Companies focusing on IVDs are regularly confronted with the need to make critical strategic decisions regarding how they commercialize their test. The novelty of the technology and its intended use are just some of the factors that govern what kind of application a firm can file to gain approval.
With the overlap of pharmaceutical R&D and diagnostic development becoming greater, drug companies must be aware of these options as well. The wrong decision can have a dramatic effect on the time to market, cost, and commercial success.
In the July issue of GEN, the article titled “Primer on Regulation of Diagnostic Assays” gave an overview on what governs IVD regulation. It also introduced the many regulatory paths a company can take, such as filing a 510(k), a premarket approval application (PMA), or for de novo reclassification. This article details what each of these options means and requires.
510(k) Premarket Notification
The principal means by which diagnostic devices obtain FDA marketing authorization is through a 510(k) premarket notification. Approximately 98% of all new devices enter the market through this route. In fiscal year 2005, the Office of In Vitro Devices (OIVD) with in the FDA cleared 520 510(k)s and approved six PMAs (OIVD—Annual Report FY 2005).
Under the 510(k) process, the company must show that the product is substantially equivalent to a predicate device. A predicate device is either be a product marketed before May 28, 1976 (the effective date of the Medical Device Amendments of 1976 (MDA)) or a device that has been 510(k)-cleared.
This requirement of showing substantial equivalence has led to the misperception that a 510(k) is similar to an ANDA. While some variation between an ANDA and the pioneer drug is permitted, the differences must be relatively minor. For example, an ANDA would not be approved for a new chemical even if it had the same intended use as the reference listed drug.
In contrast, a 510(k) could be cleared for a product using a novel diagnostic methodology that has the same intended use as the predicate device. For instance, if FDA sanctioned a mAb test for a particular intended use, a company may be able to use that product as a predicate device for a molecular diagnostic with the same proposed use.
When a device presents such technology differences, then the FDA requires appropriate clinical or scientific data demonstrating that the device is as safe and effective as a legally marketed device (21 USC 360c(i)(1)(A)(ii)). This may require both preclinical and clinical studies.
A new technology does not necessarily preclude the 510(k) mechanism. Underscoring the flexibility of this process, it may even be possible to claim substantial equivalence to a well-established laboratory method that has not undergone FDA review.
Conversely, a company seeking clearance of an old technology for a new intended use may not be able to use the 510(k) process. For example, a PSA test intended to monitor patients who have been diagnosed with prostate cancer can utilize the 510(k) process, while a PSA test intended for screening must go through the more cumbersome PMA process.
It is not always clear whether a diagnostic test intended to be used in conjunction with a drug will be eligible for a 510(k). The assay for the UGT1A1 gene, which is used in conjunction with the drug Camptosar, was cleared via a 510(k). On the other hand, the assays for HER2/neu went through the PMA process. FDA used de novo reclassification to clear the MammaPrint gene-expression profiling test for use in predicting whether existing breast cancer will metastasize. The corresponding Special Controls Guidance Document, though, specifically says “[a] gene expression profiling test system … is not intended … to select the optimal therapy for patients” (FDA—Guidance for Industry and FDA Staff, Class II Special Controls Guidance Document: Gene Expression Profiling Test System for Breast Cancer Prognosis. May 2007:4–5).
Another important aspect of the 510(k) process is that the applicant does not need to show safety and effectiveness, per se. Rather, the goal is to demonstrate substantial equivalence to the predicate device. Sometimes, this seems to be a distinction without a difference, since the company still has to prove factors like clinical sensitivity and specificity.
Once a 510(k) is submitted, FDA will review the application within 90 days. The agency can find the device substantially equivalent, allowing it to be marketed; not substantially equivalent (NSE), in effect rejected; or ask for more information. A 510(k) contains no manufacturing section and there are no preapproval inspections. Additionally, the application for an IVD will typically contain test data but dramatically less than in an NDA or PMA.
Once a 510(k) is cleared, the 510(k)-holder will need to submit a new 510(k) if it makes a significant change in intended use or a change that could affect the product’s safety or effectiveness (21 CFR 807.81(a)(3); FDA—Deciding When to Submit a 510(k) for a Change to an Existing Device. January 1997). Thus, a company that receives 510(k) clearance has more latitude than an NDA-holder or PMA-holder in modifying and updating its product and labeling.
