Voice complaints after surgery to remove all or part of the thyroid gland have been reported in a minority of patients, and can have devastating implications for these patients. The published literature generally indicates that durable voice problems occur in 1-7% of patients after thyroidectomy, but transient voice problems soon after surgery are far more common (~ 35%). The most common operative injury involves the recurrent laryngeal nerve, resulting in transient or permanent vocal fold paresis or paralysis. Other complications can involve damage to the external branch of the superior laryngeal nerve, biomechanical arytenoid or laryngeal mucosal injury related to endotracheal intubation, cricothyroid muscle injury, and mucosal congestion/edema. However, functional voice disorders have been documented in the absence of major laryngeal nerve injury with as many as 30% of patients suffering early and 14% suffering lasting functional voice changes after thyroid resection.3
Voice evaluations typically include laryngeal imaging and patient report, as well as perceptual and acoustic analysis of the voice signal. Several studies have documented possible changes in the perceptual quality of the voice as well as in the characteristics of the acoustic signal post-thyroidectomy surgery. Kark et al. (1984) reported retrospectively on post-thyroidectomy voice function in 325 patients. Permanent voice change occurred in 59 patients, characterized by fatigue, reduction in singing range, hoarseness, or a combination of these characteristics. In a second set of 38 patients studied prospectively, acoustic analyses indicated that 50% of the patients had changes in pitch range, frequency, or phonation times 1 week post-surgery; approximately 20% showed symptomatic subjective changes. By three months six patients (15%) still showed measurable changes. The most common cause of vocal dysfunction appeared to be injury to the external laryngeal nerves on one or both sides. Aluffi et al. (2001) reported an increase in the perturbation of the sustained voice signal (as measured via jitter and shimmer) in 42 patients 12-18 months post-thyroidectomy. These researchers reported that there was no objective evidence of damage to external SLN branches, and interpreted their findings as suggestive of an extralaryngeal cause of vocal dysfunction.
Get your grade
or your money back
using our Essay Writing Service!
Stojadinovic et al. (2002) reported on 54 patients 1-week and 3-months post-thyroidectomy. Results indicated that vocal symptoms were observed in 30% of the patients following thyroidectomy, and persisted in 14% of patients. Maximum phonational frequency range and vocal jitter changes from baseline were significantly associated with voice symptoms at three-months. Petto et al (2006) reported on 100 patients who underwent thyroidectomy. Results showed postoperative laryngeal alterations in 28% of the thyroidectomized patients via videolaryngoscopy, as well as subjective voice changes (29.7% of the patients) and increased values in acoustic perturbation (the Voice Turbulence Index) as compared to a non-thyroidectomy control group. In addition, voice complaints were more frequently reported via the Voice Handicap Index (VHI) in the thyroid group rather than in the control group. A study by Stojadinovic et al. (2008) reported on the ability of patient-reported and clinician-determined voice assessment to identify voice dysfunction post-thyroidectomy in fifty patients. Results indicated that the patient-reported total VHI score was most predictive of post-operative voice dysfunction. Postoperative laryngeal abnormalities identified via stroboscopic evaluation were observed in eight subjects (16%); one subject (2%) had permanent voice dysfunction >6 months after operation due to transection of the recurrent laryngeal nerve (RLN). The most commonly reported voice symptoms at 1-2 weeks after operation were hoarseness, pitch loss, and vocal fatigue. In contrast to the aforementioned reports of possible vocal dysfunction post-thyroidectomy, Akyildiz et al (2008) reported no significant change pre vs. post-surgery in acoustic perturbation measures in males, and a significant reduction in several acoustic measures including jitter, shimmer, and noise-to-harmonic ratio in female subjects. This finding was interpreted as reflecting possible improvement in vocal function for the female subjects.
Many of the post-thyroidectomy changes in voice function reported in previous studies have been obtained via non-speech samples such as sustained vowel production of phonational range tasks. Unfortunately, standard acoustic analysis of voice that involves the steady-state portion of a sustained vowel (most commonly /É‘/) may be unrepresentative of the speaker's typical voice, and measures such as jitter, shimmer, and harmonic-to-noise ratio (HNR) are only valid with sustained vowel productions produced with the intention of steady pitch and loudness - any purposeful changes in vocal pitch or loudness will be measured as increases in vocal perturbation, even though these measurements may not reflect vocal abnormality. Several authors have indicated that continuous/running speech may provide a more valid assessment of the patient's control of vocal parameters such as vocal quality, and may correlate better with perceptions of dysphonia (Qi, Hillman, & Milstein, 1999; Halberstam, 2004; Laflen, Lazarus, & Amin, 2008; Maryn, Cornthals, Van Cauwenberge, Roy, & De Bodt, in press). In addition, continuous speech incorporates important vocal attributes such as rapid voice onset /termination and variations in fundamental frequency and amplitude that may be highly relevant to the perception of dysphonia in everyday situations and to clinical decisions regarding the voice quality of the patient (Hammarberg, et al., 1980; Parsa & Jamieson, 2001).
