Abstract
Objectives:
Measures of speech-in-noise, such as the QuickSIN, are increasingly common tests of speech perception in audiologic practice. However, the effect of vestibular schwannoma (VS) on speech-in-noise abilities is unclear. Here, we compare the predictive ability of interaural QuickSIN asymmetry for detecting VS against other measures of audiologic asymmetry.
Methods:
A retrospective review of patients in our institution who received QuickSIN testing in addition to a regular audiologic battery between September 2015 and February 2019 was conducted. Records for patients with radiographically confirmed, unilateral, pretreatment VSs were identified. The remaining records excluding conductive pathologies were used as controls. The predictive abilities of various measures of audiologic asymmetry to detect VS were statistically compared.
Results:
Our search yielded 73 unique VS patients and 2423 controls. Receiver operating characteristic curve analysis showed that QuickSIN asymmetry was more sensitive and specific than pure-tone average asymmetry and word-recognition-in-quiet asymmetry for detecting VS. Multiple logistic regression analysis revealed that QuickSIN asymmetry was more predictive of VS (odds ratio [OR] = 1.23, 95% confidence interval [CI] [1.10, 1.38], p < 0.001) than pure-tone average asymmetry (OR = 1.04, 95% CI [1.00, 1.07], p = 0.025) and word-recognition-in-quiet asymmetry (OR = 1.03, 95% CI [0.99, 1.06], p = 0.064).
Conclusion:
Between-ear asymmetries in the QuickSIN appear to be more efficient than traditional measures of audiologic asymmetry for identifying patients with VS. These results suggest that speech-in noise testing could be integrated into clinical practice without hindering the ability to identify retrocochlear pathology.
Keywords: Acoustic neuroma, Audiology, Speech-in noise, Vestibular schwannoma, Word recognition
INTRODUCTION
A vestibular schwannoma (VS) is a tumor that results from a benign growth of Schwann cells surrounding the auditory-vestibular nerve. Early symptoms of a VS are unilateral or asymmetric hearing loss and/or tinnitus, and problems with balance or dizziness (Johnson & Sheehy 1966; Selesnick & Jackler 1992). Thus, audiologic assessment is crucial in the management of VSs and often is the first step taken before diagnosis of the VS. Magnetic resonance imaging (MRI) remains the gold standard for diagnosis of VS. However, audiometric results are a key variable in determining which patients are referred for imaging. As such, audiologists and physicians are trained to look for signs of VS when interpreting the results of an audiogram. On an audiologic basis, the most common symptoms of a unilateral VS are interaural asymmetries in hearing thresholds and/ or word-recognition scores in quiet (Johnson & Sheehy 1966; Selesnick & Jackler 1992). These asymmetries likely reflect compression of the auditory nerve and local secretion of pathological factors from the VS tumor growth (Dilwali et al. 2015; Sagers et al. 2019). However, the optimal protocol for identifying VS patients on the basis of audiologic results remains unclear. Most efforts have focused on asymmetries in pure-tone thresholds to date, with different protocols having varying degrees of effectiveness (Cheng & Wareing 2012; Sweeney et al. 2018; Waterval et al. 2018). The emphasis on pure-tone thresholds has occurred in part because asymmetries in word-recognition in quiet (WRQ) scores are largely ineffective at separating patients with VS from those with cochlear loss when a single list is used (Dirks et al. 1977). The efficacy of WRQ scores improve, however, when a “rollover ratio” is used (Jerger & Jerger 1971; Dirks et al. 1977). “Rollover” refers to a phenomenon when word-recognition scores decrease with increasing presentation level; this decrease was reported to be associated with the presence of retrocochlear pathology, particularly an acoustic neuroma (Jerger & Jerger 1971; Dirks et al. 1977). The “rollover ratio” referred to the ratio between the score obtained at the highest intensity relative to the highest word-recognition score overall, and ratios > 0.45 were thought to most effective at separating patients with cochlear versus retrocochlear disorders. While the idea of “rollover” suggests that asymmetries in word-recognition scores may have utility in identifying patients with possible VSs, the procedure to maximize efficiency of “rollover” is too time-prohibitive for clinical use. For example, Jerger and Jerger (1971) reported that “rollover ratios” were most effective when 6 to 8 word lists of 25 words were presented at intensities varying from ~10 dB above SRT to levels near the uncomfortable loudness level. Despite the time-prohibitive nature of rollover measurements, audiologists and physicians remain attuned to the idea of asymmetries in WRQ scores as a possible sign of VS.
WRQ has been widely examined in patients with VS, because it has been the default test of speech perception in routine audiologic assessment since the inception of audiology. In recent years, however, there is growing recognition that measures of speech-in noise (SIN) should be a routine component of audiologic practice. This occurs in part because the primary concern of most patients with hearing loss is an inability to understand speech in the presence of background noise (Le Prell & Clavier 2017), and traditional audiologic measures are poor predictors of how well patients perform in noise (Wilson 2011; Vermiglio et al. 2018; Fitzgerald et al. 2023). Thus, conventional measures of WRQ scores appear to be insensitive to a key concern of patients. However, for SIN measures to achieve widespread use in routine audiologic practice, key information is needed, such as how SIN performance is affected by auditory pathology such as VS. One possibility is that measures of SIN could be more sensitive to the presence of a VS because SIN testing has been suggested to be a “stress test” for the auditory system (Wilson et al. 2007; Wilson 2011; Vermiglio et al. 2012). By this logic, the performance on SIN measures would be more readily affected by the pathology underlying VS than traditional measures of WRQ. To date, this possibility has not been tested in any systematic manner. Moreover, it is unclear which degree of between-ear asymmetry (if any) in SIN abilities is most accurate in identifying patients with VS. Here, we address these issues by comparing audiologic data between patients with radiographically confirmed, unilateral, pretreatment VS and control patients with sensorineural hearing loss. Our first goal was to characterize WRQ and SIN performance as a function of hearing loss in patients with VS, and a group of control patients without VS. Our second goal was to compare the diagnostic ability of between-ear asymmetries in SIN performance, WRQ scores, and pure-tone thresholds in identifying patients with VS.
