Off to the data mines

When Ariana Pharma announced this summer that the WINConsortium had chosen it to develop new personalized cancer medicine softwarethrough the consortium's WINTHER trial, it wasn't just a validation ofParis-based Ariana's data management skills in life science, notes Dr. DavidMorley, Ariana's vice president of computational technology—it was also a signof how much data mining has shifted from its early beginnings in drug discoveryto something much more oriented toward clinical trials.

"It's a very exciting time for data mining, and we've seen areal move from retrospective use of it to a more proactive role," Morley tells DDNews. "In the roughly 10 years Arianahas been bringing data mining to life sciences—after years of use in otherindustries and areas like marketing, customer analysis and frauddetection—we've seen these techniques move from a very niche mode to being muchmore mainstream."

Although Ariana started off marketing data mining for drugdiscovery primarily—and still does market for that purpose—Morley says that hiscompany's Knowledge Extraction & Management (KEM) technology is overwhelminglybeing sought for clinical trial applications now, with significant action inbiomarker identification and analysis as well. And with the growth ofpersonalized medicine efforts, the biomarker area is, to some degree, mergingwith the clinical trial applications for data mining, he adds.

"The WINTHER trial is an example of how data mining hasmoved from the chemistry side primarily to a role of facilitating adaptivetrials and identifying the best responders among patients to truly tailor therapiesfor them," Morley says. "If you look from the 30,000-foot view, roughly a thirdof the time when a drug is prescribed, it works, another third of the time itdoesn't work and even worse, a third of the time, you have huge adverseeffects. Data mining enhances your percentage of success."

The WINTHER trial is being conducted by Worldwide InnovativeNetworking in personalized cancer medicine, otherwise known as the WINConsortium, and is reportedly the first clinical trial offering 100 percent ofrecruited cancer patients a therapy choice guided by individual patientbiology, and it is said to represent "a breakthrough in current oncologypractice which at best offers biology-guided therapy to 30 percent of patients.This is because most cancers are diagnosed at a late stage and for the vastmajority of patients the therapeutic choice is based on standard protocols."The WINTHER trial promises to offer personalized medicine based on individualDNA, RNA and microRNA profiles, with trial results expected in 2015.

In spring, Ariana's KEM technology was secured for theFrench government's Innovative Models Initiative, one of the largest nationalcollaborative projects in health in France, which will focus the development ofnew drugs and personalized medicine approaches in cancer by creating anindustrial-scale pipeline that characterizes, standardizes and makes use ofpredictive cancer models.

"Data mining is something that's seeing particular use andoffers particular promise in the area of oncology," Morley says.

The burgeoning data mining market in life sciences may beone reason that Research and Markets recently has been promoting athree-year-old publication titled "Pharmaceutical Data Mining," noting, "In the era of post-genomic drug development,extracting and applying knowledge from chemical, biological and clinical datais one of the greatest challenges facing the pharmaceutical industry."

Rithme, which is based in France like Ariana and deals withsimilar customers, noted that data mining "can successfully contribute in theexplanation or prediction of complex phenomenon in healthcare andpharmaceutical industries," with some of the key examples being dose-responseanalysis, cohort analysis, longitudinal and survival analysis, benefit-risk analysisand predictive modeling.

The need to rely more on data mining, particularly in thearea of clinical trials, is one that has been recognized for several years,with the Institute of Medicine noting in a 2010 report that many considerrandomized clinical trials, for so long the gold standard in pharmaceuticalresearch, "to be unsustainable as an approach to addressing the large number ofresearch questions that need to be answered because of the time and expenseinvolved." Data mining can help by evaluating trial study feasibility to beginwith, refining the process of patient enrollment and even assisting inpost-marketing pharmacovigilance.

In the end, one of the biggest reasons Ariana, Rithme andothers will continue to see a boom in data mining interest, though, is money.After all, as the report "How Big Data Can Revolutionize PharmaceuticalR&D" by the McKinsey Global Institute notes, making better use of so-called"big data" could have a $100-billion impact on the healthcare industry just in theUnited States, resulting from "optimizing innovation, improving the efficiencyof research and clinical trials and building new tools for physicians,consumers, insurers and regulators to meet the promise of more individualizedapproaches."