Enterprise Visualization in the Pediatric Environment

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
The worst advanced visualization system in the world for a pediatric setting is one wherein the only way that referring physicians from across the organization can see 3D reconstructions of diagnostic images is by physically visiting a radiologist at his or her workstation.
Stuart Royal
“Believe it or not, this is still a far too common way of conducting business—this having to look over the radiologist’s shoulder and then ask for a view that’s rotated perhaps a little more this way, and then a little more that way, before ultimately getting just the desired view,” according to Stuart Royal, MD, MS, radiologist in chief at Children’s Hospital of Alabama in Birmingham, which boasts the fifth-largest pediatric radiology department in the nation. It does not help if the enterprise visualization system in question is based on a state-of-the-art RIS/PACS solution that distributes 3D images everywhere, but requires the actual processing of the 3D images—the reconstructions themselves—to occur at a thick-client workstation located in the radiology department. That was the way that 3D reconstructions were performed at Children’s Hospital until 2007, when the institution switched from workstation processing to server processing, a vastly more efficient and effective method. “It is now possible for our radiology customers, wherever they might be in the hospital or out in the community, to access, from their particular locations, the 3D images they want, when they want, and to manipulate them in the manner they want,” Royal says. Thick as a Brick The problem with processing 3D reconstructions at the workstation—also known as the thick-client approach—is its impracticality in an enterprise with more than a handful of customers. Part of the reason is the data volume generated by 3D CT and by multiphase, contrast-enhanced, and functional MRI studies. The huge (and growing) number of slices produced by these modalities pushes the limits of an organization’s IT infrastructure, causing traffic slowdowns network wide. “With a thick client, it takes time for the image data to migrate down to the workstation,” Royal says. “Each time the user wants to manipulate the image, even just a little, the full set of data must be transmitted to the workstation.” It’s not just the need to move a heavy payload of data from a server to the workstation repeatedly that creates problems, however; it’s also the fact that 3D reconstructions performed at the workstation require exorbitant amounts of processing power. That can be expensive. The cost is not so great if the enterprise visualization solution involves a handful of workstations, but it will turn very costly once the number of workstations required extends to tens or hundreds. In a similar vein, implementing updates of the reconstruction software residing on each workstation in a sizable enterprise imaging environment can be a chore involving monumental expenditures. Server-side processing, or the thin-client approach, carries none of these liabilities. “Thin client means that all of the image manipulations take place on a central server before being transmitted to the workstation,” Royal says. “Consequently, the processed image is available for viewing at the workstation almost instantly, yet traffic on the network is not significantly increased, and in some instances, may actually be reduced.” Software updates are easily implemented, since only the programming of the server needs to be altered in order to affect all users. Whatever changes are made to the software automatically show up on each user’s computer the next time it’s used. Thin Is In At Children’s Hospital, the foundation of enterprise visualization is a highly evolved, integrated RIS/PACS that includes voice-recognition software. “We wanted enterprise-wide access to 3D images because we felt it would increase the efficiency and productivity of our radiologists and our customers,” Royal says. “We also wanted our referring physicians to have the tools necessary to perform their own 3D image manipulations, but before you can achieve good things systemically, you first must have a good, efficient solution in place. You can have the best enterprise image-distribution technology in the world, but what good is it if it takes three days to get out a report of findings?” The quest to achieve all this via server-side processing led Children’s Hospital to acquire thin-client technology from Visage Imaging, Andover, Mass. “The thin-client solution,” Royal says, “is fully integrated with our RIS/PACS, and it has been deployed in a true plug-and-play fashion. It gave us the means to expose different sets of features based on the need of the individual customer, so that we neither overwhelm anyone with a larger array of buttons and tools than necessary for the desired type of view nor shortchange anyone of the full set of advanced functions, if those are what that user prefers.” Use Is Growing With facile, enterprise-wide access to, and manipulations of, 3D reconstructions now the order of the day at Children’s Hospital, utilization of these types of images is up appreciably, Royal reports. “Our most important clinical-care applications for 3D reconstruction are neurosurgical, orthopedic, and oncological,” he says. “3D reconstructions are used by neurosurgeons routinely for evaluation of complex conditions such as skull-base lesions, vascular lesions, and tumors. They allow them to answer questions such as where the lesion is located and what its relationship is to adjacent neuroforamina, vascular structures, neural structures, or bony structures.” Once answers to those questions have been obtained, neurosurgeons then can manipulate the 3D images to help develop appropriate surgical approaches. Royal says, “The radiologist can’t plan the surgical strategy. Only the neurosurgeon can, and that’s one of the most compelling arguments for giving him or her, and all other radiology customers, a distributed 3D capability.” Pediatric orthopedic patients seen at Children’s Hospital tend to be among the most complicated cases—another reason that 3D reconstructions are so highly valued there. “Many of these cases involve congenital anomalies, malformations of bony structures, and other deformities,” Royal indicates. “They also often have trauma or tumors superimposed over those base issues. The need to understand the 3D relationships between various structures is paramount in good care.” The same holds true for oncology cases. “We need to know where the tumors are and what they’re related to, no matter whether they’re intracranial or in the spine or the periphery,” Royal says. Tools for Education 3D reconstructions are useful not just clinically, but also from an educational standpoint, Royal says. “In a medical center, you’ve got medical students, residents, and fellows. All are learners trying to understand what’s being taught to them, so it’s useful to employ this 3D application on the fly during conferences to show, for instance, the differences between what’s seen on a 2D versus 3D CT scan, and how much better you can understand what’s represented by the 2D image when it’s shown in 3D format,” he says. There also exists the need for 3D reconstructions in research. Royal says: “It’s one thing to look at an image and marvel at its beauty, but quite another to prove scientifically that this same image is the best diagnostic approach to take in order to set the stage for better outcomes.” The thin-client approach to enterprise visualization is not without controversy. “Some radiology departments worry that they may, in fact, be sealing their own fate by giving customers the ability to perform 3D reconstructions and manipulations,” Royal observes. “Once empowered, what need would a customer have for the radiology department?” Royal disputes the notion that empowering the customer is a bad idea. “In my radiology department, we have a philosophy of openness, transparency, and immediate accessibility. In other words, it’s our mission to get the information out there, enterprise-wide, in the best way possible,” he says. “There’s nothing harmful in that. Where the harm occurs is in failing—after giving out the information—to demonstrate radiology’s value by, for example, calling attention to the six, seven, or more things that nobody but radiology noticed in the image.” He continues, “We want our customers—surgeons and other clinicians—to use 3D reconstruction and manipulation technology every bit as well as we, the radiologists, do. We’ve gone to considerable lengths to train them to make sure that they can. Then, we monitor their utilization of the system so that we can refine our training accordingly. The goal is to help them derive value from the system—and, by extension, value from radiology—that continually increases.”