Objective prediction of pharyngeal swallow dysfunction in dysphagia through artificial neural network modeling

• Published in: Neurogastroenterology and motility Vol. 28; no. 3; pp. 336 - 344
• Main Authors: Kritas, S., Dejaeger, E., Tack, J., Omari, T., Rommel, N.
• Format: Journal Article
• Language: English
• Published: England Wiley Subscription Services, Inc 01.03.2016
• Subjects: Adolescent; Adult; Aged; Aged, 80 and over; Computer Simulation; Cross-Sectional Studies; Deglutition Disorders - diagnosis; Deglutition Disorders - physiopathology; dysphagia; Female; high resolution manometry; Humans; impedance measurement; Male; Middle Aged; Neural Networks (Computer); Pharynx - physiopathology; swallowing; Young Adult; high resolution manometry; dysphagia; swallowing; impedance measurement
• ISSN: 1350-1925; 1365-2982
• DOI: 10.1111/nmo.12730

Background Pharyngeal pressure‐flow analysis (PFA) of high resolution impedance‐manometry (HRIM) with calculation of the swallow risk index (SRI) can quantify swallow dysfunction predisposing to aspiration. We explored the potential use of artificial neural networks (ANN) to model the relationship between PFA swallow metrics and aspiration and to predict swallow dysfunction. Methods Two hundred consecutive dysphagia patients referred for videofluoroscopy and HRIM were assessed. Presence of aspiration was scored and PFA software derived 13 metrics and the SRI. An ANN was created and optimized over training cycles to achieve optimal classification accuracy for matching inputs (PFA metrics) to output (presence of aspiration on videofluoroscopy). Application of the ANN returned a value between 0.00 and 1.00 reflecting the degree of swallow dysfunction. Key Results Twenty one patients were excluded due to insufficient number of swallows (<4). Of 179, 58 aspirated and 27 had aspiration pneumonia history. The SRI was higher in aspirators (aspiration 24 [9, 41] vs no aspiration 7 [2, 18], p < 0.001) and patients with pneumonia (pneumonia 27 [5, 42] vs no pneumonia 8 [3, 24], p < 0.05). The ANN Predicted Risk was higher in aspirators (aspiration 0.57 [0.38, 0.82] vs no aspiration 0.13 [0.4, 0.25], p < 0.001) and in patients with pneumonia (pneumonia 0.46 [0.18, 0.60] vs no pneumonia 0.18 [0.6, 0.49], p < 0.01). Prognostic value of the ANN was superior to the SRI. Conclusions & Inferences In a heterogeneous cohort of dysphagia patients, PFA with ANN modeling offers enhanced detection of clinically significant swallowing dysfunction, probably more accurately reflecting the complex interplay of swallow characteristics that causes aspiration. Oropharyngeal dysphagia leading to aspiration is common and often causes respiratory complications. Objective assessment of the biomechanics of swallowing dysfunction in relation to aspiration is needed to understand the mechanism responsible for deglutitive aspiration. In this paper, we aimed to increase the prognostic yield of oropharyngeal pressure‐flow metrics in detection of aspiration by applying an artificial neural network (ANN) model to our data. We found that the prediction of aspiration risk was superior when the analysis was based upon an ANN model.