NEW YORK – Announcing a collaboration with the University of Maryland School of Medicine earlier this week, biotech startup Anixa Biosciences (formerly Anixa Diagnostics) now expects to commercially launch its prostate cancer confirmation assay by early 2020.
The San Jose, California-based firm, which was founded in 2017, will offer the test as an intermediate step that patients with high PSA levels can use prior to undergoing invasive tumor biopsy surgery.
Called the "Cchek Prostate Cancer Confirmation" (PCC) test, Anixa's assay uses flow cytometry and artificial intelligence to measure a patient's immune response to cancer.
"Instead of looking for a specific mutation, we're looking for a general response of the immune system and identify changes that respond to the presence of tumors," Anixa CEO Amit Kumar explained. He said that Cchek relies on a patient's immune system to act as the "main amplifier" by indirectly quantifying the number of a subtype of heterogeneous corrupted immune cells called myeloid-derived suppressor cells (MDSCs).
Myeloid cells usually expand and travel to the site of a tumor as part of the body's immune response to cancer. However, tumors can develop ways to corrupt the formation of normal myeloid cells, turning them into immune suppressors.
Chihiro Morishima, director of University of Washington's Department of Laboratory Medicine and Research Testing Services, who is not affiliated with Anixa, explained that MDSCs develop in an environment that "skews toward immunosuppression" in a patient's bloodstream. Noting that MDSCs have an immune-regulatory function in healthy patients, Morishima said that the amount of the cells is typically higher in cancer patients.
"Monitoring MDSCs is not a new idea, as people have noted that looking at these cells have been a good way to diagnose cancer and evaluate if patients can respond to therapies," Kumar said. "However, the problem is that there is a subtle visual difference between the cells of healthy and cancer patients."
Explaining that Anixa's researchers collect several milliliters of a patient's blood sample, Kumar said that the team isolates white blood cells and labels them with a cocktail of fluorescent molecule-tagged antibodies, which bind to antigens on a cell's surface. Pushing the cells through a flow cytometer, the researchers measure the fluorescence of up to 13 protein biomarkers in about 60,000 cells from the patient.
Uploading the patient data to a digital file, Anixa's researchers feed the information into a trained neural network. The technology recognizes biomarker patterns on the cell's surface to determine if the body's immune system is actively reacting to a cancer. Kumar said that the overall time needed to deliver actionable results is about two to three hours.
Although Anixa has yet to publish any of its data in peer-reviewed literature, Kumar said the firm has presented posters detailing work on the assay at several conferences and claims that the tool reaches a sensitivity and specificity of 92 percent in prostate cancer. He highlighted that the technology has also successfully detected evidence of tumors in 19 additional different cancers, including breast, lung, colon, liver, and other less common tumors in the firm's internal studies.
As part of Anixa's plans to commercialize the Cchek PCC test, the firm signed an agreement with San Jose, California-based ResearchDx in February to use its CLIA-approved labs to validate the assay as an LDT. After completing an analytical verification study at ResearchDx's lab in March and an analytical verification study in May, Anixa began a clinical validation study on the PCC test in clinical prostate cancer blood samples.
Anixa has since signed collaborations with different academic and commercial groups to use their prostate cancer samples as part of the ongoing validation study. Partners include the US Department of Veterans Affairs' Hunter Holmes McGuire Research Institute, VA Maryland Health Care System, New Jersey Urology, VA North Texas Health Care System, Potomac Urology Center, and most recently the University of Maryland School of Medicine.
"The universities and physicians are sending prostate cancer samples to ResearchDx's lab, where they're running on our systems and helping us build clinical data," Kumar explained.
Anixa hopes to collect about 500 to 600 patient samples between the different cohorts as part of the study in order to demonstrate Cchek PCC's clinical validity. Kumar anticipates launching a laboratory-developed test in early 2020, following the completion of the clinical validation study later this year.
