How will Comparative Effectiveness Research Impact Molecular Imaging?

Last year, the SNM received a $48,000 grant from the Agency for Healthcare Research and Quality (AHRQ) to develop comparative-effectiveness research (CER) of PET and other molecular imaging techniques. The primary emphasis is on the diagnosis and management of cancer patients, but both cardiology and neurology questions are being addressed. Far beyond the dollars, too, is a significant increase in intellectual capital being expended across the globe on the role of CER in molecular imaging.

Imaging presents challenges

In June 2009, 62 medical organizations signed a letter addressed to the U.S. Senate commending it for allocating more than $1 billion for comparative effectiveness research under the American Recovery and Reinvestment Act of 2009. Their primary concern in sending the letter was to ensure that CER money continues to be a "key component of any strategy designed to reform healthcare."

The signees wrote that a "federally sponsored, independent CER enterprise" would ensure that physicians deliver the most effective and appropriate care, particularly in light of the proliferation of studies "across a multitude of journals."

"In theory, comparative effectiveness research will tell you which set of procedures or which set of imaging tests is the one that will give you the best results for a given disease," says Robert Henkin, MD, professor emeritus of radiology at Loyola University Stritch School of Medicine in Chicago. "The long-term aim is to help limit the number of imaging exams done that are inappropriate."

Henkin is in charge of the evidence-based guidelines development section within the SNM. Conducting CER for imaging modalities presents challenges, he says. "Even when a well-designed trial delivers an answer about one imaging method over another, how can we ensure that those same protocols will be followed at every institution, and even within institutions?"

When drugs are tested in clinical trials and one is found to be superior to another, the answer is straightforward: the physician writes the script, the patient takes the drug. Not so with imaging. "Most imaging trials do not take into account the wide variation in imaging quality or in the experience of the person interpreting the tests," he says.

Henkin, like others, understands that medicine is moving in the direction of CER. The government wants to know how various tests, in this case, imaging exams, compare against other imaging exams in terms of outcomes. Does the use of coronary CT angiography (CCTA), for example, lead to better outcomes in acute chest pain patients compared with SPECT myocardial perfusion imaging? SPECT has a long history of validation in the literature, whereas CCTA's validation is in its infancy. But even with SPECT's impressive presence in the literature, it needs to be comparatively tested in today's healthcare environment.

"The evidence bar has been raised for cardiac CT," says U. Joseph Schoepf, MD, director of CT research and development at the Medical University of South Carolina in Charleston. "The other tests were integrated into practice at a stage when comparative effectiveness research was not required. But all medical tests should be held to the same efficacy standards."

CER trials

Example of parametric 11C-PIB (A) and 18F-FDDNP (B) BPND images in healthy control and AD patient. 11C-PIB and 18F-FDDNP scans were acquired in same subjects. In each panel, control is on the left and AD patient is on the right. High level of 18F-FDDNP binding in subcortical structures suggests nonspecific binding. (J Nucl Med. Tolboom et al. 50 (2): 191)

In that regard, several high-profile CER trials pitting molecular imaging against anatomic CT have been initiated. The RESCUE trial, begun in February, will randomize 4,300 stable angina patients to receive either CCTA or SPECT at 80 facilities, and will follow them for two years to determine if CCTA yields similar results at a reduced cost as SPECT. It is funded by AHRQ.

The PROMISE trial, sponsored by the National Heart, Lung and Blood Institute, will compare CCTA with functional stress testing such as exercise ECG, echocardiography and SPECT for patients at low to intermediate risk of CAD who present with chest pain. The $32 million trial expects to enroll 10,000 subjects from 200 sites and follow them for up to two years. The hypothesis is that information derived with anatomic testing will drive superior outcomes than information derived from functional testing.

While there are ample data on the prognostic value of SPECT and emerging promising prognostic evidence for CCTA, "prognosis is not the same as outcomes," says Schoepf. "You can't mix the two. With all that has been written about SPECT, there actually is very little in terms of outcomes."

This goes back to the subjective nature of imaging as one of the challenges in doing CER. The bottom line, however, is that this type of evidence will be increasingly required as a stepping stone to reimbursement and organizations such as SNM have begun to evaluate what it takes to do effective CER.

The evidence-based guidelines development group within the SNM will be charged with, for example, making sure journal reviewers understand how to review journal articles for scientific method—"being able to distinguish between a paper that has good evidence vs. one that has pretty pictures," Henkin says. The group also will address how to teach residents in training about developing and using evidence and will work closely with clinical trials groups regarding trial structure so that the evidence gained is high quality.

Neurology CER

A diagnosis of Alzheimer's disease is more accurately made with PET than without it, says Daniel Silverman, MD, PhD, head of the Neuronuclear Imaging section at UCLA Medical Center and associate professor of molecular and medical pharmacology at the David Geffen School of Medicine at UCLA in Los Angeles. In fact, a 2004 statement by AHRQ said as much. The relevant question, however, is "does PET benefit patients in terms of clinical outcomes?"  

To help provide answers, the Centers for Medicare & Medicaid Services (CMS) is funding the Metabolic Cerebral Imaging in Incipient Dementia (MCI-ID) trial, a randomized prospective study to test the value of PET via short- and long-term outcomes vs. MRI.

In MCI-ID, 700 patients from 50 sites will receive a PET and MRI scan, as well as neurocognitive testing. The MRI information will be released to the referring physicians at the time it is acquired. The PET information, however, will be randomized to be released at the time it is acquired or two years later. The main outcome measure is cognitive ability. The major hypotheses are that among patients whose PET results are immediately conveyed to their referring physicians, diagnoses and management plans will be positively affected, leading to more effective utilization of healthcare resources and to maintenance of cognitive and functional abilities at a higher level.

"Comparative effectiveness research is the ultimate challenge," says Silverman. "What we want in comparative effectiveness research is a real-world look at the outcomes that takes into consideration all the things that can go wrong, such as miscommunication between physicians and patients or patient compliance with medication. We want to know the impact these variables have, the net benefit to the patient."

There are data that show medication can slow the progression of Alzheimer's in those with very early signs of the disease. Not only does early intervention benefit the quality of life of the patient (and even family), but it reduces economic stress on the healthcare system, such as repeated MRI scans, time and money spent on ineffective medications, and nursing home stays, Silverman says. "With PET imaging, we can see changes in the brain long before dementia takes hold. Comparative effectiveness research will take us to the next level we need to go. Can we positively impact outcomes? It's the way all of medicine is—and should be—going." Stay tuned.