January 15, 2014 (Vol. 34, No. 2)
Gail Dutton
Specialist Company Strives to Improve Patient Selection and Treatment Outcomes
Ariana Pharma exhaustively mines and analyzes multiparametric data (biochemical, genomic, proteomic, genetic, metabolic, imaging, clinical, etc.) to identify patient responders and reduce clinical and drug development risks. The process generates actionable information using fewer biomarker signatures than traditional data-mining methods, improving both efficiency and efficacy.
“Our technology is coming at the right time to transform data into actionable knowledge,” Mohammed Afshar, Ph.D., CEO, asserts.
Ariana’s Knowledge Extraction and Management tool, KEM®, uses formal concept analysis to identify all possible relationships in the data, including hidden or complex relationships often missed by conventional statistical analyses. Relationships are then combined using logical operators to develop a short, efficient signature.
“This is one of our strengths,” David Morley, Ph.D., vp, computational technology, declares. “We identify the most efficient signatures with high specificity and sensitivity, so our predictive power is strong.”
Ariana also uses a rules-based analytics approach. Not only can clinicians get the top-line information they need, they can also drill down into the assumptions and identify the strength of the relationships for cases that are less clear-cut.
Ariana Pharma’s KEM® Query application is an intuitive, visual query tool for signature-pattern matching. KEM Query’s uses include the analysis of protein profiling data, cell-line profiling data, and transcriptomic data. The screen above gives a glimpse of the graphical user interface, which Ariana designed in collaboration with intactile DESIGN.
Collaborations
“Globally, if you look at multicomponent biomarker signatures, only four have been validated by the FDA,” Dr. Afshar points out. Ariana, collaborating with Bio-Rad Laboratories, is identifying short, high-performance biomarker signatures to develop prognostics to identify likely responders to a therapy combatting hepatitis C infections. The therapy is expected to launch in 2015. “This will be our first product to hit the market,” Dr. Afshar notes.
The Winther trial, which began recruiting last April, is evaluating the clinical value of this analytic approach. It is being conducted with the Worldwide Innovative Networking (WIN) Consortium in personalized cancer medicine. “The Winther trial will provide tailored therapy for patients based on genomics, transcription, and other data,” Dr. Morley says. “There is a huge need to push big data analytics to the patient for direct use by clinicians.” Personalized medicine drives that need.
During the Winther trial, Ariana will build a commercial system that contains an algorithm that is meant to help clinicians make decisions about cancer therapies. Here, the algorithm is being tested within the trial while it is ongoing, explains Dr. Morley: “Current methods would enable clinicians to recommend treatment for about 30% of the participants. By using the trial to validate the initial algorithm, our goal is to recommend a therapy for 100% of the enrolled patients.”
Ariana is also involved in a joint effort with Pierre Fabre, Research & Development Center, Toulouse. This collaboration, which evaluates patients with fibromyalgia, has shown that KEM is able to identify subpopulations and increase efficacy fivefold.
Last spring, KEM was tapped for the French government’s Innovative Models Initiative, one of the largest national collaborations in France. This initiative, which boasts 17 partners and a total budget of €41 million ($56 million), aims to characterize, standardize, and apply biomarkers in oncology treatments.
Ariana also is collaborating on an FDA project in pharmacogenomics, reducing the initial panel of about 26 markers to a signature combining only 9 markers, with improved performance and efficiency. “We have a large number of projects in oncology and Alzheimer’s and other diseases that demonstrate shorter, more effective signatures using this platform,” Dr. Afshar comments.
“We’re involved in seven large multinational multiyear product consortia in oncology, personalized medicine, virology, and home care,” Dr. Afshar adds. Ariana is receiving more than €4.1 million ($5.6 million) for these product development projects, as well as downstream royalties upon commercialization.
Industry Attitude Shifts
Five years ago, pharmaceutical companies feared patient stratification would reduce market size. “The FDA’s emphasis on clinical trial veracity and improving the process has given analytics a leg up,” Dr. Afshar observes. The FDA guidelines released December 2012 specified how to enrich trials by using data mining to identify the most likely responders. This acceptance of patient stratification changed trials dramatically.
“By narrowing trials to the populations that can most benefit, the numbers of patients required for a trial decreases, as does the length of each trial, while the probability of success of those trials increases. The acceptance of that approach has had a tremendous impact on our business,” Dr. Afshar points out.
“Payers are unwilling to pay for marginally effective drugs, so they are pushing for stratification,” Dr. Afshar continues. “And finally the issue of side effects has become important.” As a consequence, remarks Dr. Afshar, “Algorithms have become the heart of diagnostics.”
Analytics have changed during the past few years, too. “Traditionally, datasets had few columns, but many rows,” Dr. Morley observes. “Now, there are many more descriptors, fueled mainly by the overwhelming availability of omics and sequencing technologies, hence many columns and relatively few rows.”
The availability of richer datasets implies that there are multiple ways to build predictive models. “Unless you allow for that, the models you generate won’t be applicable to new patients,” Dr. Morley advises. Ariana’s models allow for generalization, so they can be used as routine diagnostic or prognostic tools.
Diagnostics Products Offer Growth
Ten years ago, Ariana spun out of the Institut Pasteur, where it performed data analysis in-house. Today, Ariana still performs this kind of work and fee-for-service analysis provides current revenue. The company, however, is transforming itself from a service company into a diagnostic product company. It is generating revenue through milestone payments and building a royalty stream.
“The diagnostic product model, with its software-as-a-service (SaaS) component, allows us to capture the value of the technology we develop,” Dr. Afshar explains. “This is our long-term growth strategy.”
“Our goal is to evolve this into a commercial package and, while doing that, to work on the next generation of software,” Dr. Afshar concludes. “It’s like a relay.”
Ariana Pharma
Location: 28 rue du Docteur Finlay, 75015 Paris, France
Phone: +44 (0) 208 432 3284
Website: www.arianapharma.com
Principal: Mohammed Afshar, Ph.D., CEO
Number of Employees: 19
Focus: Ariana Pharma focuses analytics on patient stratification, transforming disparate data into actionable knowledge that increases healthcare access and decreases healthcare costs. The company is evolving from an analysis provider into a developer of diagnostic tools.
