Dr. Mendelson is a Radiologist in the Department of Radiology at The Mount Sinai School of Medicine, New York, NY. Portions of this work have been presented at the following meetings:

• Mendelson DS. Image sharing: Current status, IHE and the RSNA/NIBIB image sharing project. RSNA annual meeting 2010, Chicago, IL. Dec.1, 2010.
• Mendelson DS. IHE: Interoperability: Enabling patient control. IHE Connectathon Conference, Chicago IL. Jan. 18, 2011.
• Mendelson DS. Image sharing 2011 update. American Roentgen Ray Society Annual Meeting. Chicago, IL. May 2, 2011.
• Mendelson DS. Panel discussion: Meaningful use for radiology: A vendor perspective. The Society for Imaging Informatics in Medicine Annual Meeting. Washington, DC. June 2, 2011.
• Mendelson DS. Image sharing 2011 update. The Society for Imaging Informatics in Medicine Annual Meeting. Washington, DC. June 2, 2011.

There are many good reasons for sharing imaging exams.¹ For radiologists, a previous study is highly valuable for comparison with a more recent study in which an apparently new imaging abnormality is identified. Prior comparison exams also improve the quality of interpretation and expedite clinical care. In addition, given the continuing growth in healthcare expenditures, ready access to a previous comparison study may avert the need for a more expensive or invasive procedure to work up a current finding. For instance, a pulmonary nodule that remains unchanged through 2 or 3 years of comparison studies can be declared stable and nonaggressive, rather than being investigated with CT or biopsy without those studies. Prior studies also help avoid placing additional radiation burden on the individual patient and the population.²

Patients are mobile; they often receive healthcare services at multiple sites, and their images need to move with them. They also often see a variety of specialists or seek second opinions from other radiologists. Many patients, in addition, enroll in clinical trials that require them to submit their imaging exams. Indeed, the need for access to patients’ diagnostic imaging studies continues to expand and requires an easy and secure mechanism to permit this access.

The first transition: Film to CD

Historically, film was the medium that physicians used to share exams with each other. It served its purpose, but it was expensive, clumsy, and inefficient. As a community, we all thought the compact disc (CD) was the ideal solution.^3,4 Cheap, portable, and able to hold thousands of images and multiple exams, the CD certainly was an improvement. But CDs also have problems: Many vendors employ proprietary formats that cannot be opened on standard DICOM viewers; in many cases, standards have not been followed. Also, many CDs include a viewer, but viewers differ in significant ways and often are not intuitive to new users or to those who use them only occasionally. In a clinical practice, there just isn’t time to learn how to navigate each viewer. In many offices, workstations are “locked down;” that is, as a security measure, they won’t run external programs, including DICOM viewers. Discs can also be defective and unreadable. Occasionally, even the wrong patient is included on a disc. Lastly, the entire process of opening a disc can be quite time consuming.

Make no mistake; despite their deficiencies, CDs represent a significant advance with regard to image sharing.^3,4 Current data demonstrate that the quality of care improves when historical exams are easily available on CD.⁴ The CD is likely to remain with us for some time, and we must leverage solutions that refine the workflow around the CD, making it an effective means to share data while we transition to the next era—that of cloud computing.

Extending the network: First steps onto the internetNetwork-based sharing

Once picture archive and communication systems (PACS) and radiology information systems (RIS) became commonplace, the next step was to move the images and reports over networks, outside the radiology department and directly into the hands of the clinical staff. As enterprises, in parallel with these radiology developments, extended their internal networks, this became a fairly straightforward task. Within the firewall, images were exposed on local workstations through PACS client applications or Web viewers. Clinicians soon wanted to be able to view exams in their offices, but outside the firewall. Virtual Private Network (VPN) solutions have emerged over the last few years, which provide credentialed physicians with this access. The key phrase here is “credentialed physicians;” this solution works well for physicians who are members of a system, but it breaks down for those who are part of a separate organization.

Today’s patients, being extremely mobile, often visit providers who are members of disparate healthcare systems, which don’t all have access to the same VPN solutions. Yet it is highly desirable to provide image access safely and securely with patient consent. How do we extend the network outside of the local firewall?

Off to the clouds

We share data over the Internet every day. Photographs and movies are now instantly available wherever we are. We radiologists can move medical images in the same way, albeit with appropriate security and privacy measures. As bandwidth increases, the constraints on transferring large imaging data files diminish. Medical enterprises have turned to the vendor community for solutions; and the vendors have responded.

Cloud-based computing has been well described in many places.^3,5,6 Both storage and image postprocessing can be outsourced from the local enterprise.⁵ Dr. Rasu Shrestha, the Medical Director of Interoperability & Division Chief of Radiology Informatics, The University of Pittsburgh Medical Center (UPMC), has provided an excellent summary of its applications for an article in Applied Radiology.⁶ Basically, the sharing and exchanging of images through the Internet and cloud-based services adhere to 2 overarching paradigms. The first is to provide a set of services in the cloud that move images between sites on demand. This methodology is often employed to manage point-to-point transfers. The second method includes archiving the healthcare data, including images, in the cloud so that the original source needs to send the images into the cloud just once. All subsequent sharing takes place from the cloud archive. This latter solution is the one generally employed in Health Information Exchanges (HIE) (Figure 1).

Several business goals may be achieved from these solutions.⁷ Cloud-based storage might initially be purchased as a data recovery and business continuity solution. However, once the imaging data and reports are present in the cloud, an additional set of services can be added, employing the same archive, to share images. Another growing case is for expanding a multi-institutional enterprise to employ a cloud archive not only as a business continuity solution but also as an integration point so that the multiple sites, with disparate PACS systems and archives, can share images. Primary PACS services in the cloud have arisen to provide small organizations with an economically feasible solution when the outright purchase of a PACS system is perceived as impractical or too costly. Economics and efficiency are the main drivers toward cloud computing solutions.

