AJNR. American journal of neuroradiology最新文献

Background and purpose: Brain imaging with MRI or CT is standard in screening for intracranial disease among ambulatory patients with cancer. Though MRI offers greater sensitivity, CT is frequently employed because of its accessibility, affordability, and faster acquisition time. However, the necessity of routinely performing a noncontrast CT with the contrast-enhanced study is unknown. This study evaluates the clinical and economic utility of the noncontrast portion of the brain CT examination.

Materials and methods: A board-certified neuroradiologist reviewed 737 brain CT reports from outpatients at The University of Texas MD Anderson Cancer Center who underwent contrast and noncontrast CT for cancer staging (October 2014 to March 2016) to assess if clinically meaningful findings were identified only on noncontrast CT. A Generative Pretrained Transformers-3 (GPT-3) model was then fine-tuned to extract reports with a high likelihood of unique and meaningful noncontrast findings from 1980 additional brain CT reports (January 2017 to April 2022). These reports were manually reviewed by 2 neuroradiologists, with adjudication by a third reviewer if needed. The incremental cost-effectiveness ratio of noncontrast CT inclusion was then calculated based on Medicare reimbursement and the 95% CI of the proportion of all reports in which noncontrast CT was necessary for identifying meaningful findings.

Results: Seven of 737 reports in the initial data set revealed meaningful findings unique to the noncontrast CT, all of which were hemorrhage. The GPT-3 model identified 145 additional reports with a high unique noncontrast CT finding likelihood for manual review from the second data set of 1980 reports. Nineteen of these reports were found to have unique and meaningful noncontrast CT findings. In total, 0.96% (95% CI: 0.63-1.40) of reports had meaningful findings identified only on noncontrast CT. The incremental cost-effectiveness ratio for the identification of a single meaningful finding on noncontrast CT missed on the contrast-enhanced study was $1855 to $4122.

Conclusions: In brain CT for ambulatory screening for intracranial disease in patients with cancer, noncontrast CT offers limited additional diagnostic value compared with contrast-enhanced CT alone. Considering the associated financial cost, workload, and patient radiation exposure associated with performing a noncontrast CT, contrast-enhanced brain CT alone is sufficient for cancer staging in patients with asymptomatic cancer.

Background and purpose: There is an imperfect correlation between morphologic MRI findings and radiating low back pain (LBP). Nerve irritation, visualized as glucose hypermetabolism on [¹⁸F]FDG-PET/MRI, has the potential to identify symptomatic segments. This study aimed to investigate the association of foraminal [¹⁸F]FDG uptake on PET/MRI, radiologic abnormalities, and patient outcomes.

Materials and methods: Prospectively recruited patients with radiating LBP underwent [¹⁸F]FDG-PET/MRI of the lumbar spine in this observational study. Back pain and leg/buttock pain were assessed by using the visual analog scale (0-10). Foraminal stenosis, facet joint arthropathy, and annular fissures of the disc were graded by radiologists. As part of the standard clinical care, a subset of patients received image-guided nerve root blocks, by using a steroid/anesthetic mixture, and pain on visual analog scale was noted before and after injection. Standardized tracer uptake was quantitatively assessed in all neural foramina, facet joints, and discs. Generalized estimating equations were used to investigate associations between the maximum standardized uptake value of [¹⁸F]FDG in the neural foramina, degree of stenosis (none, mild, moderate, severe), and pain, additionally adjusted for tracer uptake in the adjacent tissues, age, sex, and body mass index.

Results: A total of 110 lumbar neural foramina in 11 patients were included in the analysis. Generalized estimating equations revealed significant associations between foraminal [¹⁸F]FDG uptake and degree of foraminal stenosis (β = 0.18; 95% CI, 0.03-0.33; P = .02). In patients with unilateral radicular symptoms but bilateral stenoses on MRI, [¹⁸F]FDG uptake was significantly higher on the symptomatic side (1.64 versus 1.88; P = .002). In segments treated with image-guided nerve root block, change in pain was positively associated with foraminal [¹⁸F]FDG uptake before injection (β = 2.24; 95% CI, 0.03-4.45; P = .05) but negatively associated with degree of stenosis (β = -1.27; 95%CI -2.24 to -0.31; P = .01).

