Medicare study questions use of high-tech cancer imaging

June 21, 2010
Greg Frieherr

Greg Frieherr

Oncology NEWS International, Oncology NEWS International Vol 19 No 6, Volume 19, Issue 6

Multiple scans are now par for the course in cancer care. Experts debate whether all are essential to patient management or some should be cut to help rein in surging costs.

ABSTRACT: Multiple scans are now par for the course in cancer care. Experts debate whether all are essential to patient management or some should be cut to help rein in surging costs.

The average Medicare patient with lung cancer receives a battery of imaging studies: 11 radiographs, six CTs, a PET scan, another non-PET molecular imaging scan, an MRI, two echocardiograms, and an ultrasound-all within two years of diagnosis.

According to new research, there was a steady increase in the use and cost of imaging Medicare cancer patients from 1999 to 2006, at a rate double that of any other cost related to caring for such patients. The trend is likely continuing, said researchers from the Duke Center for Clinical and Genetic Economics in Durham, N.C., who cite emerging technologies, changing diagnostic treatment patterns, and changes in Medicare reimbursement as likely contributors to an increasing use of advanced imaging in cancer patients.

"You could easily envision a world in which we are using PET like we use CT and MRI today, with multiple scans per patient." - Kevin A. Schulman, MD

"Are all these [imaging studies] essential? Are they all of value? Is the information really meaningful?" said co-author Kevin A. Schulman, MD, director of the center. "We have to go back and ask the question: 'What is changing as a result of all this imaging?'"

Conjecture over why the use of high technology was accelerating at such a rate included the possible roles played by economic incentives to perform imaging studies and the novelty of new technologies. Also considered was whether patient outcomes were really improved through their use. The answers take on greater significance in light of the possibility that the Medicare data may only hint at the true scope of imaging use for all cancer patients in this country.

"Our data are on patients who average 76 years old," Dr. Schulman said. "If this is what they are getting, obviously younger patients might be getting more."

What Duke study determined

The Duke study looked at imaging performed over the course of nearly 101,000 cases of breast cancer, colorectal cancer, leukemia, lung cancer, non-Hodgkin's lymphoma, and prostate cancer among Medicare beneficiaries. Extensive use of imaging emerged not only in lung cancer but in lymphoma, where patients by 2006 averaged eight conventional radiographs, six CTs, a PET scan, a nuclear medicine test, an MRI, three echocardiograms, and three ultrasounds within two years of diagnosis (JAMA 303:1625-1631, 2010).

"New technology seems to be additive to the old technology. New technologies are not being substituted for the old," Dr. Schulman said.

That’s a good thing, said Larry Habelson-Wilf, MD, director of oncologic imaging for ICON (Integrated Community Oncology Network) in Jacksonville, Fla. He argued that substituting new for old technologies would mean all patients suspected of cancer would immediately go for CT or PET/CT, rather than beginning the work-up with the least invasive, least costly and least informative methods.

 “If you drop these less inclusive screening tests on the higher risk patients, in the long run you would save substantial time, patient anxiety, and cost in their total work-up,” he said. “When we scan, we stage, diagnose, and [plan] treatment for our patients all at the same time.” 

'Less reason for concern'

This affects the bottom line of Medicare costs, however, less severely than might be expected. In 2006, medical imaging expenses accounted for less than 6% of the overall cost of caring for Medicare cancer patients, according to the authors. If all imaging studies were eliminated at once, the total savings would be minimal, while physicians would be blindly making decisions about patient management. The situation would not be much better if imaging were restricted to just diagnosis and staging.


Dr. Habelson-Wilf interpreted the Duke team's conclusions in the broader context of cancer care for Medicare patients. He said he saw less reason for concern than Dr. Schulman's team does. Dr. Habelson-Wilf scoffed at the idea that the number of imaging exams ordered for cancer patients should be restricted. "The correct number is however many are needed to make the correct decision for the appropriate care of the patient," he said.


