歪み生成 otoacoustic emission検査の臨床機器における前方圧力レベル校正による再検査信頼性の向上

Standard in situ calibration using sound pressure level (SPL) measured at the probe microphone for distortion-product otoacoustic emission (DPOAE) testing can cause stimulus level errors up to 20 dB. Forward-pressure level (FPL) calibration, which is derived from a wideband acoustic immittance (WAI) measurement, enables accurate calibration in the presence of the standing-wave nulls that cause these errors. FPL and SPL calibration were compared to determine if FPL calibration made a clinically meaningful difference in test-retest reliability. Measurability, validity, and reliability were assessed on a prototype WAI and DPOAE system designed for clinical use.

Nineteen adults (38 ears) with normal hearing participated in experiment 1 (E1), which measured WAI and DPOAEs at 14 frequencies between 0.7 and 8 kHz (F2/F1 = 1.2, L1/L2 = 65/55 dB SPL) with a 4-sec sampling time. Measurements were made using the same probe fit for both calibration methods. The probe was removed and refitted, and the measurements were repeated. A within-subjects design was used. Experiment 2 (E2) was similar to E1, with 19 normal-hearing adults (33 ears), and DPOAEs measured at eight frequencies between 0.75 and 8 kHz, with a maximum 32-sec sampling time.

The maximum observed difference in the pressure response between SPL and FPL was 22 dB. Refitting the probe resulted in different stimulus pressure responses and a change in the standing-wave null frequency for many participants. DPOAE measurability (the percentage of measurements with a signal to noise ratio greater than 3 dB) was high for both calibration methods. There were no significant differences between the group mean DPOAE levels at each frequency for the two calibration methods for E1, and only a 2 dB difference at 8 kHz for E2. When comparing changes in DPOAE levels between the Test and Retest conditions in E1, many ears showed good consistency, but the range was wider with more extreme outliers (up to 15 dB) for SPL calibration. In both E1 and E2, the standard error of measurement values was similar or lower, and the intraclass correlation coefficients were similar or higher for FPL calibration, especially at higher frequencies, which were more affected by standing-wave nulls. Within-session statistically derived significant emission shift criteria illustrated the magnitude of the difference between the two calibration techniques. The improvement in significant emission shift criteria for FPL calibration ranged from 0.02 to 4.10 dB in E1 and 0.37 to 3.36 dB in E2, across frequencies.

FPL calibration provided clinically meaningful improvements in test-retest reliability with SPL, especially above 4 kHz. With the longer averaging in E2, improvements in reliability were seen down to 1 kHz. Smaller changes in DPOAEs in individual ears can be reliably detected, potentially allowing for earlier intervention with serial monitoring. Clinically significant shift criteria, however, should be established by replicating actual testing conditions using a control group from the population of interest, with the same DPOAE protocol used with patients and a more representative time (on the order of days or months) between the test and retest.