Asuragen presented data at last week's American Pancreatic Association annual meeting from two studies, one a multi-center evaluation of its miRInform Pancreas fine needle aspirate test, and the other an early study of a set of miRNAs that can be used to discern which pancreatic cysts are likely to become cancerous.
The blinded validation study of pancreatic cyst fluid specimens showed a nine-miRNA model had 89 percent sensitivity and 100 percent specificity in distinguishing low-risk cysts from high-risk cysts to inform treatment, Asuragen reported. The company is now planning a multi-center prospective validation trial to further refine the test, according to Rollie Carlson, Asuragen’s president.
In the FNA study, described in a poster presentation at the conference, Asuragen reported that adding its seven-miRNA assay — which distinguishes pancreatic ductal adenocarcinoma from benign pancreatitis and other conditions — to basic FNA cytology enhanced diagnostic accuracy by about 12 percentage points in the study of 186 subjects.
The company said that the FNA miRInform Pancreas test will be available through Asuragen’s CLIA laboratory starting Dec. 1.
"We certainly see the need for the FNA test," Carlson told Gene Silencing News this week. "But also, a number of researchers were coming to us and saying that it would be of high clinical value for early detection, to interrogate these cystic lesions, which in many cases are benign but in some cases are precursors for pancreatic cancer."
"These [cysts] show up in imaging, but they don't lend themselves well to conventional cytopathology," he said.
Accurate determination of a patient's cyst variety is necessary for appropriate treatment, according to Carlson. Patients with cysts that are more likely to be precursors to pancreatic cancer — called intraductal papillary mucinous neoplasms — that also have high-grade dysplasia require immediate resection, while those with low-grade dysplasia may not, the company reported. Other, less common cysts, such as solid pseudopapillary neoplasms and cystic neuroendocrine tumors also need to be surgically removed.
"Currently it is challenging to accurately identify the specific cyst entity and predict the grade of dysplasia without resection and precise histopathologic examination," the company said.
To investigate whether it could find miRNAs with potential to better discern between the benign and potentially cancerous cysts, Asuragen collaborated with Anirban Maitra at Johns Hopkins University School of Medicine.
The Hopkins team evaluated RNA from microdissected formalin-fixed paraffin-embedded tissue from IPMNs and from cyst fluid samples. They reported that they measured expression of 754 miRNAs using 23 FFPE samples and 15 cyst fluid samples. A separate verification set comprised 38 tissue and 50 cyst fluid samples.
The group found 26 and 37 differentially expressed miRNAs between low-grade and high-grade IPMNs, respectively, with four overlapping. Nineteen of the miRNAs selected were able to differentiate between high-grade and low-grade IPMNs, as well as identify other uncommon cysts such as solid pseudopapillary neoplasms and cystic neuroendocrine tumors, they reported.
A logistic regression model predicted cyst pathology well enough to determine appropriate treatment — either surgical resection or conservative watchful waiting — with an accuracy of 90 percent and 100 percent, respectively, according to the researchers. Asuragen wrote that this model was narrowed to nine miRNAs.
Carlson said that Asuragen is now planning a larger multi-center study to validate the results.
"We basically have a very good dataset which we want to now plug into something like what we did in PDAC FNA, a multi-center validation to determine if that profile is exactly what we think it is from that first research," he said.
He said that after presenting the work, "there seemed to be a broad interest [from potential collaborators] to be able to plug in, to provide samples, to help." He said the company is aiming for around 200 samples for the planned study. "If the performance holds up to what we saw in the initial study, 200 or 300 samples should do it."
Meanwhile, in a poster presented at the conference, Asuragen detailed their multi-center study of the FNA miRInform Pancreas test, which interrogates expression of seven proprietary miRNAs: miR-130b, -135b, -148a, -196a, -375, -96 and -24.
Earlier this year, Asuragen shared details about its miRInform Pancreas test for formalin-fixed tissue, which was commercialized in 2008 as the first miRNA diagnostic to reach the market (GSN, 4/21/2011). The FNA version had been in the works for several years, Carlson said at the time.
In last week's poster, the authors wrote that development of the FNA assay involved a panel of 95 FFPE samples used to develop the miRNA signature, which was then tested for its ability to predict PDAC status on an independent set of 186 FNA samples.
In the study, the assay diagnosed PDAC with 92.5 percent accuracy compared to 80.2 percent for FNA cytology alone in the 186-subject study. The test was also able to characterize samples with "indeterminate or suspicious FNA cytology" with an overall accuracy of about 75 percent, the researchers reported.
The researchers wrote that they developed the assay in accordance with CLIA and CAP regulations. Carlson said that Asuragen plans to start a "clinical experience" program in December to continue assessing the performance of the test. The company will also begin offering the test via its CLIA lab in December.
Aside from its pancreatic work, Asuragen has also been working on miRNA-based assays to diagnose and predict the recurrence of colon cancer and to diagnose lung cancer.
Carlson said the company will be presenting data on the colon cancer work at the Association for Molecular Pathology annual meeting next week.
The lung cancer research is ongoing, he said.
Have topics you'd like to see covered in Gene Silencing News? Contact the editor at mashford [at] genomeweb [.] com.