"The patient is not a fixed object," agreed C. Carl Jaffe, MD, a professor of radiology at the Boston Medical Center. "We can't just cluster all of our imaging at the front of the process and then blindly give [patients] therapy. It would be like going to Europe and photographing yourself as you board the plane in the U.S. It wouldn't tell you anything about your trip."

Necessary tests, but at what price?

The NCCN has established guidelines that lead physicians through the process of working up patients suspected of specific types of cancer. These guidelines account for much of what the Duke team documented in the Medicare data, according to Tomas Dvorak, MD, clinical instructor in radiation oncology at Tufts Medical Center in Boston.

CT, PET/CT, and MRI scans are necessary, Dr. Dvorak said, to diagnose and stage the patient so that decisions can be made about whether to begin palliative or curative treatments. A bone scan may be done to look for metastases in the bone. Chest x-rays are indicated every four to six months for the first two years.

"The reason is that we are looking for local or regional recurrences in the chest, where we can still re-treat or salvage patients for cure, as opposed to just giving up on them," he said.

In their paper, Dr. Schulman and colleagues raised the possibility that financial incentives contribute to the use of imaging. Dr. Habelson-Wilf acknowledged that financial incentives can encourage the use of imaging, just as they encourage the use of any medical device or drug. "You're kidding yourself if you think there is not a financial incentive somewhere along the way," he said.

The financial incentive to use imaging, however, has been trimmed substantially in recent years. Major cuts to medical imaging, specifically with the 2005 Deficit Reduction Act, were implemented in 2007.


While recent federal cutbacks have substantially restricted compensation for imaging exams, they have not prompted Barbara L. McAneny, MD, a medical oncologist and CEO of the New Mexico Cancer Center in Albuquerque, to order fewer of these exams. "When used appropriately, imaging is an incredibly useful tool," she said.

Appropriate imaging seems to be at the heart of the matter. Dr. McAneny said she is battling an insurer that insists on the performance of a CT scan prior to a PET/CT even though CTs don't tell the whole story, as a shrinking tumor may leave behind dead tissue. "I need to know if the mass is alive or dead," she said. "If you don't get a PET, you don't know."

Some examinations may be unnecessary because they are duplicative, as in the case of exams performed at one facility that are repeated at a tertiary cancer center because the initial results are not transferred. Exams may also be ordered unnecessarily by emergency room physicians, particularly if cancer patients come to the ER complaining of symptoms that might indicate a cancer recurrence.

EMR: The answer to the overuse dilemma?

Questions regarding the overall value of cancer imaging, its risk vs benefit for specific patient groups, as well as which tests and how many should be conducted, may require outcomes research to determine the answers.

Alternatively, progress may come with the widening adoption of electronic medical record (EMR) systems. These systems promise to reduce or, ultimately, eliminate redundant exams done at multiple facilities. They may also provide the means for doing the meta-analysis needed to streamline the use of imaging to achieve optimal management of cancer patients. But progress will come slowly, Dr. Jaffe said.

"There is no master EMR structure. EMR systems can't talk to each other if they are implemented piecemeal by individual institutions. We need something that crosses all institutions and has the same data fields," he said.

"If you have even a remote history of any kind of cancer and you go to the emergency room with belly pain, I guarantee you will get a CT scan," Dr. McAneny said.

Pointing fingers at imaging studies gives the impression that it is disproportionately contributing to rising healthcare costs. This may be especially misleading, as advances in imaging, particularly PET/CT, are proving useful in determining whether patients are responding to therapy. This is the kind of information oncologists need in order to decide whether to continue or change the current course of treatment, Dr. Jaffe said.

"My concern is that we not look backwards at how medicine has been done but rather at how medicine could be done," he said.

But Dr. Schulman argued that if areas of medical practice are contributing to a rise in cost, these areas need to be closely examined. "I wouldn't put the disproportionate burden on imaging," he said, "but I would definitely say it is a low-hanging fruit."

The Duke study singled out PET as being particularly ripe. From 1999 to 2006, use of the modality increased at an average annual rate of 35.9% to 53.6%, depending on the type of cancer. Lung cancer and lymphoma patients showed the largest increase.

Dr. Schulman suggested that novelty leads to overuse. But such use is not likely to continue unless there is a clinical payoff, Dr. McAneny argued.

"My use [of PET/CT] has increased because I have benefited in terms of the information it gives me so that I can take care of my patients," she said.

While the Duke researchers do not specifically call out the leap in PET imaging seen in the Medicare data, it could be inferred that the modality is an unreasonable contributor to cost. But the percentages are less damning if you consider that PET only entered the clinical mainstream in 2001. With a low baseline of use, percentage increases will necessarily be large, according to Dr. McAneny.

But Dr. Schulman said he views the increasing adoption of PET as reason for worry. "With PET we are watching the adoption of a new technology. It is an expensive technology; it does not seem to be replacing other technologies; and I don't think it has reached saturation yet. You could easily envision a world in which we are using PET like we use CT and MRI today, with multiple scans per patient."

TABLE 1Annual change in non-PET cancer imaging (1999-2006)

Ultrasound+0.7% to +7.4% per year
MRI+4.4% to +11.5% per year
Echo- cardiography+5% to +7.8% per year
Bone densitometry+6.3% to +20% per year
Conventional x-ray use–22% to –0.5% per year

But there is also the possibility that PET might have skewed the overall data regarding imaging in the Duke study, making imaging look like more of a contributor to exorbitant costs than it is. Percentage gains from an artificially low level in 1999 to routine use in 2006 naturally show a huge increase (see Table 1).

The JAMA paper raised issues with two other modalities: echocardiography and bone densitometry. Neither is directly involved in cancer imaging, but Dr. Schulman argued there are good reasons for including these in the Duke study.

"In regard to breast cancer, where you're thinking about preventing fractures, bone densitometry could be supportive care and the same goes for echocardiography, as there is a lot of toxicity [to the heart] from different chemotherapeutic regimens," he said.

Particularly worrisome to Dr. Schulman is that Medicare patients may under-represent the overall volume of imaging studies done on cancer patients. The advanced average age of patients in the Duke study may have limited the number of imaging studies prescribed for them. Younger patients, particularly those with pediatric cancers, may receive many more imaging studies over their lifetime, he said.

“Concerns must be tempered by the exquisite detail and management information imaging studies give. If you look at the risks involved, they are small.” - Steven B. Birnbaum, MD

Radiation is another factor in the use of imaging, especially for pediatric patients. If radiation exposure leads to cancer, it is likely to do so 10, 15, or 20 years down the road rather than soon after the scans are acquired, said Steven B. Birnbaum, MD, radiation safety officer at Southern New Hampshire Medical Center in Nashua. Children are the most susceptible to the possible effects of radiation because monitoring their health may require a lifetime of repeated examinations.

"But concerns must be tempered by the exquisite detail and management information imaging studies give," Dr. Birnbaum said. "If you look at the risks involved, they are small."

Radiologists are guided by the ALARA (as low as reasonably achievable) principle, which instructs that the least amount of radiation be given to accomplish the radiologic goal. Dr. Habelson-Wilf notes that ALARA is designed to minimize risk while maintaining the benefit of imaging.

TABLE 2National Bureau of Economic Research"Has medical innovation reduced cancer mortality?" can be accessed at


Advocates of medical imaging argue that imaging innovation has accounted for 40% of the reduction in cancer deaths between 1996 and 2006, citing as support research recently released by the National Bureau of Economic Research. In the April 2010 working report, author Frank Lichtenberg, PhD, a professor at Columbia University's Graduate School of Business in New York, stated that cancer imaging innovation alone reduced the U.S. cancer mortality rate by 10.2 deaths per 100,000 during this time period (see Table 2).