Automatic segmentation of hemorrhagic transformation on follow-up non-contrast CT after acute ischemic stroke

Background

Hemorrhagic transformation (HT) following reperfusion therapies is a serious complication for patients with acute ischemic stroke. Segmentation and quantification of hemorrhage provides critical insights into patients' condition and aids in prognosis. This study aims to automatically segment hemorrhagic regions on follow-up non-contrast head CT (NCCT) for stroke patients treated with endovascular thrombectomy (EVT).

Methods

Patient data were collected from 10 stroke centers across two countries. We propose a semi-automated approach with adaptive thresholding methods, eliminating the need for extensive training data and reducing computational demands. We used Dice Similarity Coefficient (DSC) and Lin's Concordance Correlation Coefficient (Lin's CCC) to evaluate the performance of the algorithm.

Results

A total of 51 patients were included, with 28 Type 2 hemorrhagic infarction (HI2) cases and 23 parenchymal hematoma (PH) cases. The algorithm achieved a mean DSC of 0.66 ± 0.17. Notably, performance was superior for PH cases (mean DSC of 0.73 ± 0.14) compared to HI2 cases (mean DSC of 0.61 ± 0.18). Lin's CCC was 0.88 (95% CI 0.79-0.93), indicating a strong agreement between the algorithm's results and the ground truth. In addition, the algorithm demonstrated excellent processing time, with an average of 2.7 s for each patient case.

Conclusion

To our knowledge, this is the first study to perform automated segmentation of post-treatment hemorrhage for acute stroke patients and evaluate the performance based on the radiological severity of HT. This rapid and effective tool has the potential to assist with predicting prognosis in stroke patients with HT after EVT.

Application of imaging modalities for endovascular thrombectomy of large core infarcts in clinical practice

Four randomized controlled trials of large infarct core volume (LICV) included three imaging modalities: non-contrast CT (NCCT)-Alberta Stroke Program Early CT Score (ASPECTS), diffusion-weighted imaging (DWI)-ASPECTS, and NCCT-ASPECTS combined with CTP (CT perfusion). However, there is no clear consensus on the optimal imaging modality for endovascular thrombectomy (EVT) trials of large core infarcts. The variety and complexity of imaging modalities make it difficult to apply them in clinical practice. By familiarizing ourselves with these imaging modalities, we can better apply them in the clinic and correctly screen patients with large core infarcts in the anterior circulation who can benefit from EVT therapy.

Association of non-contrast CT markers with long-term functional outcome in deep intracerebral hemorrhage

Objective

Hematoma expansion (HE) is the most important therapeutic target during acute care of patients with intracerebral hemorrhage (ICH). Imaging biomarkers such as non-contrast CT (NCCT) markers have been associated with increasing risk for HE. The aim of the present study was to evaluate the influence of NCCT markers with functional long-term outcome and with HE in patients with deep (basal ganglia and thalamus) ICH who represent an important subgroup of patients at the highest risk for functional deterioration with HE due to the eloquence of the affected brain region.

Methods

From our prospective institutional database, all patients maximally treated with deep ICH were included and retrospectively analyzed. NCCT markers were recorded at diagnostic imaging, ICH volume characteristics were volumetrically evaluated, and all patients received follow-up imaging within 0-48 h. We explored associations of NCCT makers with unfavorable functional outcome, defined as modified Rankin scale 4-6, after 12 months and with HE. Bias and confounding were addressed by multivariable regression modeling.

Results

In 322 patients with deep ICH, NCCT markers were distributed as follows: irregular shape: 69.6%, heterogenous density: 55.9%, hypodensities: 52.5%, island sign: 19.3%, black hole sign: 11.5%, and blend sign: 4.7%. Upon multivariable regression analyses, independent associations were documented with the functional outcome for irregular shape (aOR: 2.73, 95%CI: 1.42-5.22, p = 0.002), heterogenous density (aOR: 2.62, 95%CI: 1.40-4.90, p = 0.003) and island sign (aOR: 2.54, 95%CI: 1.05-6.14, p = 0.038), and with HE for heterogenous density (aOR: 5.01, 95%CI: 1.93-13.05, p = 0.001) and hypodensities (aOR: 3.75, 95%CI: 1.63-8.62, p = 0.002).

Conclusion

NCCT markers are frequent in deep ICH patients and provide important clinical implications. Specifically, markers defined by diverging intra-hematomal densities provided associations with a 5-times higher risk for HE and a 2.5-times higher likelihood for unfavorable functional long-term outcome. Hence, these markers allow the identification of patients with deep ICH at high risk for clinical deterioration due to HE.