In addition to ResearchDx, Anixa started partnering with the University of Pennsylvania's Wistar Institute in 2015 to further develop the Cchek technology. While Anixa developed the original platform, Kumar explained that his team began working with Wistar Immunology Professor Dmitry Gabrilovich to understand the role of MDSCs in cancer and their potential use in the diagnostic and therapeutic spaces.
Anixa has also been issued multiple patents related to the Cchek technology from the US Patent and Trademark Office since the firm was launched in 2017, with the most recent patent issued in July.
Kumar acknowledged that his team has struggled to collect clinical prostate cancer samples to train the Cchek's neural network to differentiate between controls and cancer samples. He noted that collecting appropriate clinical samples requires a lot of effort on the practical side for both Anixa and physicians working with their patients. In addition, Anixa aims to solely use samples from US-based verified commercial and academic institutions, rather than contracting groups outside the US to collect samples and mail them to ResearchDx's lab.
"The challenge with AI is that if you give it bad data, the machine-learning step won't work correctly, since it will fail to recognize differences between samples," Kumar said. "You need to make sure that the reference samples are either cancer or control-based, and the data has to be accurate for the AI to produce an equally [accurate] result."
Kumar argued that using flow cytometry instead of sequencing-based methods to detect the presence of cancer allows Anixa to minimize the cost of the assay while offering a faster turnaround time. In contrast to ctDNA-based tests that Kumar said require "complex analysis, take a lot of time and labor, and could cost between $3,000 to $4,000," he believes that the Anixa can launch the test for "a couple of hundred dollars."
Since Cchek PCC monitors the immune system using white blood cells instead of examining the tumor itself, Kumar also believes that his team can identify cancer earlier than methods that require the tumor to shed circulating tumor cells or circulating tumor DNA (ctDNA) into the bloodstream.
Because the Cchek technology has successfully detected additional cancers during Anixa's internal studies, Kumar hopes to eventually offer the tool as a pan-cancer detection assay as well. However, he acknowledged that such a test would run into problems with distinguishing the exact source of the tumor in the body.
But, Kumar noted that undisclosed firms who have been developing cancer diagnosis tools using ctDNA have reached out to Anixa to potentially couple Cchek technology with ctDNA assays, which could potentially solve that issue.
"First, we'd develop an inexpensive, confirmatory test like ours to tell if a patient has cancer based on their MDSCs or other immune cells," Kumar said. "The clinician could then use a ctDNA-based test to determine which specific mutation the patient has and identify the source of the cancer."
After launching Cchek PPC in early 2020, Anixa expects to move testing to its own CLIA-approved lab, followed by a demonstration of the assay's ability to key opinion leaders in the prostate cancer field.
Anixa is not the only firm exploring the use of measuring the body's immune response to cancer. Alameda, California-based OncoCyte has been developing an immune gene expression biomarker panel called DetermaVu to determine whether a lung nodule is malignant.
Blood-based cancer detection firm Freenome is also developing a tool that measures genomic changes in immune cells as they interact with a tumor. While full data on San Francisco-based Freenome's multi-omics approach will not be published until 2020, the firm pointed to a study published last month in BMC Cancer that showed the tool's ability to define both tumor- and immune-associated signatures that can indicate the presence of early-stage tumors.
However, UW's Morishima pointed out that attempting to detect tumors in the earliest stages using genomic changes or other immune-based techniques like quantifying MDSCs will ultimately depend on the specific type of cancer. From her team's work in studying relationships between MDSCs and liver cancer, Morishima believes that MDSCs are mostly detected in patients with large tumors and possibly during vascular invasions in liver cancer.
"Usually DNA techniques are quite sensitive, and so the ability to identify a marker of a tumor before measuring by imaging would be fantastic," Morishima said. Although, she noted that in general, "the question is whether the technology is there yet, especially if you don't have a high probability that the cancer is present, as there would have to be a lot of data validating how sensitive and specific the test is."