Use of a commercial artificial intelligence-based mammography analysis software for improving breast ultrasound interpretations

OBJECTIVES

To evaluate the use of a commercial artificial intelligence (AI)-based mammography analysis software for improving the interpretations of breast ultrasound (US)-detected lesions.

METHODS

A retrospective analysis was performed on 1109 breasts that underwent both mammography and US-guided breast biopsy. The AI software processed mammograms and provided an AI score ranging from 0 to 100 for each breast, indicating the likelihood of malignancy. The performance of the AI score in differentiating mammograms with benign outcomes from those revealing cancers following US-guided breast biopsy was evaluated. In addition, prediction models for benign outcomes were constructed based on clinical and imaging characteristics with and without AI scores, using logistic regression analysis.

RESULTS

The AI software had an area under the receiver operating characteristics curve (AUROC) of 0.79 (95% CI, 0.79-0.82) in differentiating between benign and cancer cases. The prediction models that did not include AI scores (non-AI model), only used AI scores (AI-only model), and included AI scores (integrated model) had AUROCs of 0.79 (95% CI, 0.75-0.83), 0.78 (95% CI, 0.74-0.82), and 0.85 (95% CI, 0.81-0.88) in the development cohort, and 0.75 (95% CI, 0.68-0.81), 0.82 (95% CI, 0.76-0.88), and 0.84 (95% CI, 0.79-0.90) in the validation cohort, respectively. The integrated model outperformed the non-AI model in the development and validation cohorts (p < 0.001 for both).

CONCLUSION

The commercial AI-based mammography analysis software could be a valuable adjunct to clinical decision-making for managing US-detected breast lesions.

CLINICAL RELEVANCE STATEMENT

The commercial AI-based mammography analysis software could potentially reduce unnecessary biopsies and improve patient outcomes.

KEY POINTS

• Breast US has high rates of false-positive interpretations. • A commercial AI-based mammography analysis software could distinguish mammograms having benign outcomes from those revealing cancers after US-guided breast biopsy. • A commercial AI-based mammography analysis software may improve interpretations for breast US-detected lesions.

Comparative Performance of Susceptibility Map-Weighted MRI According to the Acquisition Planes in the Diagnosis of Neurodegenerative Parkinsonism

OBJECTIVE

To evaluate the diagnostic performance of susceptibility map-weighted imaging (SMwI) taken in different acquisition planes for discriminating patients with neurodegenerative parkinsonism from those without.

MATERIALS AND METHODS

This retrospective, observational, single-institution study enrolled consecutive patients who visited movement disorder clinics and underwent brain MRI and 18F-FP-CIT PET between September 2021 and December 2021. SMwI images were acquired in both the oblique (perpendicular to the midbrain) and the anterior commissure-posterior commissure (AC-PC) planes. Hyperintensity in the substantia nigra was determined by two neuroradiologists. 18F-FP-CIT PET was used as the reference standard. Inter-rater agreement was assessed using Cohen's kappa coefficient. The diagnostic performance of SMwI in the two planes was analyzed separately for the right and left substantia nigra. Multivariable logistic regression analysis with generalized estimating equations was applied to compare the diagnostic performance of the two planes.

RESULTS

In total, 194 patients were included, of whom 105 and 103 had positive results on 18F-FP-CIT PET in the left and right substantia nigra, respectively. Good inter-rater agreement in the oblique (κ = 0.772/0.658 for left/right) and AC-PC planes (0.730/0.741 for left/right) was confirmed. The pooled sensitivities for two readers were 86.4% (178/206, left) and 83.3% (175/210, right) in the oblique plane and 87.4% (180/206, left) and 87.6% (184/210, right) in the AC-PC plane. The pooled specificities for two readers were 83.5% (152/182, left) and 82.0% (146/178, right) in the oblique plane, and 83.5% (152/182, left) and 86.0% (153/178, right) in the AC-PC plane. There were no significant differences in the diagnostic performance between the two planes (P > 0.05).

CONCLUSION

There are no significant difference in the diagnostic performance of SMwI performed in the oblique and AC-PC plane in discriminating patients with parkinsonism from those without. This finding affirms that each institution may choose the imaging plane for SMwI according to their clinical settings.

Development of statistical auto-segmentation method for diffusion restriction gray matter lesions in patients with newly diagnosed sporadic Creutzfeldt-Jakob disease

Quantification of diffusion restriction lesions in sporadic Creutzfeldt-Jakob disease (sCJD) may provide information of the disease burden. We aim to develop an automatic segmentation model for sCJD and to evaluate the volume of disease extent as a prognostic marker for overall survival. Fifty-six patients (mean age ± SD, 61.2 ± 9.9 years) were included from February 2000 to July 2020. A threshold-based segmentation was used to obtain abnormal signal intensity masks. Segmented volumes were compared with the visual grade. The Dice similarity coefficient was calculated to measure the similarity between the automatic vs. manual segmentation. Cox proportional hazards regression analysis was performed to evaluate the volume of disease extent as a prognostic marker. The automatic segmentation showed good correlation with the visual grading. The cortical lesion volumes significantly increased as the visual grade aggravated (extensive: 112.9 ± 73.2; moderate: 45.4 ± 30.4; minimal involvement: 29.6 ± 18.1 mm3) (P < 0.001). The deep gray matter lesion volumes were significantly higher for positive than for negative involvement of the deep gray matter (5.6 ± 4.6 mm3 vs. 1.0 ± 1.3 mm3, P < 0.001). The mean Dice similarity coefficients were 0.90 and 0.94 for cortical and deep gray matter lesions, respectively. However, the volume of disease extent was not associated with worse overall survival (cortical extent: P = 0.07; deep gray matter extent: P = 0.12).

Improved diagnostic performance of susceptibility-weighted imaging with compressed sensing-sensitivity encoding and neuromelanin-sensitive MRI for Parkinson's disease and atypical Parkinsonism

AIM

To verify the diagnostic performance of the loss of nigrosome-1 on susceptibility-weighted imaging (SWI) with compressed sensing-sensitivity encoding (CS-SENSE) and neuromelanin on neuromelanin-sensitive (NM) magnetic resonance imaging (MRI) for the diagnosis of Parkinson's disease (PD) and atypical Parkinsonism.

MATERIALS AND METHODS

A total of 195 patients who underwent MRI between October 2019 and February 2020, including SWI, with or without CS-SENSE, and NM-MRI, were reviewed retrospectively. Two neuroradiologists assessed the loss of nigrosome-1 on SWI and neuromelanin on the NM-MRI. The result of N-3-fluoropropyl-2-beta-carbomethoxy-3-beta-(4-iodophenyl) nortropane positron-emission tomography (PET) was set as the reference standard.

RESULTS

When CS-SENSE was applied for nigrosome-1 imaging on SWI, the non-diagnostic scan rate was lowered significantly from 19.3% (17/88) to 5.6% (6/107; p=0.004). Diagnosis of PD and atypical Parkinsonism based on the loss of nigrosome-1 on SWI and based on NM-MRI showed good diagnostic value (area under the curve [AUC] 0.821, 95% confidence interval [CI] = 0.755-0.875: AUC 0.832, 95% CI = 0.771-0.882, respectively) with a substantial inter-reader agreement (κ = 0.791 and 0.681, respectively). Combined SWI and neuromelanin had a similar discriminatory ability (AUC 0.830, 95% CI = 0.770-0.880). Similarly, the diagnosis of PD was excellent.

CONCLUSIONS

CS-SENSE may add value to the diagnostic capability of nigrosome-1 on SWI to reduce the nondiagnostic scan rates. Furthermore, loss of nigrosome-1 on SWI or volume loss of neuromelanin on NM-MRI may be helpful for diagnosing PD.

Relationship between disproportionately enlarged subarachnoid-space hydrocephalus and white matter tract integrity in normal pressure hydrocephalus

Normal pressure hydrocephalus (NPH) patients had altered white matter tract integrities on diffusion tensor imaging (DTI). Previous studies suggested disproportionately enlarged subarachnoid space hydrocephalus (DESH) as a prognostic sign of NPH. We examined DTI indices in NPH subgroups by DESH severity and clinical symptoms. This retrospective case-control study included 33 NPH patients and 33 age-, sex-, and education-matched controls. The NPH grading scales (0-12) were used to rate neurological symptoms. Patients with NPH were categorized into two subgroups, high-DESH and low-DESH groups, by the average value of the DESH scale. DTI indices, including fractional anisotropy, were compared across 14 regions of interest (ROIs). The high-DESH group had increased axial diffusivity in the lateral side of corona radiata (1.43 ± 0.25 vs. 1.72 ± 0.25, p = 0.04), and showed decreased fractional anisotropy and increased mean, and radial diffusivity in the anterior and lateral sides of corona radiata and the periventricular white matter surrounding the anterior horn of lateral ventricle. In patients with a high NPH grading scale, fractional anisotropy in the white matter surrounding the anterior horn of the lateral ventricle was significantly reduced (0.36 ± 0.08 vs. 0.26 ± 0.06, p = 0.03). These data show that DESH may be a biomarker for DTI-detected microstructural alterations and clinical symptom severity.

Development and validation of an automatic classification algorithm for the diagnosis of Alzheimer's disease using a high-performance interpretable deep learning network

OBJECTIVES

To develop and validate an automatic classification algorithm for diagnosing Alzheimer's disease (AD) or mild cognitive impairment (MCI).

METHODS AND MATERIALS

This study evaluated a high-performance interpretable network algorithm (TabNet) and compared its performance with that of XGBoost, a widely used classifier. Brain segmentation was performed using a commercially approved software. TabNet and XGBoost were trained on the volumes or radiomics features of 102 segmented regions for classifying subjects into AD, MCI, or cognitively normal (CN) groups. The diagnostic performances of the two algorithms were compared using areas under the curves (AUCs). Additionally, 20 deep learning-based AD signature areas were investigated.

RESULTS

Between December 2014 and March 2017, 161 AD, 153 MCI, and 306 CN cases were enrolled. Another 120 AD, 90 MCI, and 141 CN cases were included for the internal validation. Public datasets were used for external validation. TabNet with volume features had an AUC of 0.951 (95% confidence interval [CI], 0.947-0.955) for AD vs CN, which was similar to that of XGBoost (0.953 [95% CI, 0.951-0.955], p = 0.41). External validation revealed the similar performances of two classifiers using volume features (0.871 vs. 0.871, p = 0.86). Likewise, two algorithms showed similar performances with one another in classifying MCI. The addition of radiomics data did not improve the performance of TabNet. TabNet and XGBoost focused on the same 13/20 regions of interest, including the hippocampus, inferior lateral ventricle, and entorhinal cortex.

