A joint convolutional-recurrent neural network with an attention mechanism for detecting intracranial hemorrhage on noncontrast head CT.

To investigate the performance of a joint convolutional neural networks-recurrent neural networks (CNN-RNN) using an attention mechanism in identifying and classifying intracranial hemorrhage (ICH) on a large multi-center dataset; to test its performance in a prospective independent sample consisting of consecutive real-world patients. All consecutive patients who underwent emergency non-contrast-enhanced head CT in five different centers were retrospectively gathered. Five neuroradiologists created the ground-truth labels. The development dataset was divided into the training and validation set. After the development phase, we integrated the deep learning model into an independent center's PACS environment for over six months for assessing the performance in a real clinical setting. Three radiologists created the ground-truth labels of the testing set with a majority voting. A total of 55,179 head CT scans of 48,070 patients, 28,253 men (58.77%), with a mean age of 53.84 ± 17.64 years (range 18-89) were enrolled in the study. The validation sample comprised 5211 head CT scans, with 991 being annotated as ICH-positive. The model's binary accuracy, sensitivity, and specificity on the validation set were 99.41%, 99.70%, and 98.91, respectively. During the prospective implementation, the model yielded an accuracy of 96.02% on 452 head CT scans with an average prediction time of 45 ± 8 s. The joint CNN-RNN model with an attention mechanism yielded excellent diagnostic accuracy in assessing ICH and its subtypes on a large-scale sample. The model was seamlessly integrated into the radiology workflow. Though slightly decreased performance, it provided decisions on the sample of consecutive real-world patients within a minute.

Topographic evaluation of bithalamic infarcts: Are these due to occlusion of the artery of Percheron?

Occlusion of the artery of Percheron (AoP) causes bithalamic paramedian infarct (BTPI). Although it can be diagnosed easily in its pure form, it can be underdiagnosed in cases with concomitant extrathalamic acute infarcts (plus-BTPI) as it may be difficult to determine whether BTPI is due to occlusion of AoP or two different paramedian arteries even with conventional angiography. This study was performed to highlight plus-BTPI that could result from occlusion of AoP rather than of two distinct paramedian arteries using topographic evaluation of bithalamic infarcts. We retrospectively reviewed imaging and clinical databases for patients admitted to radiology department between 2013 and 2019. Two radiologists independently evaluated the results of imaging studies, and findings reached by consensus were used in the analysis. This retrospective review yielded 34 patients with bithalamic infarct. Each affected thalamic vascular region was investigated separately. Any patient could have more than 2 different vascular zone infarct. The affected thalamic vascular territories were paramedian (n = 24), inferolateral (n = 13), anterior (n = 10), and posterior (n = 7). When we evaluated bithalamic infarcts in terms of symmetrically affected territories, the distribution of symmetric affected territories was as: paramedian (n = 18), inferolateral (n = 2), anterior (n = 1), and posterior (n = 1). BTPI had a 4.5-fold higher frequency than the sum of symmetric involvement of other territories (p = 0.0552, OR = 4.5,95%CI 0.93-21.5). In addition, mesencephalic involvement was only observed in BTPI, and not in other patterns (p < 0.001). The fact that in bilateral thalamic infarcts the symmetric involvement of paramedic territory is significantly higher and mesencephalic involvement is seen only in BTPI can suggest that plus-BPTI may develop due to AoP occlusion rather than occlusion of two distinct paramedian arteries.

Inter-vendor performance of deep learning in segmenting acute ischemic lesions on diffusion-weighted imaging: a multicenter study.

There is little evidence on the applicability of deep learning (DL) in the segmentation of acute ischemic lesions on diffusion-weighted imaging (DWI) between magnetic resonance imaging (MRI) scanners of different manufacturers. We retrospectively included DWI data of patients with acute ischemic lesions from six centers. Dataset A (n = 2986) and B (n = 3951) included data from Siemens and GE MRI scanners, respectively. The datasets were split into the training (80%), validation (10%), and internal test (10%) sets, and six neuroradiologists created ground-truth masks. Models A and B were the proposed neural networks trained on datasets A and B. The models subsequently fine-tuned across the datasets using their validation data. Another radiologist performed the segmentation on the test sets for comparisons. The median Dice scores of models A and B were 0.858 and 0.857 for the internal tests, which were non-inferior to the radiologist's performance, but demonstrated lower performance than the radiologist on the external tests. Fine-tuned models A and B achieved median Dice scores of 0.832 and 0.846, which were non-inferior to the radiologist's performance on the external tests. The present work shows that the inter-vendor operability of deep learning for the segmentation of ischemic lesions on DWI might be enhanced via transfer learning; thereby, their clinical applicability and generalizability could be improved.

Computed Tomography Findings Predicting the Need for Surgery in Cases of Small Bowel Obstruction: Emphasis on Duodenal Distension.

