Application of drug-induced sleep endoscopy in patients treated with upper airway stimulation therapy

Troubleshooting Upper Airway Stimulation Therapy Using Drug-Induced Sleep Endoscopy

OBJECTIVE

This study assesses the utility of drug-induced sleep endoscopy (DISE) in guiding further treatment for patients with obstructive sleep apnea (OSA) who have difficulty tolerating upper airway stimulation (UAS) or have inadequate response to therapy.

STUDY DESIGN

We conducted a retrospective analysis of UAS patients at our institution who underwent DISE, post-UAS, and evaluated the efficacy of different electrode configurations and maneuvers.

SETTING

A tertiary care hospital.

METHODS

Out of 379 patients who received UAS therapy, 34 patients who underwent DISE post-UAS (DISE-UAS) were included. Palatal coupling (PC) was assessed with UAS stimulation alone, jaw thrust alone, and both simultaneously during DISE.

RESULTS

Among 34 patients, 5 had suboptimal adherence to UAS therapy, 19 had suboptimal therapy efficacy with residual OSA burden, and 10 had both. During DISE-UAS, PC was observed in 7 patients (21%) with UAS stimulation alone, 9 patients (26%) with jaw thrust alone, and 8 patients (24%) with both maneuvers combined. Notably, 10 patients (29%) did not exhibit PC with any maneuver. Based on DISE-UAS findings, 13 patients were recommended oral appliance therapy (OAT), and 8 patients underwent further surgical interventions.

CONCLUSION

DISE-UAS is a valuable adjunct in troubleshooting UAS therapy for patients intolerant to CPAP or with suboptimal therapy efficacy. This study provides an algorithm for targeted multimodality therapy based on DISE findings, facilitating personalized management approaches.

Predicting Therapeutic Response to Hypoglossal Nerve Stimulation Using Deep Learning

OBJECTIVES

To develop and validate machine learning (ML) and deep learning (DL) models using drug-induced sleep endoscopy (DISE) images to predict the therapeutic efficacy of hypoglossal nerve stimulator (HGNS) implantation.

METHODS

Patients who underwent DISE and subsequent HGNS implantation at a tertiary care referral center were included. Six DL models and five ML algorithms were trained on images from the base of tongue (BOT) and velopharynx (VP) from patients classified as responders or non-responders as defined by Sher's criteria (50% reduction in apnea-hypopnea index (AHI) and AHI < 15 events/h). Precision, recall, F1 score, and overall accuracy were evaluated as measures of performance.

RESULTS

In total, 25,040 images from 127 patients were included, of which 16,515 (69.3%) were from responders and 8,262 (30.7%) from non-responders. Models trained on the VP dataset had greater overall accuracy when compared to BOT alone and combined VP and BOT image sets, suggesting that VP images contain discriminative features for identifying therapeutic efficacy. The VCG-16 DL model had the best overall performance on the VP image set with high training accuracy (0.833), F1 score (0.78), and recall (0.883). Among ML models, the logistic regression model had the greatest accuracy (0.685) and F1 score (0.813).

CONCLUSION

Deep neural networks have potential to predict HGNS therapeutic efficacy using images from DISE, facilitating better patient selection for implantation. Development of multi-institutional data and image sets will allow for development of generalizable predictive models.

LEVEL OF EVIDENCE

N/A Laryngoscope, 2024.