CONCLUSIONS

TabNet shows high performance in AD classification and detailed interpretation of the selected regions.

CLINICAL RELEVANCE STATEMENT

Using a high-performance interpretable deep learning network, the automatic classification algorithm assisted in accurate Alzheimer's disease detection using 3D T1-weighted brain MRI and detailed interpretation of the selected regions.

KEY POINTS

• MR volumetry data revealed that TabNet had a high diagnostic performance in differentiating Alzheimer's disease (AD) from cognitive normal cases, which was comparable with that of XGBoost. • The addition of radiomics data to the volume data did not improve the diagnostic performance of TabNet. • Both TabNet and XGBoost selected the clinically meaningful regions of interest in AD, including the hippocampus, inferior lateral ventricle, and entorhinal cortex.

Amphiregulin normalizes altered circuit connectivity for social dominance of the CRTC3 knockout mouse

Social hierarchy has a profound impact on social behavior, reward processing, and mental health. Moreover, lower social rank can lead to chronic stress and often more serious problems such as bullying victims of abuse, suicide, or attack to society. However, its underlying mechanisms, particularly their association with glial factors, are largely unknown. In this study, we report that astrocyte-derived amphiregulin plays a critical role in the determination of hierarchical ranks. We found that astrocytes-secreted amphiregulin is directly regulated by cAMP response element-binding (CREB)-regulated transcription coactivator 3 (CRTC3) and CREB. Mice with systemic and astrocyte-specific CRTC3 deficiency exhibited a lower social rank with reduced functional connectivity between the prefrontal cortex, a major social hierarchy center, and the parietal cortex. However, this effect was reversed by astrocyte-specific induction of amphiregulin expression, and the epidermal growth factor domain was critical for this action of amphiregulin. These results provide evidence of the involvement of novel glial factors in the regulation of social dominance and may shed light on the clinical application of amphiregulin in the treatment of various psychiatric disorders.

Bone Age Assessment Using Artificial Intelligence in Korean Pediatric Population: A Comparison of Deep-Learning Models Trained With Healthy Chronological and Greulich-Pyle Ages as Labels

OBJECTIVE

To develop a deep-learning-based bone age prediction model optimized for Korean children and adolescents and evaluate its feasibility by comparing it with a Greulich-Pyle-based deep-learning model.

MATERIALS AND METHODS

A convolutional neural network was trained to predict age according to the bone development shown on a hand radiograph (bone age) using 21036 hand radiographs of Korean children and adolescents without known bone development-affecting diseases/conditions obtained between 1998 and 2019 (median age [interquartile range {IQR}], 9 [7-12] years; male:female, 11794:9242) and their chronological ages as labels (Korean model). We constructed 2 separate external datasets consisting of Korean children and adolescents with healthy bone development (Institution 1: n = 343; median age [IQR], 10 [4-15] years; male: female, 183:160; Institution 2: n = 321; median age [IQR], 9 [5-14] years; male: female, 164:157) to test the model performance. The mean absolute error (MAE), root mean square error (RMSE), and proportions of bone age predictions within 6, 12, 18, and 24 months of the reference age (chronological age) were compared between the Korean model and a commercial model (VUNO Med-BoneAge version 1.1; VUNO) trained with Greulich-Pyle-based age as the label (GP-based model).

RESULTS

Compared with the GP-based model, the Korean model showed a lower RMSE (11.2 vs. 13.8 months; P = 0.004) and MAE (8.2 vs. 10.5 months; P = 0.002), a higher proportion of bone age predictions within 18 months of chronological age (88.3% vs. 82.2%; P = 0.031) for Institution 1, and a lower MAE (9.5 vs. 11.0 months; P = 0.022) and higher proportion of bone age predictions within 6 months (44.5% vs. 36.4%; P = 0.044) for Institution 2.

CONCLUSION

The Korean model trained using the chronological ages of Korean children and adolescents without known bone development-affecting diseases/conditions as labels performed better in bone age assessment than the GP-based model in the Korean pediatric population. Further validation is required to confirm its accuracy.

Detection rate of brain MR and MR angiography for neuroimaging abnormality in patients with newly diagnosed left-sided infective endocarditis

We aimed to investigate the detection rate of brain MR and MR angiography for neuroimaging abnormality in newly diagnosed left-sided infective endocarditis patients with/without neurological symptoms. This retrospective study included consecutive patients with definite or possible left-sided infective endocarditis according to the modified Duke criteria who underwent brain MRI and MR angiography between March 2015 and October 2020. The detection rate for neuroimaging abnormality on MRI was defined as the number of patients with positive brain MRI findings divided by the number of patients with left-sided infective endocarditis. Positive imaging findings included acute ischemic lesions, cerebral microbleeds, hemorrhagic lesions, and infectious aneurysms. In addition, aneurysm rupture rate and median period to aneurysm rupture were evaluated on follow-up studies. A total 115 patients (mean age: 55 years ± 19; 65 men) were included. The detection rate for neuroimaging abnormality was 77% (89/115). The detection rate in patients without neurological symptoms was 70% (56/80). Acute ischemic lesions, cerebral microbleeds, and hemorrhagic lesions including superficial siderosis and intracranial hemorrhage were detected on MRI in 56% (64/115), 57% (66/115), and 20% (23/115) of patients, respectively. In particular, infectious aneurysms were detected on MR angiography in 3% of patients (4/115), but MR angiography in 5 patients (4.3%) was insignificant for infectious aneurysm, which were detected using CT angiography (n = 3) and digital subtraction angiography (n = 2) during follow-up. Among the 9 infectious aneurysm patients, aneurysm rupture occurred in 4 (44%), with a median period of aneurysm rupture of 5 days. The detection rate of brain MRI for neuroimaging abnormality in newly diagnosed left-sided infective endocarditis patients was high (77%), even without neurological symptoms (70%).

Development and validation of a deep learning-based automatic segmentation model for assessing intracranial volume: comparison with NeuroQuant, FreeSurfer, and SynthSeg

Background and purpose

To develop and validate a deep learning-based automatic segmentation model for assessing intracranial volume (ICV) and to compare the accuracy determined by NeuroQuant (NQ), FreeSurfer (FS), and SynthSeg.

Materials and methods

This retrospective study included 60 subjects [30 Alzheimer's disease (AD), 21 mild cognitive impairment (MCI), 9 cognitively normal (CN)] from a single tertiary hospital for the training and validation group (50:10). The test group included 40 subjects (20 AD, 10 MCI, 10 CN) from the ADNI dataset. We propose a robust ICV segmentation model based on the foundational 2D UNet architecture trained with four types of input images (both single and multimodality using scaled or unscaled T1-weighted and T2-FLAIR MR images). To compare with our model, NQ, FS, and SynthSeg were also utilized in the test group. We evaluated the model performance by measuring the Dice similarity coefficient (DSC) and average volume difference.

Results

The single-modality model trained with scaled T1-weighted images showed excellent performance with a DSC of 0.989 ± 0.002 and an average volume difference of 0.46% ± 0.38%. Our multimodality model trained with both unscaled T1-weighted and T2-FLAIR images showed similar performance with a DSC of 0.988 ± 0.002 and an average volume difference of 0.47% ± 0.35%. The overall average volume difference with our model showed relatively higher accuracy than NQ (2.15% ± 1.72%), FS (3.69% ± 2.93%), and SynthSeg (1.88% ± 1.18%). Furthermore, our model outperformed the three others in each subgroup of patients with AD, MCI, and CN subjects.

Conclusion

Our deep learning-based automatic ICV segmentation model showed excellent performance for the automatic evaluation of ICV.

Risk estimation for idiopathic normal-pressure hydrocephalus: development and validation of a brain morphometry-based nomogram

OBJECTIVES

To develop and validate a nomogram based on MRI features for predicting iNPH.

METHODS

Patients aged  ≥ 60 years (clinically diagnosed with iNPH, Parkinson's disease, or Alzheimer's disease or healthy controls) who underwent MRI including three-dimensional T1-weighted volumetric MRI were retrospectively identified from two tertiary referral hospitals (one hospital for derivation set and the other for validation set). Clinical and imaging features for iNPH were assessed. Deep learning-based brain segmentation software was used for 3D volumetry. A prediction model was developed using logistic regression and transformed into a nomogram. The performance of the nomogram was assessed with respect to discrimination and calibration abilities. The nomogram was internally and externally validated.

RESULTS

A total of 452 patients (mean age ± SD, 73.2 ± 6.5 years; 200 men) were evaluated as the derivation set. One hundred eleven and 341 patients were categorized into the iNPH and non-iNPH groups, respectively. In multivariable analysis, high-convexity tightness (odds ratio [OR], 35.1; 95% CI: 4.5, 275.5), callosal angle  < 90° (OR, 12.5; 95% CI: 3.1, 50.0), and normalized lateral ventricle volume (OR, 4.2; 95% CI: 2.7, 6.7) were associated with iNPH. The nomogram combining these three variables showed an area under the curve of 0.995 (95% CI: 0.991, 0.999) in the study sample, 0.994 (95% CI: 0.990, 0.998) in the internal validation sample, and 0.969 (95% CI: 0.940, 0.997) in the external validation sample.

CONCLUSION

A brain morphometry-based nomogram including high-convexity tightness, callosal angle  < 90°, and normalized lateral ventricle volume can help accurately estimate the probability of iNPH.

KEY POINTS

• The nomogram with MRI findings (high-convexity tightness, callosal angle, and normalized lateral ventricle volume) helped in predicting the probability of idiopathic normal-pressure hydrocephalus. • The nomogram may facilitate the prediction of idiopathic normal-pressure hydrocephalus and consequently avoid unnecessary invasive procedures such as the cerebrospinal fluid tap test, drainage test, and cerebrospinal fluid shunt surgery.

Detection rate of contrast-enhanced brain magnetic resonance imaging in patients with cognitive impairment

INTRODUCTION

The number of brain MRI with contrast media performed in patients with cognitive impairment has increased without universal agreement. We aimed to evaluate the detection rate of contrast-enhanced brain MRI in patients with cognitive impairment.

MATERIALS AND METHODS

This single-institution, retrospective study included 4,838 patients who attended outpatient clinics for cognitive impairment evaluation and underwent brain MRI with or without contrast enhancement from December 2015 to February 2020. Patients who tested positive for cognitive impairment were followed-up to confirm whether the result was true-positive and provide follow-up management. Detection rate was defined as the proportion of patients with true-positive results and was compared between groups with and without contrast enhancement. Individual matching in a 1:2 ratio according to age, sex, and year of test was performed.

