Editor’s note: This post is published in conjunction with the April issue of Health Affairs, which features a series of articles on Alzheimer’s disease.

In September of 2013, CMS issued its final decision memo that concluded positron emission tomography- amyloid beta (PET Aβ) imaging is “not reasonable or necessary”, finding “insufficient evidence” that use of this diagnostic tool would improve health outcomes for patients with dementia or neurodegenerative disease. As such, PET Aβ imaging to help diagnose Alzheimer’s disease (AD) is not a covered service for Medicare beneficiaries except for those enrolled in CMS-approved clinical trials.

CMS’ final decision underscores the emerging new paradigm for coverage decision-making, requiring innovators not only to demonstrate to FDA’s satisfaction that their products are effective, but also to prove to CMS and other payors that their use will improve clinical outcomes.  This paradigm will increase confidence in the value and health benefit of new technologies, although it will make the path to coverage more difficult and uncertain for diagnostic developers.

PET Aβ Coverage

PET Aβ is a relatively new technology that allows doctors to visualize beta-amyloid “plaques” in the brain. These plaques, which are comprised of beta-amyloid protein and brain cells, are found in patients with AD and other conditions, as well as in older people with normal cognition.

To visualize these plaques, doctors inject a radioactive tracer drug intravenously, which binds with beta-amyloid plaques; the radioactive signal is detected by a PET scanner that is able to produce images of the deposits.  Florbetapir, the first amyloid-specific tracer approved by FDA, was developed by Eli Lilly, which formally requested CMS coverage of PET Aβ imaging in July 2012.

Subsequently, two more amyloid-specific tracers have been approved: flutemetamol, sponsored by General Electric (approved in October 2013) and florbetaben, sponsored by Piramal (approved in March 2014).  Navidea Biopharmaceuticals is evaluating a fourth potential tracer, NAV4694, through late stage clinical trials involving patients with mild cognitive impairment and AD.

Following a spate of failed trials of promising drug candidates, CMS’ non-coverage decision for PET Aβ was a major disappointment for the Alzheimer’s community.  More than 200 comments, mostly in opposition, were submitted in response to CMS’ 2013 proposed decision memo, representing recognized scientific and medical experts in the field as well as advocates, affected individuals and their caregivers, and health professionals.

Yet, as evidenced by the deliberations of CMS’ Medicare Evidence Development and Coverage Advisory Committee (MEDCAC) meeting convened on the role of PET Aβ imaging in dementia and neurodegenerative disease, there was another camp of equally vocal supporters advising CMS to not “let the genie out of the bottle—again”.  Fundamentally, the coverage debate centers on if, how and to what extent this new technology positively impacts health outcomes for Medicare beneficiaries.

Evaluation Of Health Benefits

As detailed extensively throughout the final decision memo, CMS expects rigorous studies for PET Aβ imaging that document significant, quantifiable improvements in health outcomes, which may include changes in physician management that lead to such improvements.

Lilly demonstrated quite capably that PET Aβ accurately and reliably performs as intended, namely, measuring amyloid plaque density in the brain with high sensitivity and specificity.  However, the sponsor was handicapped in at least three ways from quantifying health outcomes to CMS’ satisfaction.

First, there is ongoing debate about the meaning and value of the amyloid biomarker, particularly the relationship between accumulation of beta-amyloid and cognitive decline.  Experts generally agree that the absence of amyloid plaques in the brain allows providers to rule out Alzheimer’s disease as a cause of cognitive impairment. As stated in florbetapir’s label, a negative study is “inconsistent with the diagnosis of AD.”  Thus, an individual with a negative PET Aβ scan likely does not have AD.

The presence of amyloid plaques alone, however, is not sufficient to rule in AD because approximately one-third of older Americans with normal cognition have amyloid deposits in their brains.  However, an amyloid positive scan for a patient with dementia in the appropriate clinical setting likely signals AD.

As such, PET Aβ imaging could be a useful addition to the evaluation of patients with cognitive impairment of unknown etiology, helping providers distinguish between common causes of dementia and to tailor their clinical approach accordingly.

Therein lies the second handicap, however, as at the time of CMS’ decision-making, there was only one study focusing on how PET Aβ imaging could change physician management.  This study found that the scan results changed the overall management plan for 86 percent of patients, including changes in intended medication management.

Although these results in a hypothetical context are helpful, CMS wants proof of  actual— as opposed to intended—changes in management, along with long-term follow up to assess the impact of such changes on patient outcomes.  To that end, an industry-sponsored clinical trial is underway that will assess actual changes in patient management as well as the association between scan status and cognitive decline.  Results from this trial are expected in December 2014.

Third, compounding the difficulty of quantifying improvements in health outcomes, no disease-modifying drugs exist for AD; current drug therapies manage symptoms of the disease and may slow progression but cannot prevent or cure AD.  The lack of effective treatments makes it more difficult to show that testing improves patient outcomes, if outcomes are measured only in terms of absolute reduction of cognitive or functional impairment.

Medical providers would argue that determining prognosis, care planning and caregiver education all improve health outcomes as well, particularly for patients with mild cognitive impairment.  CMS certainly does not disagree, but notes that items and services to address these societal concerns would not be covered benefits and thus, these concerns are “not informative” to its coverage decision.