Always on Time
Marked to Standard
Rather than relying on time-based acoustic methods which have questionable validity when applied to nonstationary signals, several authors have attempted to use spectral-based acoustic methods to analyze normal and disordered voice quality in running speech. In particular, cepstral analysis has been reported to show considerable promise as a measure of dysphonia in both sustained vowel and running speech contexts. The cepstrum has been described as a Fourier transform of the logarithm power spectrum and was originally described by Noll (1964) as a procedure for extracting the fundamental frequency from the spectrum of a sound wave. Several authors have attempted to use spectral-based acoustic methods to analyze normal and disordered voice quality in running speech. A number of studies have demonstrated the effectiveness of measures derived from cepstral analysis to quantify dysphonic voice characteristics in sustained vowel and continuous speech (Hillenbrand and Houde, 1996; Qi et al, 1999; Parsa and Jamieson, 2001; Heman-Ackah, Michael, & Goding, 2002; Heman-Ackah, Heuer, Michael, Ostrowski, Horman, Baroody, et al., 2003; Halberstam, 2004; Laflen et al., 2008). Recent work by Awan, Roy & Dromey (2009) has extended the work of Hillenbrand and Houde (1996) to include measures of low vs. high frequency spectral energy and the average variability of spectral and cepstral measures in addition to measures of the cepstral peak prominence (CPP) and demonstrated the effectiveness of spectral/cepstral measures of continuous speech as an objective treatment outcomes measure.
The purpose of this study was to evaluate possible changes in the perceptual and acoustic characteristics of continuous speech samples (CAPE-V sentences) pre- and post-thyroidectomy. Because traditional time-based acoustic measures such as jitter and shimmer are not valid with running speech samples, the acoustic analysis procedures used in this study were focused on spectral/cepstral measures.
To evaluate potential changes in perceptual and acoustic characteristics of continuous speech samples (CAPE-V sentences) before and at multiple time points after thyroidectomy.
For persons with and without negative voice outcomes soon after thyroidectomy,
Do spectral- and/or cepstral-based acoustic measures of connected speech change over time relative to pre-operative baseline measures?
Does the auditory perceptual rating of Overall Severity of dysphonia change over time?
Participants: Patients scheduled for partial or total thyroidectomy were recruited for this 6-month prospective longitudinal trial. Perceptual and acoustic analyses were conducted for 70 subjects pre-thyroidectomy; 1-4 weeks; 3 months; and 6 months after surgery. Participants included 36 women and 34 men (Mean Age: 51.3 yrs.; Range: 23-78). Subjects were also classified as Normal (n=50) vs. Negative Voice Outcome (NVO; n=20) based upon blinded Consensus Auditory Perceptual Evaluation of Voice (CAPE-V; Kempster et al., 2009), Voice Handicap Index (VHI), and Dysphonia Severity Index (DSI) scores at 1-week post-surgery.
Stimuli: Speech samples included the six sentences provided with the CAPE-V. Sentences were read aloud in a typical manner, and were digitized (16 bits, 25 kHz sampling rate) using an AKG C 420 head-set microphone (positioned 4-5 cm from lips) and the KayPentax CSL 4500.
Perceptual Analysis: Three certified SLPs who specialize in voice rated the CAPE-V recordings. Samples were presented quasi-randomly; session data were blocked for each subject. Judges listened under headphones in a sound-treated booth, and used a custom-automated CAPE-V program. Median ratings for overall severity were obtained at each time point.
Acoustic Analysis: All samples were analyzed using a Windows-based computer program developed by the first author and reported in Awan et al. (2009). Measures of the cepstral peak prominence (CPP) and the ratio of low vs. high frequency spectral energy (L/H Ratio), as well as the standard deviations for the aforementioned measures were obtained for all continuous speech samples.
70 adults scheduled for partial or total thyroidectomy
See Table 1 for demographic data
Study Time Points
After operation: 1-4 weeks, 3 months, 6 months
Six sentences provided with the Consensus Auditory Perceptual Evaluation of Voice (CAPE-V; Kempster et al., 2009)
Read aloud with habitual pitch, loudness, and quality
Digitized (16 bits, 25 kHz sampling rate) with AKG C420 head-set microphone (4-5 cm from lips) using the KayPentax Computerized Speech Lab (CSL 4500).