RESULTS

The overall detection rates of brain MRI with and without contrast media were 4.7% (57/1,203; 95% CI: 3.6%-6.1%) and 1.8% (65/3,635; 95% CI: 1.4%-2.3%), respectively (P<0.001); individual matching demonstrated similar results (4.7% and 1.9%). Among 1,203 patients with contrast media, 3.6% was only detectable with the aid of contrast media. The proportion of patients who underwent follow-up imaging or treatment for the detected lesions were significantly higher in the group with contrast media (2.0% and 0.6%, P < .001).

CONCLUSIONS

Detection rate of brain MRI for lesions only detectable with contrast media in patients with cognitive impairment was not high enough and further study is needed to identify whom would truly benefit with contrast media.

Preliminary Study for Quantitative Assessment of Sacroiliitis Activity Using Bone SPECT/CT: Comparison of Diagnostic Performance of Quantitative Parameters

Purpose

We compared the feasibility of quantitative analysis methods using bone SPECT/CT with those using planar bone scans to assess active sacroiliitis.

Methods

We retrospectively reviewed whole-body bone scans and pelvic bone SPECT/CTs of 8 patients who had clinically confirmed sacroiliitis and enrolled 24 patients without sacroiliitis as references. The volume of interest of each sacroiliac joint, including both the ilium and sacrum, was drawn. Active arthritis zone (AAZ) was defined as the zone of voxels with higher SUV than sacral mean SUV within the VOI of SI joint. Then, the following SPECT/CT quantitative parameters, SUVmax (maximum SUV), SUV50% (mean SUV in highest 50% of SUV), and SUV-AAZ, and the ratio of those values to sacral mean SUV (SUVmax/S, SUV50%/S, SUV-AAZ/S) were calculated. For the planar bone scan, the mean count ratio of SI joint/sacrum (SI/S) was conventionally measured.

Results

Most of the SPECT/CT parameters of the sacroiliitis group were significantly higher than the normal group, whereas SI/S of the planar bone scan was not significantly different between the two groups. In receiver operating characteristic curve analysis, SUV-AAZ/S showed the highest AUC of 0.992, followed by SUV50%/S and SUVmax/S. All ratio parameters of the SPECT/CT showed higher AUC values than the SUV parameters of SI joint or SI/S of the planar scan.

Conclusions

The quantitative analyses of bone SPECT/CT showed better performance in assessing active sacroiliitis than the planar bone scan. SPECT/CT parameters using the ratio of the SI joint to sacrum showed more favorable results than SUV parameters such as SUVmax, SUV50%, and SUV-AAZ.

Detection rate of MR myelography without intrathecal gadolinium in patients with newly diagnosed spontaneous intracranial hypotension

AIM

To evaluate the detection rate of magnetic resonance (MR) myelography without intrathecal gadolinium for cerebrospinal fluid (CSF) leakage in patients with newly diagnosed spontaneous intracranial hypotension (SIH) and to validate a published scoring system for predicting CSF leakage.

MATERIALS AND METHODS

This retrospective, observational, single-institution study included patients with newly diagnosed SIH between March 2015 and April 2021. Patients were included if they (a) had newly diagnosed SIH and (b) underwent initial brain MR imaging and preprocedural MR myelography with two- and three-dimensional turbo spin-echo sequences. Patients who underwent spine surgery or procedures including epidural injection and acupuncture were excluded. The detection rate was defined as the proportion of patients with a true-positive MR myelography result among all patients with confirmed CSF leakage. The interobserver agreement for the MR myelography results between two radiologists was analysed using weighted kappa statistics.

RESULTS

A total of 136 patients (mean age, 48 years; 70 women) with suspected SIH were included. Of these patients, 120 (88%, 120/136) were confirmed to have CSF leakage. Of the patients with confirmed CSF leakage, 90 (75%, 90/120) had epidural fluid collection. The detection rate of MR myelography for CSF leakage was 88% (105/120). The interobserver agreement between the two readers for detecting CSF leakage (κ = 0.76) or epidural fluid collection (κ = 0.76) on MR myelography was high. Among 24 patients with normal brain MR imaging results, 16 had CSF leakage (67%, 16/24).

CONCLUSIONS

Non-invasive MR myelography without intrathecal gadolinium should be considered to detect CSF leakage in patients with suspected SIH.

Deep learning-based diagnosis of Alzheimer's disease using brain magnetic resonance images: an empirical study

The limited accessibility of medical specialists for Alzheimer's disease (AD) can make obtaining an accurate diagnosis in a timely manner challenging and may influence prognosis. We investigated whether VUNO Med-DeepBrain AD (DBAD) using a deep learning algorithm can be employed as a decision support service for the diagnosis of AD. This study included 98 elderly participants aged 60 years or older who visited the Seoul Asan Medical Center and the Korea Veterans Health Service. We administered a standard diagnostic assessment for diagnosing AD. DBAD and three panels of medical experts (ME) diagnosed participants with normal cognition (NC) or AD using T1-weighted magnetic resonance imaging. The accuracy (87.1% for DBAD and 84.3% for ME), sensitivity (93.3% for DBAD and 80.0% for ME), and specificity (85.5% for DBAD and 85.5% for ME) of both DBAD and ME for diagnosing AD were comparable; however, DBAD showed a higher trend in every analysis than ME diagnosis. DBAD may support the clinical decisions of physicians who are not specialized in AD; this may enhance the accessibility of AD diagnosis and treatment.

Diagnostic performance of deep learning-based automatic white matter hyperintensity segmentation for classification of the Fazekas scale and differentiation of subcortical vascular dementia

Purpose

To validate the diagnostic performance of commercially available, deep learning-based automatic white matter hyperintensity (WMH) segmentation algorithm for classifying the grades of the Fazekas scale and differentiating subcortical vascular dementia.

Methods

This retrospective, observational, single-institution study investigated the diagnostic performance of a deep learning-based automatic WMH volume segmentation to classify the grades of the Fazekas scale and differentiate subcortical vascular dementia. The VUNO Med-DeepBrain was used for the WMH segmentation system. The system for segmentation of WMH was designed with convolutional neural networks, in which the input image was comprised of a pre-processed axial FLAIR image, and the output was a segmented WMH mask and its volume. Patients presented with memory complaint between March 2017 and June 2018 were included and were split into training (March 2017–March 2018, n = 596) and internal validation test set (April 2018–June 2018, n = 204).

Results

Optimal cut-off values to categorize WMH volume as normal vs. mild/moderate/severe, normal/mild vs. moderate/severe, and normal/mild/moderate vs. severe were 3.4 mL, 9.6 mL, and 17.1 mL, respectively, and the AUC were 0.921, 0.956 and 0.960, respectively. When differentiating normal/mild vs. moderate/severe using WMH volume in the test set, sensitivity, specificity, and accuracy were 96.4%, 89.9%, and 91.7%, respectively. For distinguishing subcortical vascular dementia from others using WMH volume, sensitivity, specificity, and accuracy were 83.3%, 84.3%, and 84.3%, respectively.

Conclusion

Deep learning-based automatic WMH segmentation may be an accurate and promising method for classifying the grades of the Fazekas scale and differentiating subcortical vascular dementia.

Incidence of Amyloid-Related Imaging Abnormalities in Patients With Alzheimer Disease Treated With Anti-β-amyloid Immunotherapy: A Meta-analysis

OBJECTIVES

To assess the incidence of ARIA in clinical trials of anti-Aβ immunotherapy and compare the incidence among different agents and clinical characteristics to identify possible predisposing factors for ARIA.

METHODS

The PubMed and Embase databases were searched for clinical trials of anti-Aβ immunotherapy published on or before January 12, 2022. Phase 2 or 3 randomized controlled trials reporting detailed data sufficient to assess the incidence of ARIA were selected. The pooled incidences of ARIA and subgroup analyses according to agent and ApoE-4 carrier status were calculated using the DerSimonian-Liard random-effects model. The proportion of symptomatic ARIA cases was also calculated.

RESULTS

In total, 19 eligible studies, including 24 cohorts, were identified and 9,429 patients were analyzed. The overall pooled incidence of ARIA-effusion (E) and ARIA-hemorrhage (H) was 6.5% and 7.8%, respectively. In the subgroup analysis, the incidence of ARIA was different according to anti-Aβ immunotherapy agent. The cohorts treated with aducanumab had a significantly higher incidence of ARIA-E and ARIA-H (30.7% and 30.0%, respectively; both P < 0.001) compared to cohorts from other drugs. In subgroup analysis according to ApoE-4 carrier status, the incidences of ARIA-E and ARIA-H were higher in the ApoE-4 carrier group than those in ApoE-4 non-carrier group, but there was no statistical significance ( ApoE-4 carrier vs. non-carrier, ARIA-E; 8.6% vs. 6.9%, P = 0.663, and ARIA-H; 10.5% vs. 6.6%, P = 0.398). The pooled proportion of asymptomatic ARIA, detected by routine scheduled MRI surveillances, was 80.4%.

CONCLUSIONS

The overall incidences of ARIA-E and ARIA-H were 6.5% and 7.8%, respectively, and the pooled proportion of asymptomatic ARIA was 80.4%. The cohorts treated with aducanumab showed a significantly higher incidence of ARIA-E and ARIA-H (30.7% and 30.0%) compared with other drugs.

Diagnostic yield of TOF-MRA for detecting incidental vascular lesions in patients with cognitive impairment: An observational cohort study

Objectives

The role of three-dimensional (3D) TOF-MRA in patients with cognitive impairment is not well established. We evaluated the diagnostic yield of 3D TOF-MRA for detecting incidental extra- or intracranial artery stenosis and intracranial aneurysm in this patient group.

Methods

This retrospective study included patients with cognitive impairment undergoing our brain MRI protocol from January 2013 to February 2020. The diagnostic yield of TOF-MRA for detecting incidental vascular lesions was calculated. Patients with positive TOF-MRA results were reviewed to find whether additional treatment was performed. Logistic regression analysis was conducted to identify the clinical risk factors for positive TOF-MRA findings.

Results

In total, 1,753 patients (mean age, 70.2 ± 10.6 years; 1,044 women) were included; 199 intracranial aneurysms were detected among 162 patients (9.2%, 162/1,753). A 3D TOF-MRA revealed significant artery stenoses (>50% stenosis) in 162 patients (9.2%, 162/1,753). The overall diagnostic yield of TOF-MRA was 16.8% (294/1,753). Among them, 92 patients (31.3%, 92/294) underwent either medical therapy, endovascular intervention, or surgery. In total, eighty-one patients with stenosis were prescribed with either antiplatelet medications or lipid-lowering agent. In total, fifteen patients (aneurysm: 11 patients, stenosis: 4 patients) were further treated with endovascular intervention or surgery. Thus, the “number needed to scan” was 19 for identifying one patient requiring treatment. Multivariate logistic regression analysis showed that being female (odds ratio [OR] 2.05) and old age (OR 1.04) were the independent risk factors for intracranial aneurysm; being male (OR 1.52), old age (OR 1.06), hypertension (OR 1.78), and ischemic heart disease history (OR 2.65) were the independent risk factors for significant artery stenosis.