Clinical Features, Non-Contrast CT Radiomic and Radiological Signs in Models for the Prediction of Hematoma Expansion in Intracerebral Hemorrhage

PURPOSE

Rapid identification of hematoma expansion (HE) risk at baseline is a priority in intracerebral hemorrhage (ICH) patients and may impact clinical decision making. Predictive scores using clinical features and Non-Contract Computed Tomography (NCCT)-based features exist, however, the extent to which each feature set contributes to identification is limited. This paper aims to investigate the relative value of clinical, radiological, and radiomics features in HE prediction.

METHODS

Original data was retrospectively obtained from three major prospective clinical trials ["Spot Sign" Selection of Intracerebral Hemorrhage to Guide Hemostatic Therapy (SPOTLIGHT)NCT01359202; The Spot Sign for Predicting and Treating ICH Growth Study (STOP-IT)NCT00810888] Patients baseline and follow-up scans following ICH were included. Clinical, NCCT radiological, and radiomics features were extracted, and multivariate modeling was conducted on each feature set.

RESULTS

317 patients from 38 sites met inclusion criteria. Warfarin use (p=0.001) and GCS score (p=0.046) were significant clinical predictors of HE. The best performing model for HE prediction included clinical, radiological, and radiomic features with an area under the curve (AUC) of 87.7%. NCCT radiological features improved upon clinical benchmark model AUC by 6.5% and a clinical & radiomic combination model by 6.4%. Addition of radiomics features improved goodness of fit of both clinical (p=0.012) and clinical & NCCT radiological (p=0.007) models, with marginal improvements on AUC. Inclusion of NCCT radiological signs was best for ruling out HE whereas the radiomic features were best for ruling in HE.

CONCLUSION

NCCT-based radiological and radiomics features can improve HE prediction when added to clinical features.

Epicardial fat volume, an independent risk factor for major adverse cardiovascular events, had an incremental prognostic value to myocardial perfusion imaging in Chinese populations with suspected or known coronary artery disease with a normal left ventricular ejection fraction

Background

Most coronary artery disease (CAD) patients with a normal left ventricular ejection fraction (LVEF) experience a poor prognosis. Single-photon emission computerized tomography (SPECT)-myocardial perfusion imaging (MPI), a routine examination, is useful in assessing risk and predicting major adverse cardiovascular events (MACEs) in populations with suspected or known CAD. SPECT/CT is a "one-stop shop" examination, which, through non-contrast CT, can produce attenuation correction for MPI and obtain information on coronary artery calcium (CAC) and epicardial fat volume (EFV) simultaneously. This study aims to investigate the predictive and incremental value of EFV to MPI for MACE in Chinese populations with suspected or known CAD with a normal LVEF.

Methods and results

We retrospectively studied 290 suspected or known CAD inpatients with a normal LVEF who underwent SPECT/CT between February 2014 and December 2017. Abnormal MPI was defined as a summed stress score ≥4 or summed difference score ≥2. EFV and CAC were calculated using non-contrast CT. The end date of follow-ups was in February 2022. The follow-up information was obtained from the clinical case notes of the patients or reviews of telephone calls. MACE was defined as cardiac death, late coronary revascularization ≥3 months after MPI, non-fatal myocardial infarction, angina-related rehospitalization, heart failure, and stroke. During the 76-month follow-up, the event rate was 32.0% (93/290). Univariate and multivariate Cox regression analyses concluded that high EFV (>108.3 cm3) [hazard ratio (HR): 3.3, 95% CI: 2.1-5.2, P < 0.000] and abnormal MPI (HR: 1.8, 95% CI: 1.1-2.8, P = 0.010) were independent risk factors for MACE. The event-free survival of patients with high EFV was significantly lower than that of the low EFV group (log-rank test P < 0.001). In the subgroup with normal MPI, high EFV was associated with reduced event-free survival (log-rank P < 0.01), with a higher annualized event rate (8.3% vs. 1.9%). Adding high EFV to MPI could predict MACEs more effectively, with a higher concordance index (0.56-0.69, P < 0.01), higher global chi square (7.2-41.4, P < 0.01), positive integrated discrimination improvement (0.10, P < 0.01), and net reclassification index (0.37, P < 0.01).

