MIT researchers have created a software that not only speeds up a critical preparatory step for heart surgery, but can also be used to print physical models of a patient’s heart in just three hours.
Before doctors can perform heart surgery, they must take an image of the heart via a CT scan or MRI and then perform what is known as a segmentation process.
Segmentation is when a doctor or trained technician slices the image of the heart — essentially creating borders around different parts of it — in order to create a 3D image. The 3D image then helps physicians understand the anatomical idiosyncrasies of the heart they are working with and what kind of surgery they must perform.
The segmentation process, however, is a lengthy one. The more slices, the greater the accuracy — meaning doctors or technicians can spend upwards of eight hours creating 50 to 150 slices.
“Eight hours is too long to be practical,” said Dr. Andrew Powell, a cardiologist at Boston Children’s Hospital, who leads the project’s clinical work.
The new algorithm created by MIT researchers should change that. Danielle Pace, an MIT graduate student in electrical engineering and computer science, spearheaded the development of the software that will speed up the segmentation process to convert the heart into a 3D image in just an hour and into a physical model in three.
To use the algorithm, the doctor or technician responsible for segmenting the heart only has to do a few slices (about 10 per cent) and then let the software do the rest of the work.
“MRI images can be hard in general to segment…the algorithm helps to fill the gaps,” Pace said.
Powell notes that the novelty is not the software’s ability to make a physical model, but rather that it speeds up the segmentation process and can provide a more accurate picture of the heart physicians are working with. Doctors can then choose to print the model and practice surgery or use it to create a hologram and perform virtual surgery.
“Right now, children born with heart problems have tremendous variability and there is a great need for imaging to provide surgery,” Powell said. “Now, we take two-dimensional slices of the heart to integrate into a 3D picture, but a person’s ability to do that varies, it’s hard. The algorithm lets people skip the integration step to see what the heart looks like for a more accurate diagnosis and to better plan the procedure.”
When testing the software, the researchers were able to manually segment 14 patches and let the algorithm do the rest of the work. The algorithm was able to create an image that had 90 per cent agreement with a manual segmentation of 200 slices. When the researchers only segmented three areas by hand, the algorithm did the rest and created an image with 80 per cent agreement.
Seven cardiac surgeons from Boston Children’s Hospital will practice procedures on the 3D-printed models this fall.
Business Insider Emails & Alerts
Site highlights each day to your inbox.