Normal pressure hydrocephalus (NPH) is a chronic form of hydrocephalus in older adults which is thought to be caused by obstruction of the normal flow of CSF. NPH typically presents with cognitive impairment, gait dysfunction, and urinary incontinence, and may ultimately lead to permanent brain damage. NPH may account for more than five percent of all cases of dementia; but unlike most other causes of dementia, NPH can sometimes be reversed by shunt surgery or endoscopic third ventriculostomy (ETV) which allow excess CSF to drain. But not all shunt surgeries and ETVs are successful and presently it is not known why some respond and others do not. Screening prior to surgery is often performed by a trial of CSF diversion via a lumbar puncture. If the patient responds by showing improvement in gait, then it is thought (and usually found) that shunt surgery or ETV will improve their condition on an ongoing basis. It has also been found that for patients that do not respond favorably to the lumbar puncture, shunt surgery or ETV may still help their condition. Since shunt surgery and ETV are not without risk, it is a major diagnostic obstacle to determine by some objective measure the likelihood of a positive response to therapy after a negative response to lumbar puncture screening. The proposed research will develop a new method to segment and label the ventricles and their connecting pathways from magnetic resonance images. While this task is generally straightforward in normal subjects, it represents a significant challenge in patients with enlarged and/or deformed ventricles. Segmenting and labeling the ventricles in NPH patients is the major challenge of the proposed work and it is where our major innovation arises. We will combine a patch-based tissue classification method with a registration-based multi-atlas labeling method to generate the novel algorithm. The result will be a labeling of the lateral ventricles (body and anterior, posterior, and inferior horns), the third and fourth ventricles, the cerebral aqueduct, the interventricular foramina, and the subarachnoid space. The relative volumes of these spaces will be used to evaluate where the likely blockage in CSF flow is occurring and whether shunt surgery or ETV is therefore likely to be successful, providing a potential alternative assessment to lumbar puncture. Specifically, we will 1) Develop a detailed manual delineation protocol for labeling the ventricular system; 2) Develop and evaluate an automatic ventricular segmentation and labeling algorithm; and 3) Carry out a retrospective study on NPH patients to model the ventricular system together with shunt or ETV responsiveness. This research will yield open-source software for segmentation and labeling enlarged ventricles and will provide the first image-based metrics to assist in the evaluation of potential efficacy of shunt surgery or ETV in NPH patients. Our research will also provide other researchers with the means to study ventricular enlargement in other neurodegenerative diseases such as Alzheimer's disease and Parkinson's disease, as well as in normal aging, conditions which are considered in the differential diagnosis of NPH.
This project concerns normal pressure hydrocephalus, a condition caused by enlargement of the fluid- filled spaces in the brain. The research will develop new image processing methods to help physicians determine whether shunt surgery is likely to alleviate the cognitive and motor difficulties that often accompany the condition.
| Carass, Aaron; Shao, Muhan; Li, Xiang et al. (2017) Whole Brain Parcellation with Pathology: Validation on Ventriculomegaly Patients. Patch Based Tech Med Imaging (2017) 10530:20-28 |
| Dewey, Blake E; Carass, Aaron; Blitz, Ari M et al. (2017) Efficient Multi-Atlas Registration using an Intermediate Template Image. Proc SPIE Int Soc Opt Eng 10137: |
| Ellingsen, Lotta M; Roy, Snehashis; Carass, Aaron et al. (2016) Segmentation and labeling of the ventricular system in normal pressure hydrocephalus using patch-based tissue classification and multi-atlas labeling. Proc SPIE Int Soc Opt Eng 9784: |
| Huo, Yuankai; Plassard, Andrew J; Carass, Aaron et al. (2016) Consistent cortical reconstruction and multi-atlas brain segmentation. Neuroimage 138:197-210 |
