The Future of Radiology
Screenshot image courtesy of Aperio.
“The present is big with the future,” observed author Rudyard Kipling. The novelist was not referring to 21st century diagnostic imaging; however, the insight certainly applies. As radiology heads into the second decade of the 21st century, imaging stakeholders will leverage molecular imaging and digital pathology to deliver personalized medicine. CT radiation dose will plummet as radiologists re-invent relationships with clinical colleagues and make sense of meaningful use. The coming years will require new levels of creativity, collaboration and communication. As 2010 draws to a close, Health Imaging & IT visited with leaders to reflect on the evolution of imaging over the next three to five years.

CT imaging: A dose of good news

For the last several years, CT vendors have battled over slice count. However, the paradigm is changing. “The good news is that phenomenal activity and innovation are taking place in this space. We need to increase diagnostic confidence in smart ways,” offers Dominic Smith, vice president, CT and nuclear medicine for Philips Healthcare. Vendors and professional societies have put a bulls-eye on CT dose and are targeting an array of dose reduction tools with research and development dollars.

Dose reduction is a massive technical challenge; traditional approaches degrade image quality and increase reconstruction time—which defeats the utility of CT as an efficient triage tool and may lead to additional scanning. With CT dose in the hot seat, vendors are striving to reduce dose to less than 3 milliseverts (mSv) and even less than 1 mSv for some studies, such as pediatric cardiac CT studies. Technical development, however, represents only part of the changing paradigm.

“Education and collaboration among governing bodies, clinicians and industry are going to be key,” says Joe Cooper, director of CT for Toshiba America Medical Systems. Evidence of industry collaboration is mounting, shares Ken Denison, global CT dose leader for GE Healthcare. Consider for example various Medical Imaging & Technology Alliance (MITA) initiatives:
  • Dose Check will provide an alert to CT operators when recommended radiation dose levels—as set by the healthcare provider—will be exceeded.
  • DICOM Dose Structured Report provides a standard format for sending radiation exposure information to systems such as PACS and EMRs.
  • Enhanced Protocol Security, a standard still in development, will allow healthcare providers to better control access to protocols, likely through password protection, as well as provide a record of all changes.

Vendors also are working on internal system checks; the latest scanner software has the capability to identify and recommend KV and mAs settings based on patient size and exam type. Such scan protocol optimization is an example of decision support prior to scanning, says Christian Eusemann, director, CT clinical research and development for Siemens Healthcare.

New technical developments also hone in on image reconstruction. Vendors are improving on earlier models with techniques and reconstruction algorithms to cut dose and maintain or improve image quality. Universal reconstruction algorithms that apply to all CT studies could slash dose for many studies to less than 3 mSv.

Finally, radiologists will play a key role and could assist with increased efforts to educate referring clinicians about low-dose ordering. For example, when a radiologist receives an order for a chest CT to rule out nodules, he can continue with a low-dose scan. A general chest CT requires a higher dose. “It remains to be seen how the radiologist’s role will evolve. Radiologists may prescribe how much dose is necessary for CT scanning,” suggests Cooper.

Most insiders agree that the heightened focus on dose will have a nominal impact on CT utilization. “CT will remain a diagnostic modality of choice for many clinical questions,” says Eusemann. As CT vendors solve the dose problem, they can focus on other developments.

“We’ll see a drive toward quantifiable CT over the next few years,” shares Smith. Upcoming developments will detect and differentiate tissue density for increasingly accurate and lower dose imaging. Such advances are welcome as are ongoing developments in molecular imaging and personalized medicine.

Molecular imaging edges ahead

Magnified colon polyp at CT colonography is shown in 3D. Typical radiation exposure with CTC is about 3 mSv, the equivalent of the annual natural background radiation to which someone is exposed while living in Denver. Source: David J. Vining, MD, professor of radiology at the University of Texas MD Anderson Cancer Center in Houston
“Molecular imaging is on the verge of a lot of potential developments. It is an unmatched tool with nearly limitless possibilities,” shares Landis K. Griffeth, MD, PhD, director of nuclear medicine at Baylor University Medical Center in Dallas. A great deal of research focuses on tracer development.

In fact, PET and SPECT tracers in the pipeline could open new doors in Alzheimer’s disease detection and treatment, enabling physicians to more accurately target future therapies. A similar personalized medicine model will evolve in oncologic imaging as researchers explore and vendors develop more precise agents that better diagnose specific cancers or characteristics of a specific type of cancer, track disease more accurately or enable earlier disease monitoring.  

Currently FDG-PET has relatively blanket approval for treatment monitoring of many common cancers including breast, lung and colorectal. However, FDG-PET is not a perfect solution, partially because it “is the blunt instrument of PET,” explains Griffeth. For example, FDG-PET is plagued by false positives related to inflammatory conditions, and it is not an ideal marker for small brain lesions.