Methods
Procedures
This study was approved by the institutional review board of Stanford University School of Medicine (IRB-40822). Data were obtained as part of clinical audiologic evaluations at Stanford Ear Institute (SEI). All tests were completed in a double-walled sound booth using GSI-61 (Grayson-Stadler) audiometers and either ER-3A insert earphones or Sennheiser circumaural headphones. Air-conduction and bone-conduction thresholds were obtained using the modified Hughson-Westlake method (Carhart & Jerger 1959). For purposes of data analysis, all “no-response” values were recorded as 110dB HL.
WRQ scores were tested unilaterally for each ear using NU-6 lists (Tillman & Carhart 1966). WRQ scores were computed by determining the percentage of words correctly repeated by the patient. Typically, 25 words were used to obtain WRQ scores (Auditec NU-6 Form A). In some instances, the difficulty-weighted word lists were utilized (Version II; Hurley & Sells 2003), and if a patient scored 90% or 100% across the first 10 words, testing was discontinued and the WRQ percent correct value was reported.
Speech-in-noise (SIN) scores were obtained unilaterally for each ear by averaging the results from two QuickSIN lists (Killion et al. 2004). In the QuickSIN, each list consisted of six low-context sentences, and contained five key words. Each sentence was subsequently presented with a decreasing signal-to-noise ratio (SNR) in 5dB steps beginning at +25dB, and ending at 0dB SNR (25, 20, 15, 10, 5, and 0). The output of the QuickSIN was the SNR at which 50% of key words in each sentence could be repeated. Two QuickSIN lists were presented for each patient. A single QuickSIN list requires approximately 57 seconds to complete (Wilson et al. 2007). Thus, the completion of two lists added ~ 5 to 6 minutes to the total testing time.
To minimize effects related to the presentation level, the same level was used for both WRQ and SIN measurements in each patient. The default presentation level was 70 dB HL. If part of the signal was inaudible at that level, the audiologists increased the presentation level to maximize audibility while not exceeding uncomfortable loudness level of the patient. Recorded stimuli were used for the speech material in both WRQ and SIN testing.
Participants
Identifying VS Patients
To identify patients with VS, the Stanford Translational Research Integrated Database Environment was queried for International Classification of Disease, Ninth and Tenth Revisions, Clinical Modification (ICD-9-CM and ICD-10-CM) codes for benign neoplasm of cranial nerves (ICD-9-CM 225.1 and ICD-10-CM D33.3) from 2000 to 2019. Resulting medical records were manually examined to identify patients with a diagnosis of VS. Each diagnosis was confirmed by the research team using radiologic data. Audiometric records for confirmed VS patients were obtained from the audiologic database maintained at the Stanford Ear Institute. Only data from individuals who completed bilateral pure-tone audiometry, WRQ, and QuickSIN were included in this study. In patients who had multiple audiograms before treatment, we used data from the earliest record for analysis.
Selecting Control Patients
Audiologic data through 2019 were examined from our audiologic database. Patients under the age of 18 were excluded to avoid developmental effects observed with performance on some tests of SIN ability (Holder et al. 2016). As with the VS patient cohort, we only included data from individuals who completed bilateral pure-tone audiometry, WRQ, and QuickSIN. All patients with any type of retrocochlear pathology were excluded. Patients with conductive or mixed hearing losses were excluded by examining the air-bone gap for 0.5, 1, and 2 kHz. Any patients with an air-bone gap ≥10 dB HL across all three frequencies was excluded from further analysis.
Data Analysis
Pure-tone averages (PTAs) were calculated by averaging thresholds of 0.5, 1, 2, and 3 kHz, as recommended by the American Academy of Otolaryngology-Head and Neck Surgery (Anon 1995). PTA, WRQ, and SNR asymmetries were obtained by calculating the absolute difference between ears for each respective measure. When comparing audiometric performance of VS patients to control patients, VS ears were compared to control right ears to eliminate any bias from using the better versus worse hearing ears in the control patients. The sensitivity and specificity of various asymmetry cutpoints to detect VS were computed using receiver operating characteristic (ROC) curve analysis. The area under the ROC curve (AUC-ROC) was statistically compared between measures of audiometric asymmetry.
Multiple logistic regression analysis was used to compare the measures of audiometric asymmetry in their ability to predict the presence of VS. Since PTA, WRQ, and SNR asymmetry are all measures of hearing and are thus interrelated, a regression model containing interaction terms for PTA and WRQ, PTA and SNR, and WRQ and SNR asymmetries was developed. An estimate was considered significant at α = 0.05. All analyses were performed using STATA 15 (STATA Corp, College Station, TX).
RESULTS
Our search yielded 73 unique patients with radiographically confirmed VS who received pretreatment QuickSIN testing. We subsequently identified 2423 patients without VS or conductive pathology. Demographic and clinical information is shown in Table 1.
