The Next Horizon: Radiology's Crystal Ball

The Next Horizon: Radiology's Crystal Ball

With change being the one thing we can always guarantee, the next decade for radiology will bring continued changes in imaging operations and workflow. It's no surprise that volumes will continue to rise in terms of studies and images per study, but other developments may be less predictable. "We will see more decentralization of radiology functions," forecasts Eliot Siegel, MD, chief of imaging service with Veterans Affairs (VA) Maryland Health Care System in Baltimore. One big—and quite welcome—surprise? Workflow will improve despite increasing volume, says Siegel. That's because more sites will tap into and more effectively integrate a variety of technologies that can boost workflow. And the sterile, utilitarian analog-based reading room will fade away. Personalized reading rooms geared to the digital world will take their place.

Cranking up the volume

The writing is on the wall at nearly every hospital and imaging center across the country. Mega-scanners, including 64-slice CT and its future descendents, will continue to proliferate. And Americans will continue to age. Most of the country's 80 million Baby Boomers will reach 65 by 2017; many will need more healthcare and more diagnostic radiology services. So radiology volumes will continue to increase. At the same time, reimbursement will slowly drop, says Siegel. "Radiology departments will face pressure to change," says Siegel.

In the best of times, change signifies opportunity. "There is incredible potential for workflow improvements in radiology," notes Siegel. In fact, many of the technologies that can truly reinvent workflow exist today. The primary barrier is the culture shift that must occur before true adoption. "Radiology departments need to rethink their organization and methods of practice. Technology can change the way radiology is practiced," says Siegel. The requisite culture shift, however, takes longer than technical development and investment. The climate is shifting as younger radiologists trained in the digital era enter the profession.

The next decade will be marked by decentralization of radiology practices. Medicine will begin to take advantage of anytime/anywhere image interpretation enabled by digital acquisition and transmission. "Teleradiology will thrive and explode," says Alan Schweitzer, chief technology officer with The Radiology Consulting Group in Boston. Teleradiology can solve time and staffing issues, says Schweitzer, and help facilities optimize radiologists' time.

Two exceptions to the decentralization trend are image acquisition and billing. Images will be acquired in the radiology department, but tasks such as supervision of contrast injections could be handled by physician assistants, freeing radiologists to focus on interpretation. Similarly, billing will become more centralized, which allows facilities to take advantage of economies of scale.

In the ideal future, radiologists will benefit as departments optimize existing technology, says Siegel. For example, current communication systems facilitate hands-free, voice communication over a wireless network, allowing radiologists to instantly communicate with technologists or clinicians, which can increase efficiency and improve patient care. Similarly, dictation will evolve. Digital dictation, voice recognition and structured reporting will penetrate the market. Structured reporting not only boosts workflow but also can feed into other advanced solutions like decision support, says Schweitzer.

For example, if structured reports reveal that 80 percent of CT studies ordered by a clinician are interpreted as normal, decision-support solutions can provide the necessary feedback and education to the physician.

"Decision support will be critical in the future," says Schweitzer. A variety of decision-support tools will be incorporated into PACS to help radiologists hone in on the region of interest. CAD systems will continue to evolve beyond breast and lung systems and help radiologists navigate through volumes of data to focus on the region of interest in colon, liver and other cancers. Other decision-support tools are order-based and could cut down on inappropriate exams. Another plus of automated order entry decision support? It will cut down on the number of phone calls radiologists place to physicians who have ordered an inappropriate study.

Other improvements may stem from the elimination of paper. "Many departments have implemented PACS, EMR and other solutions, but continue to rely on paper processes," observes Siegel. Sites can optimize these solutions to reduce paper and the inefficiencies associated with its use. A truly filmless, paperless environment enabled by solutions like PACS, EMR and CPOE can slash costs, improve and accelerate patient care, reduce medical errors and fuel productivity.

Another great workflow booster is integration. Today radiology is hampered by relatively poor connectivity between PACS and subspecialty systems like nuclear medicine and ultrasound workstations. But the era of the single workstation is on the horizon. "In the future, radiologists will be able to access multiple functions at a single workstation," predicts Siegel. The universal workstation promises to reduce and possibly eliminate those workflow-busting trips between specialty workstations. 3D, CAD functionality and more will be housed on the PACS viewer.

Integrating the Healthcare Enterprise (IHE) will continue to drive RIS/PACS integration, but the question of which system owns the workflow is up in the air, says Schweitzer. Schweitzer does foresee more RIS-driven business intelligence and radiology dashboard systems. "These tools can help sites slice and dice data to run the department more efficiently." For example, the site could implement scheduling changes to accommodate peak demand.

