Can AI distinguish a bone radiograph from photos of flowers or cars? Evaluation of bone age deep learning model on inappropriate data inputs.

OBJECTIVE: To evaluate the behavior of a publicly available deep convolutional neural network (DCNN) bone age algorithm when presented with inappropriate data inputs in both radiological and non-radiological domains. METHODS: We evaluated a publicly available DCNN-based bone age application. The DCNN was trained on 12,612 pediatric hand radiographs and won the 2017 RSNA Pediatric Bone Age Challenge (concordance of 0.991 with radiologist ground-truth). We used the application to analyze 50 left-hand radiographs (appropriate data inputs) and seven classes of inappropriate data inputs in radiological (i.e., chest radiographs) and non-radiological (i.e., image of street numbers) domains. For each image, we noted if (1) the application distinguished between appropriate and inappropriate data inputs and (2) inference time per image. Mean inference times were compared using ANOVA. RESULTS: The 16Bit Bone Age application calculated bone age for all pediatric hand radiographs with mean inference time of 1.1 s. The application did not distinguish between pediatric hand radiographs and inappropriate image types, including radiological and non-radiological domains. The application inappropriately calculated bone age for all inappropriate image types, with mean inference time of 1.1 s for all categories (p = 1). CONCLUSION: A publicly available DCNN-based bone age application failed to distinguish between appropriate and inappropriate data inputs and calculated bone age for inappropriate images. The awareness of inappropriate outputs based on inappropriate DCNN input is important if tasks such as bone age determination are automated, emphasizing the need for appropriate oversight at the data input and verification stage to avoid unrecognized erroneous results.

End-to-end artificial intelligence platform for the management of large vessel occlusions: A preliminary study.

OBJECTIVES: In this study, we developed a deep learning pipeline that detects large vessel occlusion (LVO) and predicts functional outcome based on computed tomography angiography (CTA) images to improve the management of the LVO patients. METHODS: A series identifier picked out 8650 LVO-protocoled studies from 2015 to 2019 at Rhode Island Hospital with an identified thin axial series that served as the data pool. Data were annotated into 2 classes: 1021 LVOs and 7629 normal. The Inception-V1 I3D architecture was applied for LVO detection. For outcome prediction, 323 patients undergoing thrombectomy were selected. A 3D convolution neural network (CNN) was used for outcome prediction (30-day mRS) with CTA volumes and embedded pre-treatment variables as inputs. RESULT: For LVO-detection model, CTAs from 8,650 patients (median age 68 years, interquartile range (IQR): 58-81; 3934 females) were analyzed. The cross-validated AUC for LVO vs. not was 0.74 (95% CI: 0.72-0.75). For the mRS classification model, CTAs from 323 patients (median age 75 years, IQR: 63-84; 164 females) were analyzed. The algorithm achieved a test AUC of 0.82 (95% CI: 0.79-0.84), sensitivity of 89%, and specificity 66%. The two models were then integrated with hospital infrastructure where CTA was collected in real-time and processed by the model. If LVO was detected, interventionists were notified and provided with predicted clinical outcome information. CONCLUSION: 3D CNNs based on CTA were effective in selecting LVO and predicting LVO mechanical thrombectomy short-term prognosis. End-to-end AI platform allows users to receive immediate prognosis prediction and facilitates clinical workflow.

Detection of Optic Disc Abnormalities in Color Fundus Photographs Using Deep Learning.

