AJNR. American journal of neuroradiology最新文献

Background and purpose: The choroid of the eye is a rare site for metastatic tumor spread, and as small lesions on the periphery of brain MRI studies, these choroidal metastases are often missed. To improve their detection, we aimed to use artificial intelligence to distinguish between brain MRI scans containing normal orbits and choroidal metastases.

Materials and methods: We present a novel hierarchical deep learning framework for sequential cropping and classification on brain MR images to detect choroidal metastases. The key innovation of this approach lies in training an orbit localization network based on a YOLOv5 architecture to focus on the orbits, isolating the structures of interest and eliminating irrelevant background information. The initial subtask of localization ensures that the input to the subsequent classification network is restricted to the precise anatomic region where choroidal metastases are likely to occur. In step 1, we trained a localization network on 386 T2-weighted brain MRI axial slices from 97 patients. Using the localized orbit images from step 1, in step 2 we trained a binary classifier network with 33 normal and 33 choroidal metastasis-containing brain MRIs. To address the challenges posed by the small data set, we used a data-efficient evolutionary strategies approach, which has been shown to avoid both overfitting and underfitting in small training sets.

Results: Our orbit localization model identified globes with 100% accuracy and a mean average precision (mAP) of intersection over union thresholds of 0.5-0.95 [mAP(0.5:0.95)] of 0.47 on held-out testing data. Similarly, the model generalized well to our step 2 data set, which included orbits demonstrating pathologies, achieving 100% accuracy and mAP(0.5:0.95) of 0.44. mAP(0.5:0.95) appeared low because the model could not distinguish left and right orbits. Using the cropped orbits as inputs, our evolutionary strategies-trained convolutional neural network achieved a testing set area under the curve of 0.93 (95% CI, 0.83-1.03), with 100% sensitivity and 87% specificity at the optimal Youden index.

Conclusions: The semiautomated pipeline from brain MRI slices to choroidal metastasis classification demonstrates the utility of a sequential localization and classification approach, and clinical relevance for identifying small, "corner-of-the-image," easily overlooked lesions. Artificial Intelligence Level of Evidence: 5B.

Background and purpose: Cerebral amyloid angiopathy (CAA) is a leading cause of intracranial hemorrhage and cognitive decline in the elderly. This study seeks to investigate the quantitative MRA features as new markers of CAA.

Materials and methods: In this cross-sectional study, consecutive patients with CAA and controls who underwent 3D time-of-flight MRA were included. Demographic and clinical data, including sex, age, diabetes, smoking, hypertension, and atrial fibrillation, were collected. Radiologic features, including the microhemorrhage classification, siderosis, and Fazekas scale, were also collected. Using in-house developed semiautomated software (VesselVoyager), quantitative MRA features, including total arterial length, number of branches, and tortuosity, were extracted. Univariable and multivariable logistic regression analyses were then performed to compare the CAA and non-CAA cohorts.

Results: Seventy-four patients were included: 43 with CAA and 31 controls. Quantitative MRA analysis showed that patients with CAA had significantly reduced total arterial length (1900 ± 1240 mm versus 2880 ± 1540 mm; P = .006). Univariable logistic regression identified total arterial length (P = .009), age (P = .012), and total branch number (P = .107) as relevant predictors of CAA. In multivariable analysis, total arterial length (OR: 0.95; 95% CI: 0.92-0.99; P = .014) and age (OR: 1.06; 95% CI: 1.01-1.12; P = .023) remained independently associated with CAA diagnosis.

Conclusions: Quantitative MRA total arterial length is inversely, independently, and significantly associated with CAA diagnosis. This could potentially serve as an imaging marker of CAA diagnosis and potentially further elucidate vascular segment involvement in patients with CAA.

Background and purpose: Diffusion MRI measures indicative of white matter integrity have consistently been shown to be altered in the state of the corticospinal tract (CST) and corpus callosum (CC) of patients with amyotrophic lateral sclerosis (ALS). However, diffusion MRI acquisitions are not routinely collected as part of the standard medical imaging of patients with ALS. T1-weighted MRI scans are commonly available in the clinical assessment of most patients with ALS. While visual inspection of these scans reveals little about the cerebral pathology of ALS, analysis of their textural patterns has identified disease-related abnormalities in patients at various stages of the disease. The present study aimed to examine the spatial and temporal profile of CST and CC degeneration in patients with ALS using texture analysis of T1-weighted MRI scans obtained at baseline and at 4- and 8-month follow-ups.

