Diagnosis of Acute Central Dizziness With Simple Clinical Information Using Machine Learning

Abstract

Background: Acute dizziness is a common symptom among patients visiting emergency medical centers. Extensive neurological examinations aimed at delineating the cause of dizziness often require experience and specialized training. We tried to diagnose central dizziness by machine learning using only basic clinical information.Methods: Patients were enrolled who had visited an emergency medical center with acute dizziness and underwent diffusion-weighted imaging. The enrolled patients were dichotomized as either having central (with a corresponding central lesion) or non-central dizziness. We obtained patient demographics, risk factors, vital signs, and presentation (non-whirling type dizziness or vertigo). Various machine learning algorithms were used to predict central dizziness. The area under the receiver operating characteristic curve (AUROC) was measured to evaluate diagnostic accuracy. The SHapley Additive exPlanations (SHAP) value was used to explain the importance of each factor.Results: Of the 4,481 visits, 414 (9.2%) were determined as central dizziness. Central dizziness patients were more often older and male and had more risk factors and higher systolic blood pressure. They also presented more frequently with non-whirling type dizziness (79 vs. 54.4%) than non-central dizziness. Catboost model showed the highest AUROC (0.738) with a 94.4% sensitivity and 31.9% specificity in the test set (n = 1,317). The SHAP value was highest for previous stroke presence (mean; 0.74), followed by male (0.33), presentation as non-whirling type dizziness (0.30), and age (0.25).Conclusions: Machine learning is feasible for classifying central dizziness using demographics, risk factors, vital signs, and clinical dizziness presentation, which are obtainable at the triage.

Machine Learning Techniques for Differential Diagnosis of Vertigo and Dizziness: A Review

“…Kim, B.J. et al [68] applied ML algorithms on simple clinical information such as demographics and medical histories, obtained at early stages or emergency centers, can perform a differential diagnosis for vertigo disorders. Such algorithms, once optimized, could assist non-expert physicians treating vertigo in the frontline.…”

“…The studies suggest that there is a need to develop a decision support system (DSS) that can cover a wide range of vertiginous diseases [39,45,68], which should be able to collect the input data into a database that may be later used to retrain models and improve accuracy. Figure 3 illustrates the suggested model of such a decision support system for disease diagnosis.…”

“…The other notations used in the Table 2 are as follows: VS-vestibular schwannoma; BPPV-benign paroxysmal positional vertigo; MD-Meniere disease; SDsudden deafness; TV-traumatic vertigo; VNE-vestibular neuritis; VES-vestibulopatia; BRV-benign recurrent vertigo; CL-central lesion; ANE-acoustic neuroma. Machine learning and statistical techniques applied to other questionnaires and multifeature databases containing relevant information about the patients' history, vestibular test findings, symptoms, etc., have also been used for creating intelligent diagnostic systems [21,24,33,37,39,40,43,45,67,68]. Ahmadi, S.A. et al [21] compared machine learning approaches on multi-feature data sets (including a standardized assessment of symptom features, cardiovascular risk factors, and detailed quantitative testing of ocular motor, vestibular, and postural function) vs. clinical scores such as HINTS, ABCD2 for differential diagnosis of vestibular stroke, and vestibular neuritis.…”

“…Kim, B.J. et al [68] compared various classification models such as SVM, logistic regression, and random forest Catboost to differentiate between central and non-central causes of dizziness by using simple clinical information such as demographics, medical history, systolic and diastolic blood pressure, and heart rate. Additionally, the Shapley additive explanations (SHAP) value was used to explain the importance of each variable in the clinical information for diagnosis.…”

A Dynamic Nomogram to Predict the Risk of Stroke in Emergency Department Patients With Acute Dizziness

“…We searched for predictors of stroke from keynote papers (3,31,(40)(41)(42)(43)(44)(45)(46)(47)(48)(49)(50). Candidate predictors were identified when the factors were consistently reported in predictive research on patients with dizziness, easily ascertained, and routine diagnostic tests for patients with stroke.…”