Ridge Diagnostics' multi-assay protein test for major depressive disorder has performed strongly in a recent pilot and replication study by a team of researchers collaborating with the neurodiagnostic company.
The research group, led by George Papakostas at Massachusetts General Hospital, published results of two small prospective studies in Molecular Psychiatry in December. In the pilot, the team evaluated the test's performance, testing 36 subjects with clinically diagnosed MDD and 43 non-depressed controls. The researchers then replicated the study in another 34 depressed subjects, and found that in both groups the test performed with roughly 90 percent sensitivity and 80 percent specificity.
Papakostas, a medical advisor to Ridge Diagnostics, told ProteoMonitor this week that he thought the results his team achieved were strong, but that further studies evaluating the panel in a clinical screening study will be necessary to confirm the performance.
"In the past a lot of people have come up with different test attempts, and statistically many of them correlated [with depression], but their power wasn’t robust enough that it could be used as a diagnostic, only as a research lead," he said.
"This is really the first time, if this is confirmed in a screening study, that a test [has performed well enough] that it could be used clinically in situations as needed."
However, he stressed, "this finding really calls for a confirmatory study — not in a research sample like ours, but in a primary care or community clinic context."
Ridge's test, MDDScore, is a multi-analyte immunoassay panel, consisting of nine protein markers from four biologic pathways, including inflammatory, metabolic, hypothalamic-pituitary-adrenal axis, and neurochemical molecules. The company measures each of the nine markers using ELISAs and incorporates them using an algorithm that scores samples on a scale of one to nine, with one representing the least likelihood of a patient having MDD and nine marking the highest chance.
“We create what we call a hyper-map, [which is] a four-dimensional map that creates a pattern that is compared to known depressed and known non-depressed [patterns],” Ridge CEO Lonna Williams explained. “That's how we report [the results] with nine being 90 percent chance of depression and two being 20 percent chance."
Some of the markers are better known and others more novel, Williams said. But all are linked to pathways involved in the pathophysiology of depression, according to the company.
In the two studies published in Molecular Psychiatry, the research team worked with patients recruited at MGH and other participating sites, including Vanderbilt University Medical Center and Cambridge Health Alliance.
In the first study, the group tested a cohort of 36 patients who met the criteria for MDD diagnosed using a method called the mini international neuropsychiatric interview. They also enrolled 43 non-depressed control subjects.
The researchers sent frozen blood samples to Ridge for testing using the company’s panel of makers. They compared the test results with the patients' MDD statuses and found that of the 36 MDD patients, 33 had a positive MDD score, defined in the study as five or above. Of the control subjects, eight had a positive score. This translated to a sensitivity and specificity of about 91 percent and 81 percent respectively, the group reported. In the replication cohort of 34 additional MDD subjects, the group observed the same performance.
“Sensitivity and specificity above 80 percent was “really good for a medical test in general, let alone for depression, where [these] things have been elusive,” Papkostas said.
Papkostas said that there are three main reasons a reliable molecular diagnostic for depression is desirable.
First, he said, the test — if it proves itself in further study — could function as a push in one direction or the other in situations where diagnosis is unclear. “In certain situations diagnosis might be obvious,” he said. “But in some, it can be complicated by a lot of presentations or confounding factors that really necessitate a second opinion. The idea is [that] the test, along with clinical presentation and the patient's experience and the opinion of the clinician, can give additional information for [them] to make a decision.”
Secondly, any accurate biologic test for depression helps the field understand what “constitutes [the disease] from a pathological perspective," he said.
Finally, Papakostas said, the test could also be a tool for reassuring patients of the conclusiveness of their diagnosis, and of the legitimacy of their depression. “Even if a diagnosis is obvious… a patient nonetheless can find it in some cases reassuring or soothing for whatever personal reasons if there is something biologic to go along with that,” he said.
According to the company website, the test result is intended to be used "as an aid in the diagnosis and management of MDD and in conjunction with other clinical and laboratory information."
Williams said the test is currently being offered in a “pilot market” to psychiatrists through Ridge’s CLIA-approved lab. But she said the company plans to expand to other specialties.
“We were very pleased with the outcomes [of the studies],” she said.
Papakostas said he considers the test still a research tool until the group’s findings can be replicated in additional clinical screening trials. He also said clinicians and developers should be wary of ways such a tool could potentially be misused.
“In medicine there is always more we don't know than that we know,” he said. “We don't know, for instance, if a person never has depression and tests positive whether they are going to develop depression or [may be fine]… or what it might mean if someone tests positive who was once depressed but has been successfully treated.”
According to the group’s report, the performance of the test clearly exceeds what other single-protein assays have been able to achieve. “Whether additional analytes not yet identified can further improve the performance of this test remains in question,” the authors wrote.
The researchers stressed in their report that their evaluations were limited by a small pool of control subjects, which did not allow them to match patients and controls by factors like age and gender. Further research is necessary to confirm the performance across different age and ethnic groups as well as to examine additional diagnostic and potential prognostic applications, they wrote.
According to Williams, Ridge has additional ongoing prospective studies but she declined to detail their designs or scope. Papakostas said he and his team have several study designs in mind that they hope may answer some of the lingering questions generated by their findings.
“We are looking for different partnerships that could make these [studies] feasible, and are hoping to launch [them] in 2012, 2013,” he said.
“Hopefully,” he said, “it will stand the test of confirmation.”
Have topics you'd like to see covered in ProteoMonitor? Contact the editor at mashford [at] genomeweb [.] com.