Dartmouth Engineering Postdoc Helps Surgeons Navigate the Brain
By Anna Fiorentino
December 2013 • CoolStuff
Just one year ago for the first time ever anywhere, Dr. David Roberts, Chief of Neurosurgery at Dartmouth-Hitchcock (D-H), used a new economical 5-minute registration method for MRI and ultrasound images that enabled him to better navigate a patient’s brain. The method was implemented by research associate Xiaoyao Fan Th’12 who works with its inventors, Dartmouth engineering professors Songbai Ji, Keith Paulsen and Alexander Hartov.
Like a GPS for neurosurgeons, ultrasound-based registration allows doctors to correlate the patient’s head in the operating room with an MRI of their brain and then display and track the location of the surgical instruments on the MRI in real time.
The D-H team has used the method successfully more than 30 times since for a variety of open cranial procedures including brain tumor resection, electrode implantation, and lobectomy.
“Especially in emergency cases, our ultrasound-based registration is cost-efficient, effective, comparably accurate, and requires minimal time and personnel resources,” says Fan. Conventional navigation systems are time-consuming, requiring extra MRIs taken the day of surgery, as well as operating room personnel to match up skin-affixed markers with the preoperative MRI.
In contrast, the new method is completely automatic: As soon as the surgeon opens the patient’s skull, the ultrasound system—armed with a broadband matrix array transducer—acquires a 3D image and a few minutes later, the outer shape of the brain is extracted from the ultrasound image to match with its counterpart in the MRI. An additional program quickly fine-tunes the result and the surgeon is ready to accurately navigate the brain.
Because of the computational power required, the ultrasound-based method remotely utilizes 248 CPUs at Thayer School.
“I ran the programs and sent the computation jobs to the remote Thayer Linux clusters via the network and about five minutes later the problem was solved and the result was retrieved back to the local computer in the OR at Dartmouth-Hitchcock Medical Center,” says Fan. “The measurement proved incredibly accurate.”
Professors Ji, Paulsen, and Hartov, as well as Roberts, filed a patent for the technology in February of 2012.
“People have talked about this for a long time, but we were the first group that accomplished this in the OR,” says Fan.comments powered by Disqus