Materials and methods: The study involved data from 64 patients with ALS and 83 healthy controls who participated in the multicenter Canadian ALS Neuroimaging Consortium (CALSNIC) project. The texture feature "autocorrelation" (autoc) was quantified along the superior-inferior course of the CST and along the anterior-posterior direction of the CC of participants.

Results: Progressive textural changes were observed within the bilateral CST, particularly in the primary motor cortex region, posterior limb of the internal capsule, and the cerebral peduncle. As the disease progressed, significant textural changes developed in the middle and anterior parts of the CC. Autoc values in these regions correlated with the degree of upper motor neuron dysfunction on neurologic examination.

Conclusions: Progressive CST and CC degeneration was characterized in ALS using a novel imaging texture analysis approach, with changes observed over an interval of 4 months.

Background and purpose: Mechanical thrombectomy is the standard of care for acute ischemic stroke due to large vessel occlusion, but the optimal combination of devices remains under investigation. The SEMTiC-01 trial evaluates the safety and efficacy of a novel, fully integrated neurothrombectomy system composed of the iNedit balloon distal access catheter, the iNdeep microcatheter, and the iNtercept stent retriever.

Materials and methods: SEMTiC-01 is a prospective, multicenter, single-arm, open-label trial conducted across 18 centers in Spain, Germany, and Belgium. A total of 175 patients with LVO stroke were enrolled and treated within 24h of symptom onset. The primary efficacy endpoint was the rate of successful reperfusion (eTICI ≥2b) in ≤3 passes. The primary safety endpoints included serious adverse events within 24h and 90-day mortality. Secondary endpoints included good functional outcome (mRS 0-2 at 90 days), early neurological improvement, and device navigability.

Results: Successful reperfusion was achieved in 75.4% of the intention-to-treat (ITT) and 83.6% of the per-protocol (PP) population. Good functional outcomes were observed in 55.8% (ITT) and 56.9% (PP), and rapid neurological improvement in 64.6% of patients. First-pass reperfusion (eTICI ≥2b) was obtained in 46.9% (ITT) and 51.4% (PP). Symptomatic intracranial hemorrhage occurred in 1.7% (ITT), and mortality at 90 days was 13.7%. Device navigability was rated "Good" in over 75% of cases for both iNedit, iNtercept and iNdeep. Balloon inflation during thrombectomy improved outcomes in M1 occlusions.

Conclusions: The triad system (iNedit, iNdeep and iNtercept) demonstrates non-inferiority compared published literature, with favorable safety and efficacy profiles. Its favourable navigability suggests it is a promising option for mechanical thrombectomy in acute ischemic stroke.

Background and purpose: In patients with large-core infarcts, the risk and clinical implications of post-treatment intracranial hemorrhage (ICH) remain poorly understood. We aimed to characterize the frequency, patterns, predictors, and prognostic relevance of post-treatment ICH in patients with large-core infarcts treated in the TENSION trial.

Methods: We performed a post hoc analysis of 253 patients with anterior circulation stroke and an Alberta Stroke Program Early Computed Tomography Score (ASPECTS) of 3-5 randomized to either mechanical thrombectomy (MT) plus best medical treatment (BMT) or BMT alone. Hemorrhages were categorized both clinically (symptomatic vs. asymptomatic) and radiologically using the Heidelberg Bleeding Classification. Predictors of parenchymal hematoma (PH) and symptomatic ICH (sICH) were identified using logistic regression. The association between bleeding severity and 90-day outcome was evaluated using multivariable models.

Results: Any ICH occurred in 45.1% of patients, more frequently after MT compared with BMT (54.4% vs. 35.9%, p=0.004), mostly asymptomatic. Among patients with any ICH, hemorrhagic infarction was associated with the highest rate of favorable outcome (34.0%) and was equally distributed across treatment arms. PH were more common after MT (23.2% vs. 9.4%, p=0.004). Predictors of PH included MT itself (aOR 2.11, CI 1.11-3.99), higher NIHSS (aOR 1.13, CI 1.04-1.23), and larger core volume (aOR 1.003, CI 1.000-1.005). No independent predictors of sICH were identified. In adjusted models, bleeding severity was not associated with poor outcome, whereas age, NIHSS, and core volume were. Importantly, MT remained independently associated with better functional outcomes, even when adjusting for hemorrhagic events. However, the benefit of MT appeared attenuated in patients who developed PH, as shown by a significant treatment interaction.

Conclusions: ICH is common in large-core stroke, particularly after MT, but is often asymptomatic and not independently linked to poor outcome. PH may reduce the benefit of MT, but overall, MT remains associated with improved functional outcomes. Distinguishing hemorrhage types is crucial in assessing post-treatment risk in this vulnerable population.

