendous amount of computer power can be
placed in a central location within a radiology department or hospital,
and that power can be accessible and shared throughout an institution. What should ultimately happen is that the routine workstations utilized by radiologists, clinicians, and technologists will become advanced visualization workstations as we define them today. “The field isn’t quite there yet, but in only a matter of years, complete access to the full dataset with advanced visualization capabilities throughout the healthcare enterprise will be routine,” comments Siegel. Advanced Visualization
at VA Essentially, the healthcare system’s environment is one where radiologists and clinicians have full access, as long as they have credentials that allow secure access. “Orthopedic surgeons can plan surgery, see complex fractures, and look at orthopedic hardware. Neurosurgeons can look at aneurysms or complex anatomy of the spine, and they can plan surgery or look at postoperative complications. In general, the nonradiologists—at least as much as the radiologist—very much appreciate the ability to be able to look at the patient from different anatomic perspectives from 3D or advanced visualization that more closely simulate their perspective in the OR [operating room] or on physical examination,” Siegel says of the benefits. Making his own prognostication for the future, Harris believes 3D and advanced visualization will become standard practice in radiology departments. “In many ways, the places that already do it enjoy advantages that other places don’t. Also, I think that referring physicians will likely choose the places that can give them the best images for what they need.” — Dan Harvey is a freelance writer based in
Wilmington, Del., and a frequent contributor to Radiology Today.
|
|
3801 Schuylkill Rd • Spring City, PA 19475 Publishers of Radiology Today All rights reserved. |