Citations

PurposeInversion algorithms used to convert acquired MR elastography wave data into material property estimates often assume that the underlying materials are locally homogeneous. Here we evaluate the impact of that assumption on stiffness estimates in gray‐matter regions of interest in brain MR elastography.MethodsWe describe an updated neural network inversion framework using finite‐difference model–derived data to train convolutional neural network inversion algorithms. Neural network inversions trained on homogeneous simulations (homogeneous learned inversions [HLIs]) or inhomogeneous simulations (inhomogeneous learned inversions [ILIs]) are generated with a variety of kernel sizes. These inversions are evaluated in a brain MR elastography simulation experiment and in vivo in a test–retest repeatability experiment including 10 healthy volunteers.ResultsIn simulation and in vivo, HLI and ILI with small kernels produce similar results. As kernel size increases, the assumption of homogeneity has a larger effect, and HLI and ILI stiffness estimates show larger differences. At each inversion's optimal kernel size in simulation (7 × 7 × 7 for HLI, 11 × 11 × 11 for ILI), ILI is more sensitive to true changes in stiffness in gray‐matter regions of interest in simulation. In vivo, there is no difference in the region‐level repeatability of stiffness estimates between the inversions, although ILI appears to better maintain the stiffness map structure as kernel size increases, while decreasing the spatial variance in stiffness estimates.ConclusionsThis study suggests that inhomogeneous inversions provide small but significant benefits even when large stiffness gradients are absent.

OBJECTIVE Magnetic resonance imaging at 7T offers improved image spatial and contrast resolution for visualization of small brain nuclei targeted in neuromodulation. However, greater image geometric distortion and a lack of compatible instrumentation preclude implementation. In this report, the authors detail the development of a stereotactic image localizer and accompanying imaging sequences designed to mitigate geometric distortion, enabling accurate image registration and surgical planning of basal ganglia nuclei. METHODS Magnetization-prepared rapid acquisition with gradient echo (MPRAGE), fast gray matter acquisition T1 inversion recovery (FGATIR), T2-weighted, and T2*-weighted sequences were optimized for 7T in 9 human subjects to visualize basal ganglia nuclei, minimize image distortion, and maximize target contrast-to-noise and signal-to-noise ratios. Extracranial spatial distortions were mapped to develop a skull-contoured image localizer embedded with spherical silicone fiducials for improved MR image registration and target guidance. Surgical plan accuracy testing was initially performed in a custom-developed MRI phantom (n = 5 phantom studies) and finally in a human trial. RESULTS MPRAGE and T2*-weighted sequences had the best measures among global measures of image quality (3.8/4, p < 0.0001; and 3.7/4, p = 0.0002, respectively). Among basal ganglia nuclei, FGATIR outperformed MPRAGE for globus pallidus externus (GPe) visualization (2.67/4 vs 1.78/4, p = 0.008), and FGATIR, T2-weighted imaging, and T2*-weighted imaging outperformed MPRAGE for substantia nigra visualization (1.44/4 vs 2.56/4, p = 0.04; vs 2.56/4, p = 0.04; vs 2.67/4, p = 0.003). Extracranial distortion was lower in the head’s midregion compared with the base and apex ( 1.17–1.33 mm; MPRAGE and FGATIR, p < 0.0001; T2-weighted imaging, p > 0.05; and T2*-weighted imaging, p = 0.013). Fiducial placement on the localizer in low distortion areas improved image registration (fiducial registration error, 0.79–1.19 mm; p < 0.0001) and targeting accuracy (target registration error, 0.60–1.09 mm; p = 0.04). Custom surgical software and the refined image localizer enabled successful surgical planning in a human trial (fiducial registration error = 1.0 mm). CONCLUSIONS A skull-contoured image localizer that accounts for image distortion is necessary to enable high-accuracy 7T imaging–guided targeting for surgical neuromodulation. These results may enable improved clinical efficacy for the treatment of neurological disease.