Premarket Approval Application
The primary alternative for IVDs is the PMA. IVDs that present higher risk, like cancer screening tests, or those for which there are no predicate devices, will often need to follow the PMA process.
The PMA pathway is more complex than the 510(k) route. The application needs to provide a reasonable assurance that the device is safe and effective for its intended use (21 USC 360e(d)(1)(A)). To do this, companies must submit valid scientific evidence (21 CFR 860.7(c)), typically derived from a clinical study.
The application also needs to include documents demonstrating compliance with the device GMP regulation (21 CFR Part 820). The PMA must also include draft labeling, information regarding product composition, preclinical data, and a summary of relevant literature (21 CFR 814.20(b)).
Once a PMA is submitted, FDA will reach a filing decision within 45 days (21 CFR 814.42(a)). The agency then has 180 days in which to complete the review. If FDA identifies major deficiencies, the 180-day review clock starts again.
Before approving a PMA, FDA will generally conduct a preapproval inspection to evaluate GMP compliance. The agency will also routinely perform bioresearch monitoring visits of one or more clinical sites. Novel IVD PMAs typically will be reviewed by an advisory panel.
When Possible, Use 510(k)s
These various procedural steps mean that the PMA process is far longer than 510(k)s. In the most recent OIVD statistics, the average IVD 510(k) was reviewed in 88 days. For PMAs, the average was 327 (17).
PMAs provide less flexibility than 510(k)’s once approval has been granted. Some product or manufacturing modifications will require approval of a PMA supplement before being implemented. Other proposed changes will need to be the subject of a 30-day notice regarding changes being effected. There are other categories of supplements as well (FDA—Draft Guidance for Industry and FDA Staff, Modifications to Devices Subject to Premarket Approval (PMA) – The PMA Supplement Decision-Making Process. March 2007). In addition, PMA holders, unlike 510(k) holders, must submit annual reports.
Given the choice, IVD companies will almost always prefer submitting a 510(k) to a PMA. The latter application can result in patent extensions and product-liability protection through pre-emption (Riegel v Medtronic, Inc., 451 F3d 104 (2nd Cir 2006), appeal docketed, No. 06-179 (U.S. August 4, 2006)) as well as provide a regulatory barrier to competitors. Even so, the 510(k)’s advantages ordinarily offset these benefits.
Sometimes, however, there is no choice. If a particular device for a particular intended use has already been approved through a PMA, then the next applicant will also need to submit a PMA, unless it can have the product “downclassified” (21 CFR Part 860), which could be a long process itself.
Some novel products will need to go through a PMA, because the intended use is deemed too risky for a 510(k). The level of risk sometimes can be mitigated by modifying the intended use, e.g., claiming the test is to be used adjunctively rather than to determine which therapeutic intervention to employ.
And sometimes the 510(k) route is foreclosed due to lack of a predicate device. In that case, the company needs to submit a PMA, unless it can utilize the de novo route.
De Novo Classification
Under the MDA, if a device were found NSE simply for lack of a predicate device, the applicant would have to file a PMA. Sometimes, the cost and time involved are too great to justify submitting a PMA, effectively keeping the product from entering the market.
To address this flaw in the statute, in 1997 Congress created the de novo classification process. This procedure essentially allows a low- or moderate-risk IVD or other diagnostic for which there is no predicate device to bypass the PMA process (FDA—New Section 513(f)(2) – Evaluation of Automatic Class III Designation, Guidance for Industry and CDRH Staff. February 1998).
The de novo process begins with submitting a 510(k). Within 30 days of FDA finding the device to be NSE for lack of a predicate, the applicant can then file a petition for reclassification. The agency has 60 days to determine whether to grant the request. If FDA does allow reclassification, then the agency must adopt special controls, which are released after approval.
This publicly available document delineates various elements that ensure the safety and effectiveness of this device. Once a class of device has gone through de novo reclassification, other companies can submit their own applications based on conformance with the special controls document. Thus, de novo reclassification facilitates the pioneer’s market entry but also paves the way for competitors.
While, this reclassification can be granted without prior consultation with FDA, the likelihood of obtaining the status is improved by discussing the route before submitting the 510(k) application.