Three certified SLPs who specialize in voice
Listened under headphones in a double-walled sound-treated booth
Custom-automated CAPE-V program
Rated parameters using visual analog scales with non-linear visual markers for severity; the median Overall Severity is reported here
This Essay is
a Student's Work
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.Examples of our work
Samples presented randomly for each speaker (4 sessions per subject block)
Windows-based computer program developed by S.N.A. (Fig. 1)
Spectral and cepstral analysis methods (Awan et al., 2009)
Cepstral peak prominence (CPP)
Ratio of low vs. high (L/H) frequency spectral energy
CPP Standard Deviation (CPP SD)
L/H Standard Deviation (L/H SD)
Repeated measures mixed-model ANOVA
Within subjects factor (time); between subjects factors (gender, and 1-wk post-op voice outcome)
A series of Repeated measures mixed-model ANOVA's (Within subjects factor (Time); Between subjects factors (Gender, and 1-wk post-op voice outcome) were computed to assess possible differences in the median CAPE-V Overall Severity ratings and acoustic measures. Results indicated that the ANOVA for Overall Severity across time was nonsignificant (F (3,198) = 0.58; p = .63). In contrast, significant interactions of Time x Voice Outcome were observed for CPP, CPP sd, and the L/H Ratio. Post-hoc analyses using Bonferroni corrected t-tests indicated (a) a significant reduction in mean CPP pre-op to 1-4 wk post-op within the NVO group (p = .013) and significant differences between NVO and normal groups at 1-4 wk post-op (p < .001), 3-mo post-op (p = .03), and 6-mo post-op (p = .002); (b) a strong trend for the CPP sd to be reduced 1-4 wk post-op as compared to pre-op or 3-mo post-op for the NVO group, and significant differences between NVO and normal groups at 1-4 wk post-op (p < .001), and 6-mo post-op (p = .015); (c) a significant increase in the L/H ratio between pre-op and 6-mo post-op for the normal group (p = .045) and a significant increase between 1-4 wk post-op and 6-mo post-op for the NVO group (p = .016)
CPP (Fig. 3)
Significant reduction in mean CPP pre-op to 1-4 wk post-op within the NVO group (mean difference = -1.12 dB; Bonferroni-adjusted p = .013)
No significant differences within the normal voice-outcome group across Time
Significant differences between NVO and normal groups at 1-4 wk post-op (p < .001), 3-mo post-op (p = .03), and 6-mo post-op (p = .002)
CPP SD (Fig. 4)
No difference across Time for either group following Bonferroni adjustments
Strong trend for CPP sd to be reduced 1-4 wk post-op as compared to pre-op or 3-mo post-op for the NVO group
Significant differences between NVO and normal groups at 1-4 wk post-op (p < .001), and 6-mo post-op (p = .015)
L/H Spectral Ratio (Fig. 5)
Significant increase in between pre-op and 6-mo post-op for the normal group (Mean Difference = 0.77 dB; Bonferroni adjusted p = .045)
Significant increase between 1-4 wk post-op and 6-mo post-op for the NVO group (Mean Difference = 1.79 dB; Bonferroni adjusted p = .016)
Changes in the overall severity of the speaking voice following thyroidectomy that appeared to be too subtle to detect via auditory-perceptual ratings were observed using spectral/cepstral analysis of continuous speech. Decrements in CPP and CPP SD soon after thyroidectomy are consistent with previous literature indicating that transient voice problems are relatively common after thyroidectomy. These findings were clearly evident for the NVO group, which included patients who were identified with voice problems at the first post-operative visit. Therefore, it appears that CPP and CPP SD are sensitive to changes that perceptual analysis alone may miss. Spectral/cepstral methods appear to provide valuable analyses of running speech samples that cannot be analyzed validly with traditional time-based acoustic measures such as jitter and shimmer during the peri-operative period in patients undergoing thyroidectomy.
Twenty of the 70 subjects (29%) were identified as having negative voice outcomes early after thyroidectomy. This determination was made using the composite algorithm provided in Table 1.
The perceived overall severity of voice did not change significantly over the course of post-thyroidectomy recovery for this group of participants, whether or not they were identified as having a negative voice outcome.
Spectral/cepstral characteristics of the speaking voice revealed significant differences from the pre-operative assessment when compared to 1-4 weeks after operation for CPP and CPP SD, 3 months after operation for CPP, and 6 months after operation for CPP, CPP SD, and L/H spectral ratio.
Changes in the overall severity of the speaking voice that appeared to be too subtle to detect following thyroidectomy were observed using spectral/cepstral analysis of continuous speech.
Decrements in CPP and CPP SD soon after thyroidectomy are consistent with previous literature indicating that transient voice problems are relatively common after thyroidectomy. These findings were clearly evident for the NVO group, which included patients who were identified with voice problems at the first post-operative visit. It appears that CPP and CPP SD are sensitive to changes that perceptual analysis alone may miss.
Spectral/cepstral methods appear to provide valuable analyses of running speech samples that cannot be analyzed validly with traditional time-based acoustic measures such as jitter and shimmer during the peri-operative period in patients undergoing thyroidectomy.
Dr. S. N. Awan has an agreement with KayPentax (Lincoln Park, NJ) regarding the development of computer software including cepstral analysis of continuous speech algorithms.
The views expressed are those of the authors and do not reflect the official policy of the Department of Army, Department of Defense, or U.S. Government.