Conclusions: Foraminal [¹⁸F]FDG uptake on PET/MRI as a surrogate marker of nerve irritation may improve differentiation between painful versus nonpainful foraminal stenosis.

Background and purpose: This study aims to provide a comprehensive comparison of the performance and reproducibility of 2 commercially available artificial intelligence (AI) software computer-aided triage and notification solutions, Vendor A (Aidoc) and Vendor B (Viz.ai), for the detection of intracranial hemorrhage (ICH) on noncontrast-enhanced head CT scans performed within a single academic institution.

Materials and methods: The retrospective analysis was conducted on a large patient cohort from multiple health care settings within a single academic institution, utilizing standardized scanning protocols. Sensitivity, specificity, false-positive (FP), and false-negative (FN) rates were evaluated for both vendors. Outputs assessed included AI-generated case-level classification.

Results: Among 4081 scans, 595 were positive for ICH. Vendor A demonstrated a sensitivity of 94.4% and specificity of 97.4%, PPV of 77.7%, and NPV of 99.5%. Vendor B showed a sensitivity of 59.5% and specificity of 99.0%, PPV of 85.5%, and NPV of 96.2%. Vendor A had 20 FNs, which primarily involved subdural and intraparenchymal hemorrhages, and 97 FPs, which appear to be related to motion artifact. Vendor B had 145 FNs, largely comprising of subdural and subarachnoid hemorrhages, and 36 FPs, which appeared to be related to motion artifact and calcified or dense lesions. Concordantly, 18 cases were FNs and 11 cases were FPs for both AI solutions.

Conclusions: The findings of this study provide valuable information for clinicians and health care institutions considering the implementation of AI software for computer-aided triage and notification in the detection of intracranial hemorrhage. The discussion encompasses the implications of the results, the importance of evaluating AI findings in context-especially in the absence of explainability tools, potential areas for improvement, and the relevance of standardized scanning protocols in ensuring the reliability of AI-based diagnostic tools in clinical practice.

Background: Molecular imaging, particularly PET, has advanced the diagnosis and management of disease by visualizing biologic processes at a cellular and molecular level. PET imaging of the brain, spine, and head/neck, summarized under the umbrella term neuro-PET, enables noninvasive diagnosis and monitoring of diseases such as dementia, epilepsy, cancer, movement, or autoimmune disorders. The increasing prevalence of these conditions, as well as new treatment options necessitating response assessment, are expected to escalate neuro-PET imaging volumes, with projections for an increase in the need for specialized imaging services. This increasing clinical need highlights existing workforce shortages and underscores the need for neuroradiologists to acquire proficiency in molecular imaging. This expanded role seeks to address the growing demand. To this end, we propose a rigorous, structured, patient-centered, and collaborative framework for expanding neuroradiologists' training and practice to include neuro-PET interpretation.

Methods: This American Society of Neuroradiology consensus statement outlines competency recommendations, training pathways, and implementation strategies to incorporate neuro-PET into neuroradiology practice. This approach is based on existing guidelines and was informed by survey data from neuroradiologists and molecular imaging subspecialists revealing current practice patterns and training needs. For neuroradiology fellows, structured training encompasses hands-on neuro-PET imaging experience, understanding the biologic and molecular basis of radiopharmaceuticals used in neuro-PET, and integrating molecular insights with anatomic data. Neuroradiologists beyond fellowship can undertake practice-based curriculum involving supervised case interpretation, standardized reader training courses, continuing medical education (CME), and peer review.

Key message: Neuroradiologists, with their in-depth expertise of central nervous system structure and function, are well positioned to meld molecular imaging data with traditional anatomic findings. They can achieve competency and should be granted practice privileges in interpreting neuro-PET studies through a comprehensive combination of structured training, hands-on clinical experience, and documented CME hours.

Background and purpose: Spinal CSF leaks cause spontaneous intracranial hypotension (SIH), characterized by orthostatic headaches, but the detection of these leaks may require specialized and invasive spinal imaging. We have noted the presence of small lateral dural CSF collections of unclear significance on digital subtraction myelography (DSM) in some of these patients suspected of having SIH. The purpose of the present study was to compare radiographic and anatomic intraoperative findings in patients with such small lateral dural CSF collections who underwent surgical exploration.

