AMP report defines a spectrum of utilities for molecular testing

Interview by Steve Halasey

Stakeholders have long debated the nature of clinical utility when applied to diagnostic testing—and especially to molecular diagnostics, with their potential to generate a wide range of ancillary data points whose clinical significance may be unknown. To resolve some of the key issues in this area, in 2014 the Association for Molecular Pathology (AMP) convened a task force on the framework for the evidence needed to demonstrate clinical utility (FEND).

The first fruits of the task force’s labor appear in a special report in the September 2016 issue of the Journal of Molecular Diagnostics.1 The report acknowledges the importance of devising appropriate methods for assessing the clinical utility of molecular diagnostics, especially in a time of change from methods-based criteria to value-based criteria. “Levels and types of evidence to properly evaluate the clinical value of molecular diagnostics merit discussions to standardize criteria,” writes the task force, “because the growing body of knowledge from genomic medicine provides many new opportunities for molecular testing to improve healthcare.” Specific recommendations of the task force are grouped under five main headings:

  • Promote patient-centered definitions of clinical utility.
  • Utilize a modified ACCE model incorporating aspects of clinical utility beyond drug selection.
  • Support multiple modalities of clinical utility evidence generation.
  • Develop professional organization–driven practice guidelines.
  • Recognize the critical role of the molecular professional in patient care.
Elaine Lyon, PhD, ARUP Laboratories.

Elaine Lyon, PhD, ARUP Laboratories.

To find out more about AMP’s recommendations for defining clinical utility as it applies to molecular diagnostics, CLP recently spoke with Elaine Lyon, PhD, medical director of genetics, genomics, and pharmacogenomics at ARUP Laboratories; cochair of the FEND task force; and senior author of the task force’s report.

CLP: The issue of clinical utility has bedeviled payors and test developers for many years. How would you describe the key problems that the clinical lab community has been wrestling with?

Elaine Lyon, PhD: One of the most important issues we’re dealing with is the ongoing shift of payor expectations from the statutory requirement that test procedures be “reasonable and necessary” to a somewhat ill-defined requirement that procedures should have demonstrable clinical utility—together with an expectation that a procedure’s utility will be supported by evidence of improved patient outcomes. Those new expectations rise to a level that laboratories haven’t previously been asked to meet, and in many ways it’s a level that labs can’t meet. Many of the molecular diagnostic procedures that AMP members perform are for diseases so rare that a statistically valid study would be almost impossible, or would at least take many years to complete.

AMP members are fully embedded in the field, and they know very well how molecular tests are being used—which is not always for diagnosis alone. So, in dealing with payors, we need to ensure that their expectations about clinical utility are expanded to include the full spectrum of applications for which molecular diagnostics are really being used.

CLP: Payors have frequently tried to use evidence about the analytical validity and clinical validity of an assay as a litmus test for determining whether the assay meets the “reasonable and necessary” criteria for coverage and payment. What’s wrong with that logic?

Lyon: The test procedures performed by clinical laboratories have always been judged against standards of analytical validity and clinical validity, and labs must meet those standards in order to maintain their certification. Analytical validity is reviewed under the requirements of the Clinical Laboratory Improvement Amendments of 1988 (CLIA), and clinical validity is addressed as part of the lab accreditation process handled by the College of American Pathologists and other accreditation agencies.

The challenge arises when payors try to use evidence about the analytical validity of a test to determine whether the test should be considered “reasonable and necessary,” even though, in reality, evidence about analytical validity can only reveal how well the test is performing in the laboratory. And it’s equally inappropriate to try to use data about clinical validity to meet a set of criteria that constitute an expanded standard for clinical utility. These approaches place laboratories in an impossible position for trying to demonstrate the clinical utility of their tests.

These categories of evidence were really created with therapeutic products in mind, and they enable sponsors to demonstrate the utility of their products by showing that they contribute to improved patient outcomes. But it’s difficult to apply those same standards to a test whose sole purpose is to establish a diagnosis for an inherited condition. Even if the test performs this function flawlessly, it may never have any effect on improving the patient’s health.

CLP: Evidence of clinical utility is sought for a wide range of medical products, including all types of in vitro diagnostics. How are the issues of validity and utility different when molecular technologies are involved?

Lyon: A laboratory test is an intermediate step whose ultimate effect depends on how the clinician uses the information. Consequently, it is hard for laboratories to demonstrate clinical utility. Clinicians may use the exact same information and test results in a variety of different ways, depending on the reason for testing. The test used to confirm the diagnosis of a disease in an affected individual can be the same as the test used to determine whether a person is likely a carrier of the disease but not necessarily affected by it.

So clinicians who order a test may use it differently for different scenarios, and they may interpret the test results differently, depending on the scenario. It’s insufficient to define a test in terms of a single specific utility, when it may actually have many different utilities based on the reason for testing and what the physician expects to do with the information.

