Tissue phenomics is a new term given to the process that begins with tissue datafication and ultimately arrives at knowledge discovery. It is a way to structure images meaningfully and extract statistical data from relevant objects, regions and textures in tissue, including the subsequent data mining procedures. This cannot be done manually and enables increased medical and scientific insights.
Tissue phenomics is more than just data collection paired with data-mining. It is a comprehensive approach that, in the first step, already relies on pathologists to guide the process of datafication. In the second step concerning the meaningful statistical data aggregation, insights from pathologists and biologists are required to help guide the data-mining tool through the endless ocean of possibilities.
In other industries where the structure of data is much more well-defined, a big data analytics tool might be able to automatically spit out an answer to a question. This doesn’t work for tissue images, however, because there are so many possibilities for structuring data and defining regions. For tissue phenomics to work, it must be guided by pathologists’ and biologists’ knowledge and hypotheses so that the data extraction is defined and limited, and, ultimately, meaningful.
That said, through tissue phenomics, a huge amount of hypotheses, including very vague ones, can be tested and the right ones can be selected. Additionally, by optimizing parameters, the best-selected hypotheses can automatically be sharpened. Furthermore, through this process, new insight will be gained, leading to novel hypotheses that, again, can be evaluated through tissue phenomics.
As such, tissue phenomics represents a combined top down/bottom up approach, where biomedical knowledge (bottom up) is used to drive a top down mining procedure that, in turn, will stimulate a new bottom up trial.
The Value of Data
Tissue phenomics opens new doors for research. With it, pathologists can conduct experiments that aren’t possible through visual tissue inspection alone, obtain data and statistics that aren’t otherwise plausible and ultimately make new and better predictions about patients, treatments and outcomes. A big data approach to tissue studies will help uncover novel diagnostics and support drug development to significantly impact patient lives.
Pathologists and medical biologists, by their nature, identify meaningful structures in images and predict clinical outcome and drug response, but if we can do that in a quantified way, we can achieve much more. Areas where significant progress is being made already include breast, prostate, colon and lung cancer, but the potential for big data and tissue phenomics extends to all cancers as well as other disease areas-having access to better data can only result in better therapies and treatment decisions.
We know that a comprehensive big data approach to medicine, which includes tissue data, is crucial to making big advancements, but the question is, “How will we get there?” The first step on the road to leveraging big data in medicine is to change the perception and promote openness for data collection-clinical as well as digital tissue data. Also, by slightly adjusting lab processes, tissue staining and workflows, thinking in terms of leveraging a machine versus only manual evaluation is helpful. Simultaneously, it’s important to remember that, while big data requires the use of more technology and an increasing level of automation, the human element is still crucial to guiding the process and making decisions as well as improving technology.
Moving to a big data approach is a paradigm shift that will start to happen naturally as the thought processes and technological approaches change in the lab. Researchers are realizing the need for more powerful tests to ensure they’re studying the right drug candidates and matching them to the right patients. We’re already starting to see movement and buzz across the industry as the concept takes root, and within a few years, the big data approach will become inextricably embedded in the framework of medicinal research and clinical practice.
Gerd Binnig, founder and chief technology officer, Definiens, was awarded the Nobel Prize in physics for his work in scanning tunneling microscopy.