Introduction: The long term effectiveness of preventive intracranial treatment of unruptured aneurysms for hemorrhage and aneurysm-related morbidity and mortality has been hypothesized. No randomized trial has been done. Propensity adjusted comparison within The International Study of Unruptured Intracranial Aneurysms was undertaken to compare long-term effectiveness. Hypothesis: The purpose of this analysis was to compare the long-term outcome in the treatment cohorts (surgery and endovascular) with those of the untreated subgroup. Methods: Patients were subdivided into the initial treatment and untreated cohorts based upon observation or treatment practices in 61 centers from 1991-1998. 1691 patients were in the observational cohort, 471 were in the endovascular cohort and 1917 patients were in the surgical cohort. The cohorts were followed for a median follow-up of 9.2 years. Outcomes were determined prospectively and with central review. The data were grouped together and analyzed to determine treatment decisions. Comparison of outcomes was done using covariate adjustment and using propensity analysis where equivalent probability groups were constructed. Similar main predictors of treatment and outcome were observed. Analysis of hemorrhage, mortality and short-term and long-term morbidity and mortality were assessed. Risk-benefit ratios were also calculated. Results: Significant differences in baseline variables between treatment and observed patients were aneurysm size, symptoms, age, prior SAH or not group, geographical region, family history, hypertension and myocardial infarction history. The results show comparability of groups using the patient propensity-based score subset. Using the propensity score method, the results showed a benefit in prevention of hemorrhage by surgery and endovascular treatment versus observation (p<.01). No difference was found between endovascular versus surgery. However when procedure-related outcomes are included, no significant difference at 1 year, 5 years or 10 years in the combined aneurysm or procedure related endpoint was evident between cohorts. Conclusions: Propensity score matching to achieve comparability resulted in balanced groups. Comparison of long-term results showed that intervention demonstrated a benefit in reducing hemorrhage risk. However, peri-procedural risk negates any short-term benefit for 10 years.

Introduction: Mortality in patients with unruptured intracranial aneurysm (UIA) is significantly higher than the general population. However a minority of the deaths are due to subarachnoid hemorrhage. Hypothesis: There are differential characteristics of patients that might predict specific causes of death that would be informative in counseling patients about aneurysm risk. Methods: 4059 patients with UIA who were enrolled prospectively between 1991 and 1998 were followed by 61 medical centers in North America, Canada, and Europe. Patients have at least one UIA, which may or may not be symptomatic. Patients were Rankin grade 1 or 2 at enrollment. Patients were followed annually from enrollment to 2007. For patients who died a death certificate search was done and information was requested from local physicians and other contacts. A National Death Index search in 2008 determined cases who died and cause of death. Foreign centers also did a national search where possible. Patients who died before 30 days post treatment or lost to follow up before 30 days post enrollment were excluded. Results: Twenty-five percent of patients died during follow-up. The life expectancy was significantly less than age-, sex-, and country-matched expected estimates. Among the total of 4004 patients with UIA who survived 30 days after enrollment, 2331 were treated and 1673 were not. There were 965 deaths after 30 days after treatment, of which 320 died from vascular disease (252 died from non-SAH vascular disease), 114 died from respiratory tract disease, 255 died from cancer, and the rest died of other or unknown reasons. Predictors of SAH death were aneurysm size, site, age and convulsive disorders at enrollment. Factors related to vascular death were UIA treatment, age, gender, race, hypertension, Rankin at discharge and diminishing influence of aneurysm characteristics over time. Respiratory deaths were related to treatment, race, smoking, age, initial Rankin and presentation with convulsive disorder. Cancer deaths were related to gender, race, smoking and age. Conclusions: In patients with UIA, an assessment of risk for death underscores the importance of age, race and gender and several modifiable risk factors. Patient management should include not only aneurysm rupture prevention but also modification of smoking and hypertension.

Introduction: Quality of life changes can occur following procedural interventions and subarachnoid hemorrhage in unruptured intracranial aneurysms (UIA). The International Study of Unruptured Intracranial Aneurysms conducted quality of life assessment using the Mini-Mental State Exam, Rankin Score, and Barthel Index at baseline and the Telephone Interview Cognitive Status Exam (TICS), Rankin Score, and Barthel Index at follow-up. Hypothesis: The purpose of this analysis was to compare the 1 year and 5 year outcome in the treatment cohorts (surgery and endovascular) with those of the untreated subgroup. Methods: Patients were subdivided into the initial treated and untreated cohorts based upon observation or treatment practices in 61 centers from 1991-1998. 1691 patients were in the observational cohort and 2388 in the treated cohort. The cohorts were followed annually with measures of neurological status, Rankin Scale, Barthel Index and cognition (TICS). Outcomes were determined prospectively. Comparison of outcomes was done using continuous and categorical data. Covariate adjustment was done using general linear models. Stratified analysis was done by Rankin Score at the time of follow-up. Results: The percent of patients with a Barthel Index score of less than 60 at 1 year in the cohort was 9.4% (treated patients, 9.7%; untreated patients, 9.2%). The percent of patients at year 1 with a score of greater than 90 was 87.6%, 87.4% for treated patients and 87.8% for untreated patients. The percent of patients with a Barthel Index score of less than 60 at 5 years in the cohort was 41.4% (treated patients, 41.4%; untreated patients, 41.4%). The percent of patients at year 5 with a score of greater than 90 was 56.3%, 56.1% for treated patients and 56.5% for untreated patients. TICS and Barthel Index scores were significantly associated with each other for patients with Rankin Scores 0-2 (P<0.0001). This was seen at both year 1 and year 5 follow-up, but not for patients with Rankin Score 3-5. At both 1 year and 5 year follow-up, after controlling for Rankin Score, the distribution of Barthel Index scores of untreated patients and treated patients did not significantly differ. At both the 1 year follow-up, baseline Rankin Score, TICS and follow-up Rankin Score were significant predictors of Barthel Score (P<0.0001). At 5 years, only the follow-up Rankin Score was associated with Barthel Score (P<0.0001). Conclusions: Barthel Index Score decreased over time, but does not differ with treatment. The Rankin Score was the main predictor at follow-up of Barthel Index Score distribution.