Conclusions

In Chinese populations with suspected or known CAD with normal LVEF, high EFV was an independent risk factor for MACE after adjusting for traditional risk factors, CAC and MPI. In subgroups with normal MPI, EFV could also improve risk stratification. Adding EFV to MPI had an incremental value for predicting MACE.

Non-Contrasted CT Radiomics for SAH Prognosis Prediction

Subarachnoid hemorrhage (SAH) denotes a serious type of hemorrhagic stroke that often leads to a poor prognosis and poses a significant socioeconomic burden. Timely assessment of the prognosis of SAH patients is of paramount clinical importance for medical decision making. Currently, clinical prognosis evaluation heavily relies on patients' clinical information, which suffers from limited accuracy. Non-contrast computed tomography (NCCT) is the primary diagnostic tool for SAH. Radiomics, an emerging technology, involves extracting quantitative radiomics features from medical images to serve as diagnostic markers. However, there is a scarcity of studies exploring the prognostic prediction of SAH using NCCT radiomics features. The objective of this study is to utilize machine learning (ML) algorithms that leverage NCCT radiomics features for the prognostic prediction of SAH. Retrospectively, we collected NCCT and clinical data of SAH patients treated at Beijing Hospital between May 2012 and November 2022. The modified Rankin Scale (mRS) was utilized to assess the prognosis of patients with SAH at the 3-month mark after the SAH event. Based on follow-up data, patients were classified into two groups: good outcome (mRS ≤ 2) and poor outcome (mRS > 2) groups. The region of interest in NCCT images was delineated using 3D Slicer software, and radiomic features were extracted. The most stable and significant radiomic features were identified using the intraclass correlation coefficient, t-test, and least absolute shrinkage and selection operator (LASSO) regression. The data were randomly divided into training and testing cohorts in a 7:3 ratio. Various ML algorithms were utilized to construct predictive models, encompassing logistic regression (LR), support vector machine (SVM), random forest (RF), light gradient boosting machine (LGBM), adaptive boosting (AdaBoost), extreme gradient boosting (XGBoost), and multi-layer perceptron (MLP). Seven prediction models based on radiomic features related to the outcome of SAH patients were constructed using the training cohort. Internal validation was performed using five-fold cross-validation in the entire training cohort. The receiver operating characteristic curve, accuracy, precision, recall, and f-1 score evaluation metrics were employed to assess the performance of the classifier in the overall dataset. Furthermore, decision curve analysis was conducted to evaluate model effectiveness. The study included 105 SAH patients. A comprehensive set of 1316 radiomics characteristics were initially derived, from which 13 distinct features were chosen for the construction of the ML model. Significant differences in age were observed between patients with good and poor outcomes. Among the seven constructed models, model_SVM exhibited optimal outcomes during a five-fold cross-validation assessment, with an average area under the curve (AUC) of 0.98 (standard deviation: 0.01) and 0.88 (standard deviation: 0.08) on the training and testing cohorts, respectively. In the overall dataset, model_SVM achieved an accuracy, precision, recall, f-1 score, and AUC of 0.88, 0.84, 0.87, 0.84, and 0.82, respectively, in the testing cohort. Radiomics features associated with the outcome of SAH patients were successfully obtained, and seven ML models were constructed. Model_SVM exhibited the best predictive performance. The radiomics model has the potential to provide guidance for SAH prognosis prediction and treatment guidance.

Combining convolutional attention mechanism and residual deformable Transformer for infarct segmentation from CT scans of acute ischemic stroke patients

Background

Segmentation and evaluation of infarcts on medical images are essential for diagnosis and prognosis of acute ischemic stroke (AIS). Computed tomography (CT) is the first-choice examination for patients with AIS.

Methods

To accurately segment infarcts from the CT images of patients with AIS, we proposed an automated segmentation method combining the convolutional attention mechanism and residual Deformable Transformer in this article. The method used the encoder-decoder structure, where the encoders were employed for downsampling to obtain the feature of the images and the decoder was used for upsampling and segmentation. In addition, we further applied the convolutional attention mechanism and residual network structure to improve the effectiveness of feature extraction. Our code is available at: https://github.com/XZhiXiang/AIS-segmentation/tree/master.

Results

The proposed method was assessed on a public dataset containing 397 non-contrast CT (NCCT) images of AIS patients (AISD dataset). The symptom onset to CT time was less than 24 h. The experimental results illustrate that this work had a Dice coefficient (DC) of 58.66% for AIS infarct segmentation, which outperforms several existing methods. Furthermore, volumetric analysis of infarcts indicated a strong correlation (Pearson correlation coefficient = 0.948) between the AIS infarct volume obtained by the proposed method and manual segmentation.

