Evaluation of the accuracy of manual and automatic scoring of a single airflow channel in patients with a high probability of obstructive sleep apnea

Abstract

SummaryBackgroundApneaLink™ (AL) is a single-channel type-4 device that measures airflow. A limited number of studies have assessed AL’s usefulness in diagnosing obstructive sleep apnea (OSA) using automated scoring alone. This study was conducted to assess the sensitivity and specificity of AL in a selected group of people with clinical suspicion of OSA, using both automatic and manual scoring and comparing the results with those obtained for polysomnography (PSG).Material/MethodsSimultaneous overnight recordings of in-laboratory PSG and AL were performed for 95 patients (mean age 46.3±12.6 yr) with a high clinical suspicion of OSA. PSG was scored manually according to the American Academy of Sleep Medicine (AASM) guidelines. AL data were analyzed automatically using a manufacturer-provided computerized algorithm. Manual scoring of the AL flow signal followed the AASM guidelines for reduction in flow.ResultsThe mean apnea hypopneas index (AHI) for PSG was 34.1±32.4/hr. The AL Auto-AHI was 20.1±25.2/h, and the AL Manual-AHI was 39.5±30.4/h. The Pearson correlation coefficients were r=0.883 between PSG-AHI and AL Auto-AHI, and r=0.966 between PSG-AHI and AL Manual-AHI. At AHIs of 5, 10, 15, and 30, the AL Auto sensitivity/specificity was 0.79/0.68, 0.70/0.89, 0.64/0.94 and 0.63/0.98, and the AL Manual sensitivity/specificity was 1.00/0.43, 1.00/0.56, 0.98/0.58 and 1.00/0.80.ConclusionsCombining auto and manual scoring of data (automatic scoring followed by manual scoring) recorded by single-channel ApneaLink™ provides good diagnostic agreement with conventional PSG recordings.

Misclassification of OSA Severity With Automated Scoring of Home Sleep Recordings

“…OSA severity can also aff ect automated scoring accuracy, which improves with increasing disease severity. 10 Not surprisingly, visual (manual) scoring of portable monitoring data is superior to automated scoring and improves the reliability of automated scoring for both type 3 [11][12][13][14][15][16] and type 4 [17][18][19][20][21][22][23] portable monitors when AHI is used to identify and classify OSA. Similar to the current fi ndings, previously examined automated algorithms tend to underestimate the AHI compared with manual scoring.…”

Validation of a sheet-shaped body vibrometer for screening of obstructive sleep apnea

“…Moreover, if the cutoff value was set at 5/h for both REI_eTST and REI_TIB, the specificities and kappa coefficients for predicting AHI ≥ 5 were clearly low with the 2 REIs. These results suggest that screening of AHI ≥ 5 with the SBV may be difficult, a problem that has been noted with wearable PMs (12,27,28). In contrast, when the REI value was set to 17.0 for REI_eTST or 15.9 for REI_TIB, the sensitivities and specificities for AHI ≥ 15 were good.…”

The SAVE Trial: Has the Importance of CPAP for Preventing Cardiovascular Events been Discounted?

“…Choosing the appropriate population is extremely important when considering the use of a portable monitoring device to diagnose OSA. The previous studies that showed positive outcomes with AL have been done on patients who had a high likelihood of having OSA, in addition to excluding patients with confounding disorders such as respiratory diseases, cardiac diseases, and cerebrovascular diseases [5,6]. These findings, therefore, raise the question of whether AL was the appropriate diagnostic tool in this study.…”

“…Furthermore, automatic algorithms were used to analyze signals [8]. A previous study that assessed the accuracy of auto and manual scoring of AL raw data in diagnosing patients with moderate-to-severe OSA after excluding patients with CVD and cerebrovascular diseases found that AL manual scoring showed excellent sensitivity, but low specificity, and AL auto scoring showed very good specificity, but low sensitivity [5]. In an important trial such as the McEvoy et al study, which has far reaching implications for patient's care, management, and prognosis, we believe that detailed information about the scoring techniques and criteria should have been provided.…”

“…This sample selection is not representative of the typical age profile of OSA sufferers. Most previous studies that validated AL in OSA patients included younger middle-aged patients [5,6,10,11]. In addition, the study recruited Caucasians and Asians.…”