BackgroundLocalized regions of left–right image intensity asymmetry (LRIA) were incidentally observed on T2‐weighted (T2‐w) and T1‐weighted (T1‐w) diagnostic magnetic resonance imaging (MRI) images. Suspicion of herpes encephalitis resulted in unnecessary follow‐up imaging. A nonbiological imaging artifact that can lead to diagnostic uncertainty was identified.PurposeTo investigate whether systematic LRIA exist for a range of scanner models and to determine if LRIA can introduce diagnostic uncertainty.Study TypeA retrospective study using the Alzheimer's Disease Neuroimaging Initiative (ADNI) data base.SubjectsOne thousand seven hundred fifty‐three (median age: 72, males/females: 878/875) unique participants with longitudinal data were included.Field Strength3T.SequencesT1‐w three‐dimensional inversion‐recovery spoiled gradient‐echo (IR‐SPGR) or magnetization‐prepared rapid gradient‐echo (MP‐RAGE) and T2‐w fluid‐attenuated inversion recovery (FLAIR) long tau fast spin echo inversion recovery (LT‐FSE‐IR). Only General Electric, Philips, and Siemens' product sequences were used.AssessmentLRIA was calculated as the left–right percent difference with respect to the mean intensity from automated anatomical atlas segmented regions. Three neuroradiologists with 37 (**), 32 (**), and 3 (**) years of experience rated the clinical impact of 30 T2‐w three‐dimensional FLAIR exams with LRIA to determine the diagnostic uncertainty. Statistical comparisons between retrospective intensity normalized T1m and original T1‐w images were made.Statistical TestsFor each image type, a linear mixed effects model was fit using LRIA scores from all scanners, regions, and participants as the outcome and age and sex as predictors. Statistical significance was defined as having a P‐value <0.05.ResultsLRIA scores were significantly different from zero on most scanners. All clinicians were uncertain or recommended definite diagnostic follow‐up in 62.5% of cases with LRIA >10%. Individuals with acute brain pathology or focal neurologic deficits are not enrolled in ADNI; therefore, focal signal abnormalities were considered false positives.Data ConclusionLRIA is system specific, systematic, creates diagnostic uncertainty, and impacts IR‐SPGR, MP‐RAGE, and LT‐FSE‐IR product sequences.Level of Evidence: 2Technical Efficacy Stage: 3