BACKGROUND: To date, deep learning-based detection of optic disc abnormalities in color fundus photographs has mostly been limited to the field of glaucoma. However, many life-threatening systemic and neurological conditions can manifest as optic disc abnormalities. In this study, we aimed to extend the application of deep learning (DL) in optic disc analyses to detect a spectrum of nonglaucomatous optic neuropathies. METHODS: Using transfer learning, we trained a ResNet-152 deep convolutional neural network (DCNN) to distinguish between normal and abnormal optic discs in color fundus photographs (CFPs). Our training data set included 944 deidentified CFPs (abnormal 364; normal 580). Our testing data set included 151 deidentified CFPs (abnormal 71; normal 80). Both the training and testing data sets contained a wide range of optic disc abnormalities, including but not limited to ischemic optic neuropathy, atrophy, compressive optic neuropathy, hereditary optic neuropathy, hypoplasia, papilledema, and toxic optic neuropathy. The standard measures of performance (sensitivity, specificity, and area under the curve of the receiver operating characteristic curve (AUC-ROC)) were used for evaluation. RESULTS: During the 10-fold cross-validation test, our DCNN for distinguishing between normal and abnormal optic discs achieved the following mean performance: AUC-ROC 0.99 (95 CI: 0.98-0.99), sensitivity 94% (95 CI: 91%-97%), and specificity 96% (95 CI: 93%-99%). When evaluated against the external testing data set, our model achieved the following mean performance: AUC-ROC 0.87, sensitivity 90%, and specificity 69%. CONCLUSION: In summary, we have developed a deep learning algorithm that is capable of detecting a spectrum of optic disc abnormalities in color fundus photographs, with a focus on neuro-ophthalmological etiologies. As the next step, we plan to validate our algorithm prospectively as a focused screening tool in the emergency department, which if successful could be beneficial because current practice pattern and training predict a shortage of neuro-ophthalmologists and ophthalmologists in general in the near future.

Radiology "forensics": determination of age and sex from chest radiographs using deep learning.

PURPOSE: To develop and test the performance of deep convolutional neural networks (DCNNs) for automated classification of age and sex on chest radiographs (CXR). METHODS: We obtained 112,120 frontal CXRs from the NIH ChestX-ray14 database performed in 48,780 females (44%) and 63,340 males (56%) ranging from 1 to 95 years old. The dataset was split into training (70%), validation (10%), and test (20%) datasets, and used to fine-tune ResNet-18 DCNNs pretrained on ImageNet for (1) determination of sex (using entire dataset and only pediatric CXRs); (2) determination of age < 18 years old or ≥ 18 years old (using entire dataset); and (3) determination of age < 11 years old or 11-18 years old (using only pediatric CXRs). External testing was performed on 662 CXRs from China. Area under the receiver operating characteristic curve (AUC) was used to evaluate DCNN test performance. RESULTS: DCNNs trained to determine sex on the entire dataset and pediatric CXRs only had AUCs of 1.0 and 0.91, respectively (p < 0.0001). DCNNs trained to determine age < or ≥ 18 years old and < 11 vs. 11-18 years old had AUCs of 0.99 and 0.96 (p < 0.0001), respectively. External testing showed AUC of 0.98 for sex (p = 0.01) and 0.91 for determining age < or ≥ 18 years old (p < 0.001). CONCLUSION: DCNNs can accurately predict sex from CXRs and distinguish between adult and pediatric patients in both American and Chinese populations. The ability to glean demographic information from CXRs may aid forensic investigations, as well as help identify novel anatomic landmarks for sex and age.

Practical value of three-dimensional high resolution magnetic resonance Vessel Wall imaging in identifying suspicious intracranial vertebrobasilar dissecting aneurysms.

