global
Variables
Utilities
CUSTOM STYLES
Research
Review
Clinical Artificial Intelligence Applications in Radiology: Neuro

Clinical Artificial Intelligence Applications in Radiology: Neuro

Authors:

Felipe Kitamura; Ian Pan; Suely Fazio Ferraciolli; Kristen W. Yeom; Nitamar Abdala

Publication

Radiologic Clinics of North America - Volume 59, Number 6 - https://doi.org/10.1016/j.rcl.2021.07.002

Download PDF

Table of contents

Heading 2
Heading 3
Heading 4
Heading 5
Heading 6

Summary

The most recent advances of AI in neuroradiology are presented.

  • These include applications related to differential diagnosis, image acquisition, prediction of genetic mutations, lesion quantification, identification of critical findings, prognostication, and others.
  • A brief review of machine learning competitions in neuroradiology is given.
  • The first case of reimbursement for an AI algorithm is described.

‍

Neuroradiology examples

The number of AI applications in neuroradiology is increasing every day, which include differential diagnosis of diseases, improvements in image acquisition (both quality and time), prediction of genetic mutations from MR imaging, segmentation of anatomy to guide interventional procedures, segmentation to quantify CNS lesions, identification of critical findings to shorten notification time, prognostication of diseases, quality assurance of patient position during image acquisition, and many

Machine Learning Competitions

ML challenges are complementary to hypothesis-driven research. Various ML competitions in the health care domain have occurred on many platforms, with Kaggle.com and Grand-challenge.org being the most well-known platforms. The Medical Image Computing and Computer Assisted Intervention Society has organized most ML competitions in neuroimaging. The Radiological Society of North America has launched competitions with the largest, publicly available, expertly annotated datasets comprising images

Centers for Medicare and Medicaid Services approval

At the end of 2020, the CMS granted Viz.ai the first New Technology Add-on Payment (NTAP) for its LVO detection algorithm. Medicare will pay up to $1040 per use in patients with stroke. The decision was based on prospective evidence that the tool improves clinical and financial outcomes, such as shortening the time to treatment and length of stay.66 However, the modified Rankin Scale (mRS) at discharge had no statistically significant improvement in this study.66 Another study (preprint) showed

Reviews

Interested readers are invited to read other published reviews, which keep growing in number every month.2,68, 69, 70, 71, 72, 73, 74, 75 One of the first, “Machine Learning Studies on Major Brain Diseases: 5-Year Trends of 2014-2018,” summarizes the evidence and current limitations of 209 articles published between 2014 and 2018 and emphasizes the limited sample size in most papers.68 Another good review to start with is “Deep Learning in Neuroradiology,” which outlines methods used to develop

Clinics care points

  • There are countless use-cases of AI in neuroradiology, most of which show no proper external validation.
  • Both external validation and prospective clinical trials demonstrating improved outcomes are lacking for neuroradiology AI software.
  • Some models have regulatory approval, but this is not a guarantee of model performance in real-world clinical settings or better health care outcomes.
  • CMS approved the first case of reimbursement for AI software, an LVO identification software that was shown to

Summary

In this article, many promising use-cases of AI in neuroradiology are presented. However, there is a large gap between proof of concept and robust, prospectively validated algorithms. Most AI applications in neuroradiology fall in the former group. Although external validation is increasingly required for medical AI journal publication, this is just the first step toward safe and meaningful use of AI, which requires prospective trials. Regulatory approval does not guarantee model performance in

Disclosure

F.C. Kitamura and I. Pan are consultants for MD.ai. F.C. Kitamura is a speaker for GE Healthcare. The other authors have nothing to disclose.

‍

Keywords:

Medicine | Neuroradiology | Reimbursement | Medical physics | Medicaid | Digitization | Artificial intelligence | Radiology | Neurology | Computer science | Health care | Psychiatry | Computer vision

References

1. Pesapane (2018) Artificial intelligence in medical imaging: threat or opportunity? Radiologists again at the forefront of innovation in medicine. Eur Radiol Exp v. 2 doi:10.1186/s41747-018-0061-6

2. Olthof (2020) Promises of artificial intelligence in neuroradiology: a systematic technographic review. Neuroradiology doi:10.1007/s00234-020-02424-w

3. Prevedello (2019) Challenges Related to Artificial Intelligence Research in Medical Imaging and the Importance of Image Analysis Competitions. Radiol Artif Intell v. 1 p. e180031 doi:10.1148/ryai.2019180031

4. Colak E, Kitamura FC, Hobbs SB, et al. The RSNA Pulmonary Embolism CT (RSPECT) Dataset. Radiol Artif Intell. 2021;0(ja):e200254. doi:10.1148/ryai.2021200254