More widely applicable agents in the research pipeline may offer an improved model. 18F-fluorothymidine (18F-FLT) may be first to cross the approval and reimbursement hurdles, but hypoxia agents such as  60/62/64Cu-labeled diacetyl-bis (N4-methylthiosemicarbazone) (60/62/64Cu-ATSM) and 18F-fluoromisonidazole (18F-FMISO)  also show promise. Upcoming hypoxia agents could prove valuable in selecting patients for therapies by better determining which cancers are responding to treatment. Radiation oncologists also could use imaging data to increase the radiation dose to the hypoxic regions that are more resistant to radiation.  “When combined with biopsy and tissue analysis, new tracers on the horizon could allow us to get a better handle on disease,” offers Umar Mahmood, MD, PhD, associate director of nuclear medicine and molecular imaging at Massachusetts General Hospital in Boston.

At the same time, experts expect that FDG-PET will continue to prove its value. “Hopefully, clinical and pharmaceutical users will go beyond scratching the surface of FDG for treatment monitoring,” says Griffeth. In the clinical arena, FDG-PET-informed early treatment monitoring or therapeutic selection may make PET more economically viable, saving healthcare dollars on the backend by better delineating and targeting patients for specific therapies.

The relatively more favorable reimbursement environment parallels pharmaceutical developments, and the two may feed each other. The number of chemotherapy options has increased dramatically in recent years, observes Mahmood. “The likelihood of response to particular treatments can vary considerably based on specific tumor characteristics. Molecular imaging to determine which of these tumor characteristics are present allows us to make better treatment decisions in individual patients. As more specific imaging markers are employed and combined with tissue analysis from biopsies, these improved decisions may be made even before initial treatment begins,” explains Mahmood. FDG PET/CT is already routinely used as a readout in lymphoma treatment regarding the responsiveness of tumors to particular therapies very early in the treatment course.

There’s also growing interest in using FDG as a surrogate end point in clinical trials. Use of FDG-PET for early monitoring in clinical trials carries dual benefits. It could hasten the development of new drugs and help expand the tracer market. “This may be an example of a case where the rising tide [PET] floats all boats [tracers],” notes Griffeth.

Similarly, PET/CT may play a key role in Alzheimer’s disease detection by imaging the amyloid plaque burden in the brain. Future therapies may help decrease this plaque burden; however, treatments are associated with costs and risks. “In five years, we might be able to predict which patients are at higher risk of subsequent overt disease. We can stratify patients based on these risks and employ therapies in the right subset of people before they have significant symptoms. If the strategy works, it will have an incredible benefit, and it will be cost-effective,” Mahmood says.

But a reality check is in order. “The personalized medicine dream,” says Griffeth, “is to develop and utilize highly sensitive and highly specific tracers for specific tumors. But there are large and small obstacles that have to be overcome.” The realities of FDA approval, reimbursement and economically feasible manufacturing and distribution channels may hinder progress.

“There could be hundreds of types of tracers for various types of breast cancers,” explains Griffeth. Each tracer requires massive investments of time and funding to progress through regulatory and commercialization channels, but companies cannot support highly individualized tracers through approval, reimbursement and distribution. “We have to match reasonably common tracers to a reasonably-sized group of tumors or target a specific type of activity,” sums Griffeth.

Finally, complementary technologies must develop. For example, physicians will use pathology data to help narrow down appropriate tracers for imaging an individual tumor. In fact, digital pathology could provide a critical link between molecular imaging and pathology.  

Dawn of digital pathology

Prototype slide scanner and image management system is designed to handle digital pathology data rates as high as 500 MB/sec. Image courtesy of Philips.
The release of DICOM Supplement 145 puts digital pathology in the imaging spotlight. The new standard for “Whole Microscopic Images” enables electronic display, sharing, storage and management of larger images usually associated with pathology, and will facilitate health information interoperability of pathology medical images, various whole slide imaging equipment manufacturers, PACS and EHRs, according to the College of American Pathologists (CAP).

Some experts are quite bullish on the integration of digital pathology and radiology. Although digital pathology development is accelerating and the technology is benefitting from lessons learned in radiology digital image management, predictions of integration may be a bit premature.   

“Radiology and pathology are the two specialties in which diagnosis is performed on the basis of image analysis and interpretation, which ties the two together in terms of image management,” explains Bob van Gemen, general manager, digital pathology, Philips Healthcare Incubator. However, the two specialties employ very different workflow models.  

Typically, radiology is involved earlier in the diagnostic process with radiology results triggering the pathology process. There are no pathology data needed for radiologists. In contrast, pathologists sometimes need radiology data and they may use the data to help come to the right diagnosis, explains van Gemen. “This model is particularly true in breast cancer cases,” says van Gemen.

Surgeons and radiation oncologists could benefit from better access to pathology data, made possible by digital pathology, for patient treatment and followup. “The advent of digital pathology will help connect pathology to the hospital … [and it will] benefit patient safety and outcomes. The improved diagnostic process will be good for patients because it should increase accuracy and improve decision-making,” states van Gemen.

In addition to impacting patient care, digital pathology will transform pathology workflow. Most pathologists exist in the digital dark ages with few digital tools for simple tasks such as measuring distance and counting cells. “Sixty to 70 percent of medical decisions are based on pathology data. It’s ironic as we talk about EMR adoption that the pathology component of medical care is just beginning to talk about digitization,” observes Jared Schwartz, MD, PhD, chief medical officer of Aperio.

