MRI of neonates

Philip Gordon, M.D. is a neonatologist whose blog, Tales from the Womb, occasionally touches on MRI-related topics. He has a recent post about the trouble with MRI in neonatology that is well worth reading. He takes issue with the emerging practice of attempting to use MRI to predict neuro-developmental outcomes in preterm infants, pointing to a paper in NEJM last year[1] as well as a more recent study at UNC Chapel Hill. He is highly skeptical:

Because the fetal brain has yet to program its language, handedness, and speech centers many functions can be rewired. This may be why we’re detecting associations between prematurity and autistic spectrum disorders. They can retool, but we have to figure out how to help them optimize the transition.

My second problem is that Dr Inder does all of her MRIs at term equivalence. Well guess what? A lot, if not most, cortical growth occurs post term and this is especially true for preterm infants who often have growth delay. In short, it’s one snapshot in time and it might not be the best timing for the gray matter of low birth weight babies. I realize they have to start somewhere, I just don’t like the fact that the glitz of MRI somehow makes people forget that the studies are really just cohort studies (one step better than a case series).

Take MRIs with a grain of salt folks. The brain is the last biological frontier. Taking snapshots of the brain to predict neurodevelopmental outcomes is the modern day equivalent of Descarte cutting it up to find the seat of the soul. It’s going to be a while before we can scan a child and predict which types of early intervention they need.

It is certainly refreshing to see a cognitive scientist with a healthy skepticism about MRI interpretation. All too often, MRI is portrayed as a panacea. However, I would also shy away from the statement that preterm screening of high-risk infants has no value in predicting outcomes. The number of comments on the NEJM paper (see citation below) is indicative of a healthy and ongoing debate. Ultimately I think that the question of doing preterm MRI is going to be governed more by cost than any scientific value anyway, so this may be a moot point.

[1] Woodward et al. Neonatal MRI to predict neurodevelopmental outcomes in preterm infants. N Engl J Med. 2006 Aug 17;355(7):685-94. PMID: 16914704.