CLP: Previous efforts to create a systematic process for evaluating molecular diagnostics have included the CDC-supported ACCE model project (named for its key criteria: analytic validity, clinical validity, clinical utility, and associated ethical, legal, and social implications), and its outgrowth initiative for the evaluation of genomic applications in practice and prevention (EGAPP). Where do the FEND recommendations fit in relation to such earlier efforts?

Lyon: I see the FEND recommendations as building on those earlier efforts. We began with the ACCE model, and introduced modifications to address the issues of analytical validity and clinical validity in line with the way they are handled through CLIA, CAP, and other accreditation organizations. Then, looking at the ethical, legal, and social implications of the tests, we attempted to give these areas better definition by way of more examples of how to use these types of tests.

When attempting to satisfy CDC’s EGAPP criteria, EGAPP reports often reach the common conclusion that there isn’t enough evidence. That doesn’t mean that there is evidence against the use of the test, it simply means that there’s not enough evidence. We recognized that this is a common problem—rooted partly in misguided efforts to generate evidence using the traditional randomized controlled trials that drug and therapeutic product manufacturers have been used to—and therefore proposed to try to find different ways to demonstrate the effectiveness and utility of clinical testing procedures.

CLP: The FEND report defines clinical utility for molecular diagnostics as “the ability of a test result to provide information to the patient, physician, and payor related to the care of the patient and his or her family members to diagnosis, monitor, prognosticate, or predict disease progression, and to inform treatment and reproductive decisions.” How does this definition differ from previous approaches to define clinical utility?

Lyon: It is a broader definition that takes into account the different ways that a test might be used, with an understanding that the test results may indicate follow-on activities that are not as simple as merely determining which drug to prescribe, or at what dose. In the case of inherited conditions, for example, we’re talking about management of the patient within a family context. When clinicians already know the natural history and course of a disease’s progression, for instance, an appropriate diagnosis may enable the clinician to predict future symptoms for which the patient and other family members should be monitored. That way, as soon as any symptoms of disease progression are detected, clinicians are already prepared to take appropriate actions to slow or halt the disease.

One of the major differences in the FEND definition is that it puts patient utility first, and emphasizes the role of testing for determining what needs to be done for the patient. Many times, the test results are important for informing other decisions that a patient may need to consider, such as reproductive decisions, how to manage care for other family members who may be affected, or even lifetime care decisions. There are more ways a patient may use the information than even what clinicians or payors may consider important.

The definition is also different in that we’ve tried to be very inclusive of all stakeholders. So, starting with the patient, the definition also addresses the needs of physicians, payors, and society at large, with the last of these including appropriate allocation of healthcare resources. So this definition tries to be inclusive and to look at how all stakeholders need to have things defined.

CLP: The FEND report observes that the clinical utility of molecular diagnostics is context-dependent. How do the varied applications of molecular diagnostics—for diagnosis, prognosis, and prediction—contribute to framing the context for such testing?

Lyon: Everything starts with an accurate diagnosis. A patient may exhibit clinical symptoms that permit a clinician to say that a patient’s condition falls within a particular group of diseases. But to identify the disease precisely—to know its inheritance pattern, what the predictions for disease progress might be, and so on—the clinician first needs to have an accurate diagnosis.

Some payors believe that it’s not necessary to use a molecular test to manage patients when they are only managing symptoms. They may suggest waiting until the parents want to have another child, and then testing the parents. But it’s very important to do appropriate testing.

Without testing an affected child, for instance, it may not be possible to interpret the parents’ test results correctly. In some cases, we can perform better and less expensive testing on the parents if we’ve first been able to confirm the genetic variants occurring in an affected child. So in cases such as this, the types of testing that need to be considered should be expanded beyond those that might be needed just for the treatment of an affected child.

CLP: The FEND report offers five top-level recommendations for assessing the clinical utility of molecular diagnostics. What is the most important change that would result from the adoption of each of those recommendations?

Lyon: I have great hope that these recommendations will help to produce better healthcare for all of us—identifying patients and their conditions, giving answers to patients, helping practitioners know what to expect and how to manage patients, making it possible to select appropriate procedures and to spare patients from unnecessary additional procedures.

The recommendations should also help payors improve their use of healthcare resources. When molecular testing results are unavailable to guide physicians toward accurate diagnoses, they must often implement therapies that are not focused on the specific needs of a patient—using broad-based chemotherapies rather than agents targeted for the specifics of a patient’s condition, for example. But having accurate molecular test results enables clinicians and payors to preserve specialized therapies for the patients that truly need and will benefit from them.

CLP: The FEND report focuses on the clinical utility of molecular testing for inherited conditions and cancer. Why were these two areas selected and what different lessons do they provide?