TABLE 1.
Summary of patient demographic and clinical information
| All Patients, n = 2496 | VS Patients, n = 73 | Control Patients, n = 2423 | |
|---|---|---|---|
| Age, mean ± SD (years) | 54.31 ± 17.15 | 61.89 ± 15.22 | 54.14 ± 17.16 |
| Female (%) | 49.72 | 50.70 | 49.70 |
| All Ears, n = 4992 | VS Ears, n = 73 | Control Right Ears, n = 2423 | |
| PTA, mean ± SD (dB HL) | 37.91 ± 10.73 | 44.78 ± 14.12 | 37.45 ± 13.15 |
| WRQ (%) | 94.37 ± 11.03 | 78.82 ± 28.29 | 94.26 ± 14.57 |
| QuickSIN SNR loss (dB SNR) | 4.91 ± 4.28 | 9.62 ± 6.90 | 4.83 ± 4.90 |
PTA indicates pure-tone average; SNR, signal-to-noise ratio; VS, vestibular schwannoma; WRQ, word-recognition in quiet.
published online ahead of print August 22, 2023.
Effects of Hearing Loss on Speech Understanding in Quiet and Noise
Tables 2 and 3 show the mean values and ranges for WRQ scores and QuickSIN SNR loss as a function of degree of pure-tone threshold elevation (PTA in dB HL). In both VS and control ears, increasing severity of hearing loss is more likely to be reflected by abnormal QuickSIN SNR performance than by abnormal WRQ performance. Here, abnormal QuickSIN performance (>3 dB SNR loss; Etymotic Research 2006) was observed beginning with slight hearing loss in the VS ears, and mild hearing loss in the controls, and continued to deteriorate with increasing degree of hearing loss. Notably, SNR losses were always greater in the VS cohort than the controls when any degree of hearing loss was present. In contrast, abnormal WRQ performance (<70%; Anon 1995; Meyer et al. 2006) was only observed with moderately severe hearing losses in the VS cohort and severe to profound losses in the control group. Notably, both the QuickSIN SNR loss and WRQ scores were poor in those patients with moderately severe or greater degrees of hearing loss.
TABLE 2.
Comparison of average WRQ scores in VS ears and control right ears
| Number of Patients | (Mean WRQ ± SD Range), % | Meyer Classification | ||||
|---|---|---|---|---|---|---|
| VS Ears | Control Right Ears | VS Ears | Control right ears | VS Ears | Control Right Ears | |
| Normal (0–15 dB HL) |
2 (2.74%) | 103 (4.25%) | 100 ± 0 (100–100) |
99.09 ± 2.61 (90–100) |
Class I | Class I |
| Slight (16–25 dB HL) |
4 (5.48%) | 311 (12.84%) | 99 ± 2 (96–100) |
98.78 ± 6.62 (10–100) |
Class I | Class I |
| Mild (26–40 dB HL) |
18 (24.66%) | 1079 (44.53%) | 97.44 ± 7.76 (68–100) |
98.37 ± 4.87 (0–100) |
Class I | Class I |
| Moderate (41–55 dB HL) |
37 (50.68%) | 711 (29.34%) | 78.05 ± 22.34 (16–100) |
94.12 ± 9.36 (16–100) |
Class I | Class I |
| Mod severe (56–70 dB HL) |
9 (12.33%) | 190 (7.84%) | 53.33 ± 35.67 (0–92) |
73.34 ± 27.09 (0–100) |
Class II | Class I |
| Severe–profound (≥71 dB HL) |
3 (4.1%1) | 29 (1.2%) | 12 ± 14.42 (0–28) |
37.17 ± 37.64 (0–100) |
Class III | Class III |
| Total | 73 (100%) | 2423 (100%) | 78.82 ± 28.29 (0–100) |
94.51 ± 14.15 (0–100) |
Class I | Class I |
Ears were categorized by degree of hearing loss (PTA in dB HL). The Meyer classification for word-recognition scores is also provided for each degree of hearing loss (Meyer et al. 2006).
PTA indicates pure-tone average; VS, vestibular schwannoma; WRQ, word-recognition in quiet.
TABLE 3.
Comparison of average QuickSIN SNR losses in VS ears and control right ears
| Number of Patients | Mean SNR Loss ± SD (Range), dB | QuickSIN SNR Loss Classification | ||||
|---|---|---|---|---|---|---|
| VS Ears | Control Right Ears | VS Ears | Control Right Ears | VS Ears | Control Right Ears | |
| Normal (0–15 dB HL) |
2 (2.74%) | 103 (4.25%) | 2.5 ± 1.41 (1.5–3.5) |
2.65 ± 2.35 (−1.5 to 13.5) |
Normal | Normal |
| Slight (16–25 dB HL) |
4 (5.48%) | 311 (12.84%) | 3.25 ± 1.71 (1–5) |
2.66 ± 2.61 (−3.5 to 18.5) |
Mild | Normal |
| Mild (26–40 dB HL) |
18 (24.66%) | 1079 (44.53%) | 4.39 ± 4.52 (0–18.5) |
3.23 ± 2.98 (−3.5 to 25.5) |
Mild | Mild |
| Moderate (41–55 dB HL) |
37 (50.68%) | 711 (29.34%) | 11.38 ± 6.03 (0–24.5) |
6.09 ± 4.78 (−3.5 to 22.5) |
Moderate | Mild |
| Mod Severe (56–70 dB HL) |
9 (12.33%) | 190 (7.84%) | 15.28 ± 7.63 (6–25.5) |
10.99 ± 6.84 (−3.5 to 25.5) |
Severe | Moderate |
| Severe-profound (≥71 dB HL) |
3 (4.1%1) | 29 (1.2%) | 15.5 ± 3.77 (11.5–19) |
16.79 ± 8.44 (−0.5 to 25.5) |
Severe | Severe |
| Total | 73 (100%) | 2423 (100%) | 9.62 ± 6.9 (0–25.5) |
4.74 ± 4.83 (−3.5 to 25.5) |
Moderate | Mild |
Ears were categorized by degree of hearing loss (PTA in dB HL). The classification of QuickSIN SNR loss is also provided for each degree of hearing loss (Etymotic Research 2006).
