IVD Technology
Magazine
IVDT Article Index
Breaking the home-use barrier for tumor markers
Glen Freiberg and Alicia Moffat
Even for a low-risk test, the path to FDA clearance can be a tangled one. But a new spirit of cooperation at DCLD could make the road a little smoother for manufacturers.
FDA regulation of tumor-marker IVDs is steadily evolving, creating new challenges in interpretation and application for industry and regulators. This article contrasts the regulatory process for tumor-marker tests with that of other IVDs and describes the impact of some recent developments—including the transition of some markers from Class III to Class II. These topics provide the essential background for understanding the regulatory pathway for FDA's recent approval of the first prescription home-use tumor marker.
 The BTA stat test (Bion Diagnostic Sciences; Redmond, WA), the first approved home-use tumor marker. |
Although tumor-marker IVDs are low-risk devices, FDA requires a significant volume of data in support of premarket submissions for such products. In fact, very few products subject to review by FDA's Office of Device Evaluation (ODE) share such a combination of low risk and high data requirements. This discrepancy has been apparent since the onset of device regulation, when the Medical Device Amendments of 1976 listed these products as Class III transitional. This category was used instead of a final classification, probably to ensure that the first premarket submissions for tumor-marker tests would receive careful review before their low-risk nature became commonly acknowledged.
Unlike most of the diagnostic products reviewed by ODE's Division of Clinical Laboratory Devices (DCLD), tumor markers have a myriad of potential indications for use. Tumor markers are unlike quantitative tests whose results correlate directly to a clinical state (e.g., glucose with diabetes, cholesterol with atherosclerosis), or qualitative (yes/no) tests for infectious diseases. By contrast, most currently marketed tumor-marker tests are more adjunctive than definitive. So, in addition to their potential for use in diagnosis, tumor markers may also have intended uses in early detection, determination of disease severity (staging or grading of tumors), disease monitoring and management, and prognosis.
Clinical Claims
When a sponsor desires to make marketing claims related to such applications, FDA may require clinical data in support of each claim. Clinical evaluations must be designed specifically to study each claimed use, which can result in somewhat different studies for each proposed use in early detection, staging, measuring therapy response or failure, or measuring therapy toxicity.
In general, IVD clinical trials are conducted as premarket notification (510(k)) correlation trials. Known substances are tested over a range, sensitivity and specificity are calculated, and precision is demonstrated. When these characteristics can be shown to be substantially equivalent to those of a predicate test, FDA will clear the test for marketing. By contrast, demonstrating the utility of a tumor-marker analyte or antigen often requires prospective studies and patient outcome data--a much more expensive and time-consuming prospect.
The sensitivity and specificity of tumor-marker tests are often lower than those of traditional clinical chemistry, microbiology, or immunology tests. Nevertheless, users believe that even such tests with relatively low sensitivity and specificity have significant adjunctive utility, contributing information to the complex algorithms used in diagnosing and monitoring cancer patients. Tumor-marker tests rarely offer the physician a simple positive or negative result that can be used in evaluating a patient. Instead, test results are combined with the patient's medical history and data from other tests, and all of this information aids the physician in formulating a clinical diagnosis, designing therapy, or creating a follow-up plan.
As an example, one of the most recently approved tumor markers, the free/total prostate-specific antigen ratio, is intended to help physicians and patients determine whether a prostate biopsy would be worthwhile for particular patients. Such a test is clearly adjunctive with regard to the patient's diagnosis, but it also holds out the possibility that some biopsies might not have to be performed. Although such a change in medical practice is only an incremental improvement, any reduction in the total number of biopsies has clinical and financial value.
For sponsors of tumor-marker submissions, having the data necessary to demonstrate such clinical utility can be an important factor in attaining the approval of FDA and its advisory panels. In the current environment at ODE, achieving product approval for a tumor-marker test requires more than merely a statement of intended use, sensitivity, and specificity. Meanwhile, outside ODE, the question of whether clinical utility should legally be part of the clearance or approval review process remains a debatable one.