One benefit of later adoption, however, is the ability to leverage previously developed technologies. For example, Philips has employed technology used in fields outside pathology, like pan and zoom, in the digital pathology development process.  

Digital pathology development and adoption are rapidly evolving. “There is an extraordinarily vigorous debate and disagreement about if pathology and radiology are on a convergence path and will function in complementary roles or if there will be a merger with a department of imaging that includes radiologists and pathologists,” shares Schwartz.

He foresees an interim step with radiologists and pathologists working side by side on advanced imaging technologies like microendoscopy that require both specialties’ expertise. The continued development of in vivo biomarkers also could nudge the intersection of the two fields and nurture further developments in personalized medicine.

The radiologist’s role in innovation

As vendors and researchers strive to nurture developments from bench to bedside, radiologists also play an important role in innovation. A very large pipeline of potentially valuable innovations like targeted molecular diagnostic imaging agents and imaging biomarkers that are predictive of drug effectiveness are in the works, begins Bruce J. Hillman, MD, department of radiology and public health sciences at University of Virginia in Charlottesville. But innovations can be slow to translate into human trials or products, and there is little incentive for companies to move forward with development. Hillman cites several factors that hinder investment in imaging innovations:
  • An unsustainable rate of healthcare inflation;
  • The global recession;
  • Cuts in imaging payments—the Deficit Reduction Act in 2005 and Patient Protection and Affordable Care Act in 2010—mean fewer equipment purchases and less re-investment; and
  • An ongoing [and perhaps, increasing] suspicion about the value of imaging in improving health.

The radiology profession, says Hillman, can help improve the outlook by putting the brakes on inappropriate imaging and promoting appropriate use of diagnostic imaging. On one hand, the aging population requires more imaging because imaging is very good at diagnosing acute disease and following chronic conditions, says Hillman. Radiologists can communicate the positive value of imaging to policymakers, payors and clinicians.

On the flip side, practice patterns like defensive imaging and self-referral for diagnostic imaging must be curbed, says Hillman. Unfortunately, true solutions—such as changing lawyers’ incentives for malpractice suits and termination of the in-office ancillary service exception—are not probable, he admits.

Radiologists may have more control over other tactics. The information and misinformation on the web and direct-to-patient advertising tend to feed patients’ beliefs that more imaging is better. At the same time, the culture of medicine and medical education has encouraged a “shotgun approach” to imaging and diagnosis that uses imaging to conserve physicians’ time. “The training and education of physicians and culture of practice [have to change],” notes Hillman. Radiologists’ training should encourage a better understanding of patient preferences and outcomes. “Misses are not the only incorrect diagnoses.” Oversensitivity to noncritical findings can encourage downstream overtesting and inappropriate or unnecessary treatment.

The final factor that may impact future innovation is the lack of understanding of appropriate imaging, says Hillman. “We don’t really know what appropriate imaging is. It’s very expensive and very difficult to show a relationship between imaging and improved health. We need to get much better at rigorously determining the value of imaging.”

The meaningful use storm

A fair number of radiologists consider meaningful use a hurdle for other physicians with carrots and sticks (i.e. financial incentives through 2015 and penalties thereafter) that don’t apply to radiology. But not so fast, says David Avrin, MD, PhD, vice chair, informatics at University of California, San Francisco. Although it’s true that the interim rule issued earlier in 2010 was set for clinicians and its requirements left radiologists by the wayside, the implications of the final rule released in September for private practice radiologists and outpatient imaging centers are less clear. That is, practices could tap into RIS/PACS to qualify for meaningful use incentives.

Private practice radiologists who own and operate imaging centers are eligible for up to $44,000 in eligible provider incentives. “For a 10- to 12-physician radiology practice, it adds up to enough to invest in PACS. The problem,” explains Avrin, “is that neither the interim or final rule for Stage 1 addressed imaging.”

With little guidance for practices that want to qualify as eligible providers, professional organizations—American College of Radiology, RSNA, Society for Imaging Informatics in Medicine and The American Board of Radiology—have identified gaps such as image sharing and order entry rules and shared their concerns with the Centers for Medicare and Medicaid Services for consideration in Stage 2 and Stage 3 rules. However, practices eyeing the early stage of incentives may have to rely on some guesswork. “Get up to speed on recent guidelines regarding HIT certification and usage requirements and ask RIS/PACS vendors where they stand in the meaningful use certification process,” recommends Avrin.

Although it’s difficult to predict what Stage 2 and Stage 3 meaningful use will target, radiology-related considerations include radiation dose tracking, image sharing and CPOE with clinical decision support. The upshot? Outpatient imaging practices need to include specific questions about vendors’ plans for features related to certification as they plan for new digital image management systems in the next several years.

Radiology has navigated stormy waters for quite a few years, and the pressure shows no sign of abating. However, the combination of technical innovations and re-engineered practice patterns could help the specialty survive, thrive and optimize patient care in the coming decade.