Appropriate Use Criteria For PET Aβ

In light of the questions regarding use of PET Aβ imaging, the Amyloid Imaging Taskforce of the Society of Nuclear Medicine and Molecular Imaging and the Alzheimer’s Association developed appropriate use criteria (AUC) for PET Aβ to provide guidance to health care providers.   The Taskforce employed a rigorous and deliberative process to identify potential indications for PET Aβ imaging, assess and rate the evidence for such indications, and finally, vote on AUC.

Consensus was reached that PET Aβ is beneficial in diagnosing and managing patients with persistent or progressive unexplained mild cognitive impairment and those with an unclear clinical presentation for AD. Additionally, a majority of members voted in favor of the AUC including patients with progressive dementia and atypically early age of onset.

Other indications, such as screening patients with a family history of AD or to determine disease severity, were rejected.  Benefits of testing in these targeted populations were defined as a change in medication management, a change in ordering other tests and the value of knowing for patients and their caregivers.

Notably, the Taskforce recommendations call for more clinical use data from patients likely to be encountered in practice as opposed to clinical trial settings, namely patients with multiple medical conditions, complicated medical histories and atypical features of dementia.

However, despite limited data, the Taskforce supports use of PET Aβ imaging for select patients in some clinical scenarios.  CMS’ assessment of PET Aβ acknowledged the Taskforce’s recommendations; however, CMS reached a different conclusion, citing these gaps in evidence as part of its rationale for Coverage with Evidence Development (CED).

Comparative Effectiveness

Concerns about the impact of PET Aβ imaging on health outcomes likely were not the only barrier to full coverage; the role of comparator products in influencing coverage decision-making by CMS has been questioned generally.  Of relevance to PET Aβ, Medicare currently covers PET scans using an older tracer, fluorodeoxyglucose-18 (FDG), to help providers distinguish between AD and frontotemporal dementia in patients who meet the diagnostic criteria for both.

Medications for patients with AD offer no benefit and may cause harm for those with frontotemporal dementia.   The question was asked whether and why a different, and possibly higher, standard was being applied to PET Aβ coverage for a comparable indication.

In response, CMS asserts that the tracers are not comparable—FDG allows assessment of brain function and is not disease-specific—and focuses again on the need for more health outcomes data for PET Aβ.  One might speculate that CMS will want to know eventually what incremental value beta-amyloid imaging could provide.

Given at least one study suggests that PET Aβ and FDG PET results correlate closely, as a newer and more expensive test, PET Aβ may have to meet a higher bar for coverage.  Future studies may also compare PET Aβ to other diagnostic tests, alone or in combination, such as tests measuring tau protein in cerebrospinal fluid or magnetic resonance imaging.

Of note, as per the 2012 draft CED guidance, even older tests may be re-evaluated, and thus data on the role of PET Aβ imaging could have broader ramifications than intended.

Moving Forward

Despite the disagreement on the coverage decision, all parties agree that additional trials will be useful, albeit difficult to meet CMS’ requirements and costly for sponsors.  Yet, that has not dissuaded the AD community from moving forward.  The Amyloid Imaging Coverage with Evidence Development (AICED) working group and approximately half a dozen academic medical centers are drafting clinical trial proposals to meet CMS’ requirements.

Study objectives must be to (1) develop better treatments or prevention strategies for AD, or, as a strategy to identify subpopulations at risk for developing AD, or (2) to resolve clinically difficult differential diagnoses where the use of PET Aβ imaging appears to improve health outcomes. Collectively, these trials will inform and hopefully end the debate about whether PET Aβ imaging leads to improved health outcomes and for whom, how and why.

Importantly, CMS raises the urgent need for trials to enroll diverse participants.  It is unfortunate that current trials have largely included white subjects only, given African Americans are 100 percent and Hispanics are 50 percent more likely to develop AD. Minority populations will disproportionately benefit from new technologies and treatments, and CMS is right to highlight this disparity.


Fundamentally, CMS wants a test that can help to diagnose or predict onset of AD with certainty and improve health outcomes.  Unfortunately, evidence demonstrating patient benefit for PET Aβ imaging is incomplete, and many believe that CMS’ decision to limit coverage to participants in clinical trials was not entirely unexpected or unfounded.

At the same time, a lot of experts agree that there is clear benefit to PET Aβ imaging in certain populations and clinical contexts. In practice, this test will be used, but outside of clinical trials, rationed to those who can afford it.

The Obama Administration has invested significant new resources in research, education and awareness, and community services for patients with AD and their caregivers.  These are critical front-end supports; however, getting new and effective products that prevent, diagnose and treat AD to the market place is the Holy Grail for patient and providers.

To that end, this Administration must support and reward innovation by providing a clear path to coverage and reimbursement for novel products.  Important steps include better defining appropriate clinical outcomes for diagnostic interventions and the types of data needed to support coverage decisions. Additionally, CMS and other payors should clarify how comparative effectiveness research and real world evidence will be integrated into coverage decision-making.

For its part, industry will need to prepare earlier and more strategically to make its case for coverage and reimbursement by conducting the range of studies needed to address questions about the efficacy and effectiveness of their products, and most importantly, the health impact.  Patient advocacy groups and providers have an increasingly critical role to play as well.

As in other areas of health policy, health care decisions are too important to be left to payors alone, but only if patients and providers who care for them are active and engaged will their voices be heard by policy makers.  In this regard, the Alzheimer’s powerful community of stakeholders must be commended for its ongoing commitment to expanding research and improving care for the five million Americans and their families living with AD.