Submissions for radiation biodosimetry medical countermeasure devices should include well-controlled analytical studies establishing device performance across the entire range of the device, the FDA says.
The devices — the majority of which are in vitro diagnostics — are used to measure how much ionized radiation individuals have been exposed to in the wake of a natural or manmade disaster, such as an improvised nuclear device. Some are nucleic acid-based, while others detect changes in protein expression. The Dec. 30 applies to RBMCDs submitted either as 510(k)s or PMAs.
Premarket evaluation of these devices is important because radiation resistance varies from patient to patient, making it key that manufacturers explain any differences in natural responsiveness that might lead to accuracy errors. These questions can be established through bench testing or literature review, the FDA says. Manufacturers should also refer to FDA guidance on IVDs and Clinical and Laboratory Standards Institute standards in designing appropriate analytical performance testing.
Benefit/risk evaluations should assess how the RBMCDs’ performance differs from that of the laboratory standard, as well as time to results, according to the guidance. Sponsors should provide a detailed description of the device with an intended use statement that specifies the nature of the analyte, the specimen types that can be tested and the specific population for which the test is intended.
RBMCD submissions should also address:
While animal data is not traditionally used in IVD submissions, it may be necessary for RBMCDs if human specimens are unavailable, the FDA says. More specifically, animal studies are acceptable if the analyte being detected is not stable in archived specimens, if there are not enough samples in specimen banks for study or if prospective trials would be unethical or yield an inadequate sample size.
If the device qualifies for animal studies, the submission must demonstrate that the animal model chosen is an adequate substitute for human response. Multiple animal models may be necessary if a single type of animal is not appropriate for all analytes assessed by the device, the FDA notes.
Animal models may also be needed to illustrate whether common drugs, such as antihypertensives or insulin, might interfere with the analytes being studied. Sponsors should also describe any ways in which animal diet and housing may have confounded the results. Finally, experiments should be designed to bridge between animal results and available human clinical information, the guidance says.