The de novo reclassification process has been used for many IVDs including tests for cystic fibrosis, circulating tumor cells in breast cancer, breast cancer prognosis, and genotyping drug metabolizing enzymes (FDA—510(k) Premarket Notification Database Search. http://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfPMN/pmn.cfm; select “Cleared for Marketing Automatic Class III Designation” from the “Type” drop down menu; Check box labeled “Cleared/Approved IVD Products”; Click “Search” button. Database updated January 8, 2008. Accessed January 16, 2008). Frequently, there will be no obvious predicate device for a companion diagnostic. In that situation, the de novo process may facilitate market entry.
Laboratory Developed Tests
Although many IVDs take the form of kits that go through the FDA review process, many other diagnostic tests enter the market as laboratory developed tests (LDTs). This is particularly true for genetic tests. To date, FDA has cleared only a handful of the more than 1,300 genetic tests now offered in the U.S.
This is because historically, FDA has not regulated LDTs. In 1992, however, the agency asserted that it did have the authority to regulate all LDTs. This was challenged via a citizen petition, which FDA denied (Gibbs, JN, Bump, CP. Developments in the Law: Questions over FDA’s Authority to Regulate Laboratory Developed Tests. Hyman, Phelps & McNamara website. www.hpm.com/devitem.cfm?RID=74.0. Accessed January 16, 2008). Until very recently, however, the agency did not try to invoke this power. Then, FDA began to invite some laboratories to come to the agency to discuss the regulatory status of their LDTs.
On September 7, 2006, FDA issued the “Draft Guidance for Industry, Clinical Laboratories, and FDA Staff on In Vitro Diagnostic Multivariate Index Assays (IVDMIAs)” (Federal Register—2006; 71:52800-52801). In this document, for the first time, FDA declared its intention to regulate LDTs that involved the testing of multiple markers and used an algorithm that could not readily be understood by physicians.
This proposal generated significant opposition. The Washington Legal Foundation filed a petition challenging FDA’s legal authority to regulate LDTs (CaseDetail: Petition to Stop FDA Regulation of Clinical Labs. www.wlf.org/Litigating/casedetail.asp?detail=451). Others criticized the agency’s failure to go through notice-and-comment rulemaking. Even those who supported the concept expressed concerns regarding unclear language in the draft guidance document.
On July 26, 2007, FDA issued a revised draft guidance document (Draft Guidance for Industry and Food and Drug Administration Staff; In Vitro Diagnostic Multivariate Index Assays; Federal Register—2007; 72:41081-41083), reiterating that LDTs are considered devices but modifying the definition and including a transition period. FDA said that IVDMIAs are expected to be a niche area, which was also disputed. Many diagnostic tests that will affect drug therapy decisions will fall outside this category. Pharmacogenomic or proteomic LDTs that use multiple markers and algorithms, however, may fall into the IVDMIA classification.
Thus, drug companies expecting that multiple-marker diagnostics could quickly enter the market through the LDT route may—depending on the outcome of the IVDMIA debate—need to reassess this strategy. Moreover, given FDA’s repeated statements that all LDTs are subject to FDA regulation, pharmaceutical companies should be careful about attaching their regulatory fate to an LDT.
FDA’s approach toward how it exercises enforcement discretion could shift again. In addition, the agency might try to regulate an LDT if a manufacturer provided significant support to the laboratory (Gutman SI, MD, Director, OIVD, CDRH. Letter to Luber JR, President, EXACT Sciences Corporation. October 11, 2007. www.fda.gov/foi/warning_letters/s6568c.pdf. Accessed January 16, 2008).
FDA also regulates laboratory tests by asserting jurisdiction over software. With increasing frequency, laboratories are acquiring data and then using a computer to analyze the data and generate a test report. FDA has stated that this software is a device subject to FDA jurisdiction (Gutman SI, MD, Director, OIVD, CDRH. Letter to Levine PJ, President and CEO, Correlogic Systems, Inc. July 12, 2004. http://www.fda.gov/cdrh/oivd/letters/071204-correlogic.html. Accessed January 16, 2008). FDA has long asserted authority to regulate software that plays a role in diagnostic testing, and has reviewed and cleared numerous software-based products.
Even if an LDT is not regulated by FDA, it is still regulated by the federal government under the Clinical Laboratory Improvement Amendments of 1988 (CLIA) (42 USC 263a). All laboratories including ones outside the U.S. that perform diagnostic tests on specimens from the U.S. must be CLIA compliant. CLIA imposes numerous requirements on laboratories, such as personnel, reporting of test results, validation, and training (42 CFR Part 493).
