Skip to main content
Premium Trial:

Request an Annual Quote

Signature Diagnostics Offers Plasma-based Targeted NGS Test for Monitoring Colorectal Cancer


German cancer diagnostics firm Signature Diagnostics has developed a plasma-based ultradeep sequencing test to monitor colorectal cancer patients and detect metastasis early.

The test, called Onco-Detect, analyzes mutations in 10 cancer genes, using circulating free DNA from plasma, in colorectal cancer patients with stage 3 or stage 4 disease.

According to André Rosenthal, Signature's CEO, the test could have applications in three areas: to analyze samples from phase 2 and 3 clinical trials of colorectal cancer drugs, to monitor colorectal cancer patients undergoing treatment for their response, and to screen patients that are currently tumor-free for evidence of micrometastases. Rosenthal presented data on Onco-Detect two weeks ago at Hanson Wade's NGS for Cancer Drug Development conference in Boston.

"Monitoring is a totally new cancer diagnostic field, which has been completely neglected, but that's where the need is," Rosenthal told Clinical Sequencing News. "The need is not only to determine the mutation profile at the beginning, before treatment, but constantly during treatment, because the tumors get resistant. And resistance means they develop new mutations and escape the drugs. But how do you know that if you have no access to the tissue?"

The company has been offering Onco-Detect as a diagnostic service for stage 3 or 4 colorectal cancer patients in Europe since this summer, charging €1,500 ($2,000) for each plasma sample sequenced, and has analyzed samples from several self-paying patients so far. The test is not currently covered by health insurance.

Signature is in discussions with several large diagnostic companies and next-gen sequencing firms to offer Onco-Detect as a lab-developed test in the US and in Asia. The ultimate goal is to develop the test into an FDA-approved kit for the US and a CE-marked kit in Europe.

The firm does not actively market the test in Europe at the moment, relying on word-of-mouth, but it is considering building a small sales team. Most patients contact the company directly right now. "There is a huge gap between oncologists' knowledge and what can be done," Rosenthal said. "The technology advances are so fast that they are not aware of it." This is likely going to change, though, as data is presented at medical conferences, he added.

In the meantime, Signature is embarking on a prospective clinical trial of Onco-Detect in Germany that will correlate mutational profiles with treatment response.

Signature Diagnostics, which is based in Potsdam near Berlin and was founded by Rosenthal in 2004, has been developing Onco-Detect and other tests using samples from a large prospective multicenter study, called Molecular Signatures in Colorectal Cancer, or MSKK, which it sponsors and conducts.

The study, which started in 2005, involves 40 primary care hospitals in eight German states and has so far recruited about 8,500 patients with colorectal cancer of all four stages. The study's biobank includes frozen tumor tissue, formalin-fixed paraffin-embedded tumor tissue, and several plasma samples for each patient, making it the largest tumor-plasma bank for colorectal cancer worldwide, Rosenthal said.

In the beginning, Signature focused on developing prognostic tissue-based tests based on gene expression profiles using Affymetrix microarray technology. But two years ago, the company switched to next-gen sequencing, purchasing Illumina's MiSeq, and has since been developing NGS-based colorectal cancer tests for prognosis, prediction of response, and monitoring.

All its tests except Onco-Detect are tissue-based: Onco-Predict-II predicts the risk of metastasis in stage 2 colorectal cancer patients, Onco-Select-III predicts outcome of stage 3 colorectal cancer patients treated with adjuvant chemotherapy, and Onco-Select-IV is a companion diagnostic test that predicts the response of patients with metastatic colorectal cancer to first-line treatment.

"But my vision is to develop all the tests [into] plasma-based [tests], because [blood] is a better tissue," Rosenthal said. "It's much more easily available, and the principle of liquid biopsy is very attractive, and the only one you can implement mid- and long-term."

In particular, a plasma-based test could be applied during and after treatment, as well as between treatment cycles, when no new tissue is available.

The median survival for metastatic colorectal cancer patients is 20 months, he explained, and during that time, "they more or less get constant treatment," while "the tumor mutation profile, and the frequency of the mutations, will change over treatment." Understanding those changes in liver metastases in particular, he said, will be "key to understanding colorectal cancer, and many other cancers."

Onco-Detect sequences 49 amplicons, a total of 5.5 kilobases, from 10 key genes known to be mutated in colorectal cancer — including APC, TP53, BRAF, and KRAS — using the MiSeq, with a coverage of at least 2,900-fold and up to 115,000-fold and a mean coverage of 20,000-fold.

Such high coverage is necessary to be able to detect mutations with a frequency below 2.5 percent, Rosenthal said, and to distinguish them reliably from sequencing errors.

The test currently uses 2 milliliters of plasma, from which circulating free DNA is isolated using Qiagen's QIAamp Circulating Nucleic Acid kit. While the yield of cfDNA varies from patient to patient, it usually ranges between 1 nanogram and 50 nanograms, which is sufficient for the test. It is a mixture of tumor and normal DNA, and the percentage of tumor DNA differs widely between patients as well.

To demonstrate that Onco-Detect works, Signature has tested it in three clinical studies, using samples from the MSKK collection.

For the first study, it sequenced matched primary tumor and plasma samples from 44 stage 3 and 4 patients to see if the same mutations are present in tumor and plasma. While the mutational profiles matched for some patients, others had additional mutations in either the tumor or the plasma sample. "What it means is that you cannot assume that you have a one-to-one mutation profile between both tissues," Rosenthal said.

In a second study, they sequenced matched primary tumor, plasma samples, and liver metastases from 16 patients and found that about a third had identical mutations in all three tissues, while some had additional mutations in one of the three tissues that did not appear in the others.

In a third study, they sequenced two plasma samples from the same stage 3 and 4 patients, one before and one after surgery. With one exception, they did not find any mutations post-surgery.

Signature is about to test Onco-Detect in a prospective multicenter study in Germany for monitoring metastatic colorectal cancer patients undergoing treatment. The study, which aims to enroll between 200 and 400 patients, will collect plasma samples every month, including before and during treatment, and correlate the mutational profile from Onco-Detect with treatment response.

Changes in the frequency or profile of mutations might indicate whether a patient responds to the treatment. "It's an additional surrogate marker for the oncologist to manage that patient," Rosenthal said. In addition, new mutations might appear. "For instance, you had a particular actionable mutation in the beginning, and suddenly you see a second actionable mutation that you were not aware of – that's important information," he said.

In stage 3 patients that are currently tumor-free, the hope is that Onco-Detect will be able to detect relapse several months earlier than conventional image-based tests, allowing doctors to react more quickly. "If you can gain six to 12 months to pick up micrometastases, then these patients could be scheduled for treatment for metastatic disease earlier than they are currently scheduled," Rosenthal said.