BACKGROUND: Direct evidence of intimal flaps, double lumen and intramural haematomas (IMH) is difficult to detect on conventional angiography in most intracranial vertebrobasilar dissecting aneurysms (VBDAs). Our purpose was to assess the value of three-dimensional high-resolution magnetic resonance vessel wall imaging (3D HRMR VWI) for identifying VBDAs. METHODS: Between August 2013 and January 2016, consecutive patients with suspicious VBDAs were prospectively enrolled to undergo catheter angiography and VWI (pre- and post-contrast). The lesion was diagnosed as definite VBDA when presenting direct signs of dissection; as possible when only presenting indirect signs; and as segmental ectasia when there was local dilation and wall thickness similar to adjacent normal artery's without mural thrombosis. RESULTS: Twenty-one patients with 27 lesions suspicious for VBDAs were finally included. Based on findings of VWI and catheter angiography, definite VBDA was diagnosed in 25 and 7 lesions (92.6%, vs 25.9%, p <  0.001), respectively; possible VBDA in 0 and 20 (0 vs 74.1%), respectively; and segmental ectasia in 2 and 0 (7.4% vs 0%), respectively. On VWI and catheter angiography, intimal flap was detected in 21 and 7 lesions (77.8% vs 25.9%, p = 0.001), respectively; double lumen sign in 18 and 7 (66.7% vs 25.9%, p = 0.003), respectively; and IMH sign in 14 and 0 (51.9% vs 0), respectively. CONCLUSIONS: 3D HRMR VWI could detect direct dissection signs more frequently than catheter angiography. This may help obtain definite diagnosis of intracranial VBDAs, and allow accurate differentiation between dissecting aneurysm and segmental ectasia as well. Further prospective study with larger sample was required to investigate the superiority of HRMR VWI for definite diagnosis of intracranial VBDAs than catheter angiography.

Deep Learning and Transfer Learning for Optic Disc Laterality Detection: Implications for Machine Learning in Neuro-Ophthalmology.

BACKGROUND: Deep learning (DL) has demonstrated human expert levels of performance for medical image classification in a wide array of medical fields, including ophthalmology. In this article, we present the results of our DL system designed to determine optic disc laterality, right eye vs left eye, in the presence of both normal and abnormal optic discs. METHODS: Using transfer learning, we modified the ResNet-152 deep convolutional neural network (DCNN), pretrained on ImageNet, to determine the optic disc laterality. After a 5-fold cross-validation, we generated receiver operating characteristic curves and corresponding area under the curve (AUC) values to evaluate performance. The data set consisted of 576 color fundus photographs (51% right and 49% left). Both 30° photographs centered on the optic disc (63%) and photographs with varying degree of optic disc centration and/or wider field of view (37%) were included. Both normal (27%) and abnormal (73%) optic discs were included. Various neuro-ophthalmological diseases were represented, such as, but not limited to, atrophy, anterior ischemic optic neuropathy, hypoplasia, and papilledema. RESULTS: Using 5-fold cross-validation (70% training; 10% validation; 20% testing), our DCNN for classifying right vs left optic disc achieved an average AUC of 0.999 (±0.002) with optimal threshold values, yielding an average accuracy of 98.78% (±1.52%), sensitivity of 98.60% (±1.72%), and specificity of 98.97% (±1.38%). When tested against a separate data set for external validation, our 5-fold cross-validation model achieved the following average performance: AUC 0.996 (±0.005), accuracy 97.2% (±2.0%), sensitivity 96.4% (±4.3%), and specificity 98.0% (±2.2%). CONCLUSIONS: Small data sets can be used to develop high-performing DL systems for semantic labeling of neuro-ophthalmology images, specifically in distinguishing between right and left optic discs, even in the presence of neuro-ophthalmological pathologies. Although this may seem like an elementary task, this study demonstrates the power of transfer learning and provides an example of a DCNN that can help curate large medical image databases for machine-learning purposes and facilitate ophthalmologist workflow by automatically labeling images according to laterality.

Readability of Patient Education Materials From RadiologyInfo.org: Has There Been Progress Over the Past 5 Years?