Conclusion

The strong correlation between the infarct segmentation obtained via our method and the ground truth allows us to conclude that our method could accurately segment infarcts from NCCT images.

[Advanced Faster RCNN: a non-contrast CT-based algorithm for detecting pancreatic lesions in multiple disease stages]

OBJECTIVE

To propose a non-contrast CT-based algorithm for automated and accurate detection of pancreatic lesions at a low cost.

METHODS

With Faster RCNN as the benchmark model, an advanced Faster RCNN (aFaster RCNN) model for pancreatic lesions detection based on plain CT was constructed. The model uses the residual connection network Resnet50 as the feature extraction module to extract the deep image features of pancreatic lesions. According to the morphology of pancreatic lesions, 9 anchor frame sizes were redesigned to construct the RPN module. A new Bounding Box regression loss function was proposed to constrain the training process of RPN module regression subnetwork by comprehensively considering the constraints of the lesion shape and anatomical structure. Finally, a detection frame was generated using the detector in the second stage. The data from a total of 728 cases of pancreatic diseases from 4 clinical centers in China were used for training (518 cases, 71.15%) and testing (210 cases, 28.85%) of the model. The performance of aFaster RCNN was verified through ablation experiments and comparison experiments with 3 classical target detection models SSD, YOLO and CenterNet.

RESULTS

The aFaster RCNN model for pancreatic lesion detection achieved recall rates of 73.64% at the image level and 92.38% at the patient level, with an average precision of 45.29% and 53.80% at the image and patient levels, respectively, which were higher than those of the 3 models for comparison.

CONCLUSION

The proposed method can effectively extract the imaging features of pancreatic lesions from non-contrast CT images to detect the pancreatic lesions.

The Association of the Spatial Location of Contrast Extravasation with Symptomatic Intracranial Hemorrhage after Endovascular Therapy in Acute Ischemic Stroke Patients

BACKGROUND

Contrast extravasation (CE) on brain non-contrast computed tomography (NCCT) after endovascular therapy (EVT) is commonly present in patients with acute ischemic stroke (AIS). Substantial uncertainties remain about the relationship between the spatial location of CE and symptomatic intracranial hemorrhage (sICH). Therefore, this study aimed to evaluate this association.

METHODS

We performed a retrospective screening on consecutive patients with AIS due to LVO (AIS-LVO) who had CE on NCCT immediately after EVT for anterior circulation large vessel occlusion (LVO). We used the Alberta stroke program early CT Score (ASPECTS) scoring system to estimate the spatial location of CE. Multivariable logistic regression was performed to achieve the risk factors of sICH.

RESULTS

In this study, 115 of 153 (75.1%) anterior circulation AIS-LVO patients had CE on NCCT. After excluding 9 patients, 106 patients were enrolled in the final analysis. In multivariate regression analysis, atrial fibrillation (AF) (adjusted OR [aOR] 6.833, 95% confidence interval [CI] 1.331-35.081, P = 0.021) and CE-ASPECTS (aOR 0.602, 95% CI 0.411-0.882 P = 0.009) were associated with sICH. In subgroup analysis, CE at the internal capsule (IC) region was an independent risk factor for sICH (aOR 5.992, 95% CI 1.010-35.543 P < 0.05). These and conventional variables were incorporated as a predict model, with AUC of 0.899, demonstrating good discrimination and calibration for sICH in this study cohort.

CONCLUSION

The spatial location of CE on NCCT immediately after EVT was an independent and strong risk factor for sICH in acute ischemic stroke patients.

Susceptibility-Weighted MRI as an Imaging Marker for Reperfusion Injury in Acute Ischemic Stroke Following Mechanical Thrombectomy

BACKGROUND/PURPOSE

Following endovascular intervention for stroke, hyperattenuated areas are common in brain parenchyma and it is difficult to differentiate on non-contrast CT whether it is contrast staining or reperfusion hemorrhage. Differentiation between contrast staining from reperfusion hemorrhage is of paramount importance for early initiation of antiplatelets and/or anticoagulants to prevent reocclusion of vessel. This study demonstrates signal characteristics of contrast-staining and reperfusion hemorrhage on susceptibility weighted MRI and its role to differentiate between two.

