IBM researchers have unveiled prototype 3D visualisation software that will enable doctors to interact with their medical data in pretty much the same way they interact with their patients. The technology, known as the Anatomic and Symbolic Mapper Engine (ASME), uses an avatar representation of the human body which the doctor can click on to trigger a medical records search relevant to that body part.
"It's like Google Earth for the body," said IBM Researcher Andre Elisseeff, who leads the healthcare projects at IBM's Zurich Research Lab. "In hopes of speeding the move toward electronic healthcare records, we've tried to make information easily accessible for healthcare providers by combining medical data with visual representation, making it as simple as possible to interact with data that can improve patient care."
If went to see my doctor today with a bad back, they would question me about any previous back problems I may have experienced, examine me both visually and physically, maybe run some tests. The doctor will also rifle through records to search for relevant information, some of it archived on computer some of it buried in the paper files. Even in the increasingly networked age we live in, the chances are that the doctor will not have immediate access to a full patient history however. Which means the patient will not have access to an immediate and 100 percent accurate diagnosis.
IBM argues that the ASME 3-D avatar changes that. By clicking on the spine the doctor would be able to see, in an instant, all the available medical history and information related to that patient's back problems. That's everything including text entries, lab results and medical images such as radiographs or MRIs. If the doctor was only interested in information related to a particular part of the spine they could zoom in, narrowing the search parameters by time or other factors.
OK, the eagle-eyed might have spotted that what we have here is actually a database and the same level of search accuracy could be achieved by any database system, providing all the information was properly archived and accessible via the network. The comparisons above are not apples with apples, the IBM scenario assumes that the information is all available so all it is really adding to the conceptual mix is a graphical search interface.
It is a good graphical search interface though.
Using advanced machine learning and state-of-the-art 3D modeling techniques, the IBM researchers overcame key technical challenges including integrating heterogeneous data sources and complex text-based information-so-called unstructured data-and linking that data to the anatomical model in a meaningful and easy-to-navigate way. ASME also uses SNOMED, the systemized nomenclature of medicine that encompasses approximately 300,000 medical terms, to create a bridge between graphical concepts and text documents.
