Skip to main content
Premium Trial:

Request an Annual Quote

Comparison Finds LDTs, FDA-Approved Cancer Gene Tests Perform Similarly


NEW YORK (360Dx) – Molecular tests developed in clinical labs and ones approved by the US Food and Drug Administration perform similarly, a recent study has found.

Members of the College of American Pathologists' molecular oncology committee examined how well laboratory-developed and FDA-approved companion diagnostic tests did on CAP proficiency testing surveys. As they reported last week in JAMA Oncology, the committee found that both test types to gauge whether a cancer patient might benefit from a targeted therapeutic were highly accurate.

While LDTs performed by a single laboratory have been subject to enforcement discretion by the FDA, the agency has been considering altering its stance to treat them like other companion diagnostics. With this study, the CAP molecular oncology committee sought to address the question of whether tests developed in the lab have similar analytic performance as those that go through FDA review.

"The recent debate on laboratory-developed tests and FDA‐approved companion diagnostics has centered upon both the regulatory and performance aspects of LDTs and we, at the College of American Pathologists, had the data through our proficiency testing programs to address the latter point — performance — that we wanted to share with our patients and other clinician colleagues," author Annette Kim from Harvard Medical School and Brigham and Women's Hospital said in an email.

In 2014, FDA released a draft guidance that described how it would bring LDTs under its oversight within a decade. Currently, the Centers for Medicare & Medicaid Services regulates LDTs under the Clinical Laboratory Improvement Amendments, which requires labs to be certified. CAP also provides accreditation.

The draft guidance, though, would phase in FDA oversight of LDTs based on the level of risk they pose: low-risk tests would still be subject to enforcement discretion, while moderate- and high-risk test would begin to be regulated as medical devices.

Some labs and pathologists have argued that they provide services and, as such, don't fall under FDA's purview, while others have said the regulations should be updated, but that it could be done through the auspices of CLIA. At the same time, some kit makers and drug developers have said that the lack of FDA oversight in the area has led to confusion.

However, FDA decided last year to wait on finalizing the draft document, citing post-election uncertainty.

To add to the debate, in their recent JAMA Oncology paper, Kim and her team analyzed proficiency-testing results for BRAF, EGFR, and KRAS testing sent back to CAP. They focused on BRAF, EGFR, and KRAS as those were the only three for which an FDA-approved companion diagnostic was available when the study began.

The team divvied up the 6,897 responses based on whether the participating labs said they'd used an LDT or an FDA-approved companion diagnostic, then compared how well the tests performed.

About 15 percent of the assays used were FDA-approved companion diagnostics and 85 percent were LDTs. Overall, both test categories exceeded 97 percent accuracy for testing all three genes, Kim and her colleagues reported.

For BRAF testing, LDTs performed a little bit better than their FDA-approved companion diagnostic counterparts, a 96.6 percent acceptable rate and 93 percent acceptable rate, respectively. The researchers traced that difference to their analysis of p.V600K, for which LDTs had an 88 percent acceptable rate to companion diagnostics' 66.1 percent acceptable rate.

Meanwhile, for EGFR, LDTs had a 97 percent acceptable rate, while the approved companion diagnostics had a 99.1 percent acceptable rate. For KRAS testing, the researchers said there was no significant difference between the acceptable rates of the two.

However, Kim and her colleagues team also reported that 60 percent of the labs that used a kit deviated from the approved protocol, rendering the tests LDTs. For instance, the labs used the kits to analyzed unapproved specimen or tumor types. This, the researchers said, likely provided the labs with greater flexibility.

"These modifications appear to be driven by the exigencies of real day‐to‐day clinical practice that requires adapting the assays to meet the needs of a variety of clinical situations that may not be accommodated by the FDA‐approved protocol," Kim said.

From their analysis, the researchers concluded that tighter regulation of LDTs might not be necessary. Kim, for instance, noted that these lab tests had excellent performance.

A spokesperson from testing firm Quest Diagnostics, who noted the company hasn't reviewed the study in detail, said that lab-developed tests in general often fill unmet clinical needs and help diagnose diseases for which no FDA-approved kit exists. "The ability of laboratories to develop custom diagnostic tests has been critical to the growth of personalized medicine and keeping pace with the changing face of disease to best serve patients and clinicians," the spokesperson added.

However, a spokesperson for Qiagen, which makes EGFR and KRAS companion diagnostics, noted that the company's tests undergo intense vetting.

"We believe labs should rely on tests that are FDA approved since they have gone through a very rigorous validation process, while also acknowledging the need for laboratory-developed tests for biomarkers where no FDA-approved test is available," the spokesperson said.

In a related editorial in JAMA Oncology, the University of Michigan Comprehensive Cancer Center's Daniel Hayes noted that the study only includes those that have participated in CAP proficiency testing. Additionally, he pointed out the number of specimens analyzed was low, that only a small portion of labs used approved diagnostics, and that many of those didn't follow the manufacturer's instructions.

Hayes also added that the tests used to assay BRAF, EGFR, and KRAS in this study are fairly straightforward and that the results of this study may not be applied to other, more complex tests that are being developed as LDTs.