Materials and methods: This retrospective cohort study included a consecutive group of patients suspected of having SIH who 1) did not have a spinal longitudinal extradural collection or CSF-venous fistula on spinal imaging; 2) underwent DSM under general anesthesia in the lateral decubitus position; and 3) underwent surgery for the finding of small lateral dural CSF collections of uncertain significance.

Results: The study group consisted of 27 patients (22 women and 5 men; mean age, 44.6 years; range, 16-72 years). DSM demonstrated a total of 31 small lateral dural CSF collections measuring 0.6-2.4 mm in diameter (mean, 1.3 mm) and 1.3-12.3 mm (mean, 3.5 mm) caudal to the origin of the neve root sleeve. Intraoperative exploration found evidence of a CSF leak in all 27 patients. One or more CSF-venous fistulas were found in 23 patients, and a pedicular type lateral CSF leak, in 4 patients. Radiographic differentiation between these 2 types of spinal CSF leaks could not be made with confidence.

Conclusions: Some patients suspected of having SIH have small lateral dural CSF collections on DSM caudal to the origin of the nerve root sleeve. We have found evidence of a CSF leak in all these patients on surgical exploration. This observation expands their treatment options.

Definitive treatment of spontaneous intracranial hypotension caused by a spinal CSF leak may be challenging with conservative management or an epidural blood patch. This study describes 2 cases of spontaneous intracranial hypotension treated with fluoroscopically-guided, catheter-delivered Onyx to seal a spinal CSF leak, with significant improvement observed during a 2-year follow-up. Digital subtraction myelography was used to localize the site of the CSF leak. Onyx was precisely delivered to the site of the CSF leak via a catheter under fluoroscopic guidance. Postprocedural spinal CT confirmed the presence of Onyx at the injected site. Both patients exhibited resolution of symptoms and CSF leaks during the 2-year follow-up. Repeat spinal MRI demonstrated a gradual and marked reduction in the spinal longitudinal epidural collection during the follow-up period. No noticeable complications were reported. These findings indicate that epidural Onyx treatment may represent a novel and promising strategy for managing spontaneous intracranial hypotension caused by a dural tear.

Background and purpose: Our purpose was to identify spatial covariance patterns of dual-phase [¹⁸F]PI-2620 PET in biomarker-confirmed Alzheimer disease (AD) and healthy control subjects by applying data-driven multivariate analysis to multimodality neuroimaging data. We also tested the ability of pattern expression values in individual subjects to predict amyloid status and evaluate the efficacy of [¹⁸F]PI-2620 as a single, universal biomarker for the amyloid/tau/neurodegeneration (A/T/N) classification.

Materials and methods: Twenty-five subjects (15 men, 10 women, mean age: 64.5 ± 10.1, range 51-89) including 15 with amyloid-positive AD and 10 amyloid-negative healthy controls were analyzed. Brain PET images of dual-phase [¹⁸F]PI-2620 (early-phase for cerebral perfusion; late-phase for tau pathology) and late-phase [¹⁸F]florbetaben for amyloid pathology were acquired in each participant alongside high-resolution brain MRI. PET images were converted into maps of standard uptake value ratio by using cerebellar GM as reference region. Spatial covariance analysis was performed separately in the standard brain space by using a well-established computing toolbox in the public domain.

Results: We identified distinct Alzheimer disease-related patterns (ADRP) of spatial covariance capturing prominent abnormal features in brain amyloid, tau, and perfusion (ADRP-amyloid, ADRP-tau, and ADRP-perfusion) underlying this disease. There was some overlap among these topographies particularly between ADRP-tau and ADRP-amyloid. ADRP expression scores of each pattern differentiated AD from healthy controls (P < .0001) with group discriminant analysis yielding an accuracy of 96% in predicting amyloid status with the combination of ADRP-tau and ADRP-perfusion scores. These scores correlated positively among themselves and with several clinical measures of disease severity in the combined subject group.

Conclusions: Analysis of AD and normal controls suggests a potential role of [¹⁸F]PI-2620 as a single biomarker for the A/T/N classification.

