We are now three decades into the revolution of molecular biology in medicine, and it has been 15 years since the human genome was cloned. These advances have led to fundamental changes in our understanding of the genetics and biology of cancer, resulting in clinically meaningful advances for patients with cancer, although not as rapid as we would like.
Society and the federal government have long recognized that medicine is not the same as the average consumer market. Therefore, for over 100 years, the Food and Drug Administration (FDA) (or its predecessor), has carefully regulated the manufacture and marketing of therapeutics, insisting that they be both effective and safe. For oncology, accomplishing this goal has been particularly challenging, since the diseases involved are so frightening, and desperate patients have justifiably clamored for new breakthroughs.
On the other hand, because the therapies have mostly been targeted towards mechanisms that are fundamental to human cells, there is only a slight margin to distinguish malignant from normal, and the toxicities have been particularly egregious. Indeed, the FDA often turns to the Oncologic Drugs Advisory Committee (ODAC) to seek advice regarding the relative merits and drawbacks of a new cancer therapeutic agent under consideration.
What about oncologic diagnostics, in particular, in vitro tumor biomarker tests? Again, as the field has gained an increasingly sophisticated understanding of the molecular genesis of cancer, its subsequent behavior, and the tools to measure it, the field of tumor biomarkers has exploded.
However, the regulatory, reimbursement, and scientific environment surrounding this field has not matured in a manner similar to new therapeutic drugs. Therefore, we read about hundreds, if not thousands, of reports of putative biomarkers every year, yet only a handful of truly helpful tumor biomarker tests exist.1
Several colleagues and I have made the argument that, in general, tumor biomarker tests are not valued to the same extent that anti-neoplastic agents are, either by clinicians, regulators, third-party payers, journal editors, or research funders. We have further argued that this lack of relative value has led to a vicious cycle with huge implications, since tumor biomarker tests are increasingly being relied upon to direct cancer therapies (the figure above). Thus, use of a bad tumor biomarker test will result in under- or overtreatment of patients with malignancies.
What Is a Bad Test?
What is a "bad" tumor biomarker test? Or rather, what constitutes a good one? There are three important semantic considerations, first put forward by the Evaluation of Genomic Applications in Practice and Prevention (EGAPP), convened by the Centers for Disease Control and Prevention to help establish evaluation guidelines for clinical genetic tests.3
Like any diagnostic, a tumor biomarker test must have analytical validity -- it must be accurate, precise, reliable, and reproducible. A tumor biomarker test should also have clinical validity, meaning that the test divides one population into two or more, based on either differences in biology or more important, in some sort of clinical outcome.
However, clinical validity does not mean the test should be ordered to care for a patient. Rather, there must be high levels of evidence to demonstrate that for a specific use context, application of the tumor biomarker test is likely to either improve a patient's outcome (overall survival or quality of life), or that the outcome is the same but derived at a lower cost. This latter sentence defines the term clinical utility.
Mechanisms to generate high levels of evidence have been proposed by several authors. As required for approval of a new drug, a prospective randomized clinical trial to "test the test" is ideal. Several trial designs to do so have been recommended.4, 5, 6 However, such trials are expensive and time consuming.
Richard Simon, Soon Paik, and I suggested that a secondary method to generate high levels of evidence would be to perform what we coined a "prospective retrospective" set of studies to determine if a tumor biomarker has clinical utility, using archived specimens from a previously conducted prospective trial that addressed the use context at hand.
Regardless of the strategy, the Institute of Medicine has recommended a road map for getting an omics-based test into the clinic, going from discovery to analytical test generation and validation to determination of clinical utility.7 Although this road map was developed in the context of omics-based tests of any sort, the concepts and principles pertain directly to tumor biomarker tests.
So what's the problem? It is difficult to move from the long-held tradition of conducting studies of convenience, in which a laboratory investigator has a good idea and a clinician has some specimens and some outcomes, to a mind-set of conducting high level, scientifically rigorous studies to determine clinical utility. The former studies are good for getting high-exposure publications and career advancement, but they are not valuable to figure out how to care for patients. However, the latter types of studies are more difficult, and cost more to complete.
Let's return to the vicious cycle.2 First, FDA approval is not required to market a tumor biomarker test in the U.S. A laboratory-developed test (LDT) can be used to direct clinical care, and billed for, as long as it is performed in a laboratory that has been certified by theCenters for Medicare and Medicaid Services (CMS) under the Clinical Laboratory Improvement Amendments (CLIA). CLIA does not address the analytical validity, clinical validity, or clinical utility of a specific test. Its charge is to ensure good laboratory practice, which it does quite well.
Three decades ago, the FDA elected to make an enforcement discretionary decision to permit use of LDTs without FDA clearance or approval. Many tumor biomarker tests are performed by individual, CLIA-approved laboratories and the results provided to clinicians without any evidence to support that the test should be used to direct patient care.
Even if a test is reviewed by the FDA, it is not authorized to insist that application of the assay in the clinic meets the criteria for clinical utility. Furthermore, the review is conducted by the Office of In Vitro Diagnostics (OIVD) in the Center for Devices and Radiological Health (CDRH), not by the Office of Hematology and Oncology Products (OHOP) in the Center for Drug Evaluation and Research (CDER).
Although the FDA does not use the EGAPP terminology, the staff in CDHR is superb at reviewing analytical and clinical validity. However, they are only charged with clearance or approval of the assay if it performs in a manner consistent with the manufacturer's intended use, which may or may not be one of clinical utility. On the other hand, CDER is expertly competent in reviewing the safety and efficacy of new drugs, with the advice of ODAC.
