At Jerusalem’s Hadassah Hospital, a team of neurosurgeons and computer scientists are developing software to plan the delicate incisions that are necessary for neurosurgery.
Operating on the brain is a bit like ice fishing. Doctors cut a hole into the skull as small as half the size of a penny and can’t necessarily see what’s below. Sometimes they create an opening as large as 70 millimeters—the size of some camera lenses–forcing doctors to cut through broad areas of bone and tissue.
Surgeons must then navigate more than 400 miles of blood vessels and delicate lobes controlling speech, sight, smell and memory every time they want to excise a tumor or relieve symptoms of Parkinson’s disease, depression and other neurological disorders.
Figuring out the best trajectory to attack a tumor or fluid-filled cyst is as much science as it is art.
“Field of image or computer-assisted surgery is a very evolving project in medicine—especially in neurosurgery,” says Dr. Yigal Shoshan, head of Hadassah Hospital’s neurosurgery department, who along with Leo Joskowicz, a Hebrew University computer science professor, is developing the software.
A complex path
Shoshan explains that some tumors are deep inside the brain and reaching it to take a biopsy or to guide a catheter for further treatment means a safe pathway needs to be found.
Typically, a doctor plans for surgery by looking at an MRI brain scan and, using his or her own knowledge and experience, chooses the best locations to cut.
When Shoshan, now 56, was a young resident 30 years ago, he didn’t have the luxury of computer printouts. He based his decision on his knowledge of anatomy and what he saw in anatomy textbooks.
“Today it is very different. We are using technology to help us localize very precise areas in the brain … and to preserve functions of the brain that are very important,” he says.
The software combines a wide array of data to create a 3-D brain map instead of the traditional two-dimensional MRI. Functional MRIs capture blood flow changes that reflect areas of heightened neural activity, rather than simply an image of the organ. Overlaying the functional MRI image is another data set from a Magnetic Resonance Angiography (MRA), which focuses on the brain’s arteries and highlights irregularities, blockages and aneurysms.
All this data is processed to generate green and red color codes representing where surgeons should and should not cut. The areas that the software analyzes as too dangerous or too close to vital parts of the brain are placed in the red zone.
Not only does it allow for safer surgeries, it potentially paves the way for more effective treatment.
“If we are inserting an electrode to the basal ganglia of the brain to treat Parkinson’s—we are talking about sub-millimetric accuracy,” he says.
In July, a team of six scientists and computer engineers tested their prototype at Hadassah Hospital. In one test case, they gave the doctors a basic MRI showing a lesion on it. They asked each of the 10 doctors to plan a surgery based on the MRI alone, and then compare their plan with the software’s suggestions. Altogether, their experiment analyzed more than 240 surgical paths.
While the final results of the study have yet to be published and the software has yet to get FDA approval for further testing, Shoshan says the results would show how the software enables doctors to pick more accurate and safer trajectories.
He says the software was especially useful with junior residents.
“The software helps young doctors to be better,” he says, noting how it closes the gap between senior- and junior-level doctors. He noticed less experienced doctors could easily correct or modify their initial surgical plans and also learn from their mistakes.
“It’s a very strong teaching tool and executing tool,” he says. “Its like a game—if you choose a very good trajectory and you get a small risk ratio, you may win.”
Shoshan came up with the idea with Joskowicz, a former IBM computer scientist, who had worked on similar software that was built to organize all sorts of data—birth records, financial data and surgeries.
In the 1990s, Joskowicz worked at IBM’s T.J. Watson Research Center in Yorktown Heights, N.Y. Researchers there developed software to customize prostheses for hip replacement surgeries. The program used data from computerized tomography scans to make individualized prosthetics based on a patient’s bone structure.
Joskowicz began working with Hadassah Hospital neurosurgeons 10 years ago, taking his previous experience and applying it to a new realm.
“It takes a long time to develop these things. It takes perseverance. A good idea is far from being enough,” Joskowicz says.
More help for busy doctors
For doctors at Hadassah, where there is a large volume of patients, the speed and accuracy of decision-making is vital. The hospital has Jerusalem’s only neurosurgery department. Every year, the unit performs an estimated 1,500 surgical procedures, hospitalizes close to 2,000 patients and cares for roughly 4,500 patients in its outpatient clinic, Shoshan says.
The stories are dramatic, says Zack Medress, a second-year Stanford University medical student who assisted in research for the project over the summer.
One partially blind six-year-old boy came to the hospital with a fluid-filled tumor at the base of his brain that impinged on his optic nerve. Surgeons wanted to drain the tumor to relieve the symptoms but it was in an area known as “high-priced real estate,” Medress says.
Hadassah’s doctors successfully operated without damaging other parts of the brain using the standard MRI-based planning method. But, if they had used the new 3-D data-synthesizing software, Medress says, it most likely would have decreased risk to the patient.
Fifteen percent of brain surgeries for benign tumors go awry from complications like internal bleeding or misplaced incisions that lead to permanent neurological damage. The worst-case scenario is death.
But recent breakthroughs in surgical technology are changing the game. The surgeries of tomorrow will be highly accurate, most likely enhanced by virtual and robotic technology. Joskowicz and Shoshan’s new software may be used to plan those future surgeries.
“There are many positive implications. The surgery will be safer and better quality for a larger population,” says Joskowicz.
Top image: Color-coded head surface model presenting the risk ratio for a given surgical trajectory. Red is analyzed to be a no-go surgical area and green means a lower-risk surgical incision area. Courtesy Dr. Yigal Shoshan.