5. Johnson (2019) MIMIC-CXR-JPG, a large publicly available database of labeled chest radiographs. arXiv v. 14 p. 1

6. Flanders (2020) Construction of a Machine Learning Dataset through Collaboration: The RSNA 2019 Brain CT Hemorrhage Challenge. Radiol Artif Intell v. 2 p. e190211 doi:10.1148/ryai.2020190211

7. Halabi (2018) The RSNA Pediatric Bone Age Machine Learning Challenge. v. 290 p. 498

8. Shih (2019) Augmenting the National Institutes of Health Chest Radiograph Dataset with Expert Annotations of Possible Pneumonia. Radiol Artif Intell v. 1 p. e180041 doi:10.1148/ryai.2019180041

9. Tsai (2021) The RSNA International COVID-19 Open Radiology Database(RICORD). research-article. Radiology v. 0 p. 203957

10. MICCAI. Multi-shell Diffusion MRI Harmonisation Challenge 2018 (MUSHAC). Grand Challenge. 2018. Available at: https://projects.iq.harvard.edu/cdmri2018/challenge. Accessed on July 29, 2021.

11. Walach

12. Crotti

13. Chen (2019) A Multichannel Deep Neural Network Model Analyzing Multiscale Functional Brain Connectome Data for Attention Deficit Hyperactivity Disorder Detection. Radiol Artif Intell v. 2 p. e190012 doi:10.1148/ryai.2019190012

14. Ding (2019) A Deep Learning Model to Predict a Diagnosis of Alzheimer Disease by Using 18 F-FDG PET of the Brain. Radiology v. 290 doi:10.1148/radiol.2018180958

15. Quon (2020) Deep Learning for Pediatric Posterior Fossa Tumor Detection and Classification: A Multi-Institutional Study. AJNR Am J Neuroradiol doi:10.3174/ajnr.A6704

16. Rauschecker (2020) Artificial Intelligence System Approaching Neuroradiologist-level Differential Diagnosis Accuracy at Brain MRI. Radiology v. 295 doi:10.1148/radiol.2020190283

17. Rudie (2020)

18. Ueda (2019) Deep learning for MR angiography: Automated detection of cerebral aneurysms. Radiology v. 290 p. 187 doi:10.1148/radiol.2018180901

19. Hagiwara (2019) Improving the Quality of Synthetic FLAIR Images with Deep Learning Using a Conditional Generative Adversarial Network for Pixel-by-Pixel Image Translation. AJNR Am J Neuroradiol v. 40 doi:10.3174/ajnr.A5927

20. Kim (2018) Improving Arterial Spin Labeling by Using Deep Learning. Radiology v. 287 doi:10.1148/radiol.2017171154

21. Ramalho (2016) T1 Signal-Intensity Increase in the Dentate Nucleus after Multiple Exposures to Gadodiamide: Intraindividual Comparison between 2 Commonly Used Sequences. AJNR Am J Neuroradiol v. 37 doi:10.3174/ajnr.A4757

22. Gong (2018) Deep learning enables reduced gadolinium dose for contrast-enhanced brain MRI. J Magn Reson Imaging v. 48 doi:10.1002/jmri.25970

23. Gupta (2019) Generation of Synthetic CT Images From MRI for Treatment Planning and Patient Positioning Using a 3-Channel U-Net Trained on Sagittal Images. Front Oncol v. 9 doi:10.3389/fonc.2019.00964

24. Montoya (2018) 3D Deep Learning Angiography (3D-DLA) from C-arm Conebeam CT. AJNR Am J Neuroradiol v. 39 doi:10.3174/ajnr.A5597

25. Sreekumari (2019) A Deep Learning-Based Approach to Reduce Rescan and Recall Rates in Clinical MRI Examinations. AJNR Am J Neuroradiol v. 40 doi:10.3174/ajnr.A5926

26. Deguchi (2020) Clinicopathological analysis of T2-FLAIR mismatch sign in lower-grade gliomas. Scientific Rep v. 10

27. Akkus (2017) Predicting Deletion of Chromosomal Arms 1p/19q in Low-Grade Gliomas from MR Images Using Machine Intelligence. J Digit Imaging v. 30 doi:10.1007/s10278-017-9984-3

28. Korfiatis (2017) Residual Deep Convolutional Neural Network Predicts MGMT Methylation Status. J Digit Imaging v. 30 doi:10.1007/s10278-017-0009-z

29. Chang (2018) Deep-learning convolutional neural networks accurately classify genetic mutations in gliomas. AJNR Am J Neuroradiol v. 39 p. 1201 doi:10.3174/ajnr.A5667