Conclusions

Our study demonstrated the potential benefit of 3D TOF-MRA, given that it showed high diagnostic yield for detecting vascular lesions in patients with cognitive impairment and the considerable number of these lesions required further treatment. A 3D TOF-MRA may be included in the routine MR protocol for the work-up of this patient population, especially in older patients and patients with vascular risk factors.

Diagnostic performance of hippocampal volumetry in Alzheimer's disease or mild cognitive impairment: a meta-analysis

OBJECTIVE

To evaluate the diagnostic performance of hippocampal volumetry for Alzheimer's disease (AD) or mild cognitive impairment (MCI).

METHODS

The MEDLINE and Embase databases were searched for articles that evaluated the diagnostic performance of hippocampal volumetry in differentiating AD or MCI from normal controls, published up to March 6, 2022. The quality of the articles was evaluated by the QUADAS-2 tool. A bivariate random-effects model was used to pool sensitivity, specificity, and area under the curve. Sensitivity analysis and meta-regression were conducted to explain study heterogeneity. The diagnostic performance of entorhinal cortex volumetry was also pooled.

RESULTS

Thirty-three articles (5157 patients) were included. The pooled sensitivity and specificity for AD were 82% (95% confidence interval [CI], 77-86%) and 87% (95% CI, 82-91%), whereas those for MCI were 60% (95% CI, 51-69%) and 75% (95% CI, 67-81%), respectively. No difference in the diagnostic performance was observed between automatic and manual segmentation (p = 0.11). MMSE scores, study design, and the reference standard being used were associated with study heterogeneity (p < 0.01). Subgroup analysis demonstrated a higher diagnostic performance of entorhinal cortex volumetry for both AD (pooled sensitivity: 88% vs. 79%, specificity: 92% vs. 89%, p = 0.07) and MCI (pooled sensitivity: 71% vs. 55%, specificity: 83% vs. 68%, p = 0.06).

CONCLUSIONS

Our meta-analysis demonstrated good diagnostic performance of hippocampal volumetry for AD or MCI. Entorhinal cortex volumetry might have superior diagnostic performance to hippocampal volumetry. However, due to a small number of studies, the diagnostic performance of entorhinal cortex volumetry is yet to be determined.

KEY POINTS

• The pooled sensitivity and specificity of hippocampal volumetry for Alzheimer's disease were 82% and 87%, whereas those for mild cognitive impairment were 60% and 75%, respectively. • No significant difference in the diagnostic performance was observed between automatic and manual segmentation. • Subgroup analysis demonstrated superior diagnostic performance of entorhinal cortex volumetry for AD (pooled sensitivity: 88%, specificity: 92%) and MCI (pooled sensitivity: 71%, specificity: 83%).

[Brain MRI-Based Artificial Intelligence Software in Patients with Neurodegenerative Diseases: Current Status]

The incidence of neurodegenerative diseases in the older population has increased in recent years. A considerable number of studies have been performed to characterize these diseases. Imaging analysis is an important biomarker for the diagnosis of neurodegenerative disease. Objective and reliable assessment and precise detection are important for the early diagnosis of neurodegenerative diseases. Artificial intelligence (AI) using brain MRI applied to the study of neurodegenerative diseases could promote early diagnosis and optimal decisions for treatment plans. MRI-based AI software have been developed and studied worldwide. Representatively, there are MRI-based volumetry and segmentation software. In this review, we present the development process of brain volumetry analysis software in neurodegenerative diseases, currently used and developed AI software for neurodegenerative disease in the Republic of Korea, probable uses of AI in the future, and AI software limitations.

Mammographically occult breast cancers detected with AI-based diagnosis supporting software: clinical and histopathologic characteristics

Background

To demonstrate the value of an artificial intelligence (AI) software in the detection of mammographically occult breast cancers and to determine the clinicopathologic patterns of the cancers additionally detected using the AI software.

Methods

By retrospectively reviewing our institutional database (January 2017–September 2019), we identified women with mammographically occult breast cancers and analyzed their mammography with an AI software that provided a malignancy score (range 0–100; > 10 considered as positive). The hot spots in the AI report were compared with the US and MRI findings to determine if the cancers were correctly marked by the AI software. The clinicopathologic characteristics of the AI-detected cancers were analyzed and compared with those of undetected cancers.

Results

Among the 1890 breast cancers, 6.8% (128/1890) were mammographically occult, among which 38.3% (49/128) had positive results in the AI analysis. Of them, 81.6% (40/49) were correctly marked by the AI software and determined as “AI-detected cancers.” As such, 31.3% (40/128) of mammographically occult breast cancers could be identified by the AI software. Of the AI-detected cancers, 97.5% were found in heterogeneously or extremely dense breasts, 52.5% were asymptomatic, 86.5% were invasive, and 29.7% had axillary lymph node metastasis. Compared with undetected cancers, the AI-detected cancers were more likely to be found in younger patients (p < 0.001), undergo neoadjuvant chemotherapy as well as mastectomy rather than breast-conserving operation (both p < 0.001), and accompany axillary lymph node metastasis (p = 0.003).

Conclusions

AI conferred an added value in the detection of mammographically occult breast cancers.

Role of White Matter Abnormalities in the Relationship Between Microbleed Burden and Cognitive Impairment in Cerebral Amyloid Angiopathy

Background: Cerebral amyloid angiopathy (CAA) often presents as cognitive impairment, but the mechanism of cognitive decline is unclear. Recent studies showed that number of microbleeds were associated with cognitive decline. Objective: We aimed to investigate how microbleeds contribute to cognitive impairment in association with white matter tract abnormalities or cortical thickness in CAA. Methods: This retrospective comparative study involved patients with probable CAA according to the Boston criteria (Aβ  + CAA) and patients with Alzheimer’s disease (Aβ  + AD), all of whom showed severe amyloid deposition on amyloid PET. Using mediation analysis, we investigated how FA or cortical thickness mediates the correlation between the number of lobar microbleeds and cognition. Results: We analyzed 30 patients with Aβ  + CAA (age 72.2±7.6, female 53.3%) and 30 patients with Aβ  + AD (age 71.5±7.6, female 53.3%). The two groups showed similar degrees of cortical amyloid deposition in AD-related regions. The Aβ  + CAA group had significantly lower FA values in the clusters of the posterior area than did the Aβ  + AD group (family-wise error-corrected p <  0.05). The correlation between the number of lobar microbleeds and visuospatial function was indirectly mediated by white matter tract abnormality of right posterior thalamic radiation (PTR) and tapetum, while lobar microbleeds and language function was indirectly mediated by the abnormality of left PTR and sagittal stratum. Cortical thickness did not mediate the association between lobar microbleeds and cognition. Conclusion: This result supports the hypothesis that microbleeds burden leads to white matter tract damage and subsequent cognitive decline in CAA.

Prognostic value of diffusion-weighted imaging in patients with newly diagnosed sporadic Creutzfeldt-Jakob disease

OBJECTIVES

To evaluate clinico-radiologic markers that predict poor overall survival (OS) in sporadic Creutzfeldt-Jakob disease (sCJD) and to develop a prognostic model.

MATERIALS AND METHODS

Patients with newly diagnosed sCJD were included who underwent diffusion-weighted imaging (DWI) from February 2000 to July 2020. The impact of 9 clinico-radiologic features on OS was analyzed using univariable and multivariable Cox proportional hazards regression model. The DWI prognostic score model was generated. The weighted kappa was calculated for interobserver agreement.

RESULTS

Sixty patients (mean age ± SD, 61.0 ± 9.7 years, 32 women) were included. Univariable analysis showed positive associations between poor OS and patient age (p = 0.003), extent of involved cortical lobes (p = 0.11), involvement of caudate nucleus (p = 0.07), and putamen (p = 0.04). Multivariable analysis demonstrated two independent prognostic factors: age ≥ 60 (HR 2.65, 95% CI, 1.41-4.98), and diffusion restriction in caudate nucleus and putamen (HR 2.24, 95% CI, 1.15-4.37). Based on these features, the DWI prognostic score model was generated: low-risk (0-1 point), intermediate-risk (2-3 points), and high-risk (4-5 points) groups. Median OS in high-risk group was 1.7 months, which was significantly shorter than those in the intermediate-risk (14.2 months) and low-risk (26.5 months) groups (p < 0.001). Interobserver agreements were excellent (κ = 0.91-0.92).

CONCLUSIONS

Our study demonstrated that age and diffusion restriction in caudate nucleus and putamen were the independent prognostic factors of poor overall survival in sporadic Creutzfeldt-Jakob disease. Our DWI prognostic score model may be useful in clinical settings for disease stratification.

KEY POINTS

• Age ≥ 60, and diffusion restriction in caudate nucleus and putamen were the independent prognostic factors of poor overall survival in sCJD. • Based on our DWI prognostic score model, median overall survival in high-risk group was 1.7 months, which was significantly shorter than those in the intermediate-risk group (14.2 months) and low-risk group (26.5 months) (p < 0.001). • The proposed DWI prognostic score model may be useful in clinical settings for disease stratification.

Association between ARID2 and RAS-MAPK pathway in intellectual disability and short stature

BACKGROUND

ARID2 belongs to the Switch/sucrose non-fermenting complex, in which the genetic defects have been found in patients with dysmorphism, short stature and intellectual disability (ID). As the phenotypes of patients with ARID2 mutations partially overlap with those of RASopathy, this study evaluated the biochemical association between ARID2 and RAS-MAPK pathway.

METHODS

The phenotypes of 22 patients with either an ARID2 heterozygous mutation or haploinsufficiency were reviewed. Comprehensive molecular analyses were performed using somatic and induced pluripotent stem cells (iPSCs) of a patient with ARID2 haploinsufficiency as well as using the mouse model of Arid2 haploinsufficiency by CRISPR/Cas9 gene editing.

RESULTS

The phenotypic characteristics of ARID2 deficiency include RASopathy, Coffin-Lowy syndrome or Coffin-Siris syndrome or undefined syndromic ID. Transient ARID2 knockout HeLa cells using an shRNA increased ERK1 and ERK2 phosphorylation. Impaired neuronal differentiation with enhanced RAS-MAPK activity was observed in patient-iPSCs. In addition, Arid2 haploinsufficient mice exhibited reduced body size and learning/memory deficit. ARID2 haploinsufficiency was associated with reduced IFITM1 expression, which interacts with caveolin-1 (CAV-1) and inhibits ERK activation.

