Fully automated unruptured intracranial aneurysm detection and segmentation from digital subtraction angiography series using an end-to-end spatiotemporal deep neural network

2019 ◽  
Author(s):  
Hailan Jin ◽  
Yin Yin ◽  
Minghui Hu ◽  
Guangming Yang ◽  
Lan Qin
Keyword(s):  
Neural Network ◽  
Intracranial Aneurysm ◽  
Digital Subtraction Angiography ◽  
Deep Neural Network ◽  
Digital Subtraction ◽  
Unruptured Intracranial Aneurysm ◽  
End To End ◽  
Aneurysm Detection

Fully automated intracranial aneurysm detection and segmentation from digital subtraction angiography series using an end-to-end spatiotemporal deep neural network

2020 ◽  
Vol 12 (10) ◽  
pp. 1023-1027
Author(s):  
Hailan Jin ◽  
Jiewen Geng ◽  
Yin Yin ◽  
Minghui Hu ◽  
Guangming Yang ◽  
...  
Keyword(s):  
Neural Network ◽  
Digital Subtraction Angiography ◽  
Network Structure ◽  
Intracranial Aneurysms ◽  
Deep Neural Network ◽  
Short Term Memory ◽  
Medical Image Segmentation ◽  
Digital Subtraction ◽  
Blind Test ◽  
End To End

BackgroundIntracranial aneurysms (IAs) are common in the population and may cause death.ObjectiveTo develop a new fully automated detection and segmentation deep neural network based framework to assist neurologists in evaluating and contouring intracranial aneurysms from 2D+time digital subtraction angiography (DSA) sequences during diagnosis.MethodsThe network structure is based on a general U-shaped design for medical image segmentation and detection. The network includes a fully convolutional technique to detect aneurysms in high-resolution DSA frames. In addition, a bidirectional convolutional long short-term memory module is introduced at each level of the network to capture the change in contrast medium flow across the 2D DSA frames. The resulting network incorporates both spatial and temporal information from DSA sequences and can be trained end-to-end. Furthermore, deep supervision was implemented to help the network converge. The proposed network structure was trained with 2269 DSA sequences from 347 patients with IAs. After that, the system was evaluated on a blind test set with 947 DSA sequences from 146 patients.ResultsOf the 354 aneurysms, 316 (89.3%) were successfully detected, corresponding to a patient level sensitivity of 97.7% at an average false positive number of 3.77 per sequence. The system runs for less than one second per sequence with an average dice coefficient score of 0.533.ConclusionsThis deep neural network assists in successfully detecting and segmenting aneurysms from 2D DSA sequences, and can be used in clinical practice.

scholarly journals Detection and Measurement of Intracranial Aneurysm Compared between Magnetic Resonance Intracranial Black Blood Vessel Imaging and Gold Standard Cerebral Digital Subtraction Angiography

2020 ◽  
Vol 11 (04) ◽  
pp. 545-551
Author(s):  
Dittapong Songsaeng ◽  
Ittichai Sakarunchai ◽  
Sakun Mongkolnaowarat ◽  
Sasithorn Harmontree ◽  
Prapaporn Pornpunyawut ◽  
...  
Keyword(s):  
Magnetic Resonance ◽  
Blood Vessel ◽  
Intracranial Aneurysm ◽  
Digital Subtraction Angiography ◽  
Sensitivity And Specificity ◽  
Gold Standard ◽  
Black Blood ◽  
Digital Subtraction ◽  
Vessel Imaging ◽  
Aneurysm Detection

Abstract Background Magnetic resonance intracranial black blood vessel imaging (MR-IBBVI) is a new noninvasive method for evaluating intracranial vessel wall pathology. No previous studies have investigated the efficacy of MR-IBBVI to determine aneurysm size. We aimed to identify the precise diagnosis of MR-IBBVI for the detection and measurement of intracranial aneurysm compared with gold standard cerebral digital subtraction angiography (cDSA). Materials and Methods The retrospective study collected patients of precoiled or postcoiled intracranial aneurysm who were treated at our institute from January 2012 to June 2019 and who had MR-IBBVI, cDSA imaging, and/or three-dimensional time-of-flight sequence of magnetic resonance angiography. The sensitivity and specificity of aneurysm detection by MR-IBBVI and the accuracy of MR-IBBVI for measuring the aneurysm and vessel size were calculated. Results One hundred and twenty patients (61% female) with 132 aneurysms were included into this study. The mean aneurysm size was 5.3 mm (range: 2.2–22.6). Sensitivity and specificity of MR-IBBVI to detect a small aneurysm were 98.74 and 91.21%, respectively. No statistically significant results were observed between MR-IBBVI and DSA for aneurysm detection or any of the evaluated measurement parameters. Conclusion MR-IBBVI is an accurate and highly sensitive method to detect and evaluate the size of an intracranial aneurysm both before and after coiling.