OBJECTIVE. As patients increasingly turn to the Internet for healthcare information, it is imperative that patient educational materials be written at an appropriate readability level. Although RadiologyInfo.org, a patient education library sponsored by the American College of Radiology (ACR) and Radiological Society of North America, was shown in 2012 to be written at levels too high for the average patient to adequately comprehend, it is unclear if there has been progress made in the past 5 years. The purpose of this study was to provide a 5-year update on the readability of patient education materials from RadiologyInfo.org. MATERIALS AND METHODS. All patient education articles available in 2017 from the ACR and RSNA-sponsored RadiologyInfo.org patient education library were reviewed. We assessed each article for readability using 6 quantitative readability scales: the Flesch-Kincaid (FK) grade level, Flesch Reading Ease, Gunnin-Fog Index, Coleman-Liau Index, Automated Readability Index, and the Simple Measure of Gobbledygook (SMOG). The number of articles with readability ≤ the 8th grade level (average reading ability of US adults) and the 6th-grade level (NIH-recommended level for patient materials) were determined. RESULTS. 131 patient education articles were reviewed. The mean readability grade level was greater than the 11th grade reading level for all readability scales. None of the articles were written at less than the 8th-grade or the 6th-grade levels. CONCLUSION. Although there has been an increasing awareness of the issue of readability of patient educational materials within the radiological community, the patient educational materials within the ACR and RSNA-sponsored RadiologyInfo.org website are still written at levels too high for the average patient. Future efforts should be made to improve the readability of those patient education materials.

Automated semantic labeling of pediatric musculoskeletal radiographs using deep learning.

BACKGROUND: An automated method for identifying the anatomical region of an image independent of metadata labels could improve radiologist workflow (e.g., automated hanging protocols) and help facilitate the automated curation of large medical imaging data sets for machine learning purposes. Deep learning is a potential tool for this purpose. OBJECTIVE: To develop and test the performance of deep convolutional neural networks (DCNN) for the automated classification of pediatric musculoskeletal radiographs by anatomical area. MATERIALS AND METHODS: We utilized a database of 250 pediatric bone radiographs (50 each of the shoulder, elbow, hand, pelvis and knee) to train 5 DCNNs, one to detect each anatomical region amongst the others, based on ResNet-18 pretrained on ImageNet (transfer learning). For each DCNN, the radiographs were randomly split into training (64%), validation (12%) and test (24%) data sets. The training and validation data sets were augmented 30 times using standard preprocessing methods. We also tested our DCNNs on a separate test set of 100 radiographs from a single institution. Receiver operating characteristics (ROC) with area under the curve (AUC) were used to evaluate DCNN performances. RESULTS: All five DCNN trained for classification of the radiographs into anatomical region achieved ROC AUC of 1, respectively, for both test sets. Classification of the test radiographs occurred at a rate of 33 radiographs per s. CONCLUSION: DCNNs trained on a small set of images with 30 times augmentation through standard processing techniques are able to automatically classify pediatric musculoskeletal radiographs into anatomical region with near-perfect to perfect accuracy at superhuman speeds. This concept may apply to other body parts and radiographic views with the potential to create an all-encompassing semantic-labeling DCNN.

Deep Learning Method for Automated Classification of Anteroposterior and Posteroanterior Chest Radiographs.

Ensuring correct radiograph view labeling is important for machine learning algorithm development and quality control of studies obtained from multiple facilities. The purpose of this study was to develop and test the performance of a deep convolutional neural network (DCNN) for the automated classification of frontal chest radiographs (CXRs) into anteroposterior (AP) or posteroanterior (PA) views. We obtained 112,120 CXRs from the NIH ChestX-ray14 database, a publicly available CXR database performed in adult (106,179 (95%)) and pediatric (5941 (5%)) patients consisting of 44,810 (40%) AP and 67,310 (60%) PA views. CXRs were used to train, validate, and test the ResNet-18 DCNN for classification of radiographs into anteroposterior and posteroanterior views. A second DCNN was developed in the same manner using only the pediatric CXRs (2885 (49%) AP and 3056 (51%) PA). Receiver operating characteristic (ROC) curves with area under the curve (AUC) and standard diagnostic measures were used to evaluate the DCNN's performance on the test dataset. The DCNNs trained on the entire CXR dataset and pediatric CXR dataset had AUCs of 1.0 and 0.997, respectively, and accuracy of 99.6% and 98%, respectively, for distinguishing between AP and PA CXR. Sensitivity and specificity were 99.6% and 99.5%, respectively, for the DCNN trained on the entire dataset and 98% for both sensitivity and specificity for the DCNN trained on the pediatric dataset. The observed difference in performance between the two algorithms was not statistically significant (p = 0.17). Our DCNNs have high accuracy for classifying AP/PA orientation of frontal CXRs, with only slight reduction in performance when the training dataset was reduced by 95%. Rapid classification of CXRs by the DCNN can facilitate annotation of large image datasets for machine learning and quality assurance purposes.