30. Fan (2019) Deep Learning-Based Automatic Segmentation of Lumbosacral Nerves on CT for Spinal Intervention: A Translational Study. AJNR Am J Neuroradiol v. 40 doi:10.3174/ajnr.A6070

31. Quon (2020) Deep Learning for Automated Delineation of Pediatric Cerebral Arteries on Pre-operative Brain Magnetic Resonance Imaging. Front Surg v. 7 doi:10.3389/fsurg.2020.517375

32. Wu (2020) Three-Plane–assembled Deep Learning Segmentation of Gliomas. Radiol Artif Intell v. 2 p. 10 doi:10.1148/ryai.2020190011

33. Cai (2020) Fully Automated Segmentation of Head CT Neuroanatomy Using Deep Learning. Radiol Artif Intell v. 2 p. 12 doi:10.1148/ryai.2020190183

34. Quon (2020) Artificial intelligence for automatic cerebral ventricle segmentation and volume calculation: a clinical tool for the evaluation of pediatric hydrocephalus. J Neurosurg Pediatr v. c p. 1

35. Mccoy (2019) Convolutional Neural Network-Based Automated Segmentation of the Spinal Cord and Contusion Injury: Deep Learning Biomarker Correlates of Motor Impairment in Acute Spinal Cord Injury. AJNR Am J Neuroradiol v. 40

36. Park (2019) Deep Learning-Assisted Diagnosis of Cerebral Aneurysms Using the HeadXNet Model. JAMA Netw open v. 2 doi:10.1001/jamanetworkopen.2019.5600

37. Federau (2020) Improved Segmentation and Detection Sensitivity of Diffusion-weighted Stroke Lesions with Synthetically Enhanced Deep Learning. Radiol Artif Intell v. 2 p. 8 doi:10.1148/ryai.2020190217

38. Prevedello (2017) Automated Critical Test Findings Identification and Online Notification System Using Artificial Intelligence in Imaging. Radiology v. 285 doi:10.1148/radiol.2017162664

39. Chilamkurthy (2018) Deep learning algorithms for detection of critical findings in head CT scans: a retrospective study. Lancet v. 392 doi:10.1016/S0140-6736(18)31645-3

40. Nie (2019) Multi-Channel 3D Deep Feature Learning for Survival Time Prediction of Brain Tumor Patients Using Multi-Modal Neuroimages. Scientific Rep v. 9 p. 1

41. Maegerlein (2019) Automated Calculation of the Alberta Stroke Program Early CT Score: Feasibility and Reliability. Radiology v. 291 doi:10.1148/radiol.2019181228

42. Nielsen (2018) Prediction of Tissue Outcome and Assessment of Treatment Effect in Acute Ischemic Stroke Using Deep Learning. Stroke v. 49 doi:10.1161/STROKEAHA.117.019740

43. Narayana (2020) Deep Learning for Predicting Enhancing Lesions in Multiple Sclerosis from Noncontrast MRI. Radiology v. 294 doi:10.1148/radiol.2019191061

44. Hassan (2020) Early experience utilizing artificial intelligence shows significant reduction in transfer times and length of stay in a hub and spoke model. Interv Neuroradiol p. 8

45. Morey (2020) Impact of Viz LVO on Time-to-Treatment and Clinical Outcomes in Large Vessel Occlusion Stroke Patients Presenting to Primary Stroke Centers. medRxiv p. 2

46. Sakai (2019) Machine learning studies on major brain diseases: 5-year trends of 2014-2018. Jpn J Radiol v. 37 doi:10.1007/s11604-018-0794-4

47. Zaharchuk (2018) Deep Learning in Neuroradiology. AJNR Am J Neuroradiol v. 39 doi:10.3174/ajnr.A5543

48. Yao (2020) Deep Learning in Neuroradiology: A Systematic Review of Current Algorithms and Approaches for the New Wave of Imaging Technology. Radiol Artif Intell v. 2 p. e190026 doi:10.1148/ryai.2020190026

49. Lui (2020) Artificial Intelligence in Neuroradiology: Current Status and Future Directions. AJNR Am J Neuroradiol v. 41

50. Rudie (2019) Emerging Applications of Artificial Intelligence in Neuro-Oncology. Radiology v. 290 doi:10.1148/radiol.2018181928

51. Kaka (2021) Artificial Intelligence and Deep Learning in Neuroradiology: Exploring the New Frontier. Can Assoc Radiol J v. 72 doi:10.1177/0846537120954293

52. Soun (2021) Artificial Intelligence and Acute Stroke Imaging. AJNR Am J Neuroradiol v. 42 doi:10.3174/ajnr.A6883

53. Attyé (2019) Artificial neuroradiology: Between human and artificial networks of neurons?. J Neuroradiol v. 46

54. Shi (2020) Artificial Intelligence in the Management of Intracranial Aneurysms: Current Status and Future Perspectives. AJNR Am J Neuroradiol v. 41 doi:10.3174/ajnr.A6468

55. Zeng (2020) Review of Deep Learning Approaches for the Segmentation of Multiple Sclerosis Lesions on Brain MRI. Front neuroinformatics v. 14 doi:10.3389/fninf.2020.610967

Download PDF

Other resources

Research Article

Diffusion-Weighted Imaging with Color-Coded Images: Towards a Reduction in Reading Time While Keeping a Similar Accuracy

The aim of this study was to develop a diagnostic tool capable of providing diffusion and apparent diffusion coefficient (ADC) map information in a single color-coded image and to assess the performance of color-coded images compared with their corresponding diffusion and ADC map. The institutional review board approved this retrospective study, which sequentially enrolled 36 head MRI scans. Diffusion-weighted images (DWI) and ADC maps were compared to their corresponding color-coded images. Four raters had their interobserver agreement measured for both conventional (DWI) and color-coded images. Differences between conventional and color-coded images were also estimated for each of the 4 raters. Cohen’s kappa and percent agreement were used. Also, paired-samples t -test was used to compare reading time for rater 1. Conventional and color-coded images had substantial or almost perfect agreement for all raters. Mean reading time of rater 1 was 47.4 seconds for DWI and 27.9 seconds for color-coded images ( P = .00007). These findings are important because they support the role of color-coded images as being equivalent to that of the conventional DWI in terms of diagnostic capability. Reduction in reading time (which makes the reading easier) is also demonstrated for one rater in this study.

Access for Free
Research Article

The RSNA Pediatric Bone Age Machine Learning Challenge

Purpose The Radiological Society of North America (RSNA) Pediatric Bone Age Machine Learning Challenge was created to show an application of machine learning (ML) and artificial intelligence (AI) in medical imaging, promote collaboration to catalyze AI model creation, and identify innovators in medical imaging. Materials and Methods The goal of this challenge was to solicit individuals and teams to create an algorithm or model using ML techniques that would accurately determine skeletal age in a curated data set of pediatric hand radiographs. The primary evaluation measure was the mean absolute distance (MAD) in months, which was calculated as the mean of the absolute values of the difference between the model estimates and those of the reference standard, bone age. Results A data set consisting of 14 236 hand radiographs (12 611 training set, 1425 validation set, 200 test set) was made available to registered challenge participants. A total of 260 individuals or teams registered on the Challenge website. A total of 105 submissions were uploaded from 48 unique users during the training, validation, and test phases. Almost all methods used deep neural network techniques based on one or more convolutional neural networks (CNNs). The best five results based on MAD were 4.2, 4.4, 4.4, 4.5, and 4.5 months, respectively. Conclusion The RSNA Pediatric Bone Age Machine Learning Challenge showed how a coordinated approach to solving a medical imaging problem can be successfully conducted. Future ML challenges will catalyze collaboration and development of ML tools and methods that can potentially improve diagnostic accuracy and patient care. © RSNA, 2018 Online supplemental material is available for this article. See also the editorial by Siegel in this issue.

Access for Free
Research Article

The Type III Secretion System (T3SS)-Translocon of Atypical Enteropathogenic Escherichia coli (aEPEC) Can Mediate Adherence

The intimin protein is the major adhesin involved in the intimate adherence of atypical enteropathogenic Escherichia coli (aEPEC) strains to epithelial cells, but little is known about the structures involved in their early colonization process. A previous study demonstrated that the type III secretion system (T3SS) plays an additional role in the adherence of an Escherichia albertii strain. Therefore, we assumed that the T3SS could be related to the adherence efficiency of aEPEC during the first stages of contact with epithelial cells. To test this hypothesis, we examined the adherence of seven aEPEC strains and their eae (intimin) isogenic mutants in the standard HeLa adherence assay and observed that all wild-type strains were adherent while five isogenic eae mutants were not. The two eae mutant strains that remained adherent were then used to generate the eae/escN double mutants (encoding intimin and the T3SS ATPase, respectively) and after the adherence assay, we observed that one strain lost its adherence capacity. This suggested a role for the T3SS in the initial adherence steps of this strain. In addition, we demonstrated that this strain expressed the T3SS at significantly higher levels when compared to the other wild-type strains and that it produced longer translocon-filaments. Our findings reveal that the T3SS-translocon can play an additional role as an adhesin at the beginning of the colonization process of aEPEC.

Access for Free