As departments implement these changes, each modification will nudge radiology closer to the workflow ideal. The other major shift, according to Siegel, is cultural rather than technical. "It's apparent that the radiologist of the near future will be a multi-tasker," states Siegel. In this case, demographics is on radiology's side. The upcoming generation of radiologists has been trained in the fast-paced digital environment and expects to interact with and multi-task using the latest technology. Unlike their older colleagues, younger radiologists are much less likely to resist technical innovation, which makes for smoother, more rapid implementations.

Going low stress

It's true that optimizing current technologies can improve workflow and productivity. But radiology is not an assembly line. "Higher productivity can create stress," shares Siegel. One way to counter stress is to design a low-stress environment. For radiology, this means the reading room.

"The reading room of the future does not have to be a sterile, high-tech environment. It can be fun, familiar and relaxing," says Siegel. For example, the Baltimore VA's reading room of the future features blue lighting, white noise generators and acoustic barriers, which may reduce stress and enhance productivity. One challenge, says Siegel, is to personalize the space especially when multiple radiologists share a single space. Potential solutions include digital systems to display personal photos and calendars.


Radiology is at a crossroads; most sites have deployed digital image management solutions. Many have or are looking into state-of-the-art image acquisition devices that generate previously unimagined amounts of data. A number of systems that can aid workflow—CAD, 3D, speech recognition and decision support—are available now. The challenge for radiology is to harness multiple solutions to manage the onslaught of data while improving the clinical utility of radiology. Welcome to the future!

The Next Generation of Breast Imaging

Mammography has been the gold standard of breast imaging for decades; however, despite technical advances it remains far from perfect and misses up to 20 percent of breast cancers. The current solution is to employ adjunctive modalities such as MR and ultrasound to augment and improve lesion detection. Digital mammography also is making a dent, particularly in dense breasted women.

Breast tomosynthesis, now in development by several vendors, represents the next generation of breast imaging solutions and could translate into improved detection. Another work-in-progress is breast CT, but tomosynthesis likely has an edge over CT and other emerging systems.

"Breast CT is not nearly as far along as tomosynthesis. Tomosynthesis will be the next major advance in breast imaging," asserts Laurie Fajardo, MD, professor and chair, department of radiology at University of Iowa in Iowa City. The technology relies on a digital detector to acquire a series of 2D images, which are post-processed and combined into one millimeter slices to display breast tissue in three dimensions.

"Tomosynthesis delivers all of the advantages of digital mammography and 3D imaging. Radiologists can view individual slices and remove the underlying and overlying tissue that can make it difficult to visualize small lesions," explains Fajardo. The end result should be improved accuracy.

In addition to clinical benefits, breast tomosynthesis should meet other departmental needs. At most sites, breast imaging operates on slim margins, so implementing cost-effective and efficient solutions that streamline operations and workflow is essential. Once again, tomosynthesis delivers. It represents more of an evolution than a revolution in breast imaging and should allow facilities to optimize current systems and processes while improving detection.

"Tomosynthesis is quite similar to digital mammography. Sites will use the same workstation, network and storage system. Implementing tomosynthesis will be like adding a new detector to an existing digital mammography system," says Fajardo. The operational impact of tomosynthesis will be minimal; scan time is the same as a traditional digital mammogram. And images will be viewed much like CT or MR studies.

The FDA could approve breast tomosynthesis in the next year or two. Adoption could be swift; the digital mammography installed base is growing rapidly, and tomosynthesis is the next logical step after implementing digital mammography.

Radiation Oncology Imaging in Evolution

Image-guided radiation therapy (IGRT) is the current model used at many radiation oncology treatment centers. But the IGRT model is just in its early stages, says Lei Dong, PhD, researcher and medical physicist with MD Anderson Cancer Center in Houston, Texas. In the next decade, radiation oncology will continue to evolve and grow resulting in improved workflow and clinical care.

"Target delineation is the most significant bottleneck in radiation oncology. Physicians need more information [to precisely and consistently define targets and normal anatomy]," says Dong. One solution may be automated target delineation and segmentation. Currently, physicians start the time-consuming target delineation process from scratch for each patient—slowing workflow and delaying treatment.

New intelligent systems like deformable image registration techniques will work backwards from normal structures to delineate the tumor. Automated techniques will enable physicians to use software to deform images and accurately and rapidly calculate the location of the tumor and normal anatomy. "This area will evolve quickly," predicts Dong.

