A radiology department
Today, there is a vast amount of radiology data available to work with, providing the opportunity to analyze images and associated clinical data to improve diagnosis. Images are becoming far more sophisticated and accurate, and we’re able to view and quantify the inner depths of the human body.
However, as we look toward tomorrow, we see the opportunity for radiologists to enhance their impact by incorporating richer data sets analyzed by machine intelligence. We are interpreting images through deep learning, while quantifying more data, adding more sources and combining modalities that were once disjointed. We are going deeper and deeper into human physiology. With digital pathology we drill-down and analyze patients at a cellular level. We can even learn how these cells are “wired” by analyzing their DNA. So, why is all of this extra perspective important? The answer lies in the richness of contextual data and how we use it.
Handheld blood testing for point of care diagnostics ( In development. The shown device is currently not available for sale.)
But, at Philips, we see the bigger trend in data, where we must not just optimize devices, but also optimize workflows that incorporate multiple data-driven services. With our current, ever-expanding ecosystem of connected devices, new data sources and increasing usage of EMRs, it has become increasingly difficult for radiologists to see it all, know it all and use it all. Thus, it is imperative to have it all linked to a core informatics data and device platform – like the Philips HealthSuite digital platform – so that we can expose the disparate data points to algorithmic intelligence to construct a more comprehensive picture. However, this presents both new challenges and new opportunities for radiologists
Intersection of Radiology and Data Science
The growing amount of information available coupled with advancing technology is directly empowering radiologists, creating a new front in the field: radiologists as clinical data scientists. As cancer rates unfortunately continue to rise and clinical decisions become more complicated, radiologists can play a bigger role.
Complex disease patterns require stitching together fragmented pieces of information across many data sets to create a better picture. Radiologists capture more longitudinal data, while at the same time going deeper by quantifying imaging data and combining this with cell level data (using digital pathology) and even deeper with genomics, thereby creating a more precise diagnosis.
In addition, each observation, patient case and treatment can add to the cumulative body of knowledge: data is aggregated, analyzed and compared for insights. For radiologists to meet the challenge, we must provide the tools that amplify their expertise through advanced algorithms and machine-learning capabilities. Take for instance Philips’ IntelliSpace Portal 8.0, which just launched at RSNA. Here you have an advanced data-sharing, analytics and visualization platform that allows the use of data-algorithms. What once took an almost insurmountable amount of time, resources and careful analyses is now available through a few clicks. It is a prime example of how the access to a wealth of data – dense, deep and wide – is now being realized to create a quantified patient.