Evaluation of Ischemic Penumbra in Stroke Patients Based on Deep Learning and Multimodal CT

In order to investigate the value of multimodal CT for quantitative assessment of collateral circulation, ischemic semidark zone, core infarct volume in patients with acute ischemic stroke (AIS), and prognosis assessment in intravenous thrombolytic therapy, segmentation model which is based on the self-attention mechanism is prone to generate attention coefficient maps with incorrect regions of interest. Moreover, the stroke lesion is not clearly characterized, and lesion boundary is poorly differentiated from normal brain tissue, thus affecting the segmentation performance. To address this problem, a primary and secondary path attention compensation network structure is proposed, which is based on the improved global attention upsampling U-Net model. The main path network is responsible for performing accurate lesion segmentation and outputting segmentation results. Likewise, the auxiliary path network generates loose auxiliary attention compensation coefficients, which compensate for possible attention coefficient errors in the main path network. Two hybrid loss functions are proposed to realize the respective functions of main and auxiliary path networks. It is experimentally demonstrated that both the improved global attention upsampling U-Net and the proposed primary and secondary path attention compensation networks show significant improvement in segmentation performance. Moreover, patients with good collateral circulation have a small final infarct area volume and a good clinical prognosis after intravenous thrombolysis. Quantitative assessment of collateral circulation and ischemic semidark zone by multimodal CT can better predict the clinical prognosis of intravenous thrombolysis.

Automatic Detection of Brain Tumor on Computed Tomography Images for Patients in the Intensive Care Unit

Patients in the intensive care unit require fast and efficient handling, including in-diagnosis service. The objectives of this study are to produce a computer-aided system so that it can help radiologists to classify the types of brain tumors suffered by patients quickly and accurately; to build applications that can determine the location of brain tumors from CT scan images; and to get the results of the analysis of the system design. The combination of the zoning algorithm with Learning Vector Quantization can increase the speed of computing and can classify normal and abnormal brains with an average accuracy of 85%.

Machine Learning-Enabled Automated Determination of Acute Ischemic Core From Computed Tomography Angiography

Background and Purpose- The availability of and expertise to interpret advanced neuroimaging recommended in the guideline-based endovascular stroke therapy (EST) evaluation are limited. Here, we develop and validate an automated machine learning-based method that evaluates for large vessel occlusion (LVO) and ischemic core volume in patients using a widely available modality, computed tomography angiogram (CTA). Methods- From our prospectively maintained stroke registry and electronic medical record, we identified patients with acute ischemic stroke and stroke mimics with contemporaneous CTA and computed tomography perfusion (CTP) with RAPID (IschemaView) post-processing as a part of the emergent stroke workup. A novel convolutional neural network named DeepSymNet was created and trained to identify LVO as well as infarct core from CTA source images, against CTP-RAPID definitions. Model performance was measured using 10-fold cross validation and receiver-operative curve area under the curve (AUC) statistics. Results- Among the 297 included patients, 224 (75%) had acute ischemic stroke of which 179 (60%) had LVO. Mean CTP-RAPID ischemic core volume was 23±42 mL. LVO locations included internal carotid artery (13%), M1 (44%), and M2 (21%). The DeepSymNet algorithm autonomously learned to identify the intracerebral vasculature on CTA and detected LVO with AUC 0.88. The method was also able to determine infarct core as defined by CTP-RAPID from the CTA source images with AUC 0.88 and 0.90 (ischemic core ≤30 mL and ≤50 mL). These findings were maintained in patients presenting in early (0-6 hours) and late (6-24 hours) time windows (AUCs 0.90 and 0.91, ischemic core ≤50 mL). DeepSymNet probabilities from CTA images corresponded with CTP-RAPID ischemic core volumes as a continuous variable with r=0.7 (Pearson correlation, P<0.001). Conclusions- These results demonstrate that the information needed to perform the neuroimaging evaluation for endovascular therapy with comparable accuracy to advanced imaging modalities may be present in CTA, and the ability of machine learning to automate the analysis.

Volumetric and Spatial Accuracy of Computed Tomography Perfusion Estimated Ischemic Core Volume in Patients With Acute Ischemic Stroke

Background and Purpose- The volume of estimated ischemic core using computed tomography perfusion (CTP) imaging can identify ischemic stroke patients who are likely to benefit from reperfusion, particularly beyond standard time windows. We assessed the accuracy of pretreatment CTP estimated ischemic core in patients with successful endovascular reperfusion. Methods- Patients from the HERMES (Highly Effective Reperfusion Evaluated in Multiple Endovascular Stroke Trials) and EXTEND-IA TNK (Tenecteplase Versus Alteplase Before Endovascular Therapy for Ischemic Stroke) databases who had pretreatment CTP, >50% angiographic reperfusion, and follow-up magnetic resonance imaging at 24 hours were included. Ischemic core volume on baseline CTP data was estimated using relative cerebral blood flow <30% (RAPID, iSchemaView). Follow-up diffusion magnetic resonance imaging was registered to CTP, and the diffusion lesion was outlined using a semiautomated algorithm. Volumetric and spatial agreement (using Dice similarity coefficient, average Hausdorff distance, and precision) was assessed, and expert visual assessment of quality was performed. Results- In 120 patients, median CTP estimated ischemic core volume was 7.8 mL (IQR, 1.8-19.9 mL), and median diffusion lesion volume at 24 hours was 30.8 mL (IQR, 14.9-67.6 mL). Median volumetric difference was 4.4 mL (IQR, 1.2-12.0 mL). Dice similarity coefficient was low (median, 0.24; IQR, 0.15-0.37). The median precision (positive predictive value) of 0.68 (IQR, 0.40-0.88) and average Hausdorff distance (median, 3.1; IQR, 1.8-5.7 mm) indicated reasonable spatial agreement for regions estimated as ischemic core at baseline. Overestimation of total ischemic core volume by CTP was uncommon. Expert visual review revealed overestimation predominantly in white matter regions. Conclusions- CTP estimated ischemic core volumes were substantially smaller than follow-up diffusion-weighted imaging lesions at 24 hours despite endovascular reperfusion within 2 hours of imaging. This may be partly because of infarct growth. Volumetric CTP core overestimation was uncommon and not related to imaging-to-reperfusion time. Core overestimation in white matter should be a focus of future efforts to improve CTP accuracy.

Smart cloud system with image processing server in diagnosing brain diseases dedicated for hospitals with limited resources

The use of medical imaging in diagnosing brain disease is growing. The challenges are related to the big size of data and complexity of the image processing. High standard of hardware and software are demanded, which can only be provided in big hospitals. Our purpose was to provide a smart cloud system to help diagnosing brain diseases for hospital with limited infrastructure. The expertise of neurologists was first implanted in cloud server to conduct an automatic diagnosis in real time using image processing technique developed based on ITK library and web service. Users upload images through website and the result, in this case the size of tumor was sent back immediately. A specific image compression technique was developed for this purpose. The smart cloud system was able to measure the area and location of tumors, with average size of 19.91 ± 2.38 cm2 and an average response time 7.0 ± 0.3 s. The capability of the server decreased when multiple clients accessed the system simultaneously: 14 ± 0 s (5 parallel clients) and 27 ± 0.2 s (10 parallel clients). The cloud system was successfully developed to process and analyze medical images for diagnosing brain diseases in this case for tumor.

Rigid motion artifact reduction in CT using frequency domain analysis

BACKGROUND

It is often unrealistic to assume that the subject remains stationary during a computed tomography (CT) imaging scan. A patient rigid motion can be decomposed into a translation and a rotation around an origin. How to minimize the motion impact on image quality is important.

OBJECTIVE

To eliminate artifacts caused by patient rigid motion during a CT scan, this study investigated a new method based on frequency domain analysis to estimate and compensate motion impact.

METHODS

Motion parameters was first determined by the magnitude correlation of projections in frequency domain. Then, the estimated parameters were applied to compensate for the motion effects in the reconstruction process. Finally, this method was extended to helical CT.

RESULTS

In fan-beam CT experiments, the simulation results showed that the proposed method was more accurate and faster on the performance of motion estimation than using Helgason-Ludwig consistency condition method (HLCC). Furthermore, the reconstructed images on both simulated and human head experiments indicated that the proposed method yielded superior results in artifact reduction.

CONCLUSIONS

The proposed method is a new tool for patient motion compensation, with a potential for practical application. It is not only applicable to motion correction in fan-beam CT imaging, but also to helical CT.

3D movement correction of CT brain perfusion image data of patients with acute ischemic stroke

INTRODUCTION

Head movement during CT brain perfusion (CTP) acquisition can deteriorate the accuracy of CTP analysis. Most CTP software packages can only correct in-plane movement and are limited to small ranges. The purpose of this study is to validate a novel 3D correction method for head movement during CTP acquisition.

METHODS

Thirty-five CTP datasets that were classified as defective due to head movement were included in this study. All CTP time frames were registered with non-contrast CT data using a 3D rigid registration method. Location and appearance of ischemic area in summary maps derived from original and registered CTP datasets were qualitative compared with follow-up non-contrast CT. A quality score (QS) of 0 to 3 was used to express the degree of agreement. Furthermore, experts compared the quality of both summary maps and assigned the improvement score (IS) of the CTP analysis, ranging from -2 (much worse) to 2 (much better).

RESULTS

Summary maps generated from corrected CTP significantly agreed better with appearance of infarct on follow-up CT with mean QS 2.3 versus mean QS 1.8 for summary maps from original CTP (P = 0.024). In comparison to original CTP data, correction resulted in a quality improvement with average IS 0.8: 17 % worsened (IS = -2, -1), 20 % remained unchanged (IS = 0), and 63 % improved (IS = +1, +2).

CONCLUSION

The proposed 3D movement correction improves the summary map quality for CTP datasets with severe head movement.