CHICAGO – Earlier this month, Indivumed introduced nRavel, an artificial intelligence-based bioinformatics platform to support precision cancer research.
The new product marries the Hamburg, Germany-based oncology firm's IndivuType multiomics database with machine learning and a series of analytics tools to help researchers uncover new drug targets and develop novel cancer treatments. IndivuType is an online resource that combines genomic, transcriptomic, proteomic, and digital histopathology data with clinical information from thousands of cancer patients.
The nRavel platform has three main components: a series of cancer disease models, an automated machine learning system that Indivumed licenses from JADBio, and a series of systems biology and oncology informatics tools.
"We think of nRavel as the core of what we like to call our flywheel of discovery" since the firm has its own multiomics dataset for studying cancer biology, said Indivumed's chief business officer, Roald Forsberg, who heads the company's IndivuType informatics business unit. "Because we have now the nRavel platform, we can derive insights from that fast and efficiently."
Indivumed has included six toolboxes on nRavel: immuno-oncology, pathways and signaling, biomarkers, genomics, signatures, and clinical analytics. Forsberg said that these are not distinct products but rather a means to categorize different aspects of tumor biology, since many of the tools overlap.
"They've grown organically based on how we've used the data ourselves and how we've engaged with customers and service projects and partnership projects," Forsberg said. He expects the makeup of each toolbox to evolve.
The company also built a series of internal processing pipelines to generate its multiomics data. This internal infrastructure, which has not been widely publicized, features a workflow integration and management system to automate omics processing.
"We have very high standards for the tissue that we extract. We also have very high standards for the data we then get out of the tissue," Forsberg said.
He half-jokingly said that Indivumed will "pester" its customers with lots of quality metrics, which Forsberg said are necessary to make the data as useful as possible.
Forsberg, who joined Indivumed a year and a half ago, called nRavel the culmination of a multiyear effort to build an informatics infrastructure, adding that nRavel is the company's way of consolidating all of its informatics pieces onto a single platform.
CEO Hartmut Juhl said it was the fulfillment of potential that he has envisioned for several years.
Since Juhl, a surgical oncologist with an interest in translational cancer research, started the company in 2002, Indivumed has developed what he called an "industrial-like standard for tissue collection in clinics" to ensure sample quality.
Only in the last couple of years has the informatics technology caught up to achieve Juhl's desire to be able to "fully decipher the cancer biology of every patient, to understand the disease, to target the cancer according to what is available," as well as to develop new compounds.
In early 2018, after Indivumed received a €40 million ($47.7 million) loan from the European Investment Bank, Juhl said that the firm would use the funding to expand the breadth of tumor data and analysis services it provides to pharmaceutical and biotechnology companies, with an eye toward eventually offering more clinically oriented resources to enable physicians to treat patients more effectively and select relevant clinical trials.
Juhl said this month that he was smiling while rereading that news three years later because his vision was now being realized.
In addition to beefing up its informatics in the interim, the firm has shifted its business model. Indivumed is reducing its dependency on out-licensing its database. Instead, it has embraced development of new targets and early drug discovery, both in house and in partnership with pharma companies.
Indivumed has founded two drug-development subsidiaries this year. Ix Therapeutics is a joint venture with Xlife Sciences established in March to develop functional antibodies, while the company teamed with Mainz (Germany) University Medical Center in May to create KHR Biotec, which seeks to create treatments for RAS mutations.
Last year, the firm created the Oncology Alliance for Individualized Medicine, or Onco AI-Med, a partnership between cancer clinics and research institutions selected based on their research activities and expertise in molecularly informed cancer care.
Indivumed also has partnerships with companies including Ultivue, Evotec, Biognosys, Personalis, as well as with numerous academic institutions.
Personalis handles all the next-generation sequencing analysis for Indivumed, uploading data to the cloud to feed Indivumed's multiomics processing pipelines. "That has been a pretty significant effort on our side to get all that up and running because of the complexity" related to dealing with proteomics, transcriptomics, phosphoproteomics, and even small RNA and whole-genome sequences, Forsberg said.
In introducing nRavel this month, Indivumed said that Merck KGaA has become an early user of the new platform. Through Indivumed, the drug company declined to comment.
Like earlier Indivumed technologies, the new nRavel platform runs on the Amazon Web Services cloud.
