MRI of Neonatal Hypoxia-Ischemia
Cerebral hypoxia-ischemia remains a major contributor to perinatal morbidity and mortality. It is estimated that between 0.2 to 0.4% of full-term infants and up to 60% of premature infants experience asphyxiation at or before birth. Research efforts are focused on the pathophysiologic mechanisms for the development of the cerebral tissue damage resulting from cerebral hypoxia-ischemia with the hope that standardized detection and treatment protocols may be developed and implemented in the human newborn. Towards this end it is necessary to identify methods to prevent these neurological deficits in the immature brain.
Our laboratory utilizes 1H and 31P NMR spectroscopy and imaging to evaluate therapeutic regimens and the temporal evolution of neonatal hypoxia-ischemia in immature rats. Proton imaging can be used to detect brain regions sensitive to vasogenic edema (T2-weighted imaging) and regions experiencing cytotoxic edema (DWI, Diffusion-Weighted Imaging, sensitive to restrictions in the translational displacement of intrinsic tissue water). The model consists of both right common carotid artery ligation followed by exposure to 3 h of hypoxia (8% O2, balance N2). A combination of both hypoxia and ischemia is needed to cause permanent damage in this neonatal model of hypoxia-ischemia. 31P spectroscopy is used to monitor high-energy metabolites during hypoxia-ischemia. The time course of these changes can be used to predict which animals will proceed to infarction. We have shown that modest hypothermia and pre-injection with dexamethasone preserves these high-energy metabolites during hypoxia-ischemia, and provides neurological protection.
We have developed an imaging probe for our 9.4T vertical bore magnet. Samples are changed from the bottom of the magnet within 5 minutes. This probe is tuned through the change in effective inductance of the RF coil with an RF shield. There is no physical-electrical connection with the outside world, providing extremely homogenous B1 fields. With the collaboration of the Chemistry Department and Drs. Springer and Latour at Brookhaven National Laboratory, we have designed and constructed a 10 cm diameter (100 G/cm) 3-axis gradient coil, that can be used in our whole body 3T magnet, for microimaging. Below our examples of in vivo 1H images of 12g (7 d old) neonatal rat pups obtained at both 400 and 125 MHz.
Diagram of intracellular space (ICS) and extracellular space (ECS) volume changes that accompany a disruption in ion homeostasis. As high energy metabolites are depleted (ATP & PCr), the transmembrane ionic pumps shut down. Osmotically obligated water causes the ICS to increase by 12% and the ECS to decrease by 50%. Reduction in the ECS causes an increase in tortuosity. This increased tortuosity causes a reduction in the apparent diffusion coefficient (ADC) of water which results in hyperintensity in DWIs.
The percent hemisphere lesion area (%HLA) was determined in two different slices in the neonatal rat pup brain (n=12) by two different methods. The %HLA of DWI hyperintensity at 1 h post-HI is compared to triphenyltetrazolium chloride (TTC) staining at 18 h post-HI. As the exposure to 8% O2 is decreased to 2 h the predictive value of DWI becomes diminished. There is an excellent correlation for animals exposed to 3 h HI:
%HLATTC = 0.79 * %HLADWI + 7.0, R2 = 0.71, p = 0.0006
Logistic regression is used to predict the probability of neuropathologic damage from MRI. Brain regions were scored as: 0 = no damage, 1 = non-cystic, and 2 = cystic infarcts. MRI parameters (ADC value, T2 hyperintensity) were normalized to the contralateral hemisphere for 7 different brain regions. These results are for ratpups (n = 12) undergoing 3 h of HI. The ADC at 1 h post-HI is a good indicator of which brains will have neuropathologic damage at 5 d post HI. T2 hyperintensity at 42 h post-HI is a good discriminator of whether infarcts will be cystic or non-cystic 5 days post-HI. These results cannot be extrapolated to shorter episodes of HI.