MATERIALS/METHODS

Between July 2017 to March 2019, 36 patients who presented with acute ischemic stroke due to large vessel occlusion underwent mechanical thrombectomy. Low-osmolar non-ionic (Iopromide 300 mg/L) iodinated contrast was used in all patients who underwent endovascular intervention. All patients underwent noncontrast CT brain and SWI on 3T MRI within 30 minutes of endovascular intervention. MRI was evaluated by two neuroradiologists. Reperfusion hemorrhage was defined as ECASS criteria II. Symptomatic ICH was defined as hemorrhagic transformation temporally related to a negative shift in NIHSS score >/=4.

RESULTS

Out of 36 patients, 15 had hyperattenuated areas in brain on NCCT. Out of 15, 13 patients had blooming on SWI, suggestive of bleed. Two patients had no blooming on SWI, suggestive of contrast staining. Two patients didnot show any hyperdensity on NCCT but blooming on SWI, suggestive of bleed.

CONCLUSION

All patients with hyperdensity on NCCT secondary to bleed showed blooming on SWI whereas those with contrast staining didnot show any signal changes on SWI. Thus, it is possible to differentiate reperfusion hemorrhage from contrast staining using SWI MRI. The significance of SWI in normal CT may be low where a small bleed maynot have any clinical significance.

A conventional-to-spectral CT image translation augmentation workflow for robust contrast injection-independent organ segmentation

PURPOSE

In cardiovascular imaging, the numerous contrast injection protocols used to enhance structures make it difficult to gather training datasets for deep learning applications supporting diverse protocols. Moreover, creating annotations on non-contrast scans is extremely tedious. Recently, spectral CT's virtual-non-contrast images (VNC) have been used as data augmentation to train segmentation networks performing on enhanced and true-non-contrast (TNC) scans alike, while improving results on protocols absent of their training dataset. However, spectral data are not widely available, making it difficult to gather specific datasets for each task. As a solution, we present a data augmentation workflow based on a trained image translation network, to bring spectral-like augmentation to any conventional CT dataset.

METHOD

The HU-to-spectral image translation network (HUSpecNet) was first trained to generate VNC from HU images, using an unannotated spectral dataset of 1830 patients. It was then tested on a second dataset of 300 spectral CT scans, by comparing generated VNC (VNC_DL ) to their true counterparts. To illustrate and compare our workflow's efficiency with true spectral augmentation, HUSpecNet was applied to a third dataset of 112 spectral scans to generate VNC_DL along HU and VNC images. Three different 3D networks (U-Net, X-Net, U-Net++) were trained for multi-label heart segmentation, following four augmentation strategies. As baselines, trainings were performed on contrasted images without (HUonly) and with conventional gray-values augmentation (HUaug). Then, the same networks were trained using a proportion of contrasted and VNC/VNC_DL images (TrueSpec/GenSpec). Each training strategy applied to each architecture was evaluated using Dice coefficients on a fourth multi-centric multi-vendor single-energy CT dataset of 121 patients, including different contrast injection protocols and unenhanced scans. The U-Net++ results were further explored with distance metrics on every label.

RESULTS

Tested on 300 full scans, our HUSpectNet translation network shows a mean absolute error of 6.70±2.83 HU between VNC_DL and VNC, while peak-signal-to-noise-ratio reaches 43.89 dB. GenSpec and TrueSpec show very close results regardless of the protocol and used architecture: mean Dice coefficients (DSC_mean ) are equal with a margin of 0.006, ranging from 0.879 to 0.938. Their performances significantly increase on TNC scans (p-values<0.017 for all architectures) compared to HUonly and HUaug, with DSC_mean of 0.448/0.770/0.879/0.885 for HUonly/HUaug/TrueSpec/GenSpec using the Unet++ architecture. Significant improvements are also noted for all architectures on chest-abdominal-pelvic scans (p-values<0.007) compared to HUonly and for pulmonary embolism scans (p-values<0.039) compared to HUaug. Using Unet++, DSC_mean reaches 0.892/0.901/0.903 for HUonly/TrueSpec/GenSpec on pulmonary embolism scans and 0.872/0.896/0.896 for HUonly/TrueSpec/GenSpec on chest-abdominal-pelvic scans.

