DermTech's melanoma identification test, MelDTect, maintained its accuracy in a platform shift from microarrays to Life Technologies' OpenArray quantitative RT-PCR system, the company showed in study results presented today at the annual meeting of the Association for Molecular Pathology.
DermTech published a proof of concept of its Epidermal Genetic Information Retrieval, or EGIR, technology earlier this year in the British Journal of Dermatology, showing that the 17-gene classifier using EGIR-harvested RNA with microarray-based amplification and analysis was able to detect in situ or invasive melanoma with 100 percent sensitivity and 88 percent specificity.
At the time, the group intended to begin evaluating PCR platforms that could improve the assay's throughput and cost-effectiveness. This week, Bill Wachsman, a University of California, San Diego, hematologist-oncologist, told PCR Insider that the company chose Life Tech's OpenArray system for the new study, which he co-authored. DermTech reported that the platform maintained the test's earlier-reported 100 percent sensitivity and showed a comparable specificity of 85 percent in the poster presentation at the AMP meeting this week.
Wachsman, who is also a member of DermTech's scientific advisory board, said the platform "works like a charm," but that DermTech hasn't yet decided officially which quantitative PCR system it intends to move forward with for commercialization.
"We looked at a handful of others," he said. "Some were rejected [initially] because we started this work more than a year ago and at that point several of them weren’t ready for prime time. Now, because they’ve improved either their sensitivity or their throughput or their cost, we may decide to revisit them." Wachsman declined to identify any potential alternative platforms.
The OpenArray platform is approved for research use only. Wachsman said that DermTech would potentially work with any RUO platform partner to help the vendor gain FDA clearance for diagnostic use.
DermTech's EGIR technology uses custom adhesive tape to non-invasively collect cells from the surface of the skin. The cells' RNA is then isolated, amplified, and analyzed by gene expression profiling.
In the new data presented this week, the researchers reported their transition of the MelDTect melanoma assay from a microarray to the OpenArray RT-PCR system, and directly compared the performance of the new PCR assay with a microarray approach using Affymetrix's U133 plus 2.0 GeneChip on 118 EGIR specimens (57 in situ and invasive melanomas and 61 nevi).
"One of the big questions that always comes up is, 'Can you translate data from a microarray platform into an assay based around quantitative PCR?'" Wachsman said. "Microarray tends to compress data top-down and bottom-up meaning you just wont get the linear dynamic range on a microarray that you do in PCR."
In their study, the group found both platforms to be 100 percent sensitive and both to be "within the same ballpark" of specificity, according to Wachsman. The OpenArray platform showed 85 percent specificity and the microarray 96 percent specificity.
The OpenArray assay also demonstrated more than a five-log dynamic range and had a coefficient of variation of less than three percent, the authors reported.
Wachsman said in a statement that the results, "especially with respect to linear dynamic range, far exceed anything we've seen using microarrays and suggest that the non-invasive assay will perform extremely well in a high-volume clinical setting."
DermTech claims the data collected so far suggest the company's EGIR-based test is better than any known melanoma detection product.
Detected and diagnosed early, melanoma is highly curable. However, Wachsman said, the currently used biopsy-and-histopathology method can be unspecific, with only about three to ten percent of suspicious biopsied lesions actually turning out to be melanoma.
"There is a lot of time and effort and expense that goes into the process of current melanoma detection," he said. "And even the more advanced optical devices coming out now only improve it slightly."
Citing the company Mela Sciences, which developed the recently US Food and Drug Administration-approved MelaFind optical identification technology, he said it "may be 97 or 98 percent sensitive, but only about 10 percent specific."
By comparison, the assay being developed by DermTech "has been 100 percent sensitive, so its not missing any melanoma, and 85 percent specific," he said.
Wachsman said biopsy interpretation can also be highly variable from one pathologist to another, with discordance rates in histopathology reads between 10 percent and 35 percent depending on the degree of difficulty of the type of melanoma under review.
"With a genomic assay its very objective, so I think that’s going to be one step up [for DermTech]," he said.
He added that in DermTech's studies, including the recently presented poster, the researchers have excluded lesions that were bleeding or ulcerated, which are "the ones most likely to be melanoma."
"If you look at the [samples] and their T-stage," Wachsman said, "The vast majority of the invasive lesions are stage one and two and there are very few T-3s and I don't even think there was a single T-4. So we are picking up the earlier-stage lesions and picking them up with great sensitivity and specificity."
Wachsman added that he believes many dermatologists will start to use the assay to look at lesions that are more difficult to diagnose visually and get a sense of how the assay performs. Then, he predicted, "it will be adopted to look at a larger percent of lesions."
According to Wachsman, DermTech is not officially married to OpenArray despite the positive results in its poster presentation, and may consider other technologies going forward.
"Whether it's OpenArray or another platform we're not sure," he said. "But we've established that we can move to a quantitative PCR platform, and since most of the platforms will have similar capacity, now that we have the methods for pre-amplification down and we know what targets we're really going to be looking at, moving forward with more validation should be pretty straightforward."
DermTech is planning to begin enrollment for a pre-validation study at clinical sites in Europe in 2012. The study will evaluate use of the MelDTect test in a broader population.
Jennifer Larson, director of business development at DermTech, told PCR Insider this week that plans are for "100 melanoma specimens and 400 or 500 non-melanoma samples," in the pre-validation study, and that the company hopes to initiate this research in about four to six months and complete it in another six to eight months.
Initially, DermTech plans to seek a CE mark in Europe and launch the test in Australia, where the high incidence of melanoma would mean a larger demand.
"[Australia] is really the epicenter of melanoma," Lawson said, "so anything that is adopted there is broadly disseminated through the rest of the world.
"While the market isn’t huge — there aren’t that many people there — melanoma affects one in every 14 individuals in Australia, so that's a huge number," she said.
Lawson said DermTech has an "FDA plan in place," for after the European validation has begun.
"We've not met with them yet," she said. "But once the Europe pre-validation study is underway, then we plan to start conversation with the FDA."
According to Wachsman, the company aims to do a second, larger validation study with an estimated 500 melanoma samples and potentially several thousand non-melanoma samples with the aim of gaining FDA approval.
Lawson said whatever study DermTech does in the US will be guided by meeting with the FDA in advance, and that the company hopes for that research to "kick off around the middle of 2013."
In addition, Wachsman said that he, with DermTech, is also doing research on how to expand the assay to offer some staging information in addition to simple identification.
He said researchers presented proof-of-principle data at the Society of Investigative Dermatology annual meeting in April showing the group could divide melanoma samples into invasive and in situ categories using another gene classifier on the same sample.
"We don't have to take a second tape strip," Wachsman said. "So if assay one for melanoma detection is positive, we would generate a risk score that the lesion is melanoma, and if it were high enough, we could then … run the second assay on the same specimen and give them information about whether it is likely to be invasive or in situ."
At this point, he said the company is trying to stay out of the actual diagnosis of the disease, which requires analysis of additional factors like micro-ulceration, mitotic indices, and measurements of the depth of invasion.
Lawson said the company is also doing research using its EGIR method to detect other diseases like prostate cancer.
She specified that the company's tape strip sampling for prostate cancer is done behind the ear, not at the tumor site.
Have topics you'd like to see covered in PCR Insider? Contact the editor at mashford [at] genomeweb [.] com.