Pediatrics presents unusual challenge for 3T technology

Throughput of five patients per day at 3T is unlikely to impress many hospital managers or radiology practice owners. But this is actually a best-case scenario for Prof. David Edwards and his team at the Hammersmith Hospital in London. Nobody is telling them to hurry up.

The 3T system with such a low level of scheduling has been installed in the neonatal intensive care unit at Hammersmith Hospital, allowing Edwards and his team to study a vulnerable-and not necessarily cooperative-group of patients.

"These children are incredibly sick," said Edwards, a professor of neonatal medicine at Imperial College in London. "They don't walk in, lie down, and then walk out again. They are often on intensive care support and can't be made to lie still. If we do five a day, we are doing very well."

Installation of the 3T unit was completed at the end of 2006. It will be used primarily for brain imaging studies. Funding has just been secured to use diffusion MRI to assess the efficacy of new drugs designed to prevent brain damage. Functional MRI studies of sick neonates are also planned, to gain more information about impaired child development.

Researchers plan to use cardiac MRI protocols to study functional anomalies in babies' hearts that could cause additional brain injury. This could include the use of magnetization tagging techniques.

"Most cardiac MRI on babies is done to look for complex congenital heart anatomy and to plan surgical procedures. Our children don't have abnormal anatomy, they have abnormal function," Edwards said. "We are feeling our way forward, so progress is a little bit slower than our brain imaging work."

Cardiac imaging has traditionally been regarded as a problem area for 3T MRI, because of the tendency for magnetic inhomogeneities to cause dark, banded, off-resonance artifacts. This is one problem that the Hammersmith team has been spared. The small size of their patients' hearts makes homogeneity problems far less likely.

Motion artifacts pose a much greater challenge for brain and cardiac studies. Because the babies are unconstrained and likely to roll around, quick sequences are essential. Practitioners may need to stop and start imaging many times. Dedicated software for motion correction helps solve some of the difficulties with image registration, but speed is still a critical parameter.

Edwards accepts that a sizable proportion of this 3T work will only ever be useful in a research context. But this is not exclusively the case. Brain MRI protocols developed at Hammersmith to examine otherwise healthy babies starved of oxygen at birth could be of considerable clinical use.

"There is a continuum between what is purely research and what comes out into the general world. To some extent the answer is not in our hands; it is in the hands of people who plan radiological services around the world," he said.