CONCLUSION

Using the proposed workflow, we trained versatile heart segmentation networks on a dataset of conventional enhanced CT scans, providing robust predictions on both enhanced scans with different contrast injection protocols and TNC scans. The performances obtained were not significantly inferior to training the model on a genuine spectral CT dataset, regardless of the architecture implemented. Using a general-purpose conventional-to-spectral CT translation network as data augmentation could therefore contribute to reducing data collection and annotation requirements for machine learning-based CT studies, while extending their range of application. This article is protected by copyright. All rights reserved.

Identifying Thrombus on Non-Contrast CT in Patients with Acute Ischemic Stroke

The hyperdense sign is a marker of thrombus in non-contrast computed tomography (NCCT) datasets. The aim of this work was to determine optimal Hounsfield unit (HU) thresholds for thrombus segmentation in thin-slice non-contrast CT (NCCT) and use these thresholds to generate 3D thrombus models. Patients with thin-slice baseline NCCT (≤2.5 mm) and MCA-M1 occlusions were included. CTA was registered to NCCT, and three regions of interest (ROIs) were placed in the NCCT, including: the thrombus, contralateral brain tissue, and contralateral patent MCA-M1 artery. Optimal HU thresholds differentiating the thrombus from non-thrombus tissue voxels were calculated using receiver operating characteristic analysis. Linear regression analysis was used to predict the optimal HU threshold for discriminating the clot only based on the average contralateral vessel HU or contralateral parenchyma HU. Three-dimensional models from 70 participants using standard (45 HU) and patient-specific thresholds were generated and compared to CTA clot characteristics. The optimal HU threshold discriminating thrombus in NCCT from other structures varied with a median of 51 (IQR: 49-55). Experts chose 3D models derived using patient-specific HU models as corresponding better to the thrombus seen in CTA in 83.8% (31/37) of cases. Patient-specific HU thresholds for segmenting the thrombus in NCCT can be derived using normal parenchyma. Thrombus segmentation using patient-specific HU thresholds is superior to conventional 45 HU thresholds.

Can we successfully predict the outcome for extracorporeal shock wave lithotripsy (ESWL) for medium size renal stones? A single-center experience

BACKGROUND

Extracorporeal shock wave lithotripsy (ESWL) is one of the most used modalities in treatment of renal stones, but its effectiveness can be influenced by many factors related to the patient or the stone itself which may affect the success of stone disintegration. The aim of our study was to investigate the predictive value of some patient and stone-related factors for ESWL success for renal stones.

METHODS

A total of 100 patients with single radiopaque renal stone 10-20 mm in diameter, undergoing ESWL were enrolled in this study. All patients had non contrast computed tomography (NCCT) done before ESWL. We evaluated body mass index (BMI), skin-to-stone distance (SSD), stone size and Hounsfield density comparing these values between stone free (SF) and residual stone (RS) groups.

RESULTS

Of the 100 patients, 70% had successful disintegration. There was no significant difference between stone free (SF) and residual stone (RS) groups as regard age or BMI. Meanwhile, there was a significant difference between SF and RS groups as regard stones' density and SSD, with higher values in RS group but there was statistically insignificant difference as regard stone size (p = 0.522). Using logistic regression analysis, we found that Hounsfield unit (HU) was better in predicting successful disintegration than SSD but without statistical significance.

CONCLUSION

HU and SSD are the independent predictive factors for ESWL outcome, and they should be considered when planning ESWL in treatment of medium size renal stones.

Detecting the Early Infarct Core on Non-Contrast CT Images with a Deep Learning Residual Network

PURPOSE

To explore a new approach mainly based on deep learning residual network (ResNet) to detect infarct cores on non-contrast CT images and improve the accuracy of acute ischemic stroke diagnosis.

METHODS

We continuously enrolled magnetic resonance diffusion weighted image (MR-DWI) confirmed first-episode ischemic stroke patients (onset time: less than 9 h) as well as some normal individuals in this study. They all underwent CT plain scan and MR-DWI scan with same scanning range, layer thickness (4 mm) and interlayer spacing (4 mm) (The time interval between two examinations: less than 4 h). Setting MR-DWI as gold standard of infarct core and using deep learning ResNet combined with a maximum a posteriori probability (MAP) model and a post-processing method to detect the infarct core on non-contrast CT images. After that, we use decision curve analysis (DCA) establishing models to analyze the value of this new method in clinical practice.

RESULTS

116 ischemic stroke patients and 26 normal people were enrolled. 58 patients were allocated into training dataset and 58 were divided into testing dataset along with 26 normal samples. The identification accuracy of our ResNet based approach in detecting the infarct core on non-contrast CT is 75.9%. The DCA shows that this deep learning method is capable of improving the net benefit of ischemic stroke patients.

