What question did this study set out to answer?

The study aims to evaluate the effectiveness of RIP-only Type III HSAT in classifying AHI severity compared to traditional polysomnography and cannula-based assessments.

May 10, 2026

0572 Conclusive AHI from RIP-Only Type III HSAT: Comparison of RIP-Based and Cannula-Based AI Analyses in a Large Multicenter Cohort

Key Points

The study aims to evaluate the effectiveness of RIP-only Type III HSAT in classifying AHI severity compared to traditional polysomnography and cannula-based assessments.
Retrospective multicenter study with 5,771 Type I/II PSGs scored by experts.
Simulated two HSAT configurations: nasal-cannula and RIP-only, processed by AI software to derive AHI.
Computed positive, negative, and overall percentage agreement for AHI thresholds.
For AHI ≥5, RIP-only HSAT had PPA 93.7%, NPA 63.5%, OPA 92.8%; nasal-cannula HSAT had PPA 91.0%, NPA 78.0%, OPA 90.6%.
For AHI ≥15, RIP-only HSAT achieved PPA 81.0%, NPA 91.1%, OPA 83.4%; nasal-cannula PPA was 78.1%, NPA 93.9%, OPA 81.7%.
Overall performance was comparable across AHI thresholds, with minimal differences between configurations.

Abstract

Abstract Introduction Type III home sleep apnea tests (HSAT) rely on respiratory flow measurement to score apneas and hypopneas. Although respiratory inductance plethysmography (RIP) has traditionally been viewed as an effort signal, contemporary implementations can capture respiratory morphology with sufficient fidelity to support RIP as a primary measurement of breathing. We evaluated whether a RIP-only Type III HSAT configuration (without nasal cannula) can deliver conclusive AHI-based severity classification comparable to full-montage attended polysomnography (PSG). Methods We performed a retrospective multicenter study using 5,771 Type I/II PSGs manually scored by registered technologists and board-certified sleep physicians. Two HSAT signal configurations were simulated from the same recordings: (1) a nasal-cannula HSAT configuration (nasal pressure, thoracoabdominal RIP, oximetry) and (2) a RIP-only HSAT configuration (thoracoabdominal RIP, oximetry). Both were processed by DeepRESP v2.0 (K252330), an FDA-cleared medical device, to score sleep/wake, respiratory events, and derive the AHI. PSG-scored AHI served as the reference. Positive-, negative-, and overall percentage agreement (PPA, NPA, OPA) with AHI thresholds (≥5, ≥15, ≥30 events/hour) were computed with 95% confidence intervals. Subgroup summaries characterized heterogeneity across age, sex, body mass index, race/ethnicity, and AHI. Results The cohort was clinically diverse: adults 18–88 years old, 35% female, 58% with body mass index ≥30 kg/m2, with AHI values spanning normal to severe. For AHI ≥5, RIP-only HSAT achieved PPA 93.7%, NPA 63.5%, OPA 92.8%, compared with nasal-cannula HSAT PPA 91.0%, NPA 78.0%, OPA 90.6%. For AHI ≥15, RIP-only HSAT achieved PPA 81.0%, NPA 91.1%, OPA 83.4%, versus nasal-cannula PPA 78.1%, NPA 93.9%, OPA 81.7%. Agreement at AHI ≥30 remained similarly high for both configurations. Across thresholds, RIP-only performance was within a few percentage points of the nasal-cannula configuration. Conclusion RIP-only Type III HSAT analyzed by AI can yield conclusive AHI-based severity classification that closely tracks attended PSG and performs comparably to a nasal-cannula-based HSAT configuration. These findings demonstrate that high-quality RIP signals—without a nasal cannula—can provide robust breathing information for AHI estimation, reducing inconclusive HSAT studies and decreasing the likelihood of repeat testing. Support (if any) None

Mark Helpful

Bookmark

Relay