Cross-Lab Data Collaborations – The Future Of Immunotherapy?
By Ralf Huss, MD
In January, UC San Francisco announced a research alliance with AbbVie, Amgen, and BMS supporting the collection and analysis of at least 500 tumor samples from more than 10 different forms of cancer to ultimately improve patients’ responses to cancer immunotherapy. Data collaborations and consortiums have been spearheaded previously both in the U.S. and in Europe, but they have often failed due to data privacy issues and challenges around data gathering and analysis across labs. Joining forces and bringing more data together is of high value to both industry and patients. However, it’s even more important for this kind of alliance to succeed since it can help identify treatment options more quickly, might open the door to more combination therapies, and is important to further the advancement of immunotherapy.
Why Now?
A few years ago the NIH developed data-sharing repositories that made a variety of data, including images, tools and assay results, accessible for use by other labs. Organizations such the National Cancer Institute (NCI) and various universities and research institutions manage the individual repositories. More recently, ASCO launched CancerLinQ to “harness Big Data” by organizing and sharing information from patients receiving cancer care globally into usable knowledge for oncologists. There are additional initiatives in Europe working to collect data in very centralized ways to improve patient care.
Up until now, however, drug development companies have shied away from spearheading their own collaborations. In the field of immunotherapy, a handful of Big Pharma companies have been racing to achieve first or second line approval of their own checkpoint inhibitors in a small set of indications, including nonsmall cell lung cancer and bladder cancer – collectively putting billions of dollars into research for a small handful of diseases. This approach is not getting them there any more quickly, however, and if one company “beats another to the punch,” the second company either needs to drop its program or conduct large and expensive trials to prove that its drug works marginally better than the already approved one.
The UCSF alliance is indicative of a shift — the industry is opening up to these kinds of partnerships, and we will no doubt start to see more of them in the next few years. Why? There is now enormous pressure on the industry around reimbursement and healthcare costs, and pharmaceutical companies need a way to bring treatments to market more quickly and efficiently. Pharmaceutical companies are realizing they can achieve more, both economically and scientifically, if they collaborate rather than only compete against each other to be first to market.
Research supports this. Many published studies indicate that combination therapies – such as a combination of drugs from two different pharma companies – can be more effective than monotherapies. This is where the science will meet the economic demand. Pharma companies are recognizing that it makes sense for them to come together and share their data to be more efficient and successful in the long term. From a scientific perspective, their drugs may work better in combination than alone, and if they can discover more effective therapies, and develop them more quickly together, they can get to market faster and reduce the costs of development. Not only do companies benefit from this approach, but patients also benefit from faster approvals and lower costs.
We can only truly optimize treatments and combinations, and effectively select the right drug for the right patient, if we bring more data together and use a real Big Data approach to drug development. Collaboration initiatives, sometimes through non-profit healthcare organizations or private foundations, are thus absolutely necessary to immunotherapy advancement, but in order for them to be successful, several challenges need to be addressed.
Collaboration Challenges & Considerations
The first challenge is data privacy. According to a Truven Health Analytics-NPR Health Poll, more than two-thirds of people are willing to share their health information anonymously with researchers to advance treatment options. However, the same survey indicated a heightened sensitivity about data security among patients, potentially due to the increase in major data breaches over the last few years. Data privacy is still an issue, and an increasing one. More and more patients are concerned about who can access their health data and how it is being used, and the industry needs to develop a strategy to address this issue (e.g., ensuring one company’s patient data doesn’t become mixed with another’s) before organizational collaboration will become standard. More than likely, effective data privacy strategies will be developed on a corporate level by the companies that provide solutions to the pharmaceutical industry.
The second challenge is around data quality. All parties involved need to be concerned that the data going into the analyses are robust, so that the data repositories are of high quality. Unfortunately, this is even more challenging for immunotherapy than it is for genetic testing.
Immunotherapy also requires an understanding of the tumor microenvironment (including not only the presence of key markers, but their location within the tumor and its environment) to measure immune response effectiveness. Appropriate quality tissue samples and technologies are needed to conduct a robust and sustainable type of analysis, so researchers can be sure that the expression of certain markers or immunological features in the microenvironment are reliable. Immunotherapy data is very dependent on the materials used, how they are collected, and how the testing is done, and presently, the industry struggles with a lack of standardization around these areas. For data collaboration to be useful, the data quality and data integration must improve.
Another challenge is that data often remains isolated; patient history and response information, clinical data, outcomes data, and all other genomics and blood data available are all being studied in isolation rather than being correlated via one standardized system. This prevents a holistic understanding of disease and can hinder findings. Data integration and data mining are increasingly critical to developing the best predictive value or stratifier for patient treatment in immunotherapy.
Immunoprofiling is an example of a possible predictive tool for cancer treatment made possible by data integration. The predictive aspect of the tumor microenvironment and the patient’s immune system signature (i.e., immunoprofile) may be able to help predict response to treatment, ultimately supporting therapy decision making. In short, it is used by researchers working in immuno-oncology to screen tissue samples for target biomarkers, using automated tissue image analysis, and measure immune response and obtain other information about the tumor microenvironment. All of this information is correlated with with patient outcomes, or tumor progression, and that information is then used to determine the choice of drugs for that patient.
Beyond immunoprofiling, however, it’s critical for any research alliance to not only look at each piece of data separately, but to integrate and mine the data. These partnerships usher in a new era of Big Data. Researchers will have the whole universe of available data without “boiling the entire ocean.” However, this means that basic as well as clinical scientists will need to think in broader terms; they will need to use technologies that enable them to look at the whole pool of data, and then use data-mining techniques to truly advance immunotherapy for patient benefit.
Regulatory Roadblocks
There are also potential challenges on the regulatory front that need to be considered. The classic FDA process, for example, requires looking at drugs individually and in comparison — not evaluating two in combination. Regulatory bodies will need to adjust to the new approaches being brought about by industry partnerships and Big Data, and structure drug-approval processes in a way that supports them. This could include fast tracking approvals, preliminary drug approvals, allowing the simultaneous testing of multiple drugs, reducing the number of patients required in clinical trials, or other new ways of approving drugs faster.
The use of more robust companion diagnostics (CDx) will also support the faster approval of drugs. Studies have shown that programs using selection biomarkers have higher success rates at each phase of development. Prioritizing biomarker studies and developing CDx are critical to more effective therapies and faster approvals in immuno-oncology in particular.
Cross-lab data-sharing, data integration, and data mining are important for the success of immunotherapy. These collaborations will enable researchers to make decisions based on larger datasets rather than only on drug response data from their own individual labs. We can only appreciate what this new alliance is trying to achieve — but we’re only at the beginning.
About the Author
Ralf Huss is chief medical officer of Definiens and has more than 20 years of training and experience in histopathology and cancer research. Prior to joining Definiens, he also served as global head of histopathology and tissue biomarkers at Roche Diagnostics.