PTA indicates pure-tone average; SNR, signal-to-noise ratio; VS, vestibular schwannoma; WRQ, word-recognition in quiet.
Similar results are observed if the criteria for abnormal WRQ scores is shifted to <88% correct. According to a binomial distribution, scores between 88% and 100% are statistically the same for a 25-word list (Carney & Schlauch 2007). If a WRQ deficit is defined as anything that is not excellent (e.g., statistically no different from 100%; see Fitzgerald et al. 2023), then WRQ deficits are only observed on average in VS patients with moderate or greater degrees of hearing loss, and in controls with moderately severe or worse hearing losses. This differs from the QuickSIN, in which deficits were observed on average in slight hearing losses in the VS patients, and in mild losses for controls. Taken together, these results suggest that QuickSIN SNR losses are more likely to be affected by both hearing loss, and the presence of VS than are WRQ scores, regardless of the criteria used to denote a WRQ deficit.
In general, VS ears have similar distribution in WRQ performance compared to control right ears, while VS ears had tended to have poorer QuickSIN SNR performance compared to control right ears. Individual patient data for WRQ scores and QuickSIN SNR loss as a function of PTA is shown in Figure 1. Here, WRQ performance was categorized per the classification scheme by Meyer et al., (2006) Class I = 100% to 70%, Class II = 69% to 50%, Class III = 49% to 1%, Class IV = 0% (see Table 2). QuickSIN SNR loss was categorized according to the QuickSIN manual (Etymotic Research 2006) normal = 0 to 3 dB SNR loss, mild = 3 to 7 dB SNR loss, moderate = 7 to 15 dB SNR loss, and severe = 15+ dB SNR loss. In VS ears, WRQ Class II-IV performance (< 70%) was observed in 23.3% of patients and only at a moderate or greater hearing loss (PTA range: 35.75–76 dB HL, Fig. 1A). In control right ears, 2.6% of patients had WRQ Class II-IV performance (Fig. 1B). In contrast, mild QuickSIN SNR loss was observed in 26.0% of VS patients (PTA range: 10.75–66.25 dB HL), moderate SNR loss in 35.6% of VS patients (PTA range: 26.75–76 dB HL), and severe SNR loss in 23.6% of VS patients (PTA range 35.75%–75%, Fig. 1C). In control right ears, mild QuickSIN SNR loss was observed in 33.2% of patients (PTA range: 6.75–76.75 dB HL), moderate SNR loss in 18.2% of patients (PTA range: 8–81 dB HL), and severe SNR loss in 5.4% patients (PTA range: 16.5–89.25 dB HL, Fig. 1D).
Fig. 1.
Comparison of PTA and measures of speech perception in the same ear. PTA is compared with WRQ in (A) VS ears, and (B) control right ears. PTA is compared with QuickSIN SNR loss in (C) VS ears and (D) control right ears. PTA indicates pure-tone average; SNR, signal-to-noise ratio; VS, vestibular schwannoma; WRQ, word-recognition in quiet.
Interaural Asymmetries in Speech Understanding in Quiet and Noise
The presence of VS was associated with greater between-ear asymmetries for PTA, WRQ, and QuickSIN SNR loss on average. The mean PTA asymmetry was 15.25 ± 11.97 dB HL for VS patients and 10.67 ± 9.82 dB HL for control patients, WRQ asymmetry was 19.78 ± 6.73% for VS patients and 6.72 ± 6.12% for control patients, and QuickSIN SNR loss asymmetry was 6.45 ± 5.08dB SNR for VS patients and 2.88 ± 2.73dB SNR for control patients.
While the presence of VS was associated with greater interaural asymmetries in all audiologic measures compared to controls, VS patients were more likely to have QuickSIN SNR asymmetry than WRQ asymmetry. Figure 2A compares WRQ between VS and contralateral ears for individual patients and Figure 2B compares QuickSIN SNR loss between VS and contralateral ears in the same patients. Significant differences in interaural WRQ performance were computed using the binomial distribution for 25-word lists (Carney & Schlauch 2007), and significant differences in interaural QuickSIN performance was indicated by >2.7 dB SNR loss (Killion et al. 2004; Etymotic Research 2006). We used these criteria because they reflect statistically significant difference in performance, and thus avoid some of the variance known to be associated with clinical measures of speech recognition. Using this methodology, only 12 of 73 (16.4%) of VS patients had significant differences in interaural WRQ performance. In contrast, 43 of 73 (58.9%) had significant differences in interaural QuickSIN performance, suggesting that QuickSIN SNR loss is more sensitive to the effects of VS than WRQ scores. Of the 41 VS ears with moderate or greater SNR losses, 36 patients (87.8%) differed significantly from the contralateral ear performance. Finally, of the 30 VS patients who did not have a significant difference in interaural QuickSIN performance, 25 (83%) had mild or moderate SNR loss in the VS ear.