CONCLUSIONS

Our deep learning residual network assisted with optimization methods is able to detect early infarct core on non-contrast CT images and has the potential to help physicians improve diagnostic accuracy in acute ischemic stroke patients.

Footprint of a Bioresorbable Vascular Scaffold in Computed Tomography Coronary Angiogram at 5-Year Follow-up

The Absorb bioresorbable vascular scaffold (Abbott Vascular) does not have an artifact on computed tomography coronary angiography (CTCA); the extent/location of the stent in situ can only be assessed by localizing its radiopaque platinum markers in a non-contrast CTCA. The characteristic appearance of BVS on CTCA should be interpreted as the footprint of a resorbed BVS, instead of a calcified plaque.

Semi-automatic measurement of intracranial hemorrhage growth on non-contrast CT

BACKGROUND

Manual segmentations of intracranial hemorrhage on non-contrast CT images are the gold-standard in measuring hematoma growth but are prone to rater variability.

AIMS

We demonstrate that a convex optimization-based interactive segmentation approach can accurately and reliably measure intracranial hemorrhage growth.

METHODS

Baseline and 16-h follow-up head non-contrast CT images of 46 subjects presenting with intracranial hemorrhage were selected randomly from the ANNEXA-4 trial imaging database. Three users semi-automatically segmented intracranial hemorrhage to measure hematoma volume for each timepoint using our proposed method. Segmentation accuracy was quantitatively evaluated compared to manual segmentations by using Dice similarity coefficient, Pearson correlation, and Bland-Altman analysis. Intra- and inter-rater reliability of the Dice similarity coefficient and intracranial hemorrhage volumes and volume change were assessed by the intraclass correlation coefficient and minimum detectable change.

RESULTS

Among the three users, the mean Dice similarity coefficient, Pearson correlation, and mean difference ranged from 76.79% to 79.76%, 0.970 to 0.980 (p < 0.001), and -1.5 to -0.4 ml, respectively, for all intracranial hemorrhage segmentations. Inter-rater intraclass correlation coefficients between the three users for Dice similarity coefficient and intracranial hemorrhage volume were 0.846 and 0.962, respectively, and the corresponding minimum detectable change was 2.51 ml. Inter-rater intraclass correlation coefficient for intracranial hemorrhage volume change ranged from 0.915 to 0.958 for each user compared to manual measurements, resulting in an minimum detectable change range of 2.14 to 4.26 ml.

CONCLUSIONS

We spatially and volumetrically validate a novel interactive segmentation method for delineating intracranial hemorrhage on head non-contrast CT images. Good spatial overlap, excellent volume correlation, and good repeatability suggest its usefulness for measuring intracranial hemorrhage volume and volume change on non-contrast CT images.

Recommendations for Processing Head CT Data

Many research applications of neuroimaging use magnetic resonance imaging (MRI). As such, recommendations for image analysis and standardized imaging pipelines exist. Clinical imaging, however, relies heavily on X-ray computed tomography (CT) scans for diagnosis and prognosis. Currently, there is only one image processing pipeline for head CT, which focuses mainly on head CT data with lesions. We present tools and a complete pipeline for processing CT data, focusing on open-source solutions, that focus on head CT but are applicable to most CT analyses. We describe going from raw DICOM data to a spatially normalized brain within CT presenting a full example with code. Overall, we recommend anonymizing data with Clinical Trials Processor, converting DICOM data to NIfTI using dcm2niix, using BET for brain extraction, and registration using a publicly-available CT template for analysis.

Evaluation of postoperative hydronephrosis following semirigid ureteroscopy: Incidence and predictors

OBJECTIVE

Hydronephrosis developing following ureteroscopy (URS) is an important issue associated with the long-term postoperative renal functions. Studies investigating the role of postoperative imaging revealed conflicting results. In this study, we aimed to determine the incidence and predictors of hydronephrosis following semirigid URS.

MATERIAL AND METHODS

We evaluated the results of 455 patients who underwent U RS a nd postoperative imaging with non-contrast computed tomography (CT). Primary endpoints of the study were to determine the frequency of development of hydronephrosis and factors associated with the development of hydronephrosis. Logistic regression analysis was used to define factors effecting on the development of hydronephrosis.

