Automated Segmentation of MR Images of Brain Tumors

Abstract

An automated brain tumor segmentation method was developed and validated against manual segmentation with three-dimensional magnetic resonance images in 20 patients with meningiomas and low-grade gliomas. The automated method (operator time, 5-10 minutes) allowed rapid identification of brain and tumor tissue with an accuracy and reproducibility comparable to those of manual segmentation (operator time, 3-5 hours), making automated segmentation practical for low-grade gliomas and meningiomas.

Serial registration of intraoperative MR images of the brain

“…This means we can use segmentation approaches that are robust and accurate but are time consuming and hence impractical to use in the operating room. In our laboratory, preoperative data is segmented with a variety of manual (Gering et al, 1999), semi-automated (Kikinis et al, 1992) or automated (Warfield et al, 1995Kaus et al, 2000) approaches. We select the most robust and accurate approach available for a given clinical application.…”

Three validation metrics for automated probabilistic image segmentation of brain tumours

“…The cases: A total of nine patients were randomly selected from a neuro-surgical database of 260 brain tumour patients, of which three had meningiomas (M), three astrocytomas (A), and three other low-grade gliomas (G) [7,55]. Visually the meningiomas enhanced better on greyscale images than the remaining two tumour types.…”

“…The fraction of target malignant class varies depending on the type of the disease. In this application, the brain tumours occupied about 10-20 per cent of the entire brain, which still comprised a large number of pixels within the tumour regions [7]. However, one may also construct a 'region of interest' (ROI), a subset of the entire brain pixels, which contains the target tumour of interest.…”

“…The automated probabilistic segmentation was the relative tumour probability of lesion per pixel with signal intensity modelled as a Gaussian mixture of the two classes, based on an initial semi-automated binary segmentation [7].…”

Dense deformation field estimation for atlas-based segmentation of pathological MR brain images

“…The patient images used in this study have been retrieved from the Surgical Planning Laboratory (SPL) of the Harvard Medical School & NSG Brain Tumor Database [20]. They consist of three SPGR T1-weighted volumes of 124 coronal slices of 256 pixels × 256 pixels and 0.9375 mm × 0.9375 mm × 1.5 mm of voxel size.…”