Deep-Learning-Based Semantic Labeling for 2D Mammography and Comparison of Complexity for Machine Learning Tasks.

Machine learning has several potential uses in medical imaging for semantic labeling of images to improve radiologist workflow and to triage studies for review. The purpose of this study was to (1) develop deep convolutional neural networks (DCNNs) for automated classification of 2D mammography views, determination of breast laterality, and assessment and of breast tissue density; and (2) compare the performance of DCNNs on these tasks of varying complexity to each other. We obtained 3034 2D-mammographic images from the Digital Database for Screening Mammography, annotated with mammographic view, image laterality, and breast tissue density. These images were used to train a DCNN to classify images for these three tasks. The DCNN trained to classify mammographic view achieved receiver-operating-characteristic (ROC) area under the curve (AUC) of 1. The DCNN trained to classify breast image laterality initially misclassified right and left breasts (AUC 0.75); however, after discontinuing horizontal flips during data augmentation, AUC improved to 0.93 (p < 0.0001). Breast density classification proved more difficult, with the DCNN achieving 68% accuracy. Automated semantic labeling of 2D mammography is feasible using DCNNs and can be performed with small datasets. However, automated classification of differences in breast density is more difficult, likely requiring larger datasets.

Case Report of Debilitating Headaches and a Coexisting Ophthalmic Artery Aneurysm: An Indication for Treatment?

OBJECTIVE: We present a case of a patient who had severe unilateral headaches related to a small, unruptured ophthalmic artery aneurysm, who experienced complete headache cessation following endovascular coiling. BACKGROUND: Small unruptured intracranial aneurysms are generally managed and followed conservatively due to minimal risk of rupture. Headaches are frequently reported in patients with intracranial aneurysms, but these aneurysms are typically considered incidental and unrelated, given the undefined association between headaches and most aneurysms. CONCLUSION: There may be some unruptured intracranial aneurysms that can cause intractable headaches and warrant interventional treatment. Future prospective studies are needed that compare pre- and post-procedure headache character and diagnosis, aneurysm characteristics such as size, location, orientation, and shape, type of aneurysm repair with materials used, and other potential risk factors for worsening post-procedure headache in order to better predict headache association to aneurysms, as well as outcomes following endovascular aneurysm treatment.

Predictors of infarct growth after endovascular therapy for acute ischemic stroke.

BACKGROUND: Intra-arterial (IA) thrombectomy for acute ischemic stroke has an excellent recanalization rate but variable outcomes. The core infarct also grows at a variable rate despite recanalization. We aim to study the factors that are associated with infarct growth after IA therapy. METHODS: We reviewed the hyperacute ischemic stroke imaging database at Cleveland Clinic for those undergoing endovascular thrombectomy of anterior circulation from 2009 to 2012. Patients with both pretreatment and follow-up magnetic resonance imaging were included. Seventy-six patients were stratified into quartiles by infarct volume growth from initial to follow-up diffusion-weighted imaging (DWI) measure by a region of interest demarcation. RESULTS: The median infarct growth of each quartile was .6 cm(3) (no-growth group), 13.8, 37, and 160.2 cm(3) (large-growth group). Pretreatment stroke severity was comparable among groups. Compared with the no-growth group, the large-growth group had larger initial infarct defined by computed tomography (CT) Alberta Stroke Program Early CT score (median 10 versus 8, P = .032) and DWI volume (mean 13.8 versus 29.2 cm(3), P = .034), lack of full collateral vessels on CT angiography (36.8% versus 0%, P = .003), and a lower recanalization rate (thrombolysis in cerebral infarction ≥2b, P = .044). The increase in infarct growth is associated with decrease in favorable outcomes defined by a modified Rankin Scale score of 0-2 at 30 days: 57.9%, 42.1%, 21.1%, and 5.3%, respectively (P < .001). DWI reversal was observed in 11 of 76 patients, translating to 82% favorable outcome. CONCLUSIONS: Infarct evolution after endovascular thrombectomy is associated with an outcome. DWI reversal or no growth translated to a favorable outcome. Small initial ischemic core, good collateral support, and better recanalization grades predict the smaller infarct growth and favorable outcome after endovascular thrombectomy.