The company has amassed a large biobank of rapidly frozen samples of tumors and adjacent normal tissues. "We transform the raw data, which is the frozen material, into a truly multiomics database," then analyze the data with AI and other bioinformatics, Juhl said.
Indivumed strives to freeze samples within 10 minutes of collection to ensure the quality of its specimens. Juhl said that this speed is essential for proteomics and transcriptomics because expression levels can change within minutes after a surgeon excises tissue.
While numerous companies are now applying AI to find new drug targets and compounds, Juhl said that Indivumed distinguishes itself by having the "right" data, namely comparable tissues and tissue data to facilitate large-scale multiomics analysis, including transcriptomics, proteomics, and even phosphoproteomics.
The firm also has built a longitudinal database for certain cohorts of patients.
"As a discovery platform, [nRavel] is a unique tool simply because already we have much different data than most others" who use the same publicly available databases "of limited value in the quality," Juhl said.
The in-house biobank and the clinical network also allow Indivumed to validate its findings rapidly.
"We can do immunohistochemistry if we want to validate where the proteins are that we've identified. We can go back and do things like immunopeptidomics," Forsberg said. "We can also do things like spatial transcriptomics on the data."
Forsberg said that the company wants to "tie the findings to the 'geography' of the tumor" to create an efficient feedback loop involving the technology, the database, and clinical researchers.
In a poster presented at the virtual American College of Clinical Oncology conference this month, researchers from Indivumed and some of its academic partners showed that with just two biomarkers, Indivumed and Onco AI-Med technology can distinguish between cancers on the left and right sides of the colon.
One of the authors of that paper was David Kerr, a professor of cancer medicine at the University of Oxford and a former president of the European Society for Medical Oncology.
Kerr first learned of Indivumed from a colleague, John Marshall, who is chief hematologist/oncologist at Medstar Georgetown University Hospital in Washington, D.C., and chair of Indivumed's clinical advisory board. Marshall was lead author of the ASCO poster.
Kerr said that he was drawn to Indivumed because the company was meticulous with its tissue collection and curation.
"For us to be able to [make findings like those described in the poster], we need to be able to collect tissue in a very clearly well-defined way," Kerr said. "We need to link that to clinical data, both demographic and outcome data, and then to do multiomic or multidimensional investigations, analysis, and studies on that tissue."
Kerr said that the nRavel technology allows cancer researchers to interrogate tissue in many different ways. "The future seems to be multiplexing, or pulling together genetics, and [seeing] what happens to DNA, what happens to RNA, what happens to protein, what happens to phosphorylation of protein," he said.
Pulling together these activities produces a more complete picture of potential biomarkers, drug targets, and patients who might benefit from certain treatments, explained Kerr, who has a longstanding interest in translational oncology.
"The integration of all of their toolboxes lends huge power to it," Kerr said. "Linking [multiomics data] with very clever bioinformatics and mathematical analysis of the data seems to be the way ahead."
While Kerr has been in cancer research for decades and has built up his own network of colleagues and collaborators, he said that the Indivumed network has helped him bring new people into the world of translational cancer medicine. He said that using tissue biology to improve the matching of patients to treatments is "the basis of precision cancer medicine."
With its full-fledged bioinformatics platform now on the market, Indivumed will now be looking to license compounds it develops through its pharma subsidiaries to larger drug companies and to enter into codevelopment deals to advance its findings to clinical trials, according to Juhl.
"In the long term, we also expect … to go into clinical trials with this [Onco AI-Med] network as well, so that we can also support the clinical development and accelerate it," Juhl said. Onco AI-Med also could give the company a foothold in molecular diagnostics because the alliance returns multiomic test results to patients, he added.
"This is still research, that's for sure," Juhl said. "But we do it with clinicians already with the hope that this is a new diagnostic approach, which you can say is a more systems biology approach [to] understanding diseases."
Juhl said that the Onco AI-Med network will be a vehicle for distributing the firm's reference database. "The long-term vision of the company is to have this reference database, which you then can utilize to give directions to patients and better therapies," he said.
Indivumed has received some undisclosed private funding since the European Investment Bank loan, and Juhl said that the firm is in negotiations for new investment.
"We [are turning] Indivumed, which is a platform technology, this development analytics platform, into a biotech drug development company," Juhl said.