Automated target registration addresses another critical issue in radiation oncology. "Seventy percent of radiation oncologists are in community practice. They aren't active in research and their learning process is limited to journal articles and annual meetings," notes Dong. Intelligent automated tools level the field.

At the same time, IGRT will continue to evolve. IGRT is not very efficient yet, says Dong. Researchers have not yet determined optimal techniques for image guidance; some patients may not require daily imaging or 3D imaging. At MD Anderson, Dong and his colleagues are comparing 3D cone-beam CT and 2D megavoltage and kilovoltage portal x-ray imaging.

"3D will come into play when soft-tissue structures are important," Dong says.

The clinical evolution will occur in parallel with other developments. "Radiation oncology has to redefine workflow to truly implement adaptive radiation therapy," states Dong. Adaptive radiation therapy, which is also referred to as image-guided adaptive radiation therapy (IGART), adjusts the patient's original treatment plan to the changes in the target captured on daily or weekly images. The current model relies on two therapists at each machine and multiple computers. Radiation oncology will move toward an integrated solution that does not require one workstation to run one piece of equipment, says Dong. Instead, one system could run multiple solutions.

Radiation oncology is on the cusp of multiple new advances. Technical advances and integrated systems could better target tumors and streamline workflow, and intelligent automated target delineation tools fulfill multiple purposes: improving workflow and patient care.

Cardiac Imaging A Tale of Two Modalities

In the last two years, 64-slice CT has revolutionized cardiac imaging by opening the door to non-invasive coronary imaging. Although the mega-scanners have generated a tremendous amount of interest and investment, cardiac imaging is not a one modality arena. MR, for example, plays a significant role in cardiac imaging. "CT and MR fill different niches. CT images coronary artery anatomy and pathology, and MR enables physicians to analyze function, wall motion perfusion and viability," details Michael Poon, MD, director of cardiology for Cabrini Medical Center in New York City.

Like other aspects of radiology, cardiac imaging will continue to evolve. Poon says CT and MR are likely to cross paths in the future with CT addressing perfusion imaging and MR peering into the coronaries, which leads to the question—which modality will emerge as the preferred modality for cardiac imaging?

Neither CT nor MR is ready to assume the dominant role today, and neither modality will completely replace the other. In fact, cardiac MR technology must advance on several levels before it is equipped to handle coronary imaging on a routine basis. "Cardiac MR is more of a battle plan. It can take five minutes to put the leads on a patient," notes Poon. In addition to longer scan times, the utility of cardiac MR is limited by image quality and relatively slow adoption.

Consequently, Poon gives CT an edge in today's environment. CT, he says, is a point-and-shoot camera, enabling techs to complete a scan in a matter of seconds. The disadvantages of CT, on the other hand, are the requisite radiation and toxic contrast. According to Poon, physicians could make a tradeoff between scan time and patient safety if MR technology evolves to provide better resolution for coronary imaging at a slightly longer scan time than 64-slice CT.

Researchers are exploring multiple avenues to improve cardiac MR. MR resolution is a function of three factors: post-sequencing technology, acquisition speed and contrast uptake. If advances in one or more areas produce better resolution, MR will have an edge over CT, Poon predicts.

But the future could evolve in a different direction, according to Anthony Mancuso, MD, professor and chairman department of radiology at University of Florida in Gainesville. "Two-hundred-fifty-six slice CT technology is clearly another revolution in anatomic cardiac imaging; it could become the dominant technology in non-invasive coronary imaging," says Mancuso.

What's more, myocardial perfusion studies may be possible via the temporal resolution capability of 256-slice scanners. "Cardiac MR is unlikely to anatomically image the coronary arteries at the level achieved by CT. If 256-slice scanners can combine anatomic imaging of the coronaries with myocardial perfusion, it will emerge as the triage test for patients presenting with coronary disease," predicts Mancuso.

A third possibility remains. Perhaps neither CT nor cardiac MR will emerge dominant in the next decade. If neither modality can effectively combine anatomic and perfusion imaging, cardiac imaging will rely on a combination of CT, MR and radionuclide studies to diagnose disease and determine the appropriate treatment.

In addition, fusion technologies such as PET/CT and CT/MR will continue to play an important role in cardiac imaging. That's because the utility of CT is limited in patients with calcification; patients with significant calcification require additional studies to complement CT. Over the next several years, researchers could determine the preferred complementary test for patients with calcification.

