JACC: Cardiac MR reformations help plan pediatric heart surgery
Top: A 3D model of hepatic flow distribution pre-surgery. Bottom: Post-surgery hepatic flow distribution options.
Image source: Georgia Tech's Ajit Yoganathan
By manipulating a patient's 3D cardiac MR images, physicians can compare how alternative approaches affect blood flow and expected outcomes, and can select the best approach for the patient before entering the operating room, according to research in this month's issue of the Journal of the American College of Cardiology: Cardiovascular Imaging.

"This tool helps us to get the best result for each patient," said co-author Mark A. Fogel, MD, an associate professor of cardiology and radiology and director of Cardiac MRI at the Children's Hospital of Philadelphia. "The team can assess the different surgical options to achieve the best blood flow and the optimum mixture of blood, so we can maximize the heart's energy efficiency."

Researchers at the Georgia Institute of Technology, collaborating with pediatric cardiologists and surgeons at the Children's Hospital of Philadelphia, described the surgical planning methodology, detailing how the tool helped them to plan the surgery of a four-year-old girl, who was born with just one functional ventricle, instead of two.

"Preoperatively determining the Fontan configuration that will achieve balanced blood flow to the lungs is very difficult and the wide variety and complexity of patients' anatomies requires an approach that is very specific and personalized," said Ajit Yoganathan, PhD, Regents' Professor in the department of biomedical engineering at Georgia Tech and Emory University in Atlanta. "With our surgical planning framework, the physicians gain a better understanding of each child's unique heart defect, thus improving the surgery outcome and recovery time."

The patient described in the study, Amanda Mayer, age four, of Staten Island, N.Y., had previously undergone all three stages of the Fontan procedure [to reshape the circulation in a way that allows oxygen] at Children's, but developed severe complications. Her oxygen saturation was very low--72 percent, compared with normal levels of at least 95 percent--indicating the possibility of abnormal connections between the veins and arteries in one of her lungs.

To improve the distribution of these liver-derived hormonal factors to both lungs, the surgeons needed to re-operate and reconfigure the patient's cardiovascular anatomy. Georgia Tech's surgical planning framework helped Thomas L. Spray, MD, chief of the division of cardiothoracic surgery at Children's, to determine the optimal surgical option.

"MRI acquires images of the child's heart without using radiation," Spray said. "Then we use the computerized technology to model different connections to simulate optimum blood flow characteristics, before we perform the surgery."

The image-based surgical planning consisted of five major steps: acquiring MRI of the child's heart at different times in the cardiac cycle, modeling the preoperative heart anatomy and blood flow, performing virtual surgeries, using computational fluid dynamics to model the proposed postoperative flow and measuring the distribution of liver-derived hormonal factors and other clinically relevant parameters for the surgeon.

Fogel collected three different types of MR images, while Yoganathan and colleagues generated a 3D model of the child's cardiovascular anatomy. From the model they reconstructed the 3D pre-operative flow fields to understand the underlying causes of the malformations.

For this patient, the team saw a highly uneven flow distribution--the left lung was receiving about 70 percent of the blood pumped out by the heart, but only 5 percent of the hepatic blood. Both observations suggested left lung malformations, but closer examination of the flow structures in this patient revealed that the competition between different vessels at the center of the original Fontan connection effectively forced all hepatic factors into the right lung, even though a vast majority of total cardiac output went to the left lung.

To facilitate the design of the surgical options that would correct this problem, Jarek Rossignac, PhD, a professor in Georgia Tech's School of Interactive Computing, developed Surgem, an interactive geometric modeling environment that allowed the surgeon to use both hands and natural gestures in 3D to grab, pull, twist and bend a 3D computer representation of the patient's anatomy. After analyzing the 3D reconstruction of the failing cardiovascular geometry, the team considered three surgical options.

The research team then performed computational fluid dynamics simulations on all three options to investigate for each how well blood would flow to the lungs and the amount of energy required to drive blood through each connection design. The measures of clinical performance allowed the cardiologists and surgeons to conduct a risk/benefit analysis, which also included factors such as difficulty of completion and potential complications.

Of the three choices, Spray favored the option that showed a slightly higher energy cost but exhibited the best performance with regards to hepatic factor distribution to the left and right lungs. Five months after the surgery, the patient showed a dramatic improvement in her overall clinical condition and oxygen saturation levels, which increased from 72 to 94 percent. Amanda Mayer is breathing easier and is now able to play actively like other children, according to her cardiologist, Donald Putman, MD, of Staten Island, N.Y.

"State-of-the-art 3D cardiac MRI married to modern biomedical engineering and applied anatomy and physiology enabled this approach," Fogel concluded

This work was funded by the National Heart, Lung and Blood Institute (NHLBI) of the National Institutes of Health (NIH).