The preceding solutions are limited in that they enable sharing through the cloud, but only within a given enterprise. Easy and secure exchange of healthcare data is a focus of Meaningful Use, a part of the American Recovery and Reinvestment Act of 2009. Image-enabled HIEs are a next step in crossing the boundaries of single enterprise data sharing. Many prototype HIEs exist, and some enable the exchange of images. Many are built on proprietary solutions and hence do not easily extend beyond the members of the exchange. It is highly desirable that HIEs be built on a set of commonly accepted standards so that technology does not represent a hurdle to participation.

Integrating the Healthcare Enterprise (IHE) provides just the answer to this latter issue. IHE profiles describe standards-based interoperability; they employ common standards and detail how to implement these standards in a fashion that is easily duplicated and extended. A specific set of IHE profiles known as Cross-Enterprise Document Sharing (XDS)⁸ describes how independent enterprises can share medical images and other data. This solution, which is increasingly being embraced internationally because it provides for a standard, well-accepted means to build an HIE, is also beginning to gain traction in the United States.

All of the above solutions focus on sharing at the enterprise level. But a group of radiologists working through the RSNA asked a different question: Could image sharing be accomplished through Personal Health Records (PHR) under the control of the consumer?

With support from the National Institute of Biomedical Imaging and Bioengineering (NIBIB), a pilot study is being designed to help answer this question. Five academic sites—The Mayo Clinic, Mount Sinai Medical Center, the University of Chicago, the University of California San Francisco, and the University of Maryland —will enable their patients to transfer their images to a PHR. From there, patients will have direct access to their imaging exams and reports. The patients can then, on the spur of the moment, sign into their PHR, launch a web viewer, and view their images (Figure 2). As an alternative, a patient can email a link to a provider that will enable the provider to look at the images on the web or download them to a local workstation. This is built upon IHE-defined profiles—again, commonly accepted standards with a strong patient security and confidentiality model. This pilot is just being deployed as this article goes to press.

Technical details of the RSNA/NIBIB project

There are 3 main components to the RSNA/NIBIB pilot project: an edge server (Figure 3), an imaging clearinghouse, and the image-enabled PHR. There is an edge device at each radiology department. The patient is registered on this device, which then waits for a message from the RIS indicating that a finalized report has been issued. The edge device then queries the PACS and packages the imaging exam and report. The “package” is then encrypted and sent to the “clearinghouse.” The package waits in the clearinghouse for the patient to request the exam from his or her PHR. When the appropriate authentication is provided, the patient can move the images and report into the PHR, where it is de-encrypted. The PHRs permit patients to view their images, read their reports, and distribute copies of each to their providers. Consumer control of their own images and a commitment to standards via IHE profiles are the governing factors. We hope this bootstraps image sharing across the nation.

What are the potential limitations of such a system? We believe that this solution will be highly desirable to many patients, but not all. It presumes one is comfortable using the Internet and with exchanging private information over the Web. Individuals capable of conducting banking or shopping on the Web should be able to participate. Those who are not facile with such technology will not be good candidates for this solution unless they employ surrogates.

In addition, there are issues with regard to trust in one’s privacy, security, and confidentiality on any Web-based service. Many consumers have concerns about these issues and will not embrace this solution at this juncture. We have devoted significant resources to ensuring a robust solution to these issues; nevertheless, some will wish to observe for a while to gain confidence that this service cannot be breached.

The operational costs of this system, similar to HIEs, are not fully understood. The cost of infrastructure is fairly straightforward, but the cost of each transaction will depend on multiple factors. We hope to use data from this pilot project to build a series of economic models that explore the feasibility of obtaining financial support from patients, payers, radiology departments, healthcare enterprises, or a combination of these stakeholders.

Lastly, this service must be easy to use, have high availability, and excellent performance. We have built the application and put in place an infrastructure to attain these goals. We must validate them in the real world.

References

1. Flanders AE. Medical image and data sharing: Are we there yet? Radiographics. 2009;29:1247-1251.
2. Prevedello LM, Sodickson AD, Andriole KP, Khorasani R. IT tools will be critical in helping reduce radiation exposure from medical imaging. J Am Coll Radiol. 2009;6:125-126.
3. Mendelson DS, Bak PRG, Menschik E, Siegel E. Informatics in radiology: Image exchange: IHE and the evolution of image sharing. Radiographics. 2008;28:1817-1833. Epub 2008 Sept 4.
4. Sodickson A, Opraseuth J, Ledbetter S. Outside imaging in emergency department transfer patients: CD import reduces rates of subsequent imaging utilization. Radiology. 2011;260:408-413. Epub 2011 Apr 19.
5. Langer SG. Challenges for data storage in medical imaging research. J Digit Imaging. 2011;24:203-207.
6. Shrestha RB. Imaging on the cloud. Applied Radiology. 2011;40:8-12.
7. Langer S. Issues surrounding PACS archiving to external third-party DICOM archives. J Digit Imaging. 2009; 22:48-52.
8. Integrating the Healthcare Enterprise: IHE IT Infrastructure Technical Framework. Vol 1 (ITITF-1), Integration Profiles. Rev 7.0. Chicago, IL: American College of Cardiology, Health Information Management Systems Society, and the Radiological Society of North America, Aug. 22, 2007. Available at: http://www.ihe.net/Technical_ Framework. Accessed July 27, 2011.