OBJECTIVE Pituitary adenoma is one of the most common primary intracranial neoplasms. Most of these tumors are soft, but up to 17% may have a firmer consistency. Therefore, knowing the tumor consistency in the preoperative setting could be helpful. Multiple imaging methods have been proposed to predict tumor consistency, but the results are controversial. This study aimed to evaluate the efficacy of MR elastography (MRE) in predicting tumor consistency and its potential use in a series of patients with pituitary adenomas. METHODS Thirty-eight patients with pituitary adenomas (≥ 2.5 cm) were prospectively evaluated with MRI and MRE before surgery. Absolute MRE stiffness values and relative MRE stiffness ratios, as well as the relative ratio of T1 signal, T2 signal, and diffusion-weighted imaging apparent diffusion coefficient (ADC) values were determined prospectively by calculating the ratio of those values in the tumor to adjacent left temporal white matter. Tumors were classified into three groups according to surgical consistency (soft, intermediate, and firm). Statistical analysis was used to identify the predictive value of the different radiological parameters in determining pituitary adenoma consistency. RESULTS The authors included 32 (84.21%) nonfunctional and 6 (15.79%) functional adenomas. The mean maximum tumor diameter was 3.7 cm, and the mean preoperative tumor volume was 16.4 cm 3 . Cavernous sinus invasion was present in 20 patients (52.63%). A gross-total resection was possible in 9 (23.68%) patients. The entire cohort’s mean absolute tumor stiffness value was 1.8 kPa (range 1.1–3.7 kPa), whereas the mean tumor stiffness ratio was 0.66 (range 0.37–1.6). Intraoperative tumor consistency was significantly correlated with absolute and relative tumor stiffness (p = 0.0087 and 0.007, respectively). Tumor consistency alone was not a significant factor for predicting gross-total resection. Patients with intermediate and firm tumors had more complications compared to patients with soft tumors (50.00% vs 12.50%, p = 0.02) and also had longer operative times (p = 0.0002). CONCLUSIONS Whereas other MRI sequences have proven to be unreliable in determining tumor consistency, MRE has been shown to be a reliable tool for predicting adenoma consistency. Preoperative knowledge of tumor consistency could be potentially useful for surgical planning, counseling about potential surgical risks, and estimating the length of operative time.

Purpose This study sought to validate whether the signal intensity ratio (SIR) of carotid intraplaque hemorrhage (IPH) was associated with acute ischemic neurologic events. Methods A retrospective review was completed of consecutive patients that underwent neck magnetic resonance angiography using magnetization prepared-rapid gradient echo (MP-RAGE) and T1-CUBE sequences between 2017 and 2020. Patients with magnetic resonance evidence of IPH were included. SIRs were measured by comparing the maximum IPH signal with the mean intramuscular signal from the adjacent sternocleidomastoid. Patients were stratified into ischemic or non-ischemic groups based on the presence of acute ipsilateral ischemic events (stroke, retinal artery occlusion). Logistic regression analysis was performed to determine if increasing IPH SIR was associated with an increased risk of ipsilateral ischemic events. Results Of 85 included patients (85 arteries), 66 were male (77.6%). Mean age was 71.0 (SD ± 11.1). There were 70 arteries with IPH that were ipsilateral to an ischemic event, and 15 that belonged to patients without an ischemic event. No association was found between increasing IPH SIR seen on MP-RAGE (odds ratio (OR): 0.82; 95% confidence interval (CI): 0.58–1.4; P = 0.43) or T1-CUBE sequences (OR: 0.85; 95% CI: 0.53–1.5; P = 0.56). Conclusions There was no association between the SIR of IPH and acute ischemia on either MP-RAGE or T1-CUBE sequences. Further investigation is required prior to widespread acceptance of SIR as a predictive imaging marker of symptomatic carotid plaque.

INTRODUCTION: The decision regarding whether to perform an interventional procedure as a strategy to prevent hemorrhage of an unruptured intracranial aneurysm (UIA) requires careful consideration of procedural risk and the UIA natural history. No randomized trial data are available. The International Study of Unruptured Intracranial Aneurysms (ISUIA) included a prospective cohort, examining hemorrhage risk and treatment risk. Hypothesis: The purpose of this analysis was to compare the factors related to treatment selection and determination of the number of hemorrhages prevented. Methods: Patients were allocated into the initial treatment and untreated cohorts based upon observation or treatment practices in 61 centers from 1991-1998. 1691 patients were in the observational cohort, 471 were in the endovascular cohort and 1917 patients were in the surgical cohort. The cohorts were followed for a median follow-up of 9.2 years. Outcomes were determined prospectively and with central review. The data were grouped together and analyzed to determine treatment decisions. A Cox proportional hazards model predicting hemorrhage developed in the observation cohort and was applied to the surgery and endovascular cohorts across the follow-up period. Results: Significant baseline variable differences between treated and observed patients were aneurysm size, symptoms, age, prior SAH group, geographical region, treatment percentage, aneurysm daughter sacs or multiple lobes, and history of hypertension, smoking and myocardial infarction. Aneurysm site and family history were not significant. Site, size, and aspirin use were significant predictors of hemorrhage. Long-term the predicted hemorrhage rates were 6.7% at 5 years and 8.0% at 10 years in the surgery group and 8.1% and 9.6% for the endovascular group, respectively. For comparison the rates in the observed cohort were 4.1% and 4.8%, respectively. Conclusions: Decisions for treatment are influenced by patient characteristics such as age and medical history, aneurysm characteristics such as size and morphology and center and regional practices. Patients in the treated cohorts were at moderately increased risk for hemorrhage compared to those in the observed cohort.