Application Areas

Disorders

Applications of automatic segmentation include assessing treatment and quantifying diagnoses for brain disorders including: Schizophrenia, Epilepsy, Multiple Sclerosis, Alzheimer's disease, Amyotrophic Lateral Sclerosis (ALS), Huntington's Disease, Obsessive Compulsive Disorder, attention-deficit hyperactivity disorder (ADHD), Tumors, Idiopathic focal dystonia, Mood disorders, Lesch-Nyhan syndrome, Normal pressure hydrocephalus, Autosomal dominant cerebellar ataxia type I, Substance Abuse, Borderline personality disorder, post traumatic stress disorder, (PTSD), Down syndrome, Autism, Usher syndrome, and HIV, among others.

In the future, clinicians will use these measurements to help diagnose patients. For example, measurement of cortical thickness, or the volume of hippocampus and entorhinal cortex is evidence of the progression of Alzheimer’s disease. For PTSD in a veteran or sexual abuse victim, or ADHD in a child, measurement of specific regions might qualify patients for special treatment.

Substrate for functional analysis

A key part of all quantitative neuromorphometric analysis involves the implicit knowledge of anatomy. This expertise can be explicitly used to create a structural basis for analyzing functional MRI (fMRI) scans. fMRI measures blood flow and this flow correlates with brain function. Looking at brain function not only helps general brain research, but also has great potential to aid the understanding of disease states. Currently, fMRI studies of multiple subjects involve averaging the scans together after positional normalization using the Talairach coordinate system. Since each brain is topographically different (especially in the cortex), both the magnitude and specificity of the sought-after fMRI signal decreases because inappropriate areas are averaged together. Anatomically identifying and precisely localizing regions of functional activation in each individual scan before averaging will increase the usefulness and statistical significance of the results. The combination of functional and structural MRI will be better able to diagnose diseases and measure response to treatments.

Improve 3D Visualization

As medical image acquisition technology progresses, scan quality is getting better but also, voxel resolution is getting smaller and the number of voxels acquired is increasing. This leads to an excess of data that requires processing before it can be useful. Without access to automated segmentation, either a great deal of time is necessary to delineate regions of interest, or else image visualization “tricks” must be used. Processing such as volume rendering result in images that show anatomy without really knowing the precise location of boundaries. Using automated segmentation can save manual analysis time and can generate surfaces that can be visualized. Automatic segmentation directly leads to improved 3D visualization.

Others

Other potential application areas involving anatomical expertise in MRI include tumor treatment, minimally invasive neurosurgery, and stroke recovery therapy. Radiologists currently make judgments about the size of brain tumors "by eye". Quantification and localization will radically improve diagnosis, treatment planning, and indications of the response to treatment for brain tumors. Knowing which brain regions are affected explicitly by a stroke lesion and also which regions will be effected because they are connected to effected regions determines what type of disabilities the patient will have and therefore what type of compensatory and re-training therapies will be needed. A quantitative neuromorphometric analysis can be used to plan surgical treatments for cerebral vascular disorders (e.g. arteriovenus malformations, aneurysms, stroke, and head trauma), and other neurological pathologies (e.g. intractable pain treatment, OCD, temporal lobe epilepsy).

< -- back                  next -- >