Fig. 2.
Comparison of interaural speech perception performance in VS patients. (A) WRQ score is compared between VS ear and contralateral ear. (B) QuickSIN SNR loss is compared between VS ear and contralateral ear. SNR indicates signal-to-noise ratio; VS, vestibular schwannoma; WRQ, word-recognition in quiet.
ROC analyses suggests that QuickSIN SNR asymmetry is more sensitive and specific at detecting VS than other measures of audiometric asymmetry. Table 4 shows the sensitivity and specificity at different asymmetry cutpoints for PTA, WRQ scores, and QuickSIN SNR loss. In general, specificity of WRQ and QuickSIN SNR loss cutpoints were higher than PTA for a given degree of sensitivity. The PTA asymmetry cutpoint of >15 dB HL (commonly cited as the AAO-HNS cutpoint; Steiger 2005; Cheng & Wareing 2012) had a sensitivity of 39.73% and a specificity of 73.55% in our cohort. There were two WRQ cutpoints (>10% and >12%) and six QuickSIN SNR loss cutpoints (>4 to >6.5 dB SNR loss) with better sensitivity and specificity in Table 3. The AUC-ROC for each measure was calculated: QuickSIN SNR loss asymmetry = 0.703; WRQ asymmetry = 0.667; PTA asymmetry = 0.615. When the AUC-ROC were statistically compared, the difference approached significance (X2 = 5.56, p = 0.062).
TABLE 4.
Sensitivity and specificity of various interaural asymmetry cutpoints for PTA, WRQ, and QuickSIN
| Cutpoint | Sensitivity, % | Specificity, % |
|---|---|---|
| PTA (dB HL) | ||
| ≥5 | 75.34 | 36.36 |
| ≥10 | 60.27 | 58.03 |
| ≥15 | 39.73 | 73.55 |
| ≥20 | 30.14 | 82.96 |
| ≥25 | 23.29 | 89.85 |
| ≥30 | 16.44 | 95.13 |
| WRQ (%) | ||
| ≥8 | 53.42 | 72.14 |
| ≥10 | 43.84 | 77.42 |
| ≥12 | 42.47 | 85.68 |
| ≥14 | 35.62 | 89.19 |
| ≥16 | 35.62 | 89.35 |
| ≥18 | 35.62 | 89.35 |
| SIN (dB SNR) | ||
| ≥2 | 75.34 | 49.69 |
| ≥2.5 | 66.94 | 60.73 |
| ≥3 | 62.10 | 65.08 |
| ≥3.5 | 59.68 | 73.30 |
| ≥4 | 58.06 | 75.82 |
| ≥4.5 | 55.65 | 80.53 |
| ≥5 | 54.03 | 81.92 |
| ≥5.5 | 49.32 | 85.60 |
| ≥6 | 43.84 | 86.67 |
| ≥6.5 | 39.73 | 88.86 |
| ≥7 | 36.99 | 89.43 |
| ≥7.5 | 31.51 | 90.84 |
| ≥8 | 30.14 | 91.46 |
| ≥8.5 | 27.40 | 92.41 |
| ≥9 | 27.40 | 92.82 |
Cell with dotted outline indicate the AAO-HNSF PTA asymmetry cutpoint. Gray highlighted cells indicate cutpoints with both greater sensitivity and specificity than the AAO-HNSF PTA asymmetry cutpoint.
PTA indicates pure-tone average; SNR, signal-to-noise ratio; VS, vestibular schwannoma; WRQ, word-recognition in quiet.
Finally, multiple logistic regression analysis showed that QuickSIN SNR loss asymmetry was most predictive of VS compared to other audiologic measures of asymmetry (Table 5). Interaction terms between all variables were considered by the model and none were found to be significant (p ≥ 0.172 in each instance). Individually, QuickSIN SNR loss asymmetry and PTA asymmetry were significantly predictive of VS (p < 0.001 and p = 0.025, respectively), while WRQ asymmetry was not (p = 0.064). QuickSIN SNR loss asymmetry had a higher odds ratio than PTA asymmetry (1.232 versus 1.039) with no overlap in the 95% confidence intervals (Table 4). In contrast, the 95% confidence intervals for PTA asymmetry and WRQ asymmetry overlapped with a minimal difference in the odds ratio (1.039 versus 1.029).
TABLE 5.
Results of the multiple logistic regression with interaction terms
| Asymmetry variable | OR (95% CI) | p |
|---|---|---|
| PTA | 1.0391 (1.0048–1.0746) | 0.025 |
| WRQ | 1.0288 (0.9940–1.0602) | 0.064 |
| SIN | 1.2321 (1.1004–1.3797) | <0.001 |
| PTA and WRQ interaction | 1.0000 (0.9986–1.0005) | 0.384 |
| PTA and SIN interaction | 0.9968 (0.9918–1.0018) | 0.210 |
| WRQ and SIN interaction | 0.9989 (0.9972–1.0005) | 0.172 |
Both PTA and QuickSIN SNR asymmetry were significantly predictive of VS. QuickSIN SNR asymmetry was more predictive than PTA asymmetry (higher OR with no overlap in 95% CI). There were no significant interactions between various measures of asymmetry.
CI indicates confidence interval; OR, odds ratio; PTA, pure-tone average; SNR, signal-to-noise ratio; VS, vestibular schwannoma; WRQ, word-recognition in quiet.
DISCUSSION
Since the inception of audiology, WRQ has been the default test of speech perception in routine audiologic testing. In contrast, it is widely known that patients with hearing loss consistently report difficulties understanding speech in the presence of background noise. As a result, the notion that SIN assessment should be part of the audiologic test battery as an adjunct or alternative measure of speech perception to WRQ has emerged. Widespread implementation of SIN testing, however, requires an understanding of how SIN performance is affected by pathology. Here, we present data that depict QuickSIN performance in patients with VS in the context of their performance on WRQ and pure-tone audiometry. To our knowledge this is the first study to report SIN performance in the setting of VS. Thus, our results are likely to be useful to both physicians and audiologists seeking to integrate SIN into their practice.
Our results suggest that between-ear asymmetries in QuickSIN SNR loss are more predictive of VS than asymmetries in WRQ scores. One interpretation of this result is that SIN may be more sensitive to disruption of auditory processing pathways than WRQ. Specifically, it is thought that SIN performance is driven by several factors including: (1) the audibility of the signal, (2) peripheral encoding, and (3) and central auditory/cognitive abilities (Plomp 1978). Thus, SIN deficits may be observed even when the signal is audible due to distortions in peripheral encoding or central signal retrieval. In contrast, WRQ performance is driven predominantly by audibility, which has led some to suggest the concept of SIN testing as a “stress test” for the auditory system as a whole (Wilson et al. 2007; Wilson 2011; Vermiglio et al. 2012). With regard to VS, these tumors are widely known to affect hearing and speech understanding through secretion of factors that are pathologic to the nerve and inner ear, and compression of the auditory nerve (Dilwali et al. 2015; Sagers et al. 2019). Our data therefore suggest that SIN performance is more sensitive than WRQ to these pathologic factors. An alternate interpretation is that the signal presentation levels used here were high enough to elicit rollover (Jerger & Jerger 1971; Dirks et al. 1977), and that rollover is more likely to be observed with measures of speech-recognition in noise than with WRQ. By this logic, the QuickSIN SNR loss would be more likely to be affected in the VS ears than the WRQ scores, and the differences in accuracy for between-ear asymmetries in the QuickSIN and WRQ scores reflect rollover-driven differences. We are unaware of any published data demonstrating a greater propensity for rollover for the QuickSIN than for WRQ, and thus cannot rule out this possible explanation for the present data.
Our results further suggest that between-ear QuickSIN SNR asymmetry is equal to or better than PTA asymmetry in identifying patients with VS. While decreased pure-tone hearing ability is associated with decreased SIN ability, performance on the QuickSIN varies widely for a given degree of hearing loss (Wilson et al. 2007; Phatak et al. 2018, 2019; Fitzgerald et al. 2023). This variance may reflect differences in peripheral encoding or central signal retrieval between patients with similar pure-tone detection abilities. When evaluating between-ear performance in the setting of VS, a unilateral tumor likely causes differences in signal audibility and peripheral encoding while having minimal effects on central signal retrieval ability. Therefore, when comparing between-ear asymmetry in PTA and QuickSIN SNR loss, it is possible that QuickSIN SNR asymmetry is more predictive of VS since both measures are affected by signal audibility but the QuickSIN may be more sensitive to deficits of peripheral encoding.
While these results show between-ear asymmetries in QuickSIN SNR loss are better than asymmetries in WRQ scores at identifying patients with VS, it is worth noting that these differences remain small, and may be influenced by the large sample size in the present study. By this logic, the sample size here would make any difference statistically significant, but the magnitude of this difference may not matter for clinical practice. Even if we assume this to be the case, these data continue to suggest that SIN could be easily integrated into clinical practice, or even replace WRQ, without hindering the ability of physicians and audiologists to diagnose or manage VS.
It is also possible that the sample of patients used here may have biased the results in other ways. For example, while we attempted to obtain the QuickSIN in addition to pure-tone audiometry and WRQ scores in as many patients as possible, some patients may not have completed the QuickSIN due to time constraints or other variables. Thus, audiologists may have been more likely to complete the QuickSIN on patients they were certain could perform the test in a reliable and timely manner. Doing so may have influenced the patient sample in two key ways. First, it is possible that patients with worse WRQ scores were less likely to be included in the study. Inclusion of these patients may therefore have improved the accuracy of between-ear asymmetries in WRQ scores. While this may be the case, it is important to note that deficits in WRQ scores are almost universally associated with even larger deficits on the QuickSIN (Fitzgerald et al. 2023). In a previous study with over 5800 patients, QuickSIN deficits were routinely observed in patients with good or excellent word-recognition abilities, while patients with difficulties on WRQ always had difficulties in noise (Fitzgerald et al. 2023). Given these observations, it is reasonable to assume that the inclusion of more patients with WRQ difficulties and asymmetries would have meant these patients would have had between-ear asymmetries in the QuickSIN as well, and thus our results may not have changed appreciably.
A second potential manner by which our sample could have been biased is that the QuickSIN was less likely to be performed in patients with more hearing loss, particularly in cases of moderate or greater hearing losses (Fitzgerald et al. 2023). By this logic, the mean SNR loss values may actually be greater than those reported here for increasing amounts of hearing loss. Data shown here and from Fitzgerald et al (2023) indicate an increasing likelihood of suboptimal WRQ scores in patients with moderate or greater losses, but such deficits are more commonly observed on the QuickSIN than for WRQ, particularly as the degree of hearing loss increases. With regard to the present study, it is unclear whether the addition of more patients with greater degrees of hearing loss would have changed the number of patients with asymmetric or symmetric losses.