Background: Leber Hereditary Optic Neuropathy (LHON) is a cause of vision loss that is presently most often diagnosed from clinical evaluation, ophthalmological examination, and genetic testing. However, imaging is being increasingly recognized by LHON specialists as tool for earlier diagnosis. Given that there is an effective treatment for LHON when implemented promptly after symptom onset, it is important for radiologists to be able to recognize and suggest LHON in the appropriate clinical context.

Purpose: In this individual participant data meta-analysis (IPDMA) study, trends in radiologic findings, additional diagnostic considerations, and treatment approaches are described in the largest cohort published to date of patients with LHON.

Data sources & study selection: The 2020 PRISMA guidelines were followed to identify all literature-based case descriptions of LHON. Inclusion criteria were the presence of neuroimaging descriptions, full-text article format, English-language, and a publication date between 2000-2025.

Data analysis: Data on patient age, imaging modality/findings, ophthalmological exams, and mitochondrial mutations were collected and nominalized for statistical analyses.

Data synthesis: 137 cases from 120 publications across 54 journals were identified. All cases involved LHON diagnoses, confirmed through genetic testing. The patient age range was 2-81 years. 22.6% of case reports described the finding of T2-hyperintensity of the optic nerves and/or chiasm. 10.9% described post-contrast enhancement of the optic nerves and/or chiasm, 5.1% described optic chiasm enlargement, and 13.1% described optic chiasm atrophy (OCA). T2-hyperintensity and post-contrast enhancement of the optic nerves and/or chiasm were disproportionately reported in females, rather than males (p=0.04 and p=0.06, respectively). T2-hyperintensity and post-contrast enhancement of the optic nerves and/or chiasm, T2-hyperintensity of the optic nerves and/or chiasm and optic chiasm enlargement, and T2-hyperintensity of the optic nerves and/or chiasm and optic chiasm enlargement were all more co-reported than expected (p<0.001 for all).

Limitations: Limitations of this study include heterogeneous data from case reports, inconsistent imaging descriptions, publication bias, and retrospective design.

Conclusion: In this IPDMA, patients of all ages, both male and female, presented with vision loss from LHON. There was a propensity for T2 -hyperintensity of the optic nerves and/or chiasm, particularly in females, and associated post-contrast enhancement of the lesions. Given the profound impact early diagnosis can make on saving patients' vision, it is important for radiologists to understand neuroimaging and clinical associations of LHON.

Background and purpose: The role of endovascular thrombectomy for acute ischemic stroke due to M2 middle cerebral artery occlusion remains uncertain, in part because angiographic recanalization rates are lower than those achieved in large-vessel occlusion. The hypoperfusion intensity ratio (HIR), a perfusion-derived marker of collateral physiology, has been associated with infarct evolution and outcomes, but its relationship with technical success of thrombectomy in M2 occlusion is unknown.

Materials and methods: In this retrospective, multinational study of the MAD-MT Consortium registry, patients with M2 MCA occlusion who underwent EVT between September 2017 and July 2023 were included. Multivariable logistic regression analysis for successful EVT (mTICI 2b-3) was performed with HIR alongside demographic and stroke characteristics.

Results: Among 111 M2 MCA occlusion patients, 23 (21%) had unsuccessful EVT and 88 (79%) had successful EVT. HIR (0.50 (0.00-0.63) vs 0.20 (0.00-0.40), p = 0.007) and NIHSS score (15 (10-18) vs 10 (7-16), p = 0.05) were significantly lower in mTICI 2b-3 patients, along with a significantly higher ASPECTS (8 (6-10) vs 9 (8-10), p = 0.03). In the multivariable logistic regression analysis, higher HIR, representing impaired collateral circulation, was independently associated with lower odds of mTICI 2b-3 (OR 0.03, 95% CI: 0.00-0.41, p = 0.009).

Conclusions: This retrospective study suggests that HIR may serve as a predictor of success in EVT and a potentially useful adjunct tool for decision-making in MeVO stroke patients. However, the limited sample size and unmeasured cofounders should be considered. These findings underscore the role that perfusion imaging-based quantitative collateral assessments can play in prognosis for these patients.

Background and purpose: Carotid webs (CWs) are shelf-like fibrous intimal projections in the carotid bulb that perturb laminar flow, promote local stasis, and predispose to embolic stroke. While morphology correlates with events, in vivo hemodynamic markers of stasis are lacking. We hypothesized that multiphase CT angiography (CTA) can quantify contrast persistence within web recess and impaired washout may be associated with stroke/transient ischemic attack (TIA).