DISCUSSION

ARID2 haploinsufficiency is associated with enhanced RAS-MAPK activity, leading to reduced IFITM1 and CAV-1 expression, thereby increasing ERK activity. This altered interaction might lead to abnormal neuronal development and a short stature.

Diagnostic performance of the medial temporal lobe atrophy scale in patients with Alzheimer's disease: a systematic review and meta-analysis

OBJECTIVE

To evaluate the diagnostic performance and reliability of the medial temporal lobe atrophy (MTA) scale in patients with Alzheimer's disease.

METHODS

A systematic literature search of MEDLINE and EMBASE databases was performed to select studies that evaluated the diagnostic performance or reliability of MTA scale, published up to January 21, 2021. Pooled estimates of sensitivity and specificity were calculated using a bivariate random-effects model. Pooled correlation coefficients for intra- and interobserver agreements were calculated using the random-effects model based on Fisher's Z transformation of correlations. Meta-regression was performed to explain the study heterogeneity. Subgroup analysis was performed to compare the diagnostic performance of the MTA scale and hippocampal volumetry.

RESULTS

Twenty-one original articles were included. The pooled sensitivity and specificity of the MTA scale in differentiating Alzheimer's disease from healthy control were 74% (95% CI, 68-79%) and 88% (95% CI, 83-91%), respectively. The area under the curve of the MTA scale was 0.88 (95% CI, 0.84-0.90). Meta-regression demonstrated that the difference in the method of rating the MTA scale was significantly associated with study heterogeneity (p = 0.04). No significant difference was observed in five studies regarding the diagnostic performance between MTA scale and hippocampal volumetry (p = 0.40). The pooled correlation coefficients for intra- and interobserver agreements were 0.85 (95% CI, 0.69-0.93) and 0.83 (95% CI, 0.66-0.92), respectively.

CONCLUSIONS

Our meta-analysis demonstrated a good diagnostic performance and reliability of the MTA scale in Alzheimer's disease.

KEY POINTS

• The pooled sensitivity and specificity of the MTA scale in differentiating Alzheimer's disease from healthy control were 74% and 88%, respectively. • There was no significant difference in the diagnostic performance between MTA scale and hippocampal volumetry. • The reliability of MTA scale was excellent based on the pooled correlation coefficient for intra- and interobserver agreements.

Extrahippocampal Radiomics Analysis Can Potentially Identify Laterality in Patients With MRI-Negative Temporal Lobe Epilepsy

Objective: The objective of the study was to investigate whether radiomics features of extrahippocampal regions differ between patients with epilepsy and healthy controls, and whether any differences can identify patients with magnetic resonance imaging (MRI)-negative temporal lobe epilepsy (TLE).

Methods: Data from 36 patients with hippocampal sclerosis (HS) and 50 healthy controls were used to construct a radiomics model. A total of 1,618 radiomics features from the affected hippocampal and extrahippocampal regions were compared with features from healthy controls and the unaffected side of patients. Using a stepwise selection method with a univariate t-test and elastic net penalization, significant predictors for identifying TLE were separately selected for the hippocampus (H+) and extrahippocampal region (H–). Each model was independently validated with an internal set of MRI-negative adult TLE patients (n = 22) and pediatric validation cohort with MRI-negative TLE (n = 20) from another tertiary center; diagnostic performance was calculated using area under the curve (AUC) of the receiver-operating-characteristic curve analysis.

Results: Forty-eight significant H+ radiomic features and 99 significant H– radiomic features were selected from the affected side of patients and used to create a hippocampus model and an extrahippocampal model, respectively. Texture features were the most frequently selected feature. Training set showed slightly higher accuracy between hippocampal (AUC = 0.99) and extrahippocampal model (AUC = 0.97). In the internal validation and external validation sets, the extrahippocampal model (AUC = 0.80 and 0.92, respectively) showed higher diagnostic performance for identifying the affected side of patients than the hippocampus model (AUC = 0.67 and 0.69).

Significance: Radiomics revealed extrahippocampal abnormality in the affected side of patients with TLE and could potentially help to identify MRI-negative TLE.

Classification of Evidence: Class IV Criteria for Rating Diagnostic Accuracy Studies.

Diagnostic performance of T2* gradient echo, susceptibility-weighted imaging, and quantitative susceptibility mapping for patients with multiple system atrophy-parkinsonian type: a systematic review and meta-analysis

OBJECTIVES

To investigate the diagnostic performance of T2*-weighted gradient echo (GRE) imaging, susceptibility-weighted imaging (SWI), or quantitative susceptibility mapping (QSM) in differentiating multiple system atrophy-parkinsonian type (MSA-P) from Parkinson's disease (PD).

METHODS

A systematic literature search through the MEDLINE and EMBASE databases was performed, starting on September 8, 2020, to identify studies evaluating the diagnostic performance of putaminal hypointensity on T2* GRE or SWI and phase shift on QSM in differentiating MSA-P from PD. The pooled sensitivity and specificity were obtained using hierarchical logistic regression modeling and hierarchical summary receiver operating characteristic (HSROC) modeling. The pooled diagnostic yields of T2* GRE, SWI, or QSM among MSA-P patients were calculated using the DerSimonian-Laird random-effects model.

RESULTS

Twelve original articles with 985 patients were finally included. SWI was performed in seven studies, T2* GRE was performed in three studies, and QSM was performed in two studies. The pooled sensitivity and specificity were 0.65 (95% CI 0.51-0.78) and 0.90 (95% CI 0.83-0.95), respectively. The area under the HSROC curve was 0.87 (95% CI 0.84-0.90). The Higgins I² statistic calculations revealed considerable heterogeneity in terms of both sensitivity (I² = 72.12%) and specificity (I² = 70.38%). The coupled forest plot revealed the threshold effect. For the nine studies in which area under the curve (AUC) was obtainable, the AUC ranged from 0.68 to 0.947, with a median of 0.819. The pooled diagnostic yield of T2* GRE, SWI, or QSM was 66% (95% CI 51-78%).

CONCLUSIONS

Putaminal hypointensity on T2* GRE or SWI and phase shift on QSM might be a promising diagnostic tool in differentiating MSA-P from PD. Further large multicenter prospective study is warranted.

KEY POINTS

• Three different index tests, definitions of positive image findings, thresholds, the way how to draw ROIs, reference standard, and MRI parameters could affect the heterogeneity of the study. • The pooled sensitivity and specificity were 0.65 (95% CI 0.51-0.78) and 0.90 (95% CI 0.83-0.95), respectively. • The pooled diagnostic yield of T2* GRE, SWI, or QSM was 66% (95% CI 51-78%).

Diagnostic Yield of Diffusion-Weighted Brain Magnetic Resonance Imaging in Patients with Transient Global Amnesia: A Systematic Review and Meta-Analysis

OBJECTIVE

To investigate the diagnostic yield of diffusion-weighted imaging (DWI) in patients with transient global amnesia (TGA) and identify significant parameters affecting diagnostic yield.

MATERIALS AND METHODS

A systematic literature search of the MEDLINE and EMBASE databases was conducted to identify studies that assessed the diagnostic yield of DWI in patients with TGA. The pooled diagnostic yield of DWI in patients with TGA was calculated using the DerSimonian-Laird random-effects model. Subgroup analyses were also performed of slice thickness, magnetic field strength, and interval between symptom onset and DWI.

RESULTS

Twenty-two original articles (1732 patients) were included. The pooled incidence of right, left, and bilateral hippocampal lesions was 37% (95% confidence interval [CI], 30-44%), 42% (95% CI, 39-46%), and 25% (95% CI, 20-30%) of all lesions, respectively. The pooled diagnostic yield of DWI in patients with TGA was 39% (95% CI, 27-52%). The Higgins I² statistic showed significant heterogeneity (I² = 95%). DWI with a slice thickness ≤ 3 mm showed a higher diagnostic yield than DWI with a slice thickness > 3 mm (pooled diagnostic yield: 63% [95% CI, 53-72%] vs. 26% [95% CI, 16-40%], p < 0.01). DWI performed at an interval between 24 and 96 hours after symptom onset showed a higher diagnostic yield (68% [95% CI, 57-78%], p < 0.01) than DWI performed within 24 hours (16% [95% CI, 7-34%]) or later than 96 hours (15% [95% CI, 8-26%]). There was no difference in the diagnostic yield between DWI performed using 3T vs. 1.5T (pooled diagnostic yield, 31% [95% CI, 25-38%] vs. 24% [95% CI, 14-37%], p = 0.31).

CONCLUSION

The pooled diagnostic yield of DWI in TGA patients was 39%. DWI obtained with a slice thickness ≤ 3 mm or an interval between symptom onset and DWI of > 24 to 96 hours could increase the diagnostic yield.

Prognostic Utility of Disproportionately Enlarged Subarachnoid Space Hydrocephalus in Idiopathic Normal Pressure Hydrocephalus Treated with Ventriculoperitoneal Shunt Surgery: A Systematic Review and Meta-analysis

BACKGROUND

Disproportionately enlarged subarachnoid space hydrocephalus is a specific radiologic marker for idiopathic normal pressure hydrocephalus. However, controversy exists regarding the prognostic utility of disproportionately enlarged subarachnoid space hydrocephalus.

PURPOSE

Our aim was to evaluate the prevalence of disproportionately enlarged subarachnoid space hydrocephalus in idiopathic normal pressure hydrocephalus and its predictive utility regarding prognosis in patients treated with ventriculoperitoneal shunt surgery.

DATA SOURCES

We used MEDLINE and EMBASE databases.

STUDY SELECTION

We searched for studies that reported the prevalence or the diagnostic performance of disproportionately enlarged subarachnoid space hydrocephalus in predicting treatment response.

DATA ANALYSIS

The pooled prevalence of disproportionately enlarged subarachnoid space hydrocephalus was obtained. Pooled sensitivity, specificity, and area under the curve of disproportionately enlarged subarachnoid space hydrocephalus to predict treatment response were obtained. Subgroup and sensitivity analyses were performed to explain heterogeneity among the studies.

DATA SYNTHESIS

Ten articles with 812 patients were included. The pooled prevalence of disproportionately enlarged subarachnoid space hydrocephalus in idiopathic normal pressure hydrocephalus was 44% (95% CI, 34%-54%). The pooled prevalence of disproportionately enlarged subarachnoid space hydrocephalus was higher in the studies using the second edition of the Japanese Guidelines for Management of Idiopathic Normal Pressure Hydrocephalus compared with the studies using the international guidelines without statistical significance (52% versus 43%, P = .38). The pooled sensitivity and specificity of disproportionately enlarged subarachnoid space hydrocephalus for prediction of treatment response were 59% (95% CI, 38%-77%) and 66% (95% CI, 57%-74%), respectively, with an area under the curve of 0.67 (95% CI, 0.63-0.71).