Addition of hyperacute MRI AIDS in patient selection, decreasing the use of endovascular stroke therapy.

BACKGROUND AND PURPOSE: The failure of recent trials to show the effectiveness of acute endovascular stroke therapy (EST) may be because of inadequate patient selection. We implemented a protocol to perform pretreatment MRI on patients with large-vessel occlusion eligible for EST to aid in patient selection. METHODS: We retrospectively identified patients with large-vessel occlusion considered for EST from January 2008 to August 2012. Patients before April 30, 2010, were selected based on computed tomography/computed tomography angiography (prehyperacute protocol), whereas patients on or after April 30, 2010, were selected based on computed tomography/computed tomography angiography and MRI (hyperacute MRI protocol). Demographic, clinical features, and outcomes were collected. Univariate and multivariate analyses were performed. RESULTS: We identified 267 patients: 88 patients in prehyperacute MRI period and 179 in hyperacute MRI period. Fewer patients evaluated in the hyperacute MRI period received EST (85 of 88, 96.6% versus 92 of 179, 51.7%; P<0.05). The hyperacute-MRI group had a more favorable outcome of a modified Rankin scale 0 to 2 at 30 days as a group (6 of 66, 9.1% versus 33 of 140, 23.6%; P=0.01), and when taken for EST (6 of 63, 9.5% versus 17 of 71, 23.9%; P=0.03). On adjusted multivariate analysis, the EST in the hyperacute MRI period was associated with a more favorable outcome (odds ratio, 3.4; 95% confidence interval, 1.1-10.6; P=0.03) and reduced mortality rate (odds ratio, 0.16; 95% confidence interval, 0.03-0.37; P<0.001). CONCLUSIONS: Implementation of hyperacute MRI protocol decreases the number of endovascular stroke interventions by half. Further investigation of MRI use for patient selection is warranted.

Direct superficial temporal artery access for middle meningeal artery embolization.

Middle meningeal artery embolization has become an important option in the management of subdural hemorrhages with multiple prospective studies demonstrating efficacy and randomized controlled trial data on the way. Access to the middle meningeal artery is usually achieved via the external carotid artery to the internal maxillary artery, then the middle meningeal artery. We report a case where a patient with symptomatic left-sided chronic subdural hemorrhage also had an external carotid artery occlusion. Direct puncture of the superficial temporal artery allowed retrograde access to the internal maxillary artery and thus the middle meningeal artery. Successful embolization of the vessel with 1:9 nBCA was performed with near total resorption of the subdural collection by 1 month postprocedure.

Scalpellum occulta: A rare case of dorsal spinal arachnoid web without the scalpel sign.

Spinal arachnoid web is a rare condition characterized by extramedullary bands of arachnoid tissue at the level of the dorsal thoracic spinal cord that may lead to progressive, permanent neurological deficits. To date, this condition has been radiographically characterized by a scalpel sign, which has been pathognomonic in all reported cases of spinal arachnoid webs. In this case, we report the first known patient with confirmed spinal arachnoid web without radiographic evidence of the scalpel sign. In reporting our finding, we encourage a higher clinical suspicion for spinal arachnoid web in patients presenting with progressive thoracic myelopathy following trauma, and radiographic evidence of ventrally displaced spinal cord and turbulent cerebrospinal fluid flow, even in the absence of a scalpel sign.

