Traditional medicine is often based on trial and error in patient care. We know from decades of experience that any given drug will work miracles in some patients and poorly or not at all in others. Variances in the disease also affect drug efficacy. For example, certain cancer tumors respond well to chemotherapy, while others build up a resistance over time. What we haven’t known is how to predict drug outcomes in advance of putting the patient through the regimen. But thanks to new developments, including and most especially big data analytics, that’s about to change.
Scientists from Mitra Biotech, a cancer diagnostics company, recently reported key findings that may pose a new way to predict clinical response to anti-cancer drugs. The results of their study are detailed in an article titled "Predicting clinical response to anti-cancer drugs using an ex vivo platform that captures tumor heterogeneity” published in the journal, Nature Communications.
“We do not limit our analysis to a specific gene or a drug pathway; rather, we examine the complete tumor ecosystem in order to predict both positive and negative clinical responses using a common set up,” said Dr. Biswanath Majumder, PhD, the lead author of this paper, in a prepared statement.
The research and what it means to healthcare and drug discovery
The research, a collaborative effort undertaken by a team from Mitra Biotech, Harvard Medical School, Dana Farber Cancer Institute, and the Broad Institute at MIT—along with other institutions in the United States and India—showed that by maintaining the complex structure and behavior of tumors in tissue culture plates, Mitra’s technology called CANScript can predict the response of individual patient tumors to anti-cancer drugs.
“The first of its kind, our technology and study show the possibility of predicting the response of patient tumors to anti-cancer drugs,” Mallikarjun Sundaram, Ph.D, MBA, co-founder, president and CEO of Mitra told me. Other founders include Sundaram’s colleagues at MIT, Harvard and the Dana Farber Cancer Institute in Boston. The company is backed by the investing firm Accel Partners.
"The outlook: speeding-up oncology drug development, reducing go-to market time and costs, and better health outcomes for cancer patients," he added.
How CANScript works
So how does the technology work? Here’s how it is explained in the report abstract:
The functional response of tumour ecosystems, engineered from 109 patients, to anticancer drugs, together with the corresponding clinical outcomes, is used to train a machine learning algorithm; the learned model is then applied to predict the clinical response in an independent validation group of 55 patients, where we achieve 100% sensitivity in predictions while keeping specificity in a desired high range. The tumour ecosystem and algorithm, together termed the CANScript technology, can emerge as a powerful platform for enabling personalized medicine.
This underscores an aspect of big data analytics that is often underappreciated: its ability to consider a much bigger picture in the analysis. It’s not just about analyzing more data, but being able to add inputs in algorithms to consider more influencing factors in the analysis.
No medical problem, you see, truly exists in a vacuum and yet in years past we’ve treated every condition as though it existed apart from the rest of the patient’s body. Big data tools make it possible to consider the entire patient—including all unique values plus the patient’s environment, extenuating circumstances, and lessons learned from all patients with similar health conditions as well as all medical knowledge throughout history and that which is discovered in real time.
In short, big data analytics can analyze a mind-boggling amount of information and serve the results directly to a physician in record time.
What it means to precision medicine
In the early days following U.S. President Obama’s announcement on the country’s new focus on precision medicine, many were concerned that futuristic research would drain funds, talent and attention from addressing more immediate and practical medical care issues.
But that’s an old way of thinking that is not at all applicable to how big data works. In the first place, big data analysis is much faster than traditional research methods. Many projects, related and unrelated, can occur simultaneously and produce results in record time. One project can look for ways to enhance clinical care immediately while also looking for new ways to treat the same ailment. We simply could not do any of this to this degree and this fast before big data tools came to be.
Critics of precision medicine tend to think in terms of years, even decades, in getting results as historically this was the case. But with big data, results can be had in days, weeks, or months.
This development underscores the truth of that as CANScript is addressing an immediate medical need in an entirely new way and making the breakthroughs quickly – precisely as everyone hoped to see happen in precision medicine. But it is also helpful in speeding future oncology drug development.
“We will be making several announcements in the coming months,” Sundaram told me, indicating that this development is just the beginning of Mitra’s advancements on this front.
You can expect other research teams and companies to similarly make rapid and radical advances in precision medicine as well.
What big data is ultimately heralding is a renaissance in medicine. Considering the amount of human suffering in the world, its arrival can’t come a moment too soon.
The nuviun blog is intended to contribute to discussion and stimulate debate on important issues in global digital health. The views are solely those of the author.