Although CLIA regulates the analytical validity of a test, the government has said that CLIA does not regulate clinical validity. This means that while CLIA oversees whether or not a test accurately measure its target, it does not control whether these results are clinically correct. This position has bolstered the argument that FDA should play a greater role in overseeing laboratories.
Hence, in developing a diagnostic strategy, pharmaceutical companies should not assume that pharmacogenomic or pharmacogenetic tests currently regulated only under CLIA will remain that way. The HHS Secretary’s Advisory Committee on Genetics, Health, and Society recently issued sweeping recommendations regarding the regulation of genetic LDTs (Secretary’s Advisory Committee on Genetics, Health, and Society. U.S. System of Oversight of Genetic Testing: A Response to the Charge of the Secretary of HHS. http://www4.od.nih.gov/oba/SACGHS/reports/SACGHS%20Draft%20Report%20on%20the%20Oversight%20of%20Genetic%20Testing%2011-5-2007.pdf. Accessed January 16, 2008).
The materials used to develop an LDT are also subject to FDA regulation as analyte specific reagents (ASRs) (21 CFR 864.4020). In some cases, ASRs are exempt from FDA clearance, e.g., if the ASR is not a kit or is sold by itself without other components. Also, the ASR supplier cannot make diagnostic claims.
Beyond these established restrictions, FDA has adopted a more controversial limit on ASRs. The issue is whether an ASR is limited to a single item, e.g., a vial containing one primer or probe, or whether a vial can contain both a primer and a probe or multiple primers and probes to identify a single condition.
On September 7, 2006, FDA issued a draft guidance on ASRs (Draft Guidance for Industry and FDA Staff; Commercially Distributed Analyte Specific Reagents (ASRs): Frequently Asked Questions; Federal Register. 2006;71:52799-52800). This document stated that the ASR definition excluded products that contain multiple moieties or multiple markers.
FDA received many comments criticizing its narrow interpretation. Manufacturers and laboratories asserted that this interpretation would have an especially negative impact on patients with rare diseases.
Nevertheless, on September 14, 2007, FDA issued a final ASR guidance document (Guidance for Industry and Food and Drug Administration Staff; Commercially Distributed Analyte Specific Reagents: Frequently Asked Questions; Availability. Federal Register. 2007; 72:52568-52570), which goes into effect in September 2008. It says that multiple probes/primers are not ASRs. Products not eligible for ASR status will need FDA marketing authorization.
This effectively means that many molecular diagnostic ASRs will not be available in the future. Some ASRs have already been discontinued due to the guidance document. Pharmaceutical companies should be aware that any diagnostic developed by a laboratory that relies on a multiple-target ASR may be in regulatory jeopardy. This is true even if the laboratory test is not an IVDMIA.
Research Use Only
Another category for in vitro assays is called research use only (RUO), which are products not intended to for diagnosis. Many companies selling RUO products will require customers to certify that they understand that they are intended solely for research purposes.
While RUO products are commonly sold by diagnostic manufacturers, and they are sometimes used off-label by laboratories to run diagnostic tests, a pharmaceutical company should not base its regulatory strategy on RUO products.
FDA has long been concerned by perceived abuses of the RUO category. Although taking enforcement action against RUO products has not been a priority for FDA, the agency’s position is clear: Applicants cannot rely upon an RUO product. Therefore, while an RUO product may help a pharmaceutical company evaluate its product, generate clinical data, or guide product development, it is very unlikely to support an NDA or BLA and even less likely to be referenced in the drug labeling.
The field of in vitro diagnostics is experiencing explosive growth, both in the number of products and technological prowess. Because of various factors including relatively low costs, shorter R&D times, and very rapid innovation, the market for novel diagnostic tests will expand dramatically. Some of these products will play a central role in healthcare by facilitating personalized medicine.
Pharmaceutical companies will inevitably be affected by these products. As HHS Secretary Michael Leavitt said at a recent meeting, “More and more of personalized medicine will be focused on diagnostics. That will be the means by which we can trigger this customization” (Marson B. Personalized Medicine Will Require Greater CDER, CDRH Cooperation. The Pink Sheet. 2007;69(39):22).
Whether drug companies choose to partner with IVD firms or not, the introduction of these new IVD products will affect marketplace and payor decisions as well as FDA’s review of drug products. Thus, it will become ever more important for pharmaceutical companies to better understand how FDA regulates IVD tests.