Top

Published in:

Open Access 01-12-2020 | Magnetic Resonance Imaging | Original Communication

Modern machine-learning can support diagnostic differentiation of central and peripheral acute vestibular disorders

Authors: Seyed-Ahmad Ahmadi, Gerome Vivar, Nassir Navab, Ken Möhwald, Andreas Maier, Hristo Hadzhikolev, Thomas Brandt, Eva Grill, Marianne Dieterich, Klaus Jahn, Andreas Zwergal

Published in: Journal of Neurology | Special Issue 1/2020

Abstract

Background

Diagnostic classification of central vs. peripheral etiologies in acute vestibular disorders remains a challenge in the emergency setting. Novel machine-learning methods may help to support diagnostic decisions. In the current study, we tested the performance of standard and machine-learning approaches in the classification of consecutive patients with acute central or peripheral vestibular disorders.

Methods

40 Patients with vestibular stroke (19 with and 21 without acute vestibular syndrome (AVS), defined by the presence of spontaneous nystagmus) and 68 patients with peripheral AVS due to vestibular neuritis were recruited in the emergency department, in the context of the prospective EMVERT trial (EMergency VERTigo). All patients received a standardized neuro-otological examination including videooculography and posturography in the acute symptomatic stage and an MRI within 7 days after symptom onset. Diagnostic performance of state-of-the-art scores, such as HINTS (Head Impulse, gaze-evoked Nystagmus, Test of Skew) and ABCD² (Age, Blood, Clinical features, Duration, Diabetes), for the differentiation of vestibular stroke vs. peripheral AVS was compared to various machine-learning approaches: (i) linear logistic regression (LR), (ii) non-linear random forest (RF), (iii) artificial neural network, and (iv) geometric deep learning (Single/MultiGMC). A prospective classification was simulated by ten-fold cross-validation. We analyzed whether machine-estimated feature importances correlate with clinical experience.

Results

Machine-learning methods (e.g., MultiGMC) outperform univariate scores, such as HINTS or ABCD², for differentiation of all vestibular strokes vs. peripheral AVS (MultiGMC area-under-the-curve (AUC): 0.96 vs. HINTS/ABCD² AUC: 0.71/0.58). HINTS performed similarly to MultiGMC for vestibular stroke with AVS (AUC: 0.86), but more poorly for vestibular stroke without AVS (AUC: 0.54). Machine-learning models learn to put different weights on particular features, each of which is relevant from a clinical viewpoint. Established non-linear machine-learning methods like RF and linear methods like LR are less powerful classification models (AUC: 0.89 vs. 0.62).

Conclusions

Established clinical scores (such as HINTS) provide a valuable baseline assessment for stroke detection in acute vestibular syndromes. In addition, machine-learning methods may have the potential to increase sensitivity and selectivity in the establishment of a correct diagnosis.

Newman-Toker DE, Hsieh Y-H, Camargo CA et al (2008) Spectrum of dizziness visits to US emergency departments: cross-sectional analysis from a nationally representative sample. Mayo Clin Proc 83:765–775. https://doi.org/10.4065/83.7.765CrossRefPubMed

Royl G, Ploner CJ, Leithner C (2011) Dizziness in the emergency room: diagnoses and misdiagnoses. Eur Neurol 66:256–263. https://doi.org/10.1159/000331046CrossRefPubMed

Tarnutzer AA, Lee S-H, Robinson KA et al (2017) ED misdiagnosis of cerebrovascular events in the era of modern neuroimaging: a meta-analysis. Neurology 88:1468–1477. https://doi.org/10.1212/WNL.0000000000003814CrossRefPubMedPubMedCentral

Atzema CL, Grewal K, Lu H et al (2016) Outcomes among patients discharged from the emergency department with a diagnosis of peripheral vertigo: outcomes in patients discharged with peripheral vestibular disorders. Ann Neurol 79:32–41. https://doi.org/10.1002/ana.24521CrossRefPubMed

Saber Tehrani AS, Kattah JC, Mantokoudis G et al (2014) Small strokes causing severe vertigo: frequency of false-negative MRIs and nonlacunar mechanisms. Neurology 83:169–173. https://doi.org/10.1212/WNL.0000000000000573CrossRefPubMedPubMedCentral