CDER may even consider whether a tumor biomarker developed in concert with a new drug (a so-called "companion diagnostic") is satisfactory for directing care with that drug. However, most new drugs do not have a reliable companion diagnostic at the time a clinical trial is initiated, and most tumor biomarkers are not specifically companion diagnostics linked to a particular drug; hence, CDER is not involved in reviewing stand-alone biomarker tests.
Taken together, these conditions result in an inconsistent regulatory environment in which FDA approval or clearance does not mean a tumor biomarker test has clinical utility, and some of the best tumor biomarkers have never been reviewed by FDA. It further means that there are many LDTs for which it is impossible to determine if they have either analytical validity or clinical utility, and yet they are used every day to direct patient care.
This regulatory inconsistency sets up the second node of the vicious cycle: inconsistent or inadequate reimbursement for a tumor biomarker test. We cannot blame a third-party payer for balking at a high cost for a test that cannot be trusted -- either because it is analytically insufficient or because the data to support clinical utility are poor, or nonexistent. On the other hand, what manufacturer would want to go through the time and expense of generating high levels of evidence for clinical utility if the subsequent payoff is too low to make it worthwhile?
The third node of the vicious cycle rests in the academic and peer-reviewed research funding environment. As for manufacturers, the payoff for academics, in this case in terms of academic promotion and prestige, is higher for new anticancer therapeutics than for new diagnostics. Moreover, the criteria for publication are much lower for tumor biomarker test manuscripts than for therapeutic trials.
No journal would publish a retrospective study in which a few patients, poorly defined, received a drug and seemed to do better. Rather, any reputable journal insists on a detailed methods section, with descriptions of patient eligibility, study conduct, and analytical methods (even to the extent of requiring that the clinical protocol be included in Supplementary Materials).
Yet, in spite of several journals endorsing calls for adherence to rigorous reporting standards (for example, the REMARK criteria)8, 9, 10 manuscripts are published weekly in high-impact journals without any thought to how the specimens were collected, how the patients were chosen, how they were treated, or who did the outcomes assessment. Such sloppiness would not be tolerated for studies of new drugs -- neither is it acceptable for putative tumor biomarker tests.
These circumstances lead to the next node in the vicious cycle: a lack of understanding and trust in the tumor biomarker test results themselves by clinicians and patients. Although this mistrust is hard to document, my own experience is that many clinicians, or their pathology counterparts, order or perform tumor biomarker tests but then ignore the results.
Further, if they look to expert guidelines bodies, such as those assembled by the American Society of Clinical Oncology or the National Comprehensive Cancer Network, they are hard pressed to find evidence-based recommendations for application of tumor biomarker tests, because high levels of evidence do not exist for most. As a consequence, the tumor biomarker tests are not valued, and we re-enter the vicious cycle.
Breaking the Cycle
How do we break this cycle? My colleagues and I have made several transformative recommendations. First, the regulatory environment for tumor biomarker tests needs to be reformed and consolidated, so that there is only one pathway for marketing by eliminating the LDT enforcement discretion rule, and so that the FDA considers both analytical validity and clinical utility.
In fact, on July 31, 2014 the FDA "notified Congress of the agency's intent to issue a draft oversight framework for LDTs based on risk to patients rather than whether they were made by a conventional manufacturer or a single laboratory." This is a major step toward unifying the regulatory requirements for diagnostics in general, and tumor biomarker tests in particular.
Second, the reimbursement community needs to recognize that use of an accurate and reliable tumor biomarker test will lead to true personalized care by getting the right drug to the right patient at the right time -- and will substantially reduce overtreatment of patients with toxic and expensive anti-neoplastic agents. In this regard, reimbursement for tumor biomarker tests should be commensurate with their value. This will incentivize device manufacturers to invest in clinical trials to deliver clinical utility data.
Third, the academic community and journals need to police themselves better, taking on the responsibility of following the IOM roadmap for development, testing, and reporting of tumor biomarker test studies. Finally, clinicians need to have the guidance of expert panels that can rely on high levels of evidence that application of a tumor biomarker test will result in better outcomes for our patients.
The stakes are high -- both in terms of financial and human costs. Our patients deserve better than what we are giving them now. I'll say it again: A bad tumor biomarker test is as bad as a bad drug. If we are going to drive patient care with tumor biomarker tests, we need to recognize that statement and make it our mantra by valuing the tests to the same extent that we do the drugs.
*A use context is the clinical situation in which the tumor biomarker is applied in order to improve the patient's outcome. Examples of use context include risk categorization, screening, differential diagnosis, prognosis, prediction of drug activity, or monitoring disease course.
Daniel F. Hayes, MD, is Clinical Director of Breast Oncology at the University of Michigan in Ann Arbor.
Hayes disclosed personal financial interest in Oncimmune, and Inbiomotion, and holds three patents. He was a consultant/advisory panel participant for Pfizer pablociclib, and was principle or co-investigator on clinical research sponsored by Janssen R&D.
ver historia personal en: www.cerasale.com.ar [dado de baja por la Cancillería Argentina por temas políticos, propio de la censura que rige en nuestro medio]//
weblog.maimonides.edu/farmacia/archives/UM_Informe_Autoevaluacion_FyB.pdf - //
weblog.maimonides.edu/farmacia/archives/0216_Admin_FarmEcon.pdf - //
www.proz.com/kudoz/english_to_spanish/art_literary/523942-key_factors.html - 65k - // www.llave.connmed.com.ar/portalnoticias_vernoticia.php?codigonoticia=17715 // www.frusculleda.com.ar/homepage/espanol/activities_teaching.htm // http://www.on24.com.ar/nota.aspx?idNot=36331 ||