SIIM: Siegel details wish list for next-generation PACS

ORLANDO, Fla.—Samuel J. Dwyer, PhD, the father of PACS, might be surprised by the slow rate of innovation in PACS, RIS and speech recognition systems, Eliot L. Siegel, MD, told the audience during the 2012 Dwyer Lecture June 8 at the annual meeting of the Society for Imaging Informatics in Medicine (SIIM). While Siegel outlined his specifications for a next-generation PACS, Paul J. Chang, MD, was much less patient, stating: “I want it now.”

In fact, Chang, vice chair of radiology informatics at the University of Chicago, argued that radiology needs next-generation technology to survive further commoditization of the specialty.

Siegel, vice chair of radiology at the University of Maryland School of Medicine in Baltimore, and colleagues are prepping to replace a PACS purchased in 1991 and looking for a smart system. The capabilities of a smart PACS differ quite dramatically from those of current models, he said.

Smart storage topped Siegel’s list of must-have features. Despite the hefty purchase price of $800,000, the medical center’s initial 1 terabyte of storage has not withstood the test of time. Current archive models force data silos and don’t allow migration from previous media. It is not user-friendly, cost-effective or efficient.

High performance, reliability and affordability are important, said Siegel. “I do not want to have to worry about storage. I want to use software to automate updating.” Currently, radiology may feel hamstrung by storage, but Siegel posed that radiology should be able to consume storage as a commodity. Smart storage should support the needs of radiology, nuclear medicine, cardiology and dermatology, and “trick” modalities into believing the archive is its own storage.

In the last two years, critical results communication has bubbled to the surface to become a top tier priority. Yet these processes remain painstakingly manual and inefficient. Siegel and colleagues, like their peers across the U.S., document phone communication of critical results in reports. But in 20 to 25 percent of cases, the referring physician does not follow up, sparking a manual chain of events with radiology checking back after one month to determine follow-up. “We need to close the communication loop. That process should be automated,” said Siegel.  

PACS inefficiencies extend into reporting. Radiologists have to dictate or type measurements made at the workstation into the report. Smart technology can improve the process. That is, the workstation could help the radiologist follow the progress of disease. Workstation-driven template reporting could allow the radiologists to record and correlate data.  

Siegel, along with every colleague exhausted by infinite mouse clicks, also anticipates the advent of new interface devices—with PACS software adapting to the capabilities provided by these devices.  

Smart PACS should include features to minimize frustrations, said Siegel. Server-side rendering could reduce or eliminate wait time associated with image delivery. While current hanging protocols force radiologists to repeat the same steps, advanced protocols could adapt to the specific study and individual radiologist’s preferences.

Also appearing on Siegel’s wish list is workstation-consumable CME. Journal articles with full DICOM datasets could be integrated into the workstation to boost learning and knowledge, offered Siegel.

An ongoing complaint in the informatics world is lack of interoperability. “We are looking for a PACS that really plays well with other apps,” said Siegel. In a nod to the millennial generation, he pondered the possibility of an iTunes store for imaging apps. He also contended that PACS should be able to output imaging data to the personal health record, and perhaps, ultimately, to Facebook.

Lastly, the need for increased efficiency is paramount in imaging. Although current business analytics tools perform admirably in terms of monitoring performance metrics such as turnaround time, they don’t make radiologists faster.

“I want clinical analytics that will help me be faster, more effective and safer,” said Siegel. An automated radiology resident could provide contextual clinical data at the point of care and present the information in a rapidly consumable manner. This graphical digital patient summary could include pathology results, surgeries and eventually proteomic and genomic information.

Despite the length and breadth of his wish list, Siegel predicted progress. “The next disruptive breakthrough will be artificial intelligence.” Artificial intelligence, like Watson technology, could represent the building blocks of the automated radiology resident.

Watson, with its capability to process 500 gigabytes per second (or the equivalent of one million books), may provide architecture to evaluate the masses of unstructured content contained in medical records. By employing approximately 100 tools to evaluate patient data and medical knowledge Watson can generate likely answers.

Watson will not replace physicians, said Siegel. However, its underlying technologies—natural language processing and artificial intelligence—will have a major impact on medicine and diagnostic imaging in the very near future, he concluded.

Although some private practice radiologists questioned the value and utility of next-generation tools for bread and butter practitioners, Chang argued that they are essential for survival. “Radiologists need to double their efficiency and demonstrate their value,” he said. The best tactic to demonstrate quality is to add clinical context. And radiologists need PACS to deliver this capability.