Validation of a Novel Semiautomated Segmentation Method for MRI Detection of Cartilage-Related Bone Marrow Lesions.

OBJECTIVE: To determine the relationship of bone marrow lesions (BMLs) with phenomena such as clinical symptoms, histological subchondral bone damage, and development of osteoarthritis, a reliable and reproducible method to localize and quantify BMLs accurately is indispensable. Therefore, the goal of the current study was to develop and validate a novel semiautomated segmentation method based on the KNN classification technique on T2-weighted (T2w) SPIR and proton density-weighted (PDw) magnetic resonance images (MRIs), as this would provide an accurate, reliable, and reproducible tool. MATERIALS AND METHODS: Twenty PDw and T2w SPIR MRIs were selected and manually segmented as a learning set for the software system. The manual segmentations were considered the gold standard. Automated segmentation based on the KNN classification technique was carried out on the same MRIs. To determine the accuracy and validity of the system, the automated segmentations were compared to the gold standard using the Dice Similarity Index (DSI). RESULTS: The KNN classification system resulted both visually and statistically in an accurate segmentation of BMLs on T2w SPIR MRIs with an excellent mean optimal DSI of 0.702 (±0.202; range, 0.409-0.908). Elimination of specific areas smaller than 10 voxels improved the accuracy. The accuracy was independent of BML size. The segmentation of BMLs on PDw MRIs was less reliable with a mean optimal DSI of 0.536 (±0.156). CONCLUSION: Although the applicability of this method is limited on PDw MRIs, the KNN classification system provides an accurate, reliable, and reproducible tool for semiautomated segmentation of BMLs in T2w SPIR MRIs of the knee.

Automated and visual scoring methods of cerebral white matter hyperintensities: relation with age and cognitive function.

BACKGROUND AND PURPOSE: A reliable scoring method for ischemic cerebral white matter hyperintensities (WMH) will help to clarify the causes and consequences of these brain lesions. We compared an automated and two visual WMH scoring methods in their relations with age and cognitive function. METHODS: MRI of the brain was performed on 154 participants of the Utrecht Diabetic Encephalopathy Study. WMH volumes were obtained with an automated segmentation method. Visual rating of deep and periventricular WMH (DWMH and PWMH) was performed with the Scheltens scale and the Rotterdam Scan Study (RSS) scale, respectively. Cognition was assessed with a battery of 11 tests. RESULTS: Within the whole study group, the association with age was most evident for the automated measured WMH volume (beta = 0.43, 95% CI = 0.29-0.57). With regard to cognition, automated measured WMH volume and Scheltens DWMH were significantly associated with information processing speed (beta = -0.22, 95% CI = -0.40 to -0.06; beta = -0.26, 95% CI = -0.42 to -0.10), whereas RSS PWMH were associated with attention and executive function (beta = -0.19, 95% CI = -0.36 to -0.02). CONCLUSION: Measurements of WMH with an automated quantitative segmentation method are comparable with visual rating scales and highly suitable for use in future studies to assess the relationship between WMH and subtle impairments in cognitive function.

Cognitive impairment in tuberous sclerosis complex is a multifactorial condition.

OBJECTIVE: In patients with tuberous sclerosis complex (TSC), associations between tuber number, infantile spasms, and cognitive impairment have been proposed. We hypothesized that the tuber/brain proportion (TBP), the proportion of the total brain volume occupied by tubers, would be a better determinant of seizures and cognitive function than the number of tubers. We investigated tuber load, seizures, and cognitive function and their relationships. METHODS: Tuber number and TBP were characterized on three-dimensional fluid-attenuated inversion recovery MRI with an automated tuber segmentation program. Seizure histories and EEG recordings were obtained. Intelligence equivalents were determined and an individual cognition index (a marker of cognition that incorporated multiple cognitive domains) was calculated. RESULTS: In our sample of 61 patients with TSC, TBP was inversely related to the age at seizure onset and to the intelligence equivalent and tended to be inversely related to the cognition index. Further, a younger age at seizure onset or a history of infantile spasms was related to lower intelligence and lower cognition index. In a multivariable analysis, only age at seizure onset and cognition index were related. CONCLUSIONS: Our systematic analysis confirms proposed relationships between tuber load, epilepsy and cognitive function in tuberous sclerosis complex (TSC), but also indicates that tuber/brain proportion is a better predictor of cognitive function than tuber number and that age at seizure onset is the only independent determinant of cognitive function. Seizure control should be the principal neurointervention in patients with TSC.

Overlapping neurologic and cognitive phenotypes in patients with TSC1 or TSC2 mutations.

OBJECTIVE: The purpose of this study was to systematically analyze the associations between different TSC1 and TSC2 mutations and the neurologic and cognitive phenotype in patients with tuberous sclerosis complex (TSC). METHODS: Mutation analysis was performed in 58 patients with TSC. Epilepsy variables, including EEG, were classified. A cognition index was determined based on a comprehensive neuropsychological assessment. On three-dimensional fluid-attenuated inversion recovery MR images, an automated tuber segmentation program detected and calculated the number of tubers and the proportion of total brain volume occupied by tubers (tuber/brain proportion [TBP]). RESULTS: As a group, patients with a TSC2 mutation had earlier age at seizure onset, lower cognition index, more tubers, and a greater TBP than those with a TSC1 mutation, but the ranges overlapped considerably. Familial cases were older at seizure onset and had a higher cognition index than nonfamilial cases. Patients with a mutation deleting or directly inactivating the tuberin GTPase activating protein (GAP) domain had more tubers and a greater TBP than those with an intact GAP domain. Patients with a truncating TSC1 or TSC2 mutation differed from those with nontruncating mutations in seizure types only. CONCLUSIONS: Although patients with a TSC1 mutation are more likely to have a less severe neurologic and cognitive phenotype than those with a TSC2 mutation, the considerable overlap between both aspects of the phenotype implies that prediction of the neurologic and cognitive phenotypes in individuals with tuberous sclerosis complex should not be based on their particular TSC1 or TSC2 mutation.