LIMITATIONS

The lack of an established method for assessing disproportionately enlarged subarachnoid space hydrocephalus using brain MR imaging served as an important cause of the heterogeneity.

CONCLUSIONS

Our meta-analysis demonstrated a relatively low prevalence of disproportionately enlarged subarachnoid space hydrocephalus in idiopathic normal pressure hydrocephalus and a poor diagnostic performance for treatment response.

Clinical Features and Brain MRI Findings in Korean Patients with AGel Amyloidosis

PURPOSE

AGel amyloidosis is systemic amyloidosis caused by pathogenic variants in the GSN gene. In this study, we sought to characterize the clinical and brain magnetic resonance image (MRI) features of Korean patients with AGel amyloidosis.

MATERIALS AND METHODS

We examined 13 patients with AGel amyloidosis from three unrelated families. Brain MRIs were performed in eight patients and eight age- and sex-matched healthy controls. Therein, we analyzed gray and white matter content using voxel-based morphometry (VBM), tract-based spatial statistics (TBSS), and FreeSurfer.

RESULTS

The median age at examination was 73 (interquartile range: 64-76) years. The median age at onset of cutis laxa was 20 (interquartile range: 15-30) years. All patients over that age of 60 years had dysarthria, cutis laxa, dysphagia, and facial palsy. Two patients in their 30s had only mild cutis laxa. The median age at dysarthria onset was 66 (interquartile range: 63.5-70) years. Ophthalmoparesis was observed in three patients. No patient presented with muscle weakness of the limbs. Axial fluid-attenuated inversion recovery images of the brain showed no significant differences between the patient and control groups. Also, analysis of VBM, TBSS, and FreeSurfer revealed no significant differences in cortical thickness between patients and healthy controls at the corrected significance level.

CONCLUSION

Our study outlines the clinical manifestations of prominent bulbar palsy and early-onset cutis laxa in 13 Korean patients with AGel amyloidosis and confirms that AGel amyloidosis mainly affects the peripheral nervous system rather than the central nervous system.

Utility of 7 Tesla Magnetic Resonance Imaging in Patients With Epilepsy: A Systematic Review and Meta-Analysis

Objective: 7 Tesla magnetic resonance imaging (MRI) enables high resolution imaging and potentially improves the detection of morphologic abnormalities in patients with epilepsy. However, its added value compared with conventional 1.5T and 3.0T MRI is unclear. We reviewed the evidence for the use of 7 Tesla MRI in patients with epilepsy and compared the detection rate of focal lesions with clinical MRI. Methods: Clinical retrospective case studies were identified using the indexed text terms "epilepsy" AND "magnetic resonance imaging" OR "MR imaging" AND "7T" OR "7 Tesla" OR "7T" in Medline (2002-September 1, 2020) and Embase (1999-September 1, 2020). The study setting, MRI protocols, qualitative, and quantitative assessment were systematically reviewed. The detection rate of morphologic abnormalities on MRI was reported in each study in which surgery was used as the reference standard. Meta-analyses were performed using a univariate random-effects model in diagnostic performance studies with patients that underwent both 7T MRI and conventional MRI. Results: Twenty-five articles were included (467 patients and 167 healthy controls) consisting of 10 case studies, 10 case-control studies, 4 case series, and 1 cohort study. All studies included focal epilepsy; 12 studies (12/25, 48%) specified the disease etiology and 4 studies reported focal but non-lesional (MRI-negative on 1.5/3.0T) epilepsy. 7T MRI showed superior detection and delineation of morphologic abnormalities in all studies. In nine comparative studies, 7T MRI had a superior detection rate of 65% compared with the 22% detection rate of 1.5T or 3.0T. Significance: 7T MRI is useful for delineating morphologic abnormalities with a higher detection rate compared with conventional clinical MRI. Most studies were conducted using a case series or case study; therefore, a cohort study design with clinical outcomes is necessary. Classification of Evidence: Class IV Criteria for Rating Diagnostic Accuracy Studies.

Diagnostic performance and interobserver agreement of the callosal angle and Evans' index in idiopathic normal pressure hydrocephalus: a systematic review and meta-analysis

OBJECTIVE

To evaluate the diagnostic performance and interobserver agreement of the callosal angle and Evans' index in idiopathic normal pressure hydrocephalus (iNPH).

METHODS

A systematic search of MEDLINE and EMBASE was performed to find studies assessing the diagnostic performance or interobserver agreement of the callosal angle and Evans' index in iNPH. Pooled sensitivity and specificity of the two radiologic indices were calculated. The area under the curve (AUC) was obtained based on a hierarchical summary receiver operating characteristic curve. The diagnostic performances of both radiologic indices were compared in subgroup analysis. To evaluate interobserver agreement, the pooled correlation coefficient was calculated.

RESULTS

Ten original articles (874 patients) were included. The pooled sensitivity and specificity of the callosal angle in the diagnosis of iNPH were 91% (95% CI, 86-94%) and 93% (95% CI, 89-96%), respectively. The pooled sensitivity and specificity of Evans' index were 96% (95% CI, 47-100%) and 83% (95% CI, 77-88%), respectively. Subgroup analysis demonstrated a significant higher specificity of the callosal angle than that of Evans' index (p < 0.01). The AUC of the callosal angle and Evans' index were 0.97 (95% CI, 0.95-0.98) and 0.87 (95% CI, 0.84-0.90), respectively. The pooled correlation coefficients for the callosal angle and Evans' index were 0.92 (95% CI, 0.82-0.96) and 0.92 (95% CI, 0.83-0.97), respectively.

CONCLUSIONS

Our meta-analysis demonstrated a high performance of the callosal angle in the diagnosis of iNPH. Evans' index showed reasonable diagnostic performance with high sensitivity but low specificity. Interobserver agreements were excellent in both radiologic indices.

KEY POINTS

• Callosal angle showed high diagnostic performance in idiopathic normal pressure hydrocephalus. • Evans' index showed reasonable diagnostic performance with high sensitivity but low specificity. • Interobserver agreements were excellent in both callosal angle and Evans' index.

De-Identification of Facial Features in Magnetic Resonance Images: Software Development Using Deep Learning Technology

Background

High-resolution medical images that include facial regions can be used to recognize the subject’s face when reconstructing 3-dimensional (3D)-rendered images from 2-dimensional (2D) sequential images, which might constitute a risk of infringement of personal information when sharing data. According to the Health Insurance Portability and Accountability Act (HIPAA) privacy rules, full-face photographic images and any comparable image are direct identifiers and considered as protected health information. Moreover, the General Data Protection Regulation (GDPR) categorizes facial images as biometric data and stipulates that special restrictions should be placed on the processing of biometric data.

Objective

This study aimed to develop software that can remove the header information from Digital Imaging and Communications in Medicine (DICOM) format files and facial features (eyes, nose, and ears) at the 2D sliced-image level to anonymize personal information in medical images.

Methods

A total of 240 cranial magnetic resonance (MR) images were used to train the deep learning model (144, 48, and 48 for the training, validation, and test sets, respectively, from the Alzheimer's Disease Neuroimaging Initiative [ADNI] database). To overcome the small sample size problem, we used a data augmentation technique to create 576 images per epoch. We used attention-gated U-net for the basic structure of our deep learning model. To validate the performance of the software, we adapted an external test set comprising 100 cranial MR images from the Open Access Series of Imaging Studies (OASIS) database.

Results

The facial features (eyes, nose, and ears) were successfully detected and anonymized in both test sets (48 from ADNI and 100 from OASIS). Each result was manually validated in both the 2D image plane and the 3D-rendered images. Furthermore, the ADNI test set was verified using Microsoft Azure's face recognition artificial intelligence service. By adding a user interface, we developed and distributed (via GitHub) software named “Deface program” for medical images as an open-source project.

Conclusions

We developed deep learning–based software for the anonymization of MR images that distorts the eyes, nose, and ears to prevent facial identification of the subject in reconstructed 3D images. It could be used to share medical big data for secondary research while making both data providers and recipients compliant with the relevant privacy regulations.

Deep-learned time-signal intensity pattern analysis using an autoencoder captures magnetic resonance perfusion heterogeneity for brain tumor differentiation

Current image processing methods for dynamic susceptibility contrast (DSC) magnetic resonance imaging (MRI) do not capture complex dynamic information of time-signal intensity curves. We investigated whether an autoencoder-based pattern analysis of DSC MRI captured representative temporal features that improves tissue characterization and tumor diagnosis in a multicenter setting. The autoencoder was applied to the time-signal intensity curves to obtain representative temporal patterns, which were subsequently learned by a convolutional neural network. This network was trained with 216 preoperative DSC MRI acquisitions and validated using external data (n = 43) collected with different DSC acquisition protocols. The autoencoder applied to time-signal intensity curves and clustering obtained nine representative clusters of temporal patterns, which accurately identified tumor and non-tumoral tissues. The dominant clusters of temporal patterns distinguished primary central nervous system lymphoma (PCNSL) from glioblastoma (AUC 0.89) and metastasis from glioblastoma (AUC 0.95). The autoencoder captured DSC time-signal intensity patterns that improved identification of tumoral tissues and differentiation of tumor type and was generalizable across centers.

Development and Validation of a Deep Learning-Based Automatic Brain Segmentation and Classification Algorithm for Alzheimer Disease Using 3D T1-Weighted Volumetric Images

BACKGROUND AND PURPOSE

Limited evidence has suggested that a deep learning automatic brain segmentation and classification method, based on T1-weighted brain MR images, can predict Alzheimer disease. Our aim was to develop and validate a deep learning-based automatic brain segmentation and classification algorithm for the diagnosis of Alzheimer disease using 3D T1-weighted brain MR images.

MATERIALS AND METHODS

A deep learning-based algorithm was developed using a dataset of T1-weighted brain MR images in consecutive patients with Alzheimer disease and mild cognitive impairment. We developed a 2-step algorithm using a convolutional neural network to perform brain parcellation followed by 3 classifier techniques including XGBoost for disease prediction. All classification experiments were performed using 5-fold cross-validation. The diagnostic performance of the XGBoost method was compared with logistic regression and a linear Support Vector Machine by calculating their areas under the curve for differentiating Alzheimer disease from mild cognitive impairment and mild cognitive impairment from healthy controls.