Reclassification
In the years after passage of the Medical Device Amendments of 1976, FDA and industry gained sufficient experience with tumor markers to consider down-classifying them and reducing the amount of clinical data required for their premarket submissions. In 1995, with the support of the American Association for Clinical Chemistry, the American Society of Clinical Pathologists, the College of American Pathologists, and the Health Industry Manufacturers Association, Centocor (now Fujirebo Diagnostics; Malvern, PA) submitted a petition requesting that serum tumor markers for monitoring and management be reclassified from Class III transitional to Class II. FDA's advisory panel accepted an expansion of Centocor's petition and recommended that tumor-marker testing of all types of human samples be down-classified to Class II (see Table I).
| Marker | Intended Use |
|---|---|
| Carcinoembryonic antigen (CEA) | Colon cancer monitoring |
| Alpha-fetoprotein (AFP) | Testicular cancer initial diagnosis and monitoring |
| CA-125 | Ovarian cancer diagnosis |
| Prostate-specific antigen (PSA) | Prostate cancer early detection and monitoring |
| Free/total PSA ratio | Aid in qualifying patients for prostate biopsy |
| CA 27.29/CA 15-3 | Breast cancer early-stage monitoring |
| Human complement factor H variant | Bladder cancer patient management |
Table I. Examples of approved tumor markers and their intended uses.
Because tumor-marker tests had been classified as transitional devices, FDA was able to down-classify them by means of an administrative order rather than having to go through notice-and-comment rule making. Centocor received the reclassification order granting Class II status in September 1996, making it possible for manufacturers to use the 510(k) route to FDA clearance from that time forward. The order was subsequently announced in the Federal Register.1
However, application of the reclassification order was somewhat restricted. The only indications for use included in the order were monitoring and management. Tumor-marker tests intended for screening, diagnosis, and other uses remain Class III devices. Similarly, the 510(k) process made available by the order could not be used to seek clearance of off-label, tumor-marker use for a product not otherwise approved as a tumor marker.
As part of its petition, Centocor also proposed a guidance document to accompany the reclassification order. FDA published that guidance with little modification, and it has since become the standard for consideration of Class II tumor markers under the 510(k) process.2 Once sponsors began using the 510(k) process and the attendant guidance, however, a number of issues of contention began to emerge. Some of these remain topics of ongoing discussion between FDA and industry, including the following:
Early 510(k) Clearances
Once FDA had issued its reclassification order for tumor-marker tests, manufacturers quickly began to take advantage of their new route to market, submitting a number of 510(k) notices. Among the first of these was the BTA stat test (Bard [now Bion] Diagnostic Sciences; Redmond, WA), an immunochromatographic assay for bladder cancer monitoring and management that was based on an earlier, manual version with the same intended uses.
Bard's original BTA test was a multistep latex agglutination test conducted using a sample well and test strip. Because the premarket submission for this new tumor marker was prepared prior to the reclassification order, it was submitted as a premarket approval (PMA) application. The application was reviewed by an FDA advisory panel and was granted approval as the first test for monitoring and management of bladder cancer patients. By contrast, the BTA stat test was ready for FDA review just after the reclassification order was issued. It was therefore possible for this version of the test to go through premarket review as a 510(k), a process that was completed comparatively quickly.
Bard's original BTA test was an analyte test measuring high-molecular-weight glycoproteins. The BTA stat test is a qualitative, single-step, lateral-flow, immunochromatographic assay, and a true antigen test that measures the amount of a human complement factor H variant in the urine. Although the substances measured by the two tests are different, the submission for the BTA stat test raised no new issues of safety or effectiveness. The fact that the two tests have the same intended use enabled FDA to determine their substantial equivalence via the 510(k) route. This approach to premarket review suggests that other new tumor markers might also receive 510(k) clearance by demonstrating their substantial equivalence for existing or new intended uses.