OBJECTIVE: The aim of the study is to retrospectively evaluate the utility of computed tomography (CT) findings, especially newly defined duodenal distension, for predicting the need to operate on small bowel obstruction (SBO) cases. METHODS: During a 51-month period, 228 patients (100 women and 128 men; mean age, 55 years) were included in this study, among 438 patients who were hospitalized with a prediagnosis of SBO. The final study population was then divided into 2 groups: a surgery group (n = 76) and a conservative group (n = 152). The CT findings of the SBO patients whose treatment decisions and outcomes were unknown were examined by 2 gastrointestinal radiologists with consensus. Statistical analyses were conducted using univariate and binary logistic regression analyses. RESULTS: According to the univariate analysis, the degree of obstruction (P = 0.001), small bowel diameter (P = 0.014), and presence of mesenteric fluid (P < 0.001), intraperitoneal free fluid (P = 0.04), intra-abdominal free gas (P < 0.001), and duodenal distension (P < 0.001) showed statistically significant differences between the surgery and conservative groups. However, there were no statistically significant group differences regarding the presence of a transition point, small bowel feces or mesenteric congestion. According to the binary logistic regression analysis, the degree of obstruction (P = 0.012), presence of mesenteric fluid (P = 0.008), intra-abdominal free gas (P = 0.019), and duodenal distension (P < 0.001) were significant predictors of the need for surgery in SBO cases. CONCLUSIONS: Duodenal distension as a CT finding predicted the need for surgery in SBO cases.

Intrathecal use of gadobutrol for gadolinium-enhanced MR cisternography in the evaluation of patients with otorhinorrhea.

PURPOSE: Intrathecal gadolinium-enhanced MR cisternography (IGE-MRC) has a high sensitivity to detect accurate localization of cerebrospinal fluid (CSF) leakage in otorhinorrhea patients. Our purpose in this study was to describe our experience in analyzing clinically suspected CSF leakage by IGE-MRC by using gadobutrol with emphasis on its safety and diagnostic performance. METHODS: We retrospectively reviewed our imaging and clinical database for the evaluation of patients admitted to our clinic with complaints of otorhinorrhea between 2017 and 2019. Two radiologists evaluated the imaging studies independently. Consensus data was used in the analysis. Medical record review and phone call were used for the follow-up. RESULTS: Of the 85 patients included in the retrospective analysis, 82 (96.5%) had rhinorrhea and 3 (3.5%) had otorrhea. Overall, 29 patients (34.1% of all patients) underwent operation for repair of the CSF leakage site. Beta-transferrin test was available and positive in 33 patients (38.8%). Five (5.9%) patients complained headaches after the procedure and complaints were resolved with increased water intake. Postprocedurally, 3 patients (3.5%) had vertigo and 1 patient (1.2%) complained nausea but spontaneous regression were observed in a few hours. None of the patients experienced a significant complication or adverse reaction during follow-up period. Sixty-seven patients (78.8%) had medical record and telephone follow-up. Mean follow-up duration with call was 14.2 months. CONCLUSION: IGE-MRC is a minimally invasive and highly sensitive imaging technique. The current results during our follow-up demonstrate the relative safety and feasibility of IGE-MRC by using gadobutrol to evaluate CSF leakage.

Topographic evaluation of medullary infarcts from the radiologist's point of view.

PURPOSE: Despite considerable published information about the clinical-radiological correlation of medullary infarcts, no study has determined whether topographic evaluations are performed accurately among researchers. Our purpose in this study was twofold: to evaluate the topographic pattern of medullary infarcts on diffusion-weighted imaging by their radiological aspect, and to assess interobserver agreement on the topographic pattern. METHODS: We retrospectively reviewed our imaging and clinical database for patients admitted to our radiology department between January 2014 and September 2019. Two radiologists evaluated the imaging studies independently. Consensus data were used in the analysis. RESULTS: The retrospective review yielded 92 patients with medullary infarction. The affected vascular territories were lateral (n = 58), anteromedial (n = 28), posterior (n = 3), and anterolateral (n = 1). Two patients had hemimedullary infarction. The rostrocaudal levels of the medullary infarct were superior (n = 34), middle (n = 31), inferior (n = 4), superior-middle (n = 13), and middle-inferior (n = 10). The medullary infarcts were divided into two types: lateral (n = 62) and medial (n = 28). The affected vascular territories differed with rostrocaudal topography of medullary infarct (p = 0.003). Excellent interobserver agreement was found for type of medullary infarct, compared with moderate for vascular territory and fair for rostrocaudal topography. The anterolateral and posterior territories were the most often misdiagnosed, while the level with the most disagreements in rostrocaudal topography was middle. CONCLUSION: The accurate topographic evaluation of a medullary infarct can be an important basis for investigating stroke etiology. However, correct topographic evaluation may not always be available and smaller territories such as anterolateral and posterior should be assessed carefully.

Hypertrophic Olivary Degeneration and Holmes Tremor: Case Report and Review of the Literature.

BACKGROUND: Hypertrophic olivary degeneration (HOD) is very rare type of degeneration that causes hypertrophy rather than atrophy. The classical presentation of HOD is palatal myoclonus. However, HOD may rarely present with Holmes tremor (HT). HT is unusual symptomatic tremor characterized by combination of rest and intention tremor. It has been reported in small case series, so far. CASE DESCRIPTION: In this study, a man aged 62 years with HOD and HT spreading to the upper and lower extremities after pontine-midbrain hemorrhage due to cavernoma was presented. CONCLUSIONS: Although pontine-midbrain hemorrhage may cause HT in the late period, HOD can be revealed on magnetic resonance imaging. Tract anatomy, especially the Guillain-Mollaret triangle, should be considered to explain the relationship between HT and HOD.