Materials and methods: We retrospectively reviewed head and neck CTA performed between January 2021 and June 2024 in adults whose radiology reports mentioned "web". CWs were confirmed by independent review. Patients were classified as stroke/TIA causing if their ischemic event was attributed to the web, and asymptomatic if the web was incidentally identified or contralateral to the presenting event. Only symptomatic patients and age- and sex-matched asymptomatic patients who underwent multiphase neck CTA were included. Region of interests (ROIs) were placed in posterior recess of CWs, proximal and distal arterial lumens. Hounsfield units (HU) were recorded in ROIs on arterial (A) and delayed (D) phases. A_web and D_web were defined as web posterior recess HU in both phases; A_ref and D_ref were defined as the averaged HU from the adjacent proximal and distal arteries. Two indices were derived: contrast-washout ratio (CWR) = (A_web - D_web)/(A_ref - D_ref); contrast-pooling index (CPI) = (D_web/D_ref) - (A_web/A_ref). Continuous variables were compared with t-test or Mann-Whitney U test, and categorical variables with Fisher exact tests.

Results: Among 190 patients with confirmed 206 webs, we identified seven symptomatic patients with ischemic stroke/TIA attributed to CW who had neck multiphase CTA and compared with seven age- and sex-matched asymptomatic patients. Baseline demographics and CW morphology did not differ between cohorts. CWR was lower in symptomatic patients than controls (median [IQR] 0.672 [0.652-0.920] vs. 0.918 [0.819-1.076]; p = 0.04), indicating reduced washout. CPI was higher in symptomatic patients (median [IQR] 0.098 [0.074-0.212] vs. -0.009 [-0.169-0.060]; p = 0.04), consistent with contrast persistence and blood stasis.

Conclusion: These findings suggest that contrast persistence metrics on multiphase CTA may reflect local stasis within CWs; however, results are preliminary and warrant validation in larger cohorts.

Background and purpose: Cerebral air embolism (CAE) is a rare but treatable cause of ischemic stroke. Clinically, CAE may be difficult to distinguish from stroke due to more typical thromboembolic causes, but accurate diagnosis is critical to initiate appropriate treatment. We aimed to define the imaging features of CAE by comparing MRI from patients with confirmed CAE to those in cardioembolic stroke due to atrial fibrillation (AF).

Materials and methods: In a retrospective, matched case-control study, CAE cases from 2012-2023 were matched 1:2 by presenting NIHSS to control patients who had stroke due to AF and were not treated with thrombolytics or thrombectomy. MRIs were reviewed by a neuroradiologist blinded to group. The primary outcome was presence of pre-specified neuroimaging features on MRI.

Results: Fourteen patients with stroke due to CAE (median age 61, 64% female, median NIHSS 12) and 28 controls with stroke due to AF (median age 81, 43% female, median NIHSS 12) were included. The predominant infarction topography in CAE patients was gyriform in 86%, punctate in 7%, and wedge-shaped in 7%, whereas in patients with stroke due to AF the predominant infarction topography was wedge-shaped in 71%, punctate in 18%, and gyriform in 11% (p<0.001). CAE patients more often presented with multiple (93% versus 50%, p=0.007) and bilateral infarctions (79% versus 43%, p=0.05). Cortical borderzone involvement was more frequent in patients with CAE compared to those with AF (86% versus 25%, p<0.001). The presence of both predominantly gyriform infarction topography and cortical borderzone involvement had a 76.6% sensitivity and 96.4% specificity for CAE.

Conclusions: CAE cause characteristic gyriform infarction patterns on MRI that are distinct from typical cardioembolic stroke. In addition, cortical borderzone predilection and multifocal infarctions were substantially more frequent in CAE. This constellation of findings, in the appropriate clinical context, should strongly suggest CAE as the mechanism of neurologic injury, and may facilitate timely identification of this uncommon but critical diagnosis.

SUMMARYSpontaneous intracranial hypotension (SIH) is a debilitating and often misdiagnosed condition. One important differential diagnosis is Chiari malformation type 1 (CM-1), which differs markedly in its pathophysiology, clinical manifestation, and treatment. We describe seven patients with SIH initially misdiagnosed as CM-1 who underwent posterior fossa decompression. Five patients developed temporary symptom relief after posterior fossa decompression and six patients developed rebound intracranial hypertension. Our findings emphasize the importance of considering SIH when evaluating low-lying cerebellar tonsils to prevent unnecessary procedures and delays in appropriate treatment.