Eagles D, Stiell IG, Clement CM et al (2008) International survey of emergency physicians’ priorities for clinical decision rules. Acad Emerg Med 15:177–182. https://doi.org/10.1111/j.1553-2712.2008.00035.xCrossRefPubMed

Zwergal A, Dieterich M (2020) Vertigo and dizziness in the emergency room. Curr Opin Neurol 33:117–125. https://doi.org/10.1097/WCO.0000000000000769CrossRefPubMed

Choi K-D, Kim J-S (2019) Vascular vertigo: updates. J Neurol 266:1835–1843. https://doi.org/10.1007/s00415-018-9040-3CrossRefPubMed

Kattah JC, Talkad AV, Wang DZ et al (2009) HINTS to diagnose stroke in the acute vestibular syndrome: three-step bedside oculomotor examination more sensitive than early MRI diffusion-weighted imaging. Stroke 40:3504–3510. https://doi.org/10.1161/STROKEAHA.109.551234CrossRefPubMedPubMedCentral

10.

Newman-Toker DE, Tehrani ASS, Mantokoudis G et al (2013) Quantitative video-oculography to help diagnose stroke in acute vertigo and dizziness: toward an ECG for the eyes. Stroke 44:1158–1161. https://doi.org/10.1161/STROKEAHA.111.000033CrossRefPubMedPubMedCentral

11.

Mantokoudis G, Saber Tehrani AS, Wozniak A et al (2015) VOR gain by head impulse video-oculography differentiates acute vestibular neuritis from stroke. Otol Neurotol 36:457–465. https://doi.org/10.1097/MAO.0000000000000638CrossRefPubMed

12.

Navi BB, Kamel H, Shah MP et al (2012) Application of the ABCD ² score to identify cerebrovascular causes of dizziness in the emergency department. Stroke 43:1484–1489. https://doi.org/10.1161/STROKEAHA.111.646414CrossRefPubMed

13.

Newman-Toker DE, Kerber KA, Hsieh Y-H et al (2013) HINTS outperforms ABCD2 to screen for stroke in acute continuous vertigo and dizziness. Acad Emerg Med 20:986–996. https://doi.org/10.1111/acem.12223CrossRefPubMed

14.

Choi J-H, Oh EH, Park M-G et al (2018) Early MRI-negative posterior circulation stroke presenting as acute dizziness. J Neurol 265:2993–3000. https://doi.org/10.1007/s00415-018-9097-zCrossRefPubMed

15.

Möhwald K, Bardins S, Müller H-H et al (2017) Protocol for a prospective interventional trial to develop a diagnostic index test for stroke as a cause of vertigo, dizziness and imbalance in the emergency room (EMVERT study). BMJ Open 7:e019073. https://doi.org/10.1136/bmjopen-2017-019073CrossRefPubMedPubMedCentral

16.

Herdman M, Gudex C, Lloyd A et al (2011) Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res 20:1727–1736. https://doi.org/10.1007/s11136-011-9903-xCrossRefPubMedPubMedCentral

17.

Jacobson GP, Newman CW (1990) The development of the dizziness handicap inventory. Arch Otolaryngol Head Neck Surg 116:424–427. https://doi.org/10.1001/archotol.1990.01870040046011CrossRefPubMed

18.

Zwergal A, Rettinger N, Frenzel C et al (2009) A bucket of static vestibular function. Neurology 72:1689–1692. https://doi.org/10.1212/WNL.0b013e3181a55ecfCrossRefPubMed

19.

Dieterich M, Brandt T (2019) Perception of verticality and vestibular disorders of balance and falls. Front Neurol 10:172. https://doi.org/10.3389/fneur.2019.00172CrossRefPubMedPubMedCentral

20.

Fawcett T (2006) An introduction to ROC analysis. Pattern Recogn Lett 27:861–874. https://doi.org/10.1016/j.patrec.2005.10.010CrossRef

21.

Johnston SC, Rothwell PM, Nguyen-Huynh MN et al (2007) Validation and refinement of scores to predict very early stroke risk after transient ischaemic attack. The Lancet 369:283–292. https://doi.org/10.1016/S0140-6736(07)60150-0CrossRef

22.

