OASIS Is Automated Statistical Inference for Segmentation with Applications to Multiple Sclerosis Lesion Segmentation in MRI

• Activity Number: 693
• Type: Contributed
• Date/Time: Thursday, August 8, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Imaging
• Author(s): Elizabeth Sweeney*+ and Russell Shinohara and Navid Shiee and Farrah Mateen and Avni Chudgar and Jennifer Cuzzocreo and Peter Calabresi and Dzung Pham and Daniel Reich and Ciprian M. Crainiceanu
• Companies: Johns Hopkins University and Univ of Pennsylvania and Henry M. Jackson Foundation and Johns Hopkins and Harvard Medical School and Johns Hopkins and Johns Hopkins and Henry M. Jackson Foundation and National Institute of Neurological Disorders & Stroke and The Johns Hopkins University
• Keywords: Brain Imaging ; Multiple Sclerosis ; OASIS ; MRI ; Logisitc Regression

Abstract

Magnetic resonance imaging (MRI) is used to detect lesions in the brains of multiple sclerosis (MS) patients. In practice, lesion load is quantified by manual inspection of MRI, which is time-consuming and associated with large inter- and intra-observer variability. We propose OASIS is Automated Statistical Inference for Segmentation (OASIS), an automated statistical method for detection of MS lesions in MRI. We use logistic regression models incorporating multiple MRI modalities to estimate voxel-level probabilities of lesion presence. Intensity-normalized MRI volumes from 131 MRI studies (98 MS subjects, 33 healthy subjects) with manual lesion segmentations were used to train and validate our model. Within this set, OASIS detected lesions with an area under the receiver-operator characteristic curve of 98% (95% CI; [96%, 99%]) at the voxel level. To further validate the method, we applied OASIS to 169 MRI studies acquired at a separate center and had an expert rate these segmentations in comparison to another method, LesionTOADS. For lesions, OASIS out-performed this method in 77% (95% CI: [71%, 83%]) of the 169 studies.