Data mining may help spot patients with fracture risk

Monday, November 27 | 11:20 a.m.-11:30 a.m. | SSC09-06 | Room E450B
In this scientific presentation, researchers from Switzerland will share how their PACS data-mining software shows potential for identifying patients at risk of osteoporotic fractures.

Seeking to test its internally developed data-mining software, called PACS-Crawler, the group from the University Hospital of Basel worked with the institution's department of orthopedic surgery to identify patients with known osteoporotic fractures. The researchers then used PACS-Crawler to identify those who had received CT studies for other indications prior to the fracture.

"Most CT scans include some osseous structures, and we were wondering if we could use the information -- already in the scan -- opportunistically for the analysis of bone density and, possibly in the future, as a predictor of fracture that could be flagged in our reports automatically," presenter Dr. Fides Schwartz told

Next, a customized software plug-in automatically extracted bone-density percentiles from these CT studies. The group then compared these results with those from a large database of normal patients.

The researchers found that the data-mining software performed well for searching, relaying, and processing big data. In addition, preliminary data showed that the patients with osteoporotic fractures demonstrated some differences in bone density compared with a normal population, she said.

What else did the group find, and what other plans do they have for PACS-Crawler? Stop by this Monday morning talk for all the details.

Page 1 of 603
Next Page