"PET performs quite well to answer questions about stress perfusion and viability, and cardiac MR performs very well," opines Poon. Currently, PET scanners are more readily available than cardiac MR; however, cardiac MR remains in its infancy and could overtake PET as more facilities embrace the modality.

Although it's difficult to clarify the future, a few items do seem certain. No one modality will completely replace another, says Poon. On the operational front, cardiac imaging and radiology will continue to merge. University of Florida, for example, is characterized by complete cooperation between radiology and cardiology. The two departments interpret studies together in a single cardiovascular reading room for radionuclide studies, CT and MR. "This is the only viable future," opines Mancuso. "There is an incidental findings issue that is only resolved by teamwork."

Cardiac imaging is young and will continue to mature over the next decade. Multidetector CT, cardiology's anatomical powerhouse, will continue to evolve and bring improved resolution and perhaps perfusion imaging. MR developments are in the works, too. Researchers are investigating multiple areas—post-processing technology, scan time and contrast uptake—to improve the utility and resolution of MR. Finally, other modalities including PET and fusion imaging will continue to complement CT and MR. 

Molecular Imaging The Next Age

"Molecular imaging will revolutionize radiology and nuclear medicine over the next five to 10 years," predicts Juri Gelovani, MD, PhD, chair of experimental diagnostic imaging at MD Anderson Cancer Center in Houston. The revolution is expected to bring clinical benefits to patients with multiple types of disease particularly cancer. Advances in molecular imaging will alter cancer diagnosis and management, and may enable physicians to detect patients at high-risk, pinpoint tumors at their earliest stages and prescribe individualized therapies.

Changes at the screening level will resemble the current model used for the BRCA1 gene. That is, non-invasive or minimally invasive tests will analyze a patient's blood, urine or sputum to detect a genetic mutation that predisposes the patient to either generic cancer development of a certain type of cancer. The next level functions like PSA screening. The patient's blood or urine will be screened for the presence of tumor biomarkers.

Early screening and detection, however, does not translate into tumor localization. "It's very hard to identify a one or two millimeter lesion with conventional approaches," explains Gelovani. Generic or specific biomarker imaging agents may be used in conjunction with diagnostic imaging to enable localization. "Imaging the biomarkers in the tissue surrounding the tumor allows us to take advantage of the volume amplification effect. In most cases, the reaction to the biomarkers is larger in tissue surrounding lesions than during tissue inflammation," says Gelovani.

How might current research translate into improved clinical care? Take for example pancreatic cancer. Approximately 30,000 Americans die of pancreatic cancer annually. Currently, no imaging modality is capable of detecting early pancreatic cancer. By the time symptoms present, 80 to 90 percent of patients have progressed to the point where the tumor is unresectable.

In the brave new world of molecular imaging, genetic screening would indicate patients at a high risk of developing pancreatic cancer such as those with familial pancreatitis or Peutz-Jeghers syndrome.

Biomarker screening must be followed by diagnostic imaging to localize the tumor and treatment to eradicate early neoplastic lesions, says Gelovani. High-risk patients would be followed with protein screening to biomarkers like amyloid A, HIP/PAP or RCAS 1 protein that correlate with pancreatic cancer. Patients might be injected with agents to enable tissue-based diagnostic imaging and detection via CT or another imaging modality. Enzymatic markers provide another means of detecting and amplifying the signal from tumors as small as one to two millimeters. "There is no first line diagnostic modality. Each modality [or agent] is unlikely to be effective on its own. A combination approach will yield the best results," sums Gelovani.

Early detection and localization, however, does not cure cancer. Treatment options will run the gamut from microscopic resection, image-guided therapy, ablation and targeted molecular therapies. In the future, molecular imaging using one to three lead biomarkers will help physicians individualize treatment by selecting the optimal therapy for each patient. Biopsied tissue could be tested to determine if the tumor is likely to respond or resist a certain treatment, or pre- and post-therapy PET scans could determine the efficacy of therapy.

Although multiple approaches are in the works and undergoing clinical trials, other advances must occur before the practical application of molecular agents—for example, patients may be injected with multiple tracers, which require shorter-lived isotopes as well as increased sensitivity and resolution of the diagnostic imaging modalities. What's more, the isotope model must change. Tracer development will need to decentralize so sites can create multiple, short-lived tracers. These sites will need mini cyclotrons to develop tracers. 

Molecular imaging is on the cusp of a revolution. New tracers and biomarkers will facilitate cancer screening, early diagnosis and individualized treatment, and the sensitivity and resolution of imaging modalities, particularly PET, MR and SPECT, must improve.