Background and purpose: Cerebral malaria (CM) is a leading cause of childhood mortality and neurologic morbidity in sub-Saharan Africa and South Asia and a strong association between diffuse brain swelling and mortality has been well-established. Our goal was to characterize patterns of cortical edema on brain MRI in children with CM and determine their association with patient outcomes.

Materials and methods: We retrospectively reviewed admission brain MR images obtained from Malawian children with clinical CM admitted at a single center from 2013-2019. Two neuroradiologists assessed the pattern of cortical edema on T1-, T2-, and DWI by using a consensus approach. The overall degree of brain volume (brain volume score [BVS]) and other brain imaging findings were also assessed, including focal signal changes in the basal ganglia, white matter, and posterior fossa. We evaluated the frequency and associations of these imaging findings with clinical outcomes at hospital discharge (deceased, alive with neurologic sequelae, or alive without neurologic sequelae).

Results: We included admission brain MRI scans from 190 children with clinical CM. Cortical edema was identified in 163 MRIs. The predominant pattern of cortical edema was diffuse cortical involvement with relative sparing of the occipital and peri-Rolandic areas: 103 (63.2%) had this pattern, whereas 37 (22.7%) had sparing of the occipital cortex only, and 23 (14.1%) had generalized cortical edema without focal sparing. The presence of occipital and peri-Rolandic sparing inversely correlated with BVS (β = -0.26, P < .001) and outcomes (OR: 0.3; 95% CI: 0.1-0.6; P = .002).

Conclusions: Pediatric CM is associated with a typical pattern of cortical edema that relatively spares the occipital and peri-Rolandic areas, which become progressively involved with more severe disease.

Background and purpose: According to the guideline, CT perfusion should be read and analyzed by using computer-aided software. This study evaluates the efficacy of artificial intelligence (AI)/machine learning-driven software in CTP imaging and the effect of neuroradiologists' interpretatios on these automated results.

Materials and methods: We conducted a retrospective, single-center cohort study from June to December 2023 at a comprehensive stroke center. A total of 132 patients suspected of having acute ischemic stroke underwent CTP using AI software. RapidAI was used for the initial analysis, with subsequent validation and adjustments made by experienced neuroradiologists. The rate of CTP marked as "nonreportable," "reportable," and "reportable with correction" by neuroradiologists was recorded. The degree of confidence in the report of basal and angio-CT scans was assessed before and after CTP visualization. Statistical analysis included logistic regression and F1 score assessments to evaluate the predictive accuracy of AI-generated CTP maps.

Results: The study found that CTP maps derived from AI software were reportable in 65.2% of cases without artifacts and improved to 87.9% reportable cases when reviewed by neuroradiologists. Key predictive factors for artifact-free CTP maps included motion parameters and the timing of contrast peak distances. There was a significant shift to higher confidence scores of the angiographic phase of CT after the results of CTP.

Conclusions: Neuroradiologists play an indispensable role in improving the reliability of CTP by interpreting and correcting AI-processed maps.

Background and purpose: CTA is a commonly used study in the setting of strangulation-related injuries due to the concern for blunt cerebrovascular injuries (BCVIs). However, patients can present with a variety of histories and symptoms, which can make screening for at-risk populations in this group difficult. Therefore, this study aimed to identify risk factors and rates of BCVIs in the setting of strangulation-related injuries to guide imaging strategies.

Materials and methods: We conducted a retrospective review of 138 patients who presented with strangulation and underwent a CTA in a level II trauma center between 2019 and 2023. Patients with CTAs were identified by using a structured search of radiology reports. Clinical, demographic, and imaging findings were analyzed to assess for potential risks of BCVIs.

Results: Among 138 patients, only 1 of 138 was found positive for a BCVI, and this patient was discharged with no complications from the injury. Two of 138 were positive for an acute cervical fracture. A variety of symptoms/findings included neck pain, loss of consciousness, bruising, tenderness, and a sore throat.

Conclusions: In patients presenting with strangulation, the rates of BCVI and even acute cervical fractures were low. These findings suggest that broad use of CTAs in this setting may be of low diagnostic value. Further studies are needed to find more relevant symptoms and signs that could suggest the need for CTAs in this patient population to best optimize diagnosis and patient safety.