Lyon: These two areas were selected because many AMP members are dealing with them directly. The American Medical Association’s current procedural terminology (CPT) codes for molecular diagnostics, implemented in 2013, are more transparent than the previously used methods-based codes. But that has led payors to question the utility of many molecular tests—and especially those for inherited conditions and cancer. AMP responded to such questioning because this was of great concern to many AMP members.

CLP: The report also notes that its recommendations can be extended to applications of molecular testing beyond inherited conditions and cancer. Are there any applications where special procedures would be needed to use the report’s recommendations?

Lyon: The broad principles apply to any diagnostics. If a lab wanted to extend application of the FEND recommendations from molecular testing to biochemical testing, and specifically for rare diseases, the same principles would likely apply. But application of the recommendations would need to be reviewed with such a specific context in mind.

CLP: How important is the use of multigene panel tests for ending the one-test-at-a-time diagnostic odyssey described in the FEND report?

Lyon: This practice is becoming very important, and it is gaining wider acceptance as we are able to demonstrate its capacity to pick up and confirm the diagnosis of individuals who would otherwise have continued on the diagnostic odyssey. Multigene panel tests are particularly useful when a patient’s symptoms overlap with a number of different diseases. Instead of testing one at a time, which prolongs the time to a diagnosis, clinicians can get to an answer and more quickly begin managing the patient appropriately.

The situation is a little different for cancer diagnostics, because biopsies are becoming smaller and molecular pathologists are trying to work with less tissue. Laboratories frequently receive just enough material to perform two or perhaps three tests. If they can’t find an answer with those first three tests, they may be able to provide only an incomplete answer—unless more material from another biopsy can be made available. So, in addition to the clinical utility aspect of such tests, there’s also a practical aspect.

CLP: The progress of liquid biopsy technologies has been quite striking in the past year. How does that new diagnostic approach fit with the recommendations of the FEND report.

Lyon: With this report, we wanted to develop a set of recommendations that could be extended to all additional applications of molecular testing. I believe it would fit well the report’s recommendations.

CLP: What will it take for the recommendations of the FEND report to be adopted by healthcare providers and the payor community? What obstacles do you foresee?

Lyon: AMP has undertaken this project in order to engage all of the stakeholders in molecular testing for inherited conditions and cancer, which is a very important goal. We need to educate patients, providers, and payors, but we also need to learn from them. The need for education runs both ways. And then, we need to advocate for the best interests of the patient.

One of the key obstacles to the adoption of the FEND recommendations is simply that there are so many points of view, and trying to be all-inclusive can sometimes be very difficult. AMP’s goal is to make this important testing accessible to all who need it. By doing so, we believe that we will have appropriate testing and follow-up, and overall be able to improve healthcare for everyone.

CLP: Are there other points that our readers should take away from looking at the FEND report?

Lyon: Even with considering only inherited conditions and cancer, the issues involved in the use of molecular testing are extremely varied. Using a molecular test to arrive at an appropriate diagnosis that can help to predict patient outcomes is very different from using a molecular test as a companion diagnostic in order to select an appropriate therapy.

One of the reasons the field has moved so fast is the fact that molecular techniques are now fairly standard. Once a gene or gene variant is demonstrated to be clinically valid—meaning that it is causative for a particular disease—laboratories can move very quickly to implement appropriate testing.

For example, mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene are known to be causative for cystic fibrosis. Consequently, laboratories can move very quickly to implement molecular testing for appropriate patients. But they may not be able to provide the types of evidence that payors want to justify the use of such testing. Assessing the clinical utility of such testing isn’t always possible through the use of methods intended to study the utility of therapeutic products.

The issues that arise when selecting and using molecular diagnostics for different inherited conditions and types of cancer can be very different from one another, and they all need to be addressed. The FEND report provides a broad overview of some of the ways that we need to start looking at these issues. Hopefully it will enable the laboratory community to engage patients and to keep the discussions among stakeholders moving, while also providing a platform for advocacy on behalf of patient access to appropriate testing.

CLP: Will this be a major topic of discussion at the upcoming AMP annual meeting?

Lyon: In some way or another, it surely will be. Engaging with payors; assessing the value of molecular diagnostics; providing high-quality, medically useful testing; and developing the critical role of molecular pathologists are all high-level topics that are woven through sessions at the annual meeting. The FEND recommendations may not be the subject of a specific session, but it’s a fair certainty that various aspects of the recommendations will be addressed during the meeting. We will also be hosting several upcoming webinars. More information can be found on AMP’s Web site.

Steve Halasey is chief editor of CLP.


  1. Special article: the spectrum of clinical utilities in molecular pathology testing procedures for inherited conditions and cancer; a report of the Association for Molecular Pathology. J Mol Diagn. 2016;18(5):605–619; doi: 10.1016/j.jmoldx.2016.05.007.