Contribution of Reformatted Multislice Temporal Computed Tomography Images in the Planes of Stenvers and Pöschl to the Diagnosis of Superior Semicircular Canal Dehiscence.

OBJECTIVE: In the diagnosis of superior semicircular canal dehiscence (SSCD), computed tomography (CT) is the only imaging method. The aims of the study were to show that reformat images are more accurate than standard planes for diagnosis of SSCD and to determine the prevalence of SSCD. METHODS: The retrospective review yielded 1309 temporal CTs performed in our radiology department for any reason. Two radiologist interpreted CTs in standard planes collaboratively. Patients with SSCD were reinterpreted in Pöschl and Stenvers planes by 2 radiologists separately. RESULTS: Statistical analysis was made by accepting that 2 radiologists diagnosis were accurate in Pöschl plane. Coronal plane sensitivity 86%, specificity 64%, Stenvers plane sensitivity 96%, and specificity 52% have been found in the mean result of 2 observers (P < 0.001). CONCLUSIONS: In the diagnosis of SSCD, standard and Stenvers planes can cause false-negative and false-positive diagnoses. Interpretation in Pöschl plane can significantly increase sensitivity, specificity, negative, and positive predictive values for diagnosing dehiscence.

Cerebral Abscess Following Endovascular Treatment of Aneurysm: Report of 2 Cases and Review of the Literature.

INTRODUCTION: Endovascular treatment is growing in popularity globally as an important treatment option for intracranial aneurysms. Cerebral infection as a complication of endovascular treatment of aneurysms is sufficiently rare that only 6 isolated cerebral abscess cases have been reported thus far. CASE REPORT: In this report, we present 2 cerebral abscesses from 3 institutions, which developed after coil embolization of anterior communicating artery and middle cerebral artery aneurysms. CONCLUSION: This report was written to discuss the possible mechanisms underlying cerebral abscess after endovascular treatment of aneurysms and to suggest potential treatment and prevention methods.

De Novo Development of a Cerebral Arteriovenous Malformation: Case Report and Review of the Literature.

BACKGROUND: Cerebral arteriovenous malformations (AVMs) have been commonly regarded as congenital. However, this suspected origin has been challenged by reports of de novo cerebral AVM. CASE DESCRIPTION: We have described a 25-year-old man without any known history of cerebrovascular disease, in whom cranial imaging demonstrated the de novo appearance of a pial AVM. Initial magnetic resonance imaging at 11 years of age had revealed the presence of a 2-cm parafalcine arachnoid cyst. Computed tomography and magnetic resonance imaging performed 14 years later showed a new cerebral AVM in the left frontal lobe, which was then confirmed angiographically. CONCLUSIONS: The findings from our case report and from 9 similar reports challenge the traditional theory that AVMs constitute congenital lesions.

The value of hepatic diffusion-weighted MR imaging in demonstrating hepatic congestion secondary to pulmonary hypertension.

BACKGROUND: Congestive hepatomegaly might be the first sign for pulmonary hypertension. Apparent diffusion coefficient (ADC) value obtained with quantitative diffusion-weighted magnetic resonance imaging (DW-MRI) is affected by liver fibrosis and perfusion. We aimed to evaluate the diagnostic value of DW-MRI in cooperation with biochemical markers, ultrasonography (US) and echocardiography (TTE) in determining the degree of hepatic congestion secondary to pulmonary hypertension (PHT). METHODS: 35 patients with PHT and 26 control subjects were included in the study. PHT was diagnosed if pulmonary artery systolic pressure (PASP) was measured above 35 mmHg with TTE. Study group was classified into mild and moderate PHT. DW-MRI was performed with b-factors of 0, 500 and 1000 sec/mm(2). Mean ADC, ADC-II (Average of the ADC values of right lobe anterior and posterior segments), US, TTE and blood biochemical parameters of both groups were compared. RESULTS: There exists a positive correlation between liver size and the diameters of vena cava inferior, right atrium, right hepatic vein(RHV), mid-hepatic vein(MHV), left hepatic vein(LHV) (p < 0.01). There was a positive correlation between PASP and RHV, MHV, LHV. The patients had lower ejection fractions (p < 0.01) and higher LDH (p < 0.01) and ALP (p < 0.05) levels than the control group. The ADC values of the patients with moderate PASP were higher than those with a mild PASP (p < 0.05). Mean ADC was higher in patients with moderate PHT compared to control group (p = 0.009). There was a positive correlation between PASP and ADC values of right lobe posterior segment of the liver (p < 0.05). The ADC-II and mean ADC values of the patients with moderate PASP were higher than those of the control group (p < 0.01). CONCLUSIONS: Congestion due to moderate PHT might be diagnosed with DW-MRI. As PASP increase; mean ADC and ADC-II values increase.