The Promise, Mystery, and Perils of Stenting for Symptomatic Internal Jugular Vein Stenosis: A Case Series.

BACKGROUND AND OBJECTIVES: Cerebral venous outflow disorders (CVDs) secondary to internal jugular vein (IJV) stenosis are becoming an increasingly recognized cause of significant cognitive and functional impairment in patients. There are little published data on IJV stenting for this condition. This study aims to report on procedural success. METHODS: A single-center retrospective analysis was performed on patients with CVD that underwent IJV stenting procedures. RESULTS: From 2019 to 2023, 29 patients with CVD underwent a total of 33 IJV stenting procedures. Most patients (20; 69%) had an underlying connective tissue disorder diagnosis. The mean age of the included patients was 36.3 years (SD 12.4), 24 were female (82.8%), and all were Caucasian except for 2 patients (27; 93.0%). Twenty-eight procedures (85%) involved isolated IJV stenting under conscious sedation, whereas 5 procedures (15%) involved IJV stenting and concomitant transverse sinus stenting under general anesthesia. Thirteen (39%) patients underwent IJV stenting after open IJV decompression and styloidectomy. Three patients had stents placed for stenosis below the C1 tubercle, one of which was for carotid compression. Periprocedural complications occurred in 11 (33%), including intracardiac stent migration in 1 patient, temporary shoulder pain/weakness in 5 (15%), and persistent and severe shoulder pain/weakness in 2 patients (6%). Approximately 75% of patients demonstrated improvement after stenting although only 12 patients (36%) had durable improvement over a mean follow-up of 4.5 months (range 6 weeks-3.5 years). CONCLUSION: Our experience, along with early published studies, suggests that there is significant promise to IJV revascularization techniques in these patients; however, stenting carries a high complication rate, and symptom recurrence is common. Most neurointerventionalists should not be performing IJV stenting unless they have experience with these patients and understand technical nuances (stent sizing, anatomy, patient selection), which can maximize benefit and minimize risk.

Differential Assessment of Internal Jugular Vein Stenosis in Patients Undergoing CT and MRI with Contrast.

OBJECTIVE: Internal Jugular Vein Stenosis (IJVS) is hypothesized to play a role in the pathogenesis of diverse neurological diseases. We sought to evaluate differences in IJVS assessment between CT and MRI in a retrospective patient cohort. METHODS: We included consecutive patients who had both MRI of the brain and CT of the head and neck with contrast from 1 June 2021 to 30 June 2022 within the same admission. The degree of IJVS was categorized into five grades (0-IV). RESULTS: A total of 35 patients with a total of 70 internal jugular (IJ) veins were included in our analysis. There was fair intermodality agreement in stenosis grades (κ = 0.220, 95% C.I. = [0.029, 0.410]), though categorical stenosis grades were significantly discordant between imaging modalities, with higher grades more frequent in MRI (χ2 = 27.378, p = 0.002). On CT-based imaging, Grade III or IV stenoses were noted in 17/70 (24.2%) IJs, whereas on MRI-based imaging, Grade III or IV stenoses were found in 40/70 (57.1%) IJs. Among veins with Grade I-IV IJVS, MRI stenosis estimates were significantly higher than CT stenosis estimates (77.0%, 95% C.I. [35.9-55.2%] vs. 45.6%, 95% C.I. [35.9-55.2%], p < 0.001). CONCLUSION: MRI with contrast overestimates the degree of IJVS compared to CT with contrast. Consideration of this discrepancy should be considered in diagnosis and treatment planning in patients with potential IJVS-related symptoms.