The example of the two Bard BTA tests indicates that the regulatory pathway for a new tumor marker depends heavily on timing and the flexibility of FDA policy at the time the sponsor is ready to approach the agency.
Advancing tumor-marker technologyWhen scientists at Bard (now Bion) Diagnostic Sciences (BDS) set out to create an improved BTA test, they didn't expect to discover a previously unknown mediator of selective growth advantage for tumor cells. The original BTA test was a latex agglutination assay that detected the presence of high-molecular-weight glycoproteins in the urine of some bladder cancer patients. The second-generation BTA stat test is a lateral-flow, immunochromatographic assay that uses two different murine monoclonal antibodies (MAbs) to qualitatively detect a variant of human complement factor H (FH) in urine. Normally synthesized in the liver and found in blood but not urine, FH plays a key role in inhibiting the alternative complement pathway, which functions to lyse cells recognized as foreign by the host. Targets of the alternative pathway are not specifically selected by the immune response, but instead have similar physical structures common to infectious bacterial agents that have plagued mankind for thousands of generations. The pathway's lack of specificity, however, means that normal human tissues can be attacked by another complement protein, activated C3 (also known as C3b). C3b focuses complement proteins onto the surface of a target cell or organism, resulting in lysis and death of the target. Characterization of the FH antigen used in the BTA stat test and corresponding quantitative ELISA (the BTA Trak assay) represented the first report of an association of FH with cancer. The MAbs (X52.1 and X13.2) were originally generated against partially purified urine proteins from patients with active bladder tumors confirmed by histology. Using X13.2 to purify urine samples from bladder cancer patients, it was possible to isolate the antigen for characterization. In contrast to what is observed for most proteins, SDS-PAGE revealed that the antigen exhibited an apparent molecular-weight shift (from 138 kD to 151 kD) under reducing conditions. This clue, along with partial amino acid sequences of the isolated antigen, led researchers to identify the antigen as FH or a very similar protein. Scientists hypothesize that cancer-cell expression of FH helps the cells escape lysis by the alternative complement pathway, giving them a selective growth advantage. To test this hypothesis, BDS researchers have recently designed and tested in vitro models for complement-mediated lysis of bladder cancer cell lines. Of three bladder cancer cell lines studied, one was found not to produce FH and is readily lysed by the alternative complement pathway; but two lines do produce FH and are resistant to complement-mediated lysis. Addition of MAb X52.1 to the resistant cell lines has been observed to overcome this resistance, resulting in lysis of the cancer cells. These intriguing results are expected to lead to new findings regarding the role of the complement system in tumor formulation and proliferation. |
Prescription Home Use
Although FDA's reclassification order moved tumor markers for management and monitoring from Class III to Class II, their labeling continues to carry the required prescription legend found in 21 CFR 801.109, "Caution: U.S.A. Federal law restricts this device to sale and distribution by or on the order of a physician, or to a clinical laboratory, and use is restricted to, by, or on the order of a physician."
There is some disagreement about what this statement is intended to mean. One could argue that it permits a physician to provide the product to a patient for home use. In fact, it may mean just that. It is nevertheless unclear whether FDA would permit a product to be used in this way unless it had been specifically approved for such an intended use. In considering an expansion of the intended uses for its BTA stat test, Bard chose to formally apply for FDA approval of prescription home use.
The regulatory premise for prescription home use of the BTA stat test is related to its potential utility in the logistics of patient management. Patients who are being monitored for bladder cancer usually receive a flexible cystoscopy exam in the doctor's office. If a suspicious area is seen, a rigid cystoscopy is then scheduled, usually in a site different from the doctor's office, and surgical intervention may follow. By having the qualitative results of the BTA stat test via home-use testing, the physician can know in advance of a scheduled visit whether a recurrence may be present. This information enables the physician to schedule a rigid cystoscopy in the first instance, thereby saving the patient an invasive procedure—and the time and money associated with that procedure.