RESULTS

In a total of 4 datasets, 1099, 212, 711, and 705 eligible patients were included. Compared with the linear Support Vector Machine and logistic regression, XGBoost significantly improved the prediction of Alzheimer disease (P < .001). In terms of differentiating Alzheimer disease from mild cognitive impairment, the 3 algorithms resulted in areas under the curve of 0.758-0.825. XGBoost had a sensitivity of 68% and a specificity of 70%. In terms of differentiating mild cognitive impairment from the healthy control group, the 3 algorithms resulted in areas under the curve of 0.668-0.870. XGBoost had a sensitivity of 79% and a specificity of 80%.

CONCLUSIONS

The deep learning-based automatic brain segmentation and classification algorithm allowed an accurate diagnosis of Alzheimer disease using T1-weighted brain MR images. The widespread availability of T1-weighted brain MR imaging suggests that this algorithm is a promising and widely applicable method for predicting Alzheimer disease.

Comparison of Dynamic Contrast-Enhancement Parameters between Gadobutrol and Gadoterate Meglumine in Posttreatment Glioma: A Prospective Intraindividual Study

BACKGROUND AND PURPOSE

Differences in molecular properties between one-molar and half-molar gadolinium-based contrast agents are thought to affect parameters obtained from dynamic contrast-enhanced imaging. The aim of our study was to investigate differences in dynamic contrast-enhanced parameters between one-molar nonionic gadobutrol and half-molar ionic gadoterate meglumine in patients with posttreatment glioma.

MATERIALS AND METHODS

This prospective study enrolled 32 patients who underwent 2 20-minute dynamic contrast-enhanced examinations, one with gadobutrol and one with gadoterate meglumine. The model-free parameter of area under the signal intensity curve from 30 to 1100 seconds and the Tofts model-based pharmacokinetic parameters were calculated and compared intraindividually using paired t tests. Patients were further divided into progression (n = 12) and stable (n = 20) groups, which were compared using Student t tests.

RESULTS

Gadobutrol and gadoterate meglumine did not show any significant differences in the area under the signal intensity curve or pharmacokinetic parameters of K ^trans, V_e, V_p, or K_ep (all P > .05). Gadobutrol showed a significantly higher mean wash-in rate (0.83 ± 0.64 versus 0.29 ± 0.63, P = .013) and a significantly lower mean washout rate (0.001 ± 0.0001 versus 0.002 ± 0.002, P = .02) than gadoterate meglumine. Trends toward higher area under the curve, K ^trans, V_e, V_p, wash-in, and washout rates and lower K_ep were observed in the progression group in comparison with the treatment-related-change group, regardless of the contrast agent used.

CONCLUSIONS

Model-free and pharmacokinetic parameters did not show any significant differences between the 2 gadolinium-based contrast agents, except for a higher wash-in rate with gadobutrol and a higher washout rate with gadoterate meglumine, supporting the interchangeable use of gadolinium-based contrast agents for dynamic contrast-enhanced imaging in patients with posttreatment glioma.

Incorporating diffusion- and perfusion-weighted MRI into a radiomics model improves diagnostic performance for pseudoprogression in glioblastoma patients

BACKGROUND

Pseudoprogression is a diagnostic challenge in early posttreatment glioblastoma. We therefore developed and validated a radiomics model using multiparametric MRI to differentiate pseudoprogression from early tumor progression in patients with glioblastoma.

METHODS

The model was developed from the enlarging contrast-enhancing portions of 61 glioblastomas within 3 months after standard treatment with 6472 radiomic features being obtained from contrast-enhanced T1-weighted imaging, fluid-attenuated inversion recovery imaging, and apparent diffusion coefficient (ADC) and cerebral blood volume (CBV) maps. Imaging features were selected using a LASSO (least absolute shrinkage and selection operator) logistic regression model with 10-fold cross-validation. Diagnostic performance for pseudoprogression was compared with that for single parameters (mean and minimum ADC and mean and maximum CBV) and single imaging radiomics models using the area under the receiver operating characteristics curve (AUC). The model was validated with an external cohort (n = 34) imaged on a different scanner and internal prospective registry data (n = 23).

RESULTS

Twelve significant radiomic features (3 from conventional, 2 from diffusion, and 7 from perfusion MRI) were selected for model construction. The multiparametric radiomics model (AUC, 0.90) showed significantly better performance than any single ADC or CBV parameter (AUC, 0.57-0.79, P < 0.05), and better than a single radiomics model using conventional MRI (AUC, 0.76, P = 0.012), ADC (AUC, 0.78, P = 0.014), or CBV (AUC, 0.80, P = 0.43). The multiparametric radiomics showed higher performance in the external validation (AUC, 0.85) and internal validation (AUC, 0.96) than any single approach, thus demonstrating robustness.

CONCLUSIONS

Incorporating diffusion- and perfusion-weighted MRI into a radiomics model improved diagnostic performance for identifying pseudoprogression and showed robustness in a multicenter setting.

Disrupted Functional and Structural Connectivity in Angelman Syndrome

BACKGROUND AND PURPOSE

This work investigated alterations in functional connectivity (FC) and associated structures in patients with Angelman syndrome (AS) by using integrated quantitative imaging analysis and connectivity measures.

MATERIALS AND METHODS

We obtained 3T brain MR imaging, including resting-state functional MR imaging, diffusion tensor imaging, and 3D T1-weighted imaging from children with AS (n = 14) and age- and sex-matched controls (n = 28). The brains of patients with AS were analyzed by measuring FC, white matter microstructural analysis, cortical thickness, and brain volumes; these were compared with brains of controls.

RESULTS

Interregional FC analysis revealed significantly reduced intra- and interhemispheric FC, especially in the basal ganglia and thalamus, in patients with AS. Significant reductions in fractional anisotropy were found in the corpus callosum, cingulum, posterior limb of the internal capsules, and arcuate fasciculus in patients with AS. Quantitative structural analysis also showed gray matter volume loss of the basal ganglia and diffuse WM volume reduction in AS compared with the control group.

CONCLUSIONS

This integrated quantitative MR imaging analysis demonstrated poor functional and structural connectivity, as well as brain volume reduction, in children with AS, which may explain the motor and language dysfunction observed in this well-characterized neurobehavioral phenotype.

Altered Structural Network in Newly Onset Childhood Absence Epilepsy

BACKGROUND AND PURPOSE

Recent quantitative neuroimaging studies of childhood absence epilepsy (CAE) have identified various structural abnormalities that might be involved in the onset of absence seizure and associated cognitive and behavioral functions. However, the neuroanatomical alterations specific to CAE remain unclear, and so this study investigated the regional alterations of brain structures associated with newly diagnosed CAE.

METHODS

Surface and volumetric magnetic resonance imaging data of patients with newly diagnosed CAE (n=18) and age-matched healthy controls (n=18) were analyzed using Free-Surfer software. A group comparison using analysis of covariance was performed with significance criteria of p<0.05 and p<0.01 in global and regional analyses, respectively.

RESULTS

Compared with control subjects, the patients with CAE had smaller total and regional volumes of cortical gray-matter (GM) in the right rostral middle frontal, right lateral orbitofrontal, and left rostral middle frontal regions, as well as in the right precentral, right superior, middle, left middle, and inferior temporal gyri. The cortex in the right posterior cingulate gyrus and left medial occipital region was significantly thicker in patients with CAE than in controls.

CONCLUSIONS

Patients with CAE showed a reduced bilateral frontotemporal cortical GM volume and an increased posterior medial cortical thickness, which are associated with the default mode network. These structural changes can be suggested as the neural basis of the absence seizures and neuropsychiatric comorbidities in CAE.

Perilesional and homotopic area activation during proverb comprehension after stroke

INTRODUCTION

The mechanism of functional recovery in right hemisphere (RH) stroke patients when attempting to comprehend a proverb has not been identified. We previously reported that there is bilateral hemisphere involvement during proverb comprehension in the normal population. However, the underlying mechanisms of proverb comprehension following a right middle cerebral artery (MCA) infarction have not yet been fully elucidated.

METHODS

We here compared the brain regions activated by literal sentences and by opaque or transparent proverbs in right MCA infarction patients using functional magnetic resonance imaging (fMRI). Experimental stimuli included 18 opaque proverbs, 18 transparent proverbs, and 18 literal sentences that were presented pseudorandomly in 1 of 3 predesigned sequences.

RESULTS

Fifteen normal adults and 17 right MCA infarction patients participated in this study. The areas of the brain in the stroke patients involved in understanding a proverb compared with a literal sentence included the right middle frontal gyrus (MFG) and frontal pole, right anterior cingulate gyrus/paracingulate gyrus and left inferior frontal gyrus (IFG), middle temporal gyrus (MTG), precuneus, and supramarginal gyrus (SMG). When the proverbs were presented to these stroke patients in the comprehension tests, the left supramarginal, postcentral gyrus, and right paracingulate gyrus were activated for the opaque proverbs compared to the transparent proverbs.

CONCLUSIONS

These findings suggest that the functional recovery of language in stroke patients can be explained by perilesional activation, which is thought to arise from the regulation of the excitatory and inhibitory neurotransmitter system, and by homotopic area activation which has been characterized by decreased transcallosal inhibition and astrocyte involvement.

Early treadmill exercise increases macrophage migration inhibitory factor expression after cerebral ischemia/reperfusion

The neuroprotective function of macrophage migration inhibitory factor (MIF) in ischemic stroke was rarely evaluated. This study aimed to investigate the effects of early treadmill exercise on recovery from ischemic stroke and to determine whether these effects are associated with the expression levels of MIF and brain-derived neurotrophic factor (BDNF) in the ischemic area. A total of 40 male Sprague-Dawley rats were randomly assigned to the ischemia and exercise group [middle cerebral artery occlusion (MCAO)-Ex, n = 10), ischemia and sedentary group (MCAO-St, n = 10), sham-surgery and exercise group (Sham-Ex, n = 10), or sham-surgery and sedentary group (Sham-St, n = 10). The MCAO-Ex and MCAO-St groups were subjected to MCAO for 60 minutes, whereas the Sham-Ex and Sham-St groups were subjected to an identical operation without MCAO. Rats in the MCAO-Ex and Sham-Ex groups then ran on a treadmill for 30 minutes once a day for 5 consecutive days. After reperfusion, the hanging time tested by the wire hang test was longer and the relative fractional anisotropy determined by MRI was higher in the peri-infarct region of the MCAO-Ex group compared with the MCAO-St group. The expression levels of MIF and BDNF in the peri-infarct region were upregulated in the MCAO-Ex group. Increased MIF and BDNF levels were positively correlated with relative fractional anisotropy changes in the peri-infarct region. There was no significant difference in the levels of MIF and BDNF in the peri-infarct region between the Sham-Ex and Sham-St groups. Our study demonstrated that early exercise (initiated 48 hours after the MCAO) could improve motor and neuronal recovery after ischemic stroke. Furthermore, the increased levels of MIF and BDNF in the peri-infarct region (penumbra) may be one of the mechanisms of enhanced neurological function recovery. All experiments were approved by the Institutional Animal Care and Use Committee in Asan Medical Center in South Korea (2016-12-126).