scholarly journals Value of Three-Dimensional Digital Subtraction Angiography for Detection and Classification of Intracranial Aneurysm Remnants After Clipping

2021 ◽  
Author(s):  
Serge Marbacher ◽  
Matthias Halter ◽  
Deborah R Vogt ◽  
Jenny C Kienzler ◽  
Christian T J Magyar ◽  
...  
Keyword(s):  
Intracranial Aneurysm ◽  
Digital Subtraction Angiography ◽  
Diagnostic Yield ◽  
Three Dimensional ◽  
P Value ◽  
Digital Subtraction ◽  
Patient Demographics ◽  
Included Patient ◽  
2D And 3D

Abstract BACKGROUND The current gold standard for evaluation of the surgical result after intracranial aneurysm (IA) clipping is two-dimensional (2D) digital subtraction angiography (DSA). While there is growing evidence that postoperative 3D-DSA is superior to 2D-DSA, there is a lack of data on intraoperative comparison. OBJECTIVE To compare the diagnostic yield of detection of IA remnants in intra- and postoperative 3D-DSA, categorize the remnants based on 3D-DSA findings, and examine associations between missed 2D-DSA remnants and IA characteristics. METHODS We evaluated 232 clipped IAs that were examined with intraoperative or postoperative 3D-DSA. Variables analyzed included patient demographics, IA and remnant distinguishing characteristics, and 2D- and 3D-DSA findings. Maximal IA remnant size detected by 3D-DSA was measured using a 3-point scale of 2-mm increments. RESULTS Although 3D-DSA detected all clipped IA remnants, 2D-DSA missed 30.4% (7 of 23) and 38.9% (14 of 36) clipped IA remnants in intraoperative and postoperative imaging, respectively (95% CI: 30 [ 12, 49] %; P-value .023 and 39 [23, 55] %; P-value = <.001), and more often missed grade 1 (< 2 mm) clipped remnants (odds ratio [95% CI]: 4.3 [1.6, 12.7], P-value .005). CONCLUSION Compared with 2D-DSA, 3D-DSA achieves a better diagnostic yield in the evaluation of clipped IA. Our proposed method to grade 3D-DSA remnants proved to be simple and practical. Especially small IA remnants have a high risk to be missed in 2D-DSA. We advocate routine use of either intraoperative or postoperative 3D-DSA as a baseline for lifelong follow-up of clipped IA.

scholarly journals An End to End Deep Neural Network for Iris Recognition

2020 ◽  
Vol 174 ◽  
pp. 505-517
Author(s):  
Qingqiao Hu ◽  
Siyang Yin ◽  
Huiyang Ni ◽  
Yisiyuan Huang
Keyword(s):  
Neural Network ◽  
Iris Recognition ◽  
Deep Neural Network ◽  
End To End

PI-Net: An End-to-End Deep Neural Network for Bidirectionally and Directly Fusing Point Clouds With Images

2021 ◽  
Vol 6 (4) ◽  
pp. 8647-8654
Author(s):  
Qi Wang ◽  
Jian Chen ◽  
Jianqiang Deng ◽  
Xinfang Zhang
Keyword(s):  
Neural Network ◽  
Deep Neural Network ◽  
Point Clouds ◽  
End To End

AirNet-SNL: End-to-End Training of Iterative Reconstruction and Deep Neural Network Regularization for Sparse-Data XPCI CT

2021 ◽  
Author(s):  
Dennis J. Lee ◽  
John Mulcahy-Stanislawczyk ◽  
Edward Jimenez ◽  
Derek West ◽  
Ryan Goodner ◽  
...  
Keyword(s):  
Neural Network ◽  
Iterative Reconstruction ◽  
Deep Neural Network ◽  
Sparse Data ◽  
End To End
Sign in / Sign up

Export Citation Format

Share Document