Greg Freiherr, Industry Consultant
Greg Freiherr, Industry Consultant

Greg Freiherr has reported on developments in radiology since 1983. He runs the consulting service, The Freiherr Group.

Blog | Greg Freiherr, Industry Consultant | Radiology Imaging| June 22, 2016

How Data Mining Can Personalize the Practice of Medicine

Screening, imaging, big data

Revolutions need data. They need to explain why the present is not working as well as it could, so that a better route might be mapped. But how to gather that information? Data mining may light the way.

Data mining uses algorithms to connect the “dots” found in large databases. Through the analyses of these databases, referred to as Big Data, it is possible to uncover patterns that show how one event leads to another; visually documenting facts so that new associations might be made; even — possibly — developing models that predict how changes might affect the future.

Anyone who doubts the power of data mining — and the modeling that can come from it — need look no further than the increased accuracy of weather forecasts in recent years. Similar potential exists for mining medical data, particularly screening data.

In April, a paper published in the Journal of Digital Imaging published an article described the use of a data mining tool to examine screening trial data regarding prostate, lung, colorectal and ovarian (PLCO) cancers. The authors stated that “data mining the PLCO dataset for clinical decision support can optimize the use of limited healthcare resources, focusing screening on patients for whom the benefit to risk ratio is the greatest and most efficacious.”

In February the American Journal of Roentgenology published research described the use of structured reporting and the development of large databases for use in mining breast imaging data. The Mount Sinai Health System research team concluded that “robust computing power creates great opportunity for data scientists and breast imagers to collaborate to improve breast cancer detection and optimize screening algorithms.”

The potential of new data technologies, exemplified by data mining, may “facilitate outcomes research and precision medicine,” wrote the Mount Sinai researchers.

Because it generates such enormous quantities of data, medical imaging — particularly its evolution into enterprise imaging — could be a fount of information for this coming revolution. Through data mining, patterns may appear that will allow truly best practices to emerge, ones that fine-tune patient care protocols and lead to consistency in healthcare that promotes wellness and increased efficiency.

A paper presented at last year’s meeting of the Society for Imaging Informatics in Medicine (SIIM) posited that Big Data mining could even be used to create individualized patient management strategies. (SIIM 2016 will be held June 29 to July 1).

“The rise of ‘Big Data,’ in concordance with the recent publication of massive clinical datasets such as the results of the PLCO screening trial and the National Lung Screening Trial, allows the development of mining tools to more precisely target such recommendations,” wrote University of Maryland radiologist Arjun Sharma.

“Using these matched data, an individualized diagnostic decision-support system can personalize imaging, testing, follow-up intervals, intervention and prognosis,” Sharma wrote.

His claim was supported by results from the use of a Web-based application that analyzed demographics, risk factors, and screening examination results. Such analyses can render outcomes that might be used to guide patient-specific treatment and care management strategies, he wrote. Incorporating this outcome information into imaging software, Sharma concluded, could provide “a data-driven personalized approach to cancer screening.”

Sharma suggested that the use of such applications could maximize economic and clinical efficacy through the early identification of patients. Whether this potential is achieved has yet to be demonstrated, at least on a wide scale.

Attempting to do so, however, may be the path to the future.

Editor's note: This is the final blog in a four-part series on screening. The first blog, “Screening: How New Looks at Old Modalities Might Turn Imaging Upside Down,” can be found here. The second blog, “Why Developing Multiple Screening Technologies is a Must,” can be found here. The third blog, "How the Trump Candidacy Might Energize Men’s Healthcare,” can be found here.

Related Content

IBM collected a dataset of 52,936 images from 13,234 women who underwent at least one mammogram between 2013 and 2017.

IBM collected a dataset of 52,936 images from 13,234 women who underwent at least one mammogram between 2013 and 2017, and who had health records for at least one year prior to the mammogram. The algorithm was trained on 9,611 mammograms. Image courtesy of Radiology.

Feature | Artificial Intelligence | July 19, 2019 | Michal Chorev
Breast cancer is the global leading cause of cancer-related deaths in women, and the most commonly diagnosed cancer...
Advances in long-length digital radiography are creating opportunities for visualization during spinal surgery, as well as pre- and post-operatively. Image courtesy of Fujifilm Medical Systems

Advances in long-length digital radiography are creating opportunities for visualization during spinal surgery, as well as pre- and post-operatively. Image courtesy of Fujifilm Medical Systems

Feature | Digital Radiography (DR) | July 19, 2019 | By Greg Freiherr
Recent advances in...
Videos | Artificial Intelligence | July 12, 2019
Khan Siddiqui, M.D., founder and CEO of HOPPR, discusses the economic advantages and costs presented by...
Videos | Digital Pathology | July 11, 2019
Toby Cornish, M.D., Ph.D., associate professor and medical director of informatics at the University of Colorado Scho
FDA Clears Koios DS Breast 2.0 AI-based Software
News | Ultrasound Women's Health | July 11, 2019
Koios Medical announced its second 510(k) clearance from the U.S. Food and Drug Administration (FDA).
Videos | Enterprise Imaging | July 09, 2019
ITN Associate Editor Jeff Zagoudis speaks with Don Dennison
Videos | Enterprise Imaging | July 08, 2019
ITN Associate Editor Jeff Zagoudis speaks with Don Dennison
Infervision Releases InferTEST Program at SIIM 2019
News | Artificial Intelligence | July 08, 2019
Infervision announced their InferTEST program at the recent Society for Imaging Informatics in Medicine (SIIM)...
Hyland Healthcare Adds ImageNext Imaging Workflow Optimizer to Enterprise Imaging Suite
Technology | Enterprise Imaging | July 05, 2019
Hyland Healthcare is launching ImageNext, a vendor-neutral imaging workflow optimizer that combines intelligent imaging...