Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Assessing Patient Eligibility for Inspire Therapy through Machine Learning and Deep Learning Models

Mohsena Chowdhury, Tejas Vyas, Rahul Alapati, Andrés M Bur, Guanghui Wang

TL;DR

The authors address the challenge of selecting suitable candidates for Inspire therapy by predicting responder status using DISE endoscopy videos from BOT and VP regions and associated clinical features. They benchmark six deep learning models and five classical ML algorithms across three datasets from 127 patients, finding that VP-based images generally yield higher predictive accuracy and that DenseNet-169 performs best among DL models. Logistic Regression on clinical data remains competitive, with an accuracy around 0.685 and F1 of 0.813, while AUC values indicate room for improvement and emphasize complementary strengths between DL and ML approaches. The work highlights the potential of imaging- and data-driven approaches to refine candidacy decisions and motivates future multimodal fusion to enhance predictive performance for Inspire therapy eligibility.

Abstract

Inspire therapy is an FDA-approved internal neurostimulation treatment for obstructive sleep apnea. However, not all patients respond to this therapy, posing a challenge even for experienced otolaryngologists to determine candidacy. This paper makes the first attempt to leverage both machine learning and deep learning techniques in discerning patient responsiveness to Inspire therapy using medical data and videos captured through Drug-Induced Sleep Endoscopy (DISE), an essential procedure for Inspire therapy. To achieve this, we gathered and annotated three datasets from 127 patients. Two of these datasets comprise endoscopic videos focused on the Base of the Tongue and Velopharynx. The third dataset composes the patient's clinical information. By utilizing these datasets, we benchmarked and compared the performance of six deep learning models and five classical machine learning algorithms. The results demonstrate the potential of employing machine learning and deep learning techniques to determine a patient's eligibility for Inspire therapy, paving the way for future advancements in this field.

Assessing Patient Eligibility for Inspire Therapy through Machine Learning and Deep Learning Models

TL;DR

The authors address the challenge of selecting suitable candidates for Inspire therapy by predicting responder status using DISE endoscopy videos from BOT and VP regions and associated clinical features. They benchmark six deep learning models and five classical ML algorithms across three datasets from 127 patients, finding that VP-based images generally yield higher predictive accuracy and that DenseNet-169 performs best among DL models. Logistic Regression on clinical data remains competitive, with an accuracy around 0.685 and F1 of 0.813, while AUC values indicate room for improvement and emphasize complementary strengths between DL and ML approaches. The work highlights the potential of imaging- and data-driven approaches to refine candidacy decisions and motivates future multimodal fusion to enhance predictive performance for Inspire therapy eligibility.

Abstract

Inspire therapy is an FDA-approved internal neurostimulation treatment for obstructive sleep apnea. However, not all patients respond to this therapy, posing a challenge even for experienced otolaryngologists to determine candidacy. This paper makes the first attempt to leverage both machine learning and deep learning techniques in discerning patient responsiveness to Inspire therapy using medical data and videos captured through Drug-Induced Sleep Endoscopy (DISE), an essential procedure for Inspire therapy. To achieve this, we gathered and annotated three datasets from 127 patients. Two of these datasets comprise endoscopic videos focused on the Base of the Tongue and Velopharynx. The third dataset composes the patient's clinical information. By utilizing these datasets, we benchmarked and compared the performance of six deep learning models and five classical machine learning algorithms. The results demonstrate the potential of employing machine learning and deep learning techniques to determine a patient's eligibility for Inspire therapy, paving the way for future advancements in this field.
Paper Structure (8 sections, 3 figures, 6 tables)

This paper contains 8 sections, 3 figures, 6 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Dataset
  4. Methods
  5. Experiments
  6. Evaluation of Deep Learning Models
  7. Evaluation of Machine Learning Models
  8. Conclusion

Figures (3)

  • Figure 1: The VOTE score criteria and classification kastoer2018comparisonqian2016classification.
  • Figure 2: Sample image frames of four patients sampled from their corresponding BOT and VP sequences. Two patients are responders (1st row) and two are non-responders (2nd row).
  • Figure 3: The precision and recall of all models on BOT (left) and VP (right) datasets.