Hyperoxia-Induced ΔR1

Background and Purpose- Acceleration of longitudinal relaxation under hyperoxic challenge (ie, hyperoxia-induced ΔR₁) indicates oxygen accumulation and reflects baseline tissue oxygenation. We evaluated the feasibility of hyperoxia-induced ΔR₁ for evaluating cerebral oxygenation status and degree of ischemic damage in stroke. Methods- In 24-hour transient stroke rat models (n=13), hyperoxia-induced ΔR₁, ischemic severity (apparent diffusion coefficient [ADC]), vasogenic edema (R₂), total and microvascular blood volume (superparamagnetic iron oxide-driven ΔR₂* and ΔR₂, respectively), and glucose metabolism activity (¹⁸F-fluorodeoxyglucose uptake on positron emission tomography) were measured. The distribution of these parameters according to hyperoxia-induced ΔR₁ was analyzed. The partial pressure of tissue oxygen change during hyperoxic challenge was measured using fiberoptic tissue oximetry. In 4-hour stroke models (n=6), ADC and hyperoxia-induced ΔR₁ was analyzed with 2,3,5-triphenyltetrazolium chloride staining being a criterion of infarction. Results- Ischemic hemisphere showed significantly higher hyperoxia-induced ΔR₁ than nonischemic brain in a pattern depending on ADC. During hyperoxic challenge, ischemic hemisphere demonstrated uncontrolled increase of partial pressure of tissue oxygen, whereas contralateral hemisphere rapidly plateaued. Ischemic hemisphere also demonstrated significant correlation between hyperoxia-induced ΔR₁ and R₂. Hyperoxia-induced ΔR₁ showed a significant negative correlation with ¹⁸F-fluorodeoxyglucose uptake. The ADC, R₂, ΔR₂, and ¹⁸F-fluorodeoxyglucose uptake showed a dichotomized distribution according to the hyperoxia-induced ΔR₁ as their slopes and values were higher at low hyperoxia-induced ΔR₁ (<50 ms^-1) than at high ΔR₁. In 4-hour stroke rats, the distribution of ADC according to the hyperoxia-induced ΔR₁ was similar with 24-hour stroke rats. The hyperoxia-induced ΔR₁ was greater in the infarct area (47±10 ms^-1) than in peri-infarct area (16±4 ms^-1; P<0.01). Conclusions- Hyperoxia-induced ΔR₁ adequately indicates cerebral oxygenation and can be a feasible biomarker to classify the degree of ischemia-induced damage in neurovascular function and metabolism in stroke brain.

Correction to: Contribution of Zinc-Dependent Delayed Calcium Influx via TRPC5 in Oxidative Neuronal Death and its Prevention by Novel TRPC Antagonist

After the publication of this work errors were noticed in Fig. 3b and 4d.

Radiomics Analysis of Gadoxetic Acid-enhanced MRI for Staging Liver Fibrosis

Purpose To develop and validate a radiomics-based model for staging liver fibrosis by using gadoxetic acid-enhanced hepatobiliary phase MRI. Materials and Methods In this retrospective study, 436 patients (mean age, 51 years; age range, 18-86 years; 319 men [mean age, 51 years; age range, 18-86 years]; 117 women [mean age, 50 years; age range, 18-79 years]) with pathologic analysis-proven liver fibrosis who underwent gadoxetic acid-enhanced MRI from June 2015 to December 2016 were randomized in a three-to-one ratio into development (n = 329) and test (n = 107) cohorts, respectively. In the development cohort, a model was developed to calculate radiomics fibrosis index (RFI) by using logistic regression with elastic net regularization to differentiate stage F3-F4 from stage F0-F2. Optimal RFI cutoffs to diagnose clinically significant fibrosis (stage F2-F4), advanced fibrosis (stage F3-F4), and cirrhosis (stage F4) were determined by receiver operating characteristic curve analysis. In the test cohort, the diagnostic performance of RFI was compared with that of normalized liver enhancement, aspartate transaminase-to-platelet ratio index (APRI), and fibrosis-4 index by using the Obuchowski index. Results In the test cohort, RFI (Obuchowski index, 0.86) significantly outperformed normalized liver enhancement (Obuchowski index, 0.77; P < .03), APRI (Obuchowski index, 0.60; P < .001), and fibrosis-4 index (Obuchowski index, 0.62; P < .001) for staging liver fibrosis. By using the cutoffs, RFI had sensitivities and specificities as follows: 81% (95% confidence interval: 71%, 89%) and 78% (95% confidence interval: 63%, 89%) for diagnosing stage F2-F4, respectively; 79% (95% confidence interval: 67%, 88%) and 82% (95% confidence interval: 69%, 91%), respectively, for diagnosing stage F3-F4; and 92% (95% confidence interval: 79%, 98%) and 75% (95% confidence interval: 62%, 83%), respectively, for diagnosing stage F4. Conclusion Radiomics analysis of gadoxetic acid-enhanced hepatobiliary phase images allows for accurate diagnosis of liver fibrosis. © RSNA, 2018 Online supplemental material is available for this article.

Reliability of fast magnetic resonance imaging for acute ischemic stroke patients using a 1.5-T scanner

OBJECTIVES

To determine whether fast scanned MRI using a 1.5-T scanner is a reliable method for the detection and characterization of acute ischemic stroke in comparison with conventional MRI.

METHODS

From May 2015 to June 2016, 862 patients (FLAIR, n = 482; GRE, n = 380; MRA, n = 190) were prospectively enrolled in the study, with informed consent and under institutional review board approval. The patients underwent both fast (EPI-FLAIR, ETL-FLAIR, TR-FLAIR, EPI-GRE, parallel-GRE, fast CE-MRA) and conventional MRI (FLAIR, GRE, time-of-flight MRA, fast CE-MRA). Two neuroradiologists independently assessed agreements in acute and chronic ischemic hyperintensity, hyperintense vessels (FLAIR), microbleeds, susceptibility vessel signs, hemorrhagic transformation (GRE), stenosis (MRA), and image quality (all MRI), between fast and conventional MRI. Agreements between fast and conventional MRI were evaluated by generalized estimating equations. Z-scores were used for comparisons of the percentage agreement among fast FLAIR sequences and fast GRE sequences and between conventional and fast MRA.

RESULTS

Agreements of more than 80% were achieved between fast and conventional MRI (ETL-FLAIR, 96%; TR-FLAIR, 97%; EPI-GRE, 96%; parallel-GRE, 98%; fast CE-MRA, 86%). ETL- and TR-FLAIR were significantly superior to EPI-FLAIR in the detection of acute ischemic hyperintensity and hyperintense vessels, while parallel-GRE was significantly superior to EPI-GRE in the detection of susceptibility vessel sign (p value < 0.05 for all). There were no significant differences in the other scores and image qualities (p value > 0.05).

CONCLUSIONS

Fast MRI at 1.5 T is a reliable method for the detection and characterization of acute ischemic stroke in comparison with conventional MRI.

KEY POINTS

• Fast MRI at 1.5 T may achieve a high intermethod reliability in the detection and characterization of acute ischemic stroke with a reduction in scan time in comparison with conventional MRI.

Deep Learning-Based Computer-Aided Diagnosis System for Localization and Diagnosis of Metastatic Lymph Nodes on Ultrasound: A Pilot Study

BACKGROUND

The presence of metastatic lymph nodes is a prognostic indicator for patients with thyroid carcinomas and is an important determinant of clinical decision making. However, evaluating neck lymph nodes requires experience and is labor- and time-intensive. Therefore, the development of a computer-aided diagnosis (CAD) system to identify and differentiate metastatic lymph nodes may be useful.

METHODS

From January 2008 to December 2016, we retrieved clinical records for 804 consecutive patients with 812 lymph nodes. The status of all lymph nodes was confirmed by fine-needle aspiration. The datasets were split into training (263 benign and 286 metastatic lymph nodes), validation (30 benign and 33 metastatic lymph nodes), and test (100 benign and 100 metastatic lymph nodes). Using the VGG-Class Activation Map model, we developed a CAD system to localize and differentiate the metastatic lymph nodes. We then evaluated the diagnostic performance of this CAD system in our test set.

RESULTS

In the test set, the accuracy, sensitivity, and specificity of our model for predicting lymph node malignancy were 83.0%, 79.5%, and 87.5%, respectively. The CAD system clearly detected the locations of the lymph nodes, which not only provided identifying data, but also demonstrated the basis of decisions.

CONCLUSION

We developed a deep learning-based CAD system for the localization and differentiation of metastatic lymph nodes from thyroid cancer on ultrasound. This CAD system is highly sensitive and may be used as a screening tool; however, as it is relatively less specific, the screening results should be validated by experienced physicians.

Development and Validation of a Deep Learning System for Staging Liver Fibrosis by Using Contrast Agent-enhanced CT Images in the Liver

Purpose To develop and validate a deep learning system (DLS) for staging liver fibrosis by using CT images in the liver. Materials and Methods DLS for CT-based staging of liver fibrosis was created by using a development data set that included portal venous phase CT images in 7461 patients with pathologically confirmed liver fibrosis. The diagnostic performance of the DLS was evaluated in separate test data sets for 891 patients. The influence of patient characteristics and CT techniques on the staging accuracy of the DLS was evaluated by logistic regression analysis. In a subset of 421 patients, the diagnostic performance of the DLS was compared with that of the radiologist's assessment, aminotransferase-to-platelet ratio index (APRI), and fibrosis-4 index by using the area under the receiver operating characteristic curve (AUROC) and Obuchowski index. Results In the test data sets, the DLS had a staging accuracy of 79.4% (707 of 891) and an AUROC of 0.96, 0.97, and 0.95 for diagnosing significant fibrosis (F2-4), advanced fibrosis (F3-4), and cirrhosis (F4), respectively. At multivariable analysis, only pathologic fibrosis stage significantly affected the staging accuracy of the DLS (P = .016 and .013 for F1 and F2, respectively, compared with F4), whereas etiology of liver disease and CT technique did not. The DLS (Obuchowski index, 0.94) outperformed the radiologist's interpretation, APRI, and fibrosis-4 index (Obuchowski index range, 0.71-0.81; P ˂ .001) for staging liver fibrosis. Conclusion The deep learning system allows for accurate staging of liver fibrosis by using CT images. © RSNA, 2018 Online supplemental material is available for this article.