As Douglas correctly pointed out:
“Radiology text and image data account for a significant portion of patient electronic medical record information in hospital archives, yet few means exist to aid radiologists in extracting relevant information from these databases.
Granted that there are Teaching PACS and MIRC (Medical Imaging Resource Center, see here), there are so much potential lying in the archives of medical imaging information systems (like RIS, CVIS and PACS etc) simply because its very difficult to get the data of interest with ease.
Apply Google’s search enginee into the world of Data Mining for Radiology and you will get Radsearch, an open source Google-like interface wih Google’s core search technologies to perform the data mining for RIS.
“This tool is different from methods we’ve used to search our RIS in the past. It lets the user skip the step of first creating a custom-designed SQL query, a step that took considerable time and expense because our information systems staff was needed to perform it”
Now this is precisely what Radiologist have been asking for years – intelligent “free text” searches, I reckon this will speed up research in the Radiology arena by years, now all they have to do is extend this to other medical discipline. 🙂