The future of healthcare is cognitive

This article was originally published here

We are on the cusp of better harnessing information to find optimal treatments for individual patients, but the smart systems that will make this possible are still being developed. These systems will collect data (e.g. medical images, genetics, and other test results) from many patients around the globe, allowing physicians to use real-life evidence to identify treatments that have worked for similar patients.

Medical imaging is the area of medicine that is rich in standardized, structured data, yet much of the value of this data is lost because it is not easy to accurately measure. The measurement process is manual, time-consuming, and labor-intensive, and the tools available to radiologists have limited computing power and are often difficult to use. Because of the time required to quantify images, and the volume of images radiologists need to read each day, accurate measurement of all the regions of interest is often skipped. Clinically, this translates into diagnostic decisions that are mostly based on visual assessments, with techniques that vary from physician to physician. This qualitative analysis results in unacceptably high variability and great difficulty in accurately tracking disease from scan to scan.

If we can’t track disease accurately, then patients may continue to receive ineffective treatments for far longer than needed. This incurs unnecessary costs and can have a negative impact on the patient’s prognosis. Also, without accurate tracking, we cannot identify early enough whether a stable condition has worsened or changed, making it much harder to catch issues early.

This is why we need much faster, yet accurate and consistent solutions to quantify medical images and track changes between scans. In the near future, we believe that consistent, automated measurement of disease will enable more precise diagnoses and better treatment decisions. Automated analysis will also allow physicians to focus more of their time on patients rather than performing tedious tasks and building reports manually.

Not only will patients benefit, but it will also help advance medical knowledge as a whole. Once automated and consistent image quantification is established, then big data analytics tools can be built from these quantified images, outcomes and other relevant clinical data. Aggregating the data of many patients with the same condition can lead to a better understanding of how their disease progressed under different treatments, and then we can begin to predict which treatments are more likely to result in positive outcomes for individual patients.

What is needed is a system that can automatically create accurate measurements in images, track these measurements across several scans for each patient, and aggregate this data from around the world. This type of solution is now feasible, as cloud infrastructure and Artificial Intelligence (AI) have both matured. With the cloud, we can aggregate data and can also access far more computing power than what is available at the hospital. With AI, we can accurately automate the identification of regions of interest in scans and their measurement. Furthermore, the combination of cloud capacity and AI also enables several new acquisition techniques that only used for research today because the file sizes they generate are too big for current systems. A powerful GPU cloud backbone is an enabler on both the visualization and AI quantification for these enormous, information-rich files.

The advent of secure, cloud-based imaging analytics has already begun to transform medical practice by providing timely, valuable information that was very difficult to access before. This new technology is being used in clinical settings today to better inform care decisions and improve outcomes, while enabling more efficient workflows across teams of healthcare providers.

Photo: nambitomo, Getty Images 

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply