The Department of Neurosurgery has received a Grace Woodward Grant for 2007-2008, entitled, “CT as a Substitute for MRI in Temporal Lobe Epilepsy Diagnosis in Sub-Saharan Africa”. We have formed a collaboration with colleagues at the CURE Children’s Hospital of Uganda in Mbale, Uganda, where an active Pediatric Neurosurgery surgery program is developing strategies to advance the treatment of children with hydrocephalus and epilepsy.

The incidence of temporal lobe epilepsy in malaria prone regions of the world is high. This is presumably related to damage to the brain from the high incidence of febrile illnesses in children in such regions. Certain infections, such as malaria, are nearly universal in such populations. In the industrialized countries, MRI technology has been the mainstay of the diagnosis of the scarring and shrinking of the deep parts of the temporal lobes of the brain in such patients. Such asymmetric scarring, when consistent with the clinical signs and scalp EEG characteristics of a deep temporal lobe epilepsy focus, carries high prognostic value that a surgical resection of the deep temporal structures will have a significant benefit to reduce seizures and improve the quality of life. The risk to benefit evaluation of the medical versus surgical treatment of such illnesses, in parts of the world where drug maintenance and pharmacological laboratory monitoring are often impractical, is different from the industrialized countries. Despite the substantial literature supporting the use of MRI for volume measurements of the deep structures of the temporal lobe, there appears to be no previous work using CT for this diagnosis. Although MRI is generally not available in the developing world, the less expensive CT is now becoming increasingly prevalent.

A related issue for East Africa is post-infectious hydrocephalus in early childhood. Neonatal infections, occurring predominantly within the first 3 months of life, account for the majority of hydrocephalus cases seen there. Maintaining shunt hardware, and managing the early postoperative complications, and the late emergent shunt obstructions in children who become shunt dependent, weakens the risk-benefit balance for shunt insertions. A growing effort to use endoscopic fluid diversions, especially third ventriculostomy, is being applied by surgical centers in East Africa. Nevertheless, the question of whether these children so treated do as well as children treated with shunts remains unanswered. In addition, as our technological approaches to hydrocephalus treatment increase in sophistication, one must ask whether the organisms causing post-infectious hydrocephalus can be identified and prevented. We know from recent literature that cerebral malaria seems not to be an active agent with such obstructive hydrocephalus, and that HIV seropositivity is below 5% in such infants. A hunt for the causative organisms is warranted.

We have 3 collaborative projects currently underway.

First, we have an effort to develop the image analysis strategies to use CT as a substitute for MRI in patients with temporal lobe epilepsy. Our first efforts are to use landmarks to measure volume in regions of the brain that we can clearly outline on CT, rather than the small grey matter structures, such as the the hippocampus, which can only be visualized well on MRI. Our basic strategies including standard volumetric image analysis - using contours on stacks of CT scans (planimetry) or surface area estimation (Cavalieri method) on such images. Yet in a broader sense, we know that even ‘unilateral’ temporal lobe epilepsy is often a more widespread disease - if you damage part of the brain, everything connected to that damaged part will be affected. We suspect that the more widespread are the abnormalities in the brain of such patients, the more likely that the epileptic seizures can and do arise from more than one structure, and the less likely that surgical resection of a small affected part (such as the hippocampus) will significantly help with the patient’s seizures. So in addition to our standard volumetric analysis, we are exploring whether the symmetries within the brain, between left and right temporal lobes, and between the hemispheres, might be a more sensitive indicator of seizure focus localization as well as a gauge of potential surgical outcome.

A related imaging project involves the evaluation of hydrocephalus. We are developing methods of breaking an image up, whether CT or MRI, into volumes of brain and fluid. We presently use head circumference to quantify the magnitude and outcome of treatment of children with hydrocephalus. But this is not the metric which would best reflect the results of the disease or its treatment. We are developing two new growth curves - one for brain, and one for fluid within the head. Such growth curves, based on populations of normal children, will provide a more sensitive and direct measure of the effect of hydrocephalus treatment - demonstrating that the brain is re-acquiring a more normal growth curve. This is critical if, for instance, following internal fluid diversion the majority of the fluid is not drained out as is typical following shunt insertion.

Lastly, we are in the process of working out the logistics to begin screening cerebrospinal fluid samples in children with post-infectious hydrocephalus for fragments of nucleic acids which would tell us the class of organisms which caused these illnesses. Although we can treat the fluid blockages, the damage to the brain from such infections can never be undone. We will work out a combination of onsite sample preparation, and the introduction of technologies such as polymerase chain reaction analysis, in order to identify these organisms.

These projects have a range of implications. We suspect that the high incidence of both infantile hydrocephalus, as well as temporal lobe epilepsy later in childhood, are related to prior infections in this part of the developing world. Nevertheless, only the infants who present within weeks of recovery from such infections are going to offer us the chance at organism identification, and we have no reason to presume that the organisms which cause these different conditions are the same. Nevertheless, the disease spectrum and living conditions in rural East Africa do not respect political borders, and we anticipate that our findings will have applicability to countries beyond Uganda. In terms of image analysis, the strategy of using less detailed and fuzzy images to make more accurate diagnosis and treatment decisions is a universally applicable strategy. Image technology, whether MRI or CT, is expensive, and requiring less expensive hardware by improving analysis serves as a model for containing costs in both developing or industrialized societies.

These projects are a collaborative effort between faculty and trainees across Penn State, as well as our colleagues at other universities and at medical centers, including:

Penn State Hershey: Department of Neurosurgery (Drs. Steven Schiff, Robert Harbaugh, James McInerney, Kenneth Hill), Department of Radiology (Drs. Dan Nguyen, Kevin Moser), Department of Neurology (Dr. Matthew Eccher).

Penn State University Park: Department of Bioengineering (Dr. Andrew Webb), Department of Computer Science and Engineering (Dr. Yanxi Liu, Joao Soares), Department of Engineering Science and Mechanics (Dr. Corina Drapaca),

And beyond Penn State: Dr. Benjamin Warf (Dupont Children’s Hospital, Delaware), Dr. Warren Boling (University of West Virginia), Dr. John Mugamba (CURE Children’s Hospital of Uganda), and Dr. Kachinga Sichizya (CURE Children’s Hospital of Zambia).

Steven J. Schiff Penn State Center for Neural Engineering