Today’s precision medicine is based on the potential of biomarkers, stratifying patient cohorts and guiding the selection of indications to ensure the success of today’s most promising drugs. What is less known is that biomarkers can help discover the utility of next-generation therapies in the future. Ideally, the biomarker discovery platform is intended as an end-to-end solution that not only reduces the risk of clinical development, but also promotes early detection and supports translational research. Linking these different stages in the drug development life cycle can bring about change and create important data and information feedback.
It is estimated that the probability of successful drug development driven by biomarkers is more than 10 times that of no biomarkers. This is particularly prominent in oncology, where the clinical trial failure rate (combined at all stages) is estimated to be about 96%. These biomarker discovery platforms can help solve the problems of high failure rates in clinical trials and high R&D costs. To help bridge the gap from successful early detection to translating evidence to improving clinical outcomes, new opportunities need to be identified to treat complex diseases that have few options today.
Let us take oncology as an example. Cancer drug development will start from more and more public data sets (such as The Cancer Genome Atlas (TCGA) and Encyclopedia of Cancer Cell Lines (CCLE), And a large number of molecular and clinical data from patients captured through clinical trials and clinical practice. Often, translational research teams will release new genetic features that claim to have predictive, prognostic, or diagnostic potential for specific diseases, mechanisms, or drug classes. Sadly, these rarely work in a clinical setting. Utilizing these data resources and previous analysis work is not easy. You must take a cautious and systematic approach to QA/QC, and take time to understand the data and models before you hope to apply them to new therapies and patient cohorts. The ultimate goal is to use public and proprietary data to build predictive models for the next generation of life-saving treatments.
So, in order to increase the chances of clinical and commercial success, what are the key areas that healthcare companies need to focus on? Identifying a scalable data solution is a necessary step. Leading opinion leaders in the field of data science/healthcare AI agree that the FAIR (Searchable, Accessible, Interoperable and Repeatable) guidelines should provide information for research data management and IT systems. In addition to excellent data management, state-of-the-art bioinformatics and data science best practices are also essential. The development of biomarkers requires the analysis of data generated in the entire life cycle of drug development and in the real environment. When the data itself is optimally processed, the analysis of these data is most likely to succeed. The FAIR guidelines help ensure that data is processed and organized in such a way to achieve the best data science results through human and machine actionable insights.
In addition, it is important to answer questions such as what predictive data will people use or collect? Or, how to better rank potential targets that are more likely to be useful for safe tumor types to achieve success? These answers will enable companies to determine solutions to make more informed biological decisions. Artificial intelligence and machine learning can find a needle in a haystack, but they are only as useful as the questions we want to answer.
The biomedical community is increasingly relying on data science, focusing on new and creative ways to treat diseases. Combining computational biology and artificial intelligence-based methods, the goal is to identify new targets, pair them with effective chemistry, and define biomarkers to optimize the positioning of new therapies. This is a way to improve drug discovery and ultimately change the lives of patients with diseases that require effective treatment.