RESULTS

Postoperative non-contrast CT revealed hydronephrosis in 81 (17.8%) patients. Stone-free status was achieved in 415 (91.2%) patients. Univariate analysis revealed history of ipsilateral URS (p=0.001), duration of operation (p=0.022), presence of multiple stones (p=0.001), and occurrence of a renal colic episode postoperatively (p=0.013) as the parameters associated with increased risk of postoperative hydronephrosis. In the multivariate analysis, history of ipsilateral URS (OR: 2.724, p=0.017) and presence of multiple stones (OR: 2.116, p=0.032) were found to be the independent prognostic markers of developing postoperative hydronephrosis.

CONCLUSION

Ipsilateral hydronephrosis following URS develops in a significant number of patients. In patients with history of ipsilateral hydronephrosis and multiple stones, risk of development of postoperative hydronephrosis is higher, therefore physicians should be keep these parameters in mind in the decision making process of selective imaging postoperatively.

Cortical venous thrombosis presenting with subarachnoid haemorrhage

BACKGROUND

Our study retrospectively reviewed the presentation, neuro-radiological findings, and outcomes of eight adult patients presenting at our institution with subarachnoid haemorrhage (SAH), which was subsequently proven to be due to cortical venous thrombosis (CVT).

METHODS

We reviewed the case records and neuroimaging findings of eight patients diagnosed with SAH and CVT over a span of two years at our institution, a tertiary care centre in Western India. All details pertaining to their presentation, clinical findings, neuroimaging, management, and outcome following therapy with anticoagulants were collected until patient discharge.

RESULTS

There were a total of eight patients, with the average age being 34 years (range 25-42). Only one patient was female. Six patients had a history of recent binge drinking. None of the patients had a past or family history of common risk factors for thrombosis. All patients presented acutely, with headache (n=6) and seizures (n=6) being the most common presenting features, occurring in three-quarters of the patients examined. Non-contrast computed tomography (NCCT) was the initial imaging study for all but one of the patients and showed cortical SAH (cSAH) without basilar haemorrhage. Magnetic resonance imaging/magnetic resonance venography (MRI/MRV) confirmed the underlying CVT. Unfractionated heparin was used in all cases. Seven patients improved and were discharged on oral anticoagulation. The eighth patient died.

CONCLUSION

Localised cSAH with sparing of basal cisterns can be a presentation for CVT. In patients with cSAH, MRI/MRV can be useful to make a diagnosis of CVT. Anticoagulation for CVT, even in the presence of SAH was related to seven out of eight patients being discharged.

Detection of cardiac myxomas with non-contrast chest CT

BACKGROUND

Cardiac myxomas are sources of systemic embolism. Currently a large volume of chest CT and calcium-scoring CT scans are performed without contrast injection.

PURPOSE

To evaluate the diagnostic capability of non-contrast CT covering heart in detecting cardiac myxomas.

MATERIAL AND METHODS

This retrospective study included 36 non-contrast CT scans of 36 consecutive patients (16 men, 20 women) who underwent CT scan before surgery for left atrial myxomas and 20 patients without myxoma as a control group. Two independent readers who were blinded to medical information reviewed non-contrast CT scans of 36 patients with cardiac myxomas and 20 scans in the control group patients. They determined the presence of lesions suspicious of myxomas using a five-point scale. The other reader measured attenuation number in the non-calcific areas of the tumors and sizes of the masses on the non-contrast CT images.

RESULTS

The average attenuation of cardiac myxoma (22.5 Hounsfield units [HU]; range, 8.9-32.9 HU) and adjacent unopacified blood (44.6 HU; range, 31.5-57 HU) were significantly different (P < 0.001). Twelve cardiac myxomas (31.6%) had internal calcification and all of them were detected by both of readers. Cardiac myxomas were measured smaller on non-contrast CT (mean, 3.5 cm; range, 1.1-9.7 cm) than on pathologic specimens (mean, 4.1 cm, 1.4-10.0 cm) (P < 0.001). Considering grade 3-5 on a five-grade scale as the detectability, the sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of non-contrast CT in detecting cardiac myxomas were 88.8%/86.1%, 95.0%/100%, 96.9%/100%, 82.6%/80.0%, and 91.1%/91.1%, by reader 1 and reader 2, respectively and there was good inter-observer reliability (kappa value = 0.92, P = 0.157).

CONCLUSION

Non-contrast CT scan is useful for detecting cardiac myxomas. Therefore, radiologists should be familiar with imaging findings of cardiac myxomas on non-contrast CT.