Josephson SA, Sidney S, Pham TN et al (2008) Higher ABCD2 score predicts patients most likely to have true transient ischemic attack. Stroke 39:3096–3098. https://doi.org/10.1161/STROKEAHA.108.514562CrossRefPubMed

23.

Friedman J, Hastie T, Tibshirani R (2010) Regularization paths for generalized linear models via coordinate descent. J Stat Soft. https://doi.org/10.18637/jss.v033.i01

24.

Kim S-J, Koh K, Lustig M et al (2007) An interior-point method for large-scale -regularized least squares. IEEE J Sel Top Signal Process 1:606–617. https://doi.org/10.1109/JSTSP.2007.910971CrossRef

25.

Hastie T, Tibshirani R, Friedman J (2009) The elements of statistical learning. Springer, New YorkCrossRef

26.

Criminisi A, Konukoglu E, Shotton J (2011) Decision forests for classification, regression, density estimation. Manifold Learning and Semi-Supervised Learning, Microsoft Technical Report

27.

Yanase J, Triantaphyllou E (2019) A systematic survey of computer-aided diagnosis in medicine: past and present developments. Expert Syst Appl 138:112821. https://doi.org/10.1016/j.eswa.2019.112821CrossRef

28.

Krafczyk S, Tietze S, Swoboda W et al (2006) Artificial neural network: a new diagnostic posturographic tool for disorders of stance. Clin Neurophysiol 117:1692–1698. https://doi.org/10.1016/j.clinph.2006.04.022CrossRefPubMed

29.

Pradhan C, Wuehr M, Akrami F et al (2015) Automated classification of neurological disorders of gait using spatio-temporal gait parameters. J Electromyograph Kinesiol 25:413–422. https://doi.org/10.1016/j.jelekin.2015.01.004CrossRef

30.

Ahmadi S-A, Vivar G, Frei J et al (2019) Towards computerized diagnosis of neurological stance disorders: data mining and machine learning of posturography and sway. J Neurol 266:108–117. https://doi.org/10.1007/s00415-019-09458-yCrossRefPubMed

31.

Lin D, Vasilakos AV, Tang Y, Yao Y (2016) Neural networks for computer-aided diagnosis in medicine: a review. Neurocomputing 216:700–708. https://doi.org/10.1016/j.neucom.2016.08.039CrossRef

32.

Bronstein MM, Bruna J, LeCun Y et al (2017) Geometric deep learning: going beyond Euclidean data. IEEE Signal Process Mag 34:18–42. https://doi.org/10.1109/MSP.2017.2693418CrossRef

33.

Parisot S, Ktena SI, Ferrante E et al (2018) Disease prediction using graph convolutional networks: application to autism spectrum disorder and Alzheimer’s disease. Med Image Anal 48:117–130. https://doi.org/10.1016/j.media.2018.06.001CrossRefPubMed

34.

Kazi A, Shekarforoush S, Arvind Krishna S et al (2019) Graph convolution based attention model for personalized disease prediction. In: Shen D, Liu T, Peters TM, et al. (eds) Medical image computing and computer assisted intervention – MICCAI 2019. Springer, Cham, pp 122–130CrossRef

35.

Kazi A, Krishna SA, Shekarforoush S et al (2019) Self-attention equipped graph convolutions for disease prediction. In: 2019 IEEE 16th international symposium on biomedical imaging (ISBI 2019). IEEE, Venice, Italy, pp 1896–1899

36.

Vivar G, Zwergal A, Navab N, Ahmadi S-A (2018) Multi-modal disease classification in incomplete datasets using geometric matrix completion. In: Stoyanov D, Taylor Z, Ferrante E, et al. (eds) Graphs in biomedical image analysis and integrating medical imaging and non-imaging modalities. Springer International Publishing, Cham, pp 24–31CrossRef

37.

Little RJ, D’Agostino R, Cohen ML et al (2012) The prevention and treatment of missing data in clinical trials. N Engl J Med 367:1355–1360. https://doi.org/10.1056/NEJMsr1203730CrossRefPubMedPubMedCentral

38.

Vivar G, Kazi A, Zwergal A, et al Simultaneous imputation and disease classification in incomplete medical datasets using Multigraph Geometric Matrix Completion (MGMC). arXiv:2005.06935v1 [cs.LG]

39.

Pedregosa F, Varoquaux G, Gramfort A et al (2011) Scikit-learn: machine learning in python. J Mach Learn Res 12:2825–2830

40.

Paszke A, Gross S, Massa F, et al (2019) PyTorch: an imperative style, high-performance deep learning library. In: Wallach H, Larochelle H, Beygelzimer A, et al (eds) Advances in neural information processing systems 32. Curran Associates, Inc., pp 8026–8037

41.

Bishop CM (2006) Pattern recognition and machine learning (information science and statistics). Springer, New York

42.

Molnar C (2019) Interpretable machine learning: a guide for making black box models explainable

43.

Breiman L, Friedman JH, Olshen RA, Stone CJ (1984) Classification and regression trees, ed. 1. Chapman and Hall/CRC, New York

44.

Tarnutzer AA, Berkowitz AL, Robinson KA et al (2011) Does my dizzy patient have a stroke? A systematic review of bedside diagnosis in acute vestibular syndrome. Can Med Assoc J 183:E571–E592. https://doi.org/10.1503/cmaj.100174CrossRef

45.

Brandt T, Dieterich M (2017) The dizzy patient: don’t forget disorders of the central vestibular system. Nat Rev Neurol 13:352–362. https://doi.org/10.1038/nrneurol.2017.58CrossRefPubMed

46.

Baier B, Stoeter P, Dieterich M (2009) Anatomical correlates of ocular motor deficits in cerebellar lesions. Brain 132:2114–2124. https://doi.org/10.1093/brain/awp165CrossRefPubMed

47.

Zwergal A, Möhwald K, Salazar Lopez E et al (2020) A prospective analysis of lesion-symptom relationships in acute vestibular and ocular motor stroke. Front Neurol. (accepted)

48.

Glasauer S, Dieterich M, Brandt T (2018) Neuronal network-based mathematical modeling of perceived verticality in acute unilateral vestibular lesions: from nerve to thalamus and cortex. J Neurol 265:101–112. https://doi.org/10.1007/s00415-018-8909-5CrossRefPubMed

49.

Carmona S, Martínez C, Zalazar G et al (2016) The diagnostic accuracy of truncal ataxia and HINTS as cardinal signs for acute vestibular syndrome. Front Neurol. https://doi.org/10.3389/fneur.2016.00125CrossRefPubMedPubMedCentral

Title: Modern machine-learning can support diagnostic differentiation of central and peripheral acute vestibular disorders
Authors: Seyed-Ahmad Ahmadi
Gerome Vivar
Nassir Navab
Ken Möhwald
Andreas Maier
Hristo Hadzhikolev
Thomas Brandt
Eva Grill
Marianne Dieterich
Klaus Jahn
Andreas Zwergal
Publication date: 01-12-2020
Publisher: Springer Berlin Heidelberg
Keywords: Magnetic Resonance Imaging
Magnetic Resonance Imaging
Stroke
Spontaneous Nystagmus
Nystagmus
Gaze-Evoked Nystagmus
Published in: Journal of Neurology / Issue Special Issue 1/2020
Print ISSN: 0340-5354
Electronic ISSN: 1432-1459
DOI: https://doi.org/10.1007/s00415-020-09931-z

At a glance: The ONWARDS insulin icodec trials

Springer Medicine

Modern machine-learning can support diagnostic differentiation of central and peripheral acute vestibular disorders

Abstract

Background

Methods

Results

Conclusions

At a glance: The ONWARDS insulin icodec trials

Springer Medicine

Abstract

Background

Methods

Results

Conclusions

Please log in to get access to this content

Other articles of this Special Issue 1/2020

Primary or secondary chronic functional dizziness: does it make a difference? A DizzyReg study in 356 patients

Comparison of three video head impulse test systems for the diagnosis of bilateral vestibulopathy

Computerized clinical decision system and mobile application with expert support to optimize management of vertigo in primary care: study protocol for a pragmatic cluster-randomized controlled trial

Cognitive functions in acute unilateral vestibular loss

Bilateral vestibulopathy: beyond imbalance and oscillopsia

Modulatory effects of magnetic vestibular stimulation on resting-state networks can be explained by subject-specific orientation of inner-ear anatomy in the MR static magnetic field