For Bard, approval of home use would have another benefit: automatic waived status under the Clinical Laboratory Improvement Amendments of 1988 (CLIA). Bard and Boehringer Mannheim (Indianapolis) lobbied heavily for inclusion of automatic waived status for such tests in the FDA Modernization Act of 1997 (FDAMA).3 Waived status would enable physicians to conduct the BTA stat test without having to satisfy the more-extensive requirements for a moderate- or high-complexity clinical laboratory.
DCLD responded quickly to Bard's first suggestion of a prescription home-use submission, in December 1997. Within a week, DCLD director Steven Gutman arranged for the submission to be discussed at a panel meeting in February 1998. Bard then sent to FDA its 510(k) submission, including a description of the proposal, its clinical utility, and a draft of the labeling insert to be used in patient home-use packages.
Bard's presentation at the panel meeting was brief. Since the product had been previously cleared for marketing, there was no need to review clinical performance or other matters already covered in the original clearance. The panel voiced no objection to the concept, understood the utility, and made several labeling suggestions. After the meeting, FDA provided the panel's suggestions to human factors specialists at CDRH's Office of Health and Industry Programs (OHIP), who subsequently supplied comments on the product labeling.
The BTA stat test is performed by placing five drops of urine into the well of the immunochromatographic test device, and reading a line after 5 minutes has elapsed. The 510(k) presented to the panel included the results of a clinical trial that demonstrated extremely high accuracy and near-perfect correlation between layperson and professional use of the test. However, this study had originally been designed to comply with a guidance issued by the Centers for Disease Control and Prevention on data necessary to request waived CLIA status. It compared the accuracy and precision of nonlaboratory personnel working in the medical industry to that of laboratory professionals. Ultimately, FDA requested additional trial data using a layperson demographic more closely representing the intended population.
Working closely with reviewers from DCLD's immunology branch, the sponsor developed a protocol for evaluating performance of the BTA stat test in the hands of additional bladder cancer patients. During the study design process the home-use labeling was exhaustively reviewed and enhancements were made to ensure that the instructions could be easily understood by persons with a variety of educational backgrounds.
Study subjects answered a multiple-choice questionnaire testing their understanding of the test procedure and soliciting their views on the test's ease of use. Participants rated both the readability of the instructions and the test's ease of use very high. DCLD and OHIP reviewers used the results of the questionnaire to suggest further labeling enhancements. After the results of the study had been submitted, DCLD granted clearance to market the BTA stat test for prescription home use within 90 days.
Conclusion
Innovative ideas sometimes travel a rocky road to regulatory clearance or abandonment. The development and approval of new tumor markers had been slow, expensive, and time consuming. But, in the spirit of FDAMA, DCLD has demonstrated a willingness to work with sponsors who present innovative ideas.
Clearance of the BTA stat test for prescription home use is a good example of the new spirit at DCLD, which has also been apparent for other tumor-marker submissions as well as for new indications for the use of genetic probes. Hopefully, such cooperation and process streamlining will remain part of ODE's progress toward the "least burdensome" review process called for in FDAMA.
References
1. Federal Register, 62 FR 242: 66003–66005 (December 17, 1997).
2. "Guidance Document for the Submission of Tumor-Associated Antigen Premarket Notifications (510(k)), to FDA" (Rockville, MD: Division of Clinical Laboratory Devices, Office of Device Evaluation, Center for Devices and Radiological Health, FDA, 1996).
3. FDA Modernization Act of 1997 (S.830), Public Law 105-115 (1997).
Glen Freiberg is a member of IVD Technology's editorial advisory board and was formerly with Bard Diagnostic Sciences (Redmond, WA); he is now with Gen-Probe Inc. (San Diego). Alicia Moffat was formerly a regulatory affairs specialist at Bard Diagnostic Sciences.