It is worth noting that in our study, lower sensitivity and higher specificity values were reported for various PTA and WRQ cutpoints compared to previously published values. With regard to PTA, it is important to note that our control cohort included all individuals without VS or a conductive component to their hearing loss. As such, this control cohort encompassed individuals having audiograms for a wide array of auditory pathologies including but not limited to presbycusis, Meniere’s, noise exposure, etc. It is possible that the control group included patients with undiagnosed retrocochlear pathology. In contrast, the control cohort in many other investigates consisted of individuals who underwent MRI for asymmetric sensorineural hearing loss (Cheng & Wareing 2012; Ahsan et al. 2015). Since clinical judgment was used to select patients for MRI referral, a control cohort obtained in this way would yield in higher sensitivity and lower specificity for detecting VS than our control cohort of all-comers with sensorineural hearing loss. The current control cohort was chosen to better represent the ability of SIN in the primary audiologic screening for VS. Another factor which could have influenced the ROC analyses in our study versus other work is the presentation level used to measure WRQ scores. Specifically, the presentation level is often not consistent between or within institutions, particularly when approaches such as “30 or 40 dB HL above SRT” are used; such variance can influence WRQ scores, which could subsequently impact the results of the ROC analyses. Our protocol attempted to maximize audibility in all instances in order to maximize WRQ performance. Reducing audibility may yield more sizeable differences in the presence of VS, which would cause decreased sensitivity and increase specificity.
CONCLUSION
The present study demonstrates that asymmetries in QuickSIN SNR loss are somewhat more accurate than asymmetries in word-recognition or the AAO-PTA in the audiologic screening for VS. This result suggests that physicians and audiologists could transition to routine SIN assessment without hindering the ability to diagnose or manage auditory pathologies such as VS and may even improve their ability to flag at-risk patients. This finding adds to a small but growing body of the literature that measures of SIN performance can be integrated into routine audiologic practice, and may better address patient concerns about speech perception in the presence of background noise.
ACKNOWLEDGMENTS
We would like to thank Christian Bourdon, Brianne Davis, Cory Hillis, Rebecca Howard, Angela Huang, Lauren Jacobs, Amanda Murphy, Mateel Musallam, Sarah Pirko, Goutham Telukuntla, Sunny Yoon, Justin Cha, Rachael Jocewicz, Veronica Koo, Grace Nance, Devon Palumbo, Michael Smith, Soumya Venkitakrishan, and Madison Wageck for collecting these data as part of their normal clinical activities. We would also like to thank Gerald Popelka for his role in creating the database used to store these data. We would also like to thank Robert Jackler, Nikolas Blevins, Peter Santa Maria, John Oghalai, Jennifer Alyono, George Shorago, John Shinn, Joanne Branzuela, Neil Fulgencio, and Konstantina Stankovic for their administrative support of this project. Finally, we would also like to thank Mona Taliaferro for her generous support of this project.
Footnotes
The authors have no conflicts of interest to declare.
Z.J.Q. assisted in data analysis and manuscript preparation. Y.V. assisted with study design and interpretation of the data. S.G. assisted in data collection, analysis, preparation of figures, and assisted with manuscript preparation. E.T. assisted with data collection and analysis. N.S.A. assisted with data collection and analysis. N.B. assisted with study design, interpretation of the data, and manuscript preparation. M.F. designed and implemented the study, assisted in data collection, analysis, preparation of figures, interpretation of data and played the lead role in manuscript preparation.
REFERENCES
- Anon. (1995). Committee on hearing and equilibrium guidelines for the evaluation of hearing preservation in acoustic neuroma (vestibular schwannoma). American Academy of Otolaryngology-Head and Neck Surgery Foundation, INC. Otolaryngol Head Neck Surg, 113, 179–180. [DOI] [PubMed] [Google Scholar]
- Ahsan S. F., Standring R., Osborn D. A., Peterson E., Seidman M., Jain R. (2015). Clinical predictors of abnormal magnetic resonance imaging findings in patients with asymmetric sensorineural hearing loss. JAMA Otolaryngol Head Neck Surg, 141, 451–456. [DOI] [PubMed] [Google Scholar]
- Carhart R., & Jerger J. F. (1959). Preferred method for clinical determination of pure-tone thresholds. J Speech Hear Disord, 24, 330–345. [Google Scholar]
- Carney E., & Schlauch R. S. (2007). Critical difference table for word recognition testing derived using computer simulation. J Speech Lang Hear Res, 50, 1203–1209. [DOI] [PubMed] [Google Scholar]
- Cheng T. C., & Wareing M. J. (2012). Three-year ear, nose, and throat cross-sectional analysis of audiometric protocols for magnetic resonance imaging screening of acoustic tumors. Otolaryngol Head Neck Surg, 146, 438–447. [DOI] [PubMed] [Google Scholar]
- Dilwali S., Landegger L. D., Soares V. Y. R., Deschler D. G., Stankovic K. M. (2015). Secreted factors from human vestibular schwannomas can cause cochlear damage. Sci Rep, 5, 18599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dirks D. D., Kamm C., Bower D., Betsworth A. (1977). Use of performance-intensity functions for diagnosis. J Speech Hear Disord, 42, 408–415. [DOI] [PubMed] [Google Scholar]
- Etymotic Research. (2006). QuickSIN: Speech in noise test. Etymotic Research.
- Fitzgerald M., Gianakas S. P., Qian Z. J., et al. (2023). Preliminary guidelines for replacing word-recognition in quiet with speech-in-noise in the routine audiologic test battery. Ear Hear, 44, 1548–1561. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Holder J. T., Sheffield S. W., Gifford R. H. (2016). Speech understanding in children with normal hearing: sound field normative data for BabyBio, BKB-SIN, and QuickSIN. Otol Neurotol, 37, e50–e55. [DOI] [PubMed] [Google Scholar]
- Hurley R. M., & Sells J. P. (2003). An abbreviated word recognition protocol based on item difficulty. Ear Hear, 24, 111. [DOI] [PubMed] [Google Scholar]
- Jerger J., & Jerger S. (1971). Diagnostic significance of PB word functions. Arch Otolaryng, 93, 573–580. [DOI] [PubMed] [Google Scholar]
- Johnson E. W., & Sheehy J. L. (1966). Audiological aspects of the diagnosis of acoustic neuromas. J Neurosurg, 24, 621–628. [DOI] [PubMed] [Google Scholar]
- Killion M. C., Niquette P. A., Gudmundsen G. I., Revit L. J., Banerjee S. (2004). Development of a quick speech-in-noise test for measuring signal-to-noise ratio loss in normal-hearing and hearing-impaired listeners. J Acoust Soc Am, 116, 2395–2405. [DOI] [PubMed] [Google Scholar]
- Le Prell C. G., & Clavier O. H. (2017). Effects of noise on speech recognition: Challenges for communication by service members. Hear Res, 349, 76–89. [DOI] [PubMed] [Google Scholar]
- Meyer T. A., Canty P. A., Wilkinson E. P., Hansen M. R., Rubinstein J. T., Gantz B. J. (2006). Small acoustic neuromas: Surgical outcomes versus observation or radiation. Otol Neurotol, 27, 380–392. [DOI] [PubMed] [Google Scholar]
- Phatak S. A., Sheffield B. M., Brungart D. S., Grant K. W. (2018). Development of a test battery for evaluating speech perception in complex listening environments: Effects of sensorineural hearing loss. Ear Hear, 39, 449–456. [DOI] [PubMed] [Google Scholar]
- Phatak S. A., Brungart D. S., Zion D. J., Grant K. W. (2019). Clinical assessment of functional hearing deficits: Speech-in-noise performance. Ear Hear, 40, 426–436. [DOI] [PubMed] [Google Scholar]
- Plomp R. (1978). Auditory handicap of hearing impairment and the limited benefit of hearing aids. J Acoust Soc Am, 63, 533–549. [DOI] [PubMed] [Google Scholar]
- Sagers J. E., Sahin M. I., Moon I., Ahmed S. G., Stemmer-Rachamimov A., Brenner G. J., Stankovic K. M. (2019). NLRP3 inflammasome activation in human vestibular schwannoma: Implications for tumor-induced hearing loss. Hear Res, 381, 107770. [DOI] [PubMed] [Google Scholar]
- Selesnick S. H., & Jackler R. K. (1992). Clinical manifestations and audiologic diagnosis of acoustic neuromas. Otolaryngol Clin North Am, 25, 521–551. [PubMed] [Google Scholar]
- Steiger J. R. (2005). Audiologic referral criteria: Sample clinic guidelines. Hear J, 58, 38–42. [Google Scholar]
- Sweeney A. D. Carlson M. L. Shepard N. T. McCracken D. J. Vivas, E. X.Neff B. A. Olson J. J. (2018). Congress of neurological surgeons systematic review and evidence-based guidelines on otologic and audiologic screening for patients with vestibular schwannomas. Neurosurgery, 82, E29–E31. [DOI] [PubMed] [Google Scholar]
- Tillman T. W., & Carhart R. (1966). An expanded test for speech discrimination utilizing CNC monosyllabic words. Northwestern University Auditory Test No. 6 SAM-TR-66-55. Tech Rep SAM-TR :1–12. 10.21236/ad0639638. [DOI] [PubMed]
- Vermiglio A. J., Soli S. D., Freed D. J., Fisher L. M. (2012). The relationship between high-frequency pure-tone hearing loss, hearing in noise test (HINT) thresholds, and the articulation index. J Am Acad Audiol, 23, 779–788. [DOI] [PubMed] [Google Scholar]
- Vermiglio A. J., Soli S. D., Fang X. (2018). An argument for self-report as a reference standard in audiology. J Am Acad Audiol, 29, 206–222. [DOI] [PubMed] [Google Scholar]
- Waterval J., Kania R., Somers T. (2018). EAONO position statement on vestibular schwannoma: Imaging assessment. What are the indications for performing a screening MRI scan for a potential vestibular schwannoma? J Int Adv Otol, 14, 95–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wilson R. H. (2011). Clinical experience with the words-in-noise test on 3430 veterans: Comparisons with pure-tone thresholds and word recognition in quiet. J Am Acad Audiol, 22, 405–423. [DOI] [PubMed] [Google Scholar]
- Wilson R. H., McArdle R. A., Smith S. L. (2007). An evaluation of the BKB-SIN, HINT, QuickSIN, and WIN materials on listeners with normal hearing and listeners with hearing loss. J Speech Lang Hear Res, 50, 844–856. [DOI] [PubMed] [Google Scholar]