Vocal assessment in patients submited to CO2 laser cordectomy

Summary Aim To evaluate voice outcomes in patients with early glottic carcinoma treated by CO2 laser cordectomy. Method 15 patients with glottic Tis and T1 squamous cell carcinoma treated with CO2 laser were analyzed. The assessment consisted of perceptual voice analysis, objective voice evaluation and video-laryngo-stroboscopic exam. In addition, patients rated their voices and completed the Voice related Quality of Life (VR-QOL) questionnaire. The results were compared with those obtained in a matched control group. Results Most of the patients presented some degree of hoarseness on perceptual voice analysis, mainly rough and breathy voices. Their acoustic evaluation compared with the control group showed a small increase in fundamental frequency, but with no statistically significant difference, and the values of jitter, shimmer and noise to harmonic ratio were worse and statistically significant. As regards to video-laryngo-stroboscopic findings, better results were achieved in the less extensive resection group. Patients have had minimal repercussion in their life quality in respect to voice. Conclusions: In spite of voice alterations in patients submitted to cordectomy by CO2 laser, functional results are acceptable, with minimal repercussion in their quality of life. Avaliação da voz em pacientes submetidos à cordectomia com laser de CO2.


INTRODUCTION
Cancer of the larynx is the second most frequent tumor of the upper airway and upper digestive tract, second only to oral cavity tumors. Brazil occupies the second place in the world in the incidence of cancer of the larynx. This disease affects mostly males in the sixth and seventh decades of life 1 . Most of the malignant tumors of the larynx (95%) are squamous cell carcinomas (SCCs) which belong to the epithelial lineage 2 .
Initial glottic carcinomas, classified as in situ carcinomas (Tis) and T1, may be treated by open surgery (cervical access), radiotherapy and laryngeal microsurgery with conventional instruments or by CO 2 laser. These approaches provide satisfactory local control and similar survival rates [3][4][5] .
Cordectomy is the surgical procedure used for early glottic tumors, where the affected vocal fold is resected at varying depths.
The advantages of CO 2 laser compared to other treatment strategies include lower morbidity compared to radiotherapy side effects (mucositis, xerostomia and dental problems), lower cost as patients do not require traqueostomy, improved homeostasis, and the fact that CO 2 laser does not exclude other forms of treatment if the disease recurs locally, regionally or at a distance 5,19 .
When different treatment strategies offer similar cure rates, resulting voice quality becomes paramount in the therapeutic decision. There are few published papers in medical literature about voice following laser cordectomy for the treatment of laryngeal tumors.
Most of the studies compare cordectomy with radiotherapy, and results are controversial. Some authors have reported worse outcomes in patients undergoing laser surgery, which may limit its use [21][22][23] . Other studies, however, have stated that voice quality in patients undergoing CO 2 laser cordectomy is similar to that in radiotherapy patients [24][25][26][27] . CO 2 laser cordectomy provides improved results in the voice quality of patients, compared to conventional cordectomy [28][29][30] . Few studies specifically on voice assessment in patients that have undergone CO 2 laser cordectomy have been published to date.
The factors that stimulated this study where the importance of preserving the phonatory function when treating early glottic tumors and the lack of guidelines on voice quality in patients undergoing CO 2 laser cordectomy for the treatment of laryngeal tumors.
Therefore, the aim of this study was to assess voice quality in patients with stages Tis and T1 glottic carcinoma undergoing CO 2 laser cordectomy.

METHODS
Between January and May 2004 we assessed fifteen patients (12 males and 3 females) who had undergone CO 2 laser cordectomy for the treatment of early vocal fold SCC. Early glottic SCC was defined in this study as patients with vocal fold Tis and T1 tumors. Patients were referred to us from the Head and Neck Surgery Unit at Sao Paulo University -Paulista School of Medicine.
Ages varied from 43 to 82 years, average 63.1 years. Average follow-up time was 14.6 months, varying from 3 to 38 months.
Patients were staged, based on data in their charts, using the clinical classification of the International Union Against Cancer (UICC) and the American Joint Committee on Cancer (AJCC) 2002.
Three patients were staged as TisN0 and twelve as T1aN0. Types of cordectomy were classified into five types according to the European Laryngological Society 6 guidelines, namely type I -sub-epithelial resection, type II -subligamentar resection, type III -transmuscular resection, type IV -total cordectomy, and type V -extended cordectomy. Type V is subdivided into Va (resection of part of the contra-lateral vocal fold), Vb (resection of the arytenoid), Vc (resection of the ventricular band), and Vd (resection of part of the subglottis). Three patients underwent type I cordectomy (Tis), five patients underwent type II cordectomy, four patients underwent type III cordectomy, and three patients underwent type IV cordectomy. No patient required type V cordectomy.
Surgery was done under general anesthesia. A 400mm lens microscope was coupled to a Visograf® CO 2 laser unit. A 5 to 7 Watts superpulse mode 0.5 mm spot was used. Safety margins were 1 to 2mm. Following tumor resection upper, lower, anterior, posterior and deep margins were sent to frozen section examination, to assure surgical safety margins.
All patients included in this study had been operated at least three months before enrollment, and were referred to post-operative speech therapy. No patient showed signs of local and/or regional recurrence and none had a second primary tumor.
Patients underwent subjective and objective voice assessment and videolaryngostroboscopic evaluation between the 3rd and 38th post-operative month. Subjective testing was done by perceptive-auditory voice analysis according to the GRBASI scale 32  GRBASI's scale assesses the global impression of dysphonia (G) considering hoarseness (R), soprosity (B), asteny (A), tension (S) and instability (I), which are scored 0 to 3, 0 (no change), 1 (mildly altered), 2 (moderately altered) and 3 (severely altered). V-RQOL is a protocol whereby patients relate their voice to quality of life in a 0 to 100 score where 100 is improved quality of life 35,36 . In a general voice assessment, patients classify their voice in a 0 to 5 scale where 0 is impaired voice and 5 is normal or near normal voice.
Computer analysis was made based on open and isolated phonation of the sustained vowel /e/ picked up by an AKG model C410 professional microphone at a standard 5 cm distance from the patient. Voice recording was done by the Computerized Speech Lab -CSL, model 4300B, and analyzed by the Kay Elemetrics Corp.'s MDVP (Multi-Dimensional Voice Program) voice analysis program.
Each patient was asked to emit an isolated and sustained vowel /e/. A voice sample was considered as the most stable 3 second period, whenever possible eliminating the beginning and the end of sound emission or a maximum period of voice stability. The following acoustic measurement parameters were used in this study: fundamental frequency, the perturbation quotients Jitter (variations of frequency) and Shimmer (variations of amplitude), and noise measurements (harmonics-to-noise ratio).
Computer analysis compared fundamental frequency, jitter, shimmer, and harmonics-to-noise ratio with a patient control group (average age 57.6 years) that had normal larynxes, and no complaints about voice.
Videolaryngostroboscopy was used to analyze the following parameters: 1. Glottic coaptation: complete or incomplete 2. Laryngeal vestibule constriction: normal or with hyperconstriction.
3. Presence of vibration: absent, glottic, supraglottic or combined. Vibration was considered as absent if there was any area of akinesia in the resected glottic region, with no supraglottic compensation. The Stata 7.0 program was used for statistical analysis. Data had a normal distribution allowing us to use Student's t-test for non-related samples to compare numerical data in the study and the control groups. The significance level was 0.10.

RESULTS
Student's t-test was used to compare the average fundamental frequency of males in the study group and the control group (139.3Hz). The descriptive value was 0.2008, which suggests no differences between both groups. There were not sufficient females in the study group to apply statistical tests. Descriptively, however, the 226.24 Hz average value may be compared to the average control group value (202,23 Hz).
Student's t-test for independent samples demonstrated that average jitter in patients was higher than in the control group (p = 0.001), that average shimmer in patients was higher than in the control group (p = 0.0007), and that patients had a higher average harmonics-to-noise ratio compared to the control group (p = 0.053).

DISCUSSION
There is no agreement about voice quality after different treatment strategies for glottic tumors, particularly Tis and T1 tumors. Important factors influencing the decision for a specific treatment procedure, except for voice quality, include oncological results, cost and morbidity.
There are few papers in literature assessing voice following CO 2 laser cordectomy, and only rarely is there a comparison with a control group of normal individuals; there are no such papers published in Brazil to date. The importance of voice is evident as we measure voice changes in patients undergoing CO 2 laser cordectomy compared to normal individuals and to how much this affects their lives. Furthermore, future comparisons may be made both within groups undergoing similar treatment and between groups subjected to other strategies such as radiotherapy and conventional surgery.
In our patients age ranged from 43 to 82 years (average 63.1 years), with 80% males. This is similar to other published studies, evidence that malignant laryngeal tumors are more frequent in men in the sixth and seventh decades of life 1 .
Although all patients were staged as Tis and T1a, the type of cordectomy varied in the group. Less extensive cordectomies, classified as type I and II, were done in 53% of patients. Type III and IV cordectomies were done in the remaining 47%.
Subjective analysis demonstrated that the majority of patients had some degree of dysphonia, usually hoarseness and soprosity. These results are not surprising, as hoarseness (R) is related to vocal crepitation and bitonality due to irregular vibration of the vocal folds, and soprosity (B)    is related to air escape during phonation. One naturally expects a scar and local irregularity when cordectomy is done, as part of the vocal fold is removed. Patient opinions about their voice following treatment were also of interest in this study. Using Sittel et al's (1998) 33 scale, six patients (40%) who judged their voices as normal or near normal, four patients (26.7%) judged their voices as good for communicating but pathological, three patients (20%) judged their voices as reasonable, and only two (13.3%) judged their voices as impaired, but understandable. No patient classified their voice as very dysfunctional.
The impact of voice on the patient's quality of life should be a constant concern in the treatment of laryngeal tumors. This concept, however, is recent; Zeitels (1995) 12 states that the current standard of phonomicrosurgery is to obtain oncological radicality with voice quality that causes the least change in quality of life. Classical assessments of health treatment valued only the presence or absence of disease following treatment, which is important, but insufficient to assess the impact on life in general 37 . O V-RQOL was developed specifically to correlate how much voice change may affect the patient's quality of life 36 . In this study the average V-RQOL was 88.17. These results show that although most patients undergoing CO 2 laser cordectomy had some degree of dysphonia, the effect on activities of daily life was minor. It is important to underline that no patient in this study used his or her voice professionally, which reduces the impact of voice changes on the quality of life. In fact, the V-RQOL is not a sensitive method to assess professional voice users 36 , as minor changes in a singer, for instance, may compromise his or her career, whereas the same change in a person that does not use his or her voice professionally may cause minimal impact. Computer voice acoustic analysis using the Kay Elemetrics Corp. Computerized Speech Lab-CSL model 4300 MDVP (Multi-Dimensional Voice Program) voice analysis program showed that, except for the fundamental frequency, all voice acoustic analysis parameters in this study -percentage of jitter and shimmer, and harmonics-to-noise ratio -are significantly affected by CO 2 laser cordectomy (p<0.1) comparing results of the surgery group to controls.
The average fundamental frequency for operated male patients was higher compared to controls, although this difference was not statistically significant. There were not sufficient females in the study group to apply statistical tests, but a minor increase of 24.01 Hz in the fundamental frequency was seen compared to controls (202.23Hz). Similar results have also been reported by McGuirt et al. (1992) 31 with a slightly higher average fundamental frequency compared to a normal control group. Voice tends to become more acute due to the removal of tissue and resulting compensation mechanisms. The main factors affecting the fundamental frequency are the length, mass and tension of the vocal fold; the lower the mass, the higher the frequency. Furthermore these patients frequently compensate with hyperconstriction and tension which can also increase the fundamental frequency.
Both jitter and shimmer were significantly higher in the study group compared to controls. These results reflect increased aperiodicity in the glottic cycle and sound wave variability. The harmonics-to-noise ratio is directly related to voice quality 36 . It analyzes aperiodic components of sound and is considered a good correlate to what we consider as dysphonia. The study group average was higher compared to controls and was statistically significant (p = 0.053).
Results of objective voice analysis clearly suggest that voice quality in patients undergoing CO 2 laser cordectomy is compromised compared to individuals with no voice-related problems. A further issue is that when these patients seek medical help, they already have some degree of dysphonia by having a glottic tumor