Skeptics want to see real evidence for virtual imaging
From the April ACP Observer, copyright © 2007 by the American College of Physicians.
By Susan FitzGerald
With imaging centers advertising their services directly to consumers, physicians are increasingly faced with requests from patients to order imaging tests to check for serious conditions such as heart disease and cancer, or to follow-up on tests the patients have obtained on their own. While these tests have the potential to catch diseases at an early, more treatable stage, some physicians worry that there's not enough evidence to justify the high cost.
Primary care doctors play an important role in helping patients sort through the costs, benefits and potential negative repercussions of imaging procedures.
A radiology clinic ad, for example, might suggest that a patient has nothing to lose by getting a CT scan—leaving out that the $1,200 fee may not be covered by their insurance, or that incidental findings that require further testing may ultimately found to be benign.
Thin slice thickness CT scan detects a small left upper lung nodule (arrow) Further diagnostic information, such as a needle biopsy, would reveal the significance of the lesion.
"The new technology sounds better and patients want it," said Robert Smith, PhD, director of screening for the American Cancer Society. "But the whole point is not to simply placate patients by sending them for a CT scan."
Weighing the costs and benefits
David Axelrod, ACP Member, a general internist at Thomas Jefferson University Hospital in Philadelphia, understands patients' interest in new imaging technology, but he also tries to explain to them the value of evidence-based medicine.
"Sometimes, it's a bad idea to get studies when there's a potential for harm from potential follow-up interventions. The results may lead to unnecessary follow-up tests and anxiety, and there's no evidence that the test improves the health outcome," he said.
A key concern is whether the very tests that patients think will give them peace of mind will instead turn up inconclusive or misleading information, prompting anxiety and the need for more tests and even biopsies and other invasive procedures, which carry risk.
Experts say primary care doctors need to discuss the pros and cons of any given test (including cost) so patients are not caught by surprise should an ambiguous result emerge. A lung CT scan, for instance, may pick up an incidental nodule that is worrisome to the patient but is in fact benign.
"The idea of screening always sounds intuitively attractive," said Raymond Gibbons, MD, a Mayo Clinic cardiologist and president of the American Heart Association. "But there are many things in medicine that seem to logically make sense that aren't such a good thing when they are scrutinized."
What's new is news
CT scans for lung cancer became big news in October 2006 when the New England Journal of Medicine published a study of more than 31,000 people, primarily smokers and ex-smokers, suggesting that, used routinely, spiral CT scans might have the potential to improve the life expectancies of lung cancer patients by detecting tumors early enough to treat effectively. For people found to have early stage lung cancer, the estimated 10-year survival rate was 88%; for those who had surgery within the month of diagnosis, the estimated 10-year survival rate was 92%. Since 85% or more of the 160,000 lung cancer cases annually in the U.S. occur in smokers or ex-smokers, the study raises the question of whether past or current smoking should trigger an annual CT scan.
"If you find it early and treat it early, you have a very high 10-year survival [rate]," said Claudia I. Henschke, MD, professor of radiology at Weill Medical College of Cornell University, who headed the study. She elaborated on those results in the January/February issue of ACP Observer, available online.
But other experts said that the findings, while promising, are not conclusive because the study was not a randomized, controlled trial.
"It is a single-arm observational study that was large, but in absence of a control group one cannot logically say the improvement in survival is an indication that with screening, fewer people will die from lung cancer," said David Midthun, FACP, a Mayo Clinic pulmonologist. He said that no form of screening has been proven to prevent mortality from lung cancer.
He and other experts were also critical of the design of the CT study because it used survival, not mortality, as its main measure.
C. Carl Jaffe, MD, diagnostic imaging branch chief at the National Cancer Institute, said there could be bias if a scan discovers slow-growing disease sooner, rather than later. It may seem that you live longer when actually you don't.
"Just because you pick up something early doesn't necessarily mean you end up with a superior outcome later," he said.
To wit, the March 7, 2007 issue of JAMA included a study of 3,246 asymptomatic current or former smokers who were CT screened and followed for a median of 3.9 years. The screened patients had three times the expected rate of new lung cancer diagnoses, and 10 times the number of lung resections, but there was no significant difference in advanced lung cancer or mortality in screened patients compared with what researchers expected to occur in high-risk populations, based on validated prediction models.
The authors of the study concluded that there is no benefit to CT screening for lung cancer, and that extra surgical treatment prompted by screening can be harmful.
An NCI-sponsored screening trial of more than 50,000 people that is comparing CT scans to conventional X-rays to detect lung cancer may provide more answers. However, those results are not expected for several more years.
The American Cancer Society's Dr. Smith said that for now there is insufficient evidence to recommend widespread screening of people with a history of smoking, but primary care doctors should be prepared to consider the pros and cons of CT scans for any given patient who is seeking to be tested.
Scans experimental, not evidential
Virtual colonoscopy (or CT colonography) has also caught the eye of patients, who may not be eager to have a sigmoidoscope snaked though their colon.
A 2003 NEJM study showed that CT scanning was about as effective as conventional colonoscopy at detecting polyps. But another study published in JAMA in 2004 showed virtual colonoscopy was 39% effective in identifying patients with one or more lesions at least 6 mm in size; and 55% effective with patients with one or more lesions that were at least 10 mm.
"There still needs to be more studies done to show the accuracy," said Esther Wei, ScD, associate epidemiologist at Brigham and Women's Hospital and Harvard Medical School.
"Some research has shown that with very experienced radiologists, these tests are sensitive and specific for finding large polyps and finding colon cancer, but for smaller polyps it's a little less clear," she said. "Smaller polyps, which the CT colonography may not detect, are less likely to turn into cancer, so whether this is a huge disadvantage is still unclear."
Researchers are waiting for results of a large NCI-sponsored study comparing traditional scoping to CT colonography.
And, virtual colonoscopies are not hassle free. Patients still must fast and take strong laxatives. While there is not the risk of bleeding or perforation, CT scans involve radiation exposure. And, if the virtual scan identifies suspicious lesions, the patient still must get a colonoscopy to remove the polyps, and that could mean another day of laxatives.
Conflicting claims in cardiology
There's also a lot of interest in the possible use of MRI to pick up signs of heart disease, in part because the test does not involve radiation or the need for contrast dye, which may be harmful for kidney patients.
Dr. Gibbons, the AHA president, said his organization remains skeptical at this point about the routine use of MR and CT scans to check for narrowed or blocked arteries because the effectiveness of the technologies is not proven. These scans might be useful to evaluate patients in the middle range of risk for heart disease based on Framingham criteria, said Dr. Gibbons, but more studies are needed to see how best the technology can be used.
Warren Manning, FACP, chief of non-invasive cardiac imaging at Beth Israel Deaconess Medical Center in Boston, said, "The concern about widespread utilization of coronary CT is the potential for a large number of false positives in low- and moderate-risk patients, which may lead to these patients then being referred for an invasive angiogram."
Another downside: CT scans expose patients to twice the radiation of conventional coronary angiograms and a similar dose of iodinated contrast.
There are also questions about what to do with the results, noted Dr. Manning. Just because an imaging test shows an artery is narrowed doesn't necessarily mean that the patient's symptoms are from the narrowing, or that the patient would benefit from an intervention.
Although not all experts agree, many said too much evidence is lacking to recommend that physicians start using imaging routinely.
"You have to weigh the implications of screening—the costs, the risks—versus the demonstrated benefits," Dr. Gibbons said. "The major limitation at this time is the absence of any outcome data to show that this approach is going to lead to meaningful benefits."
Contact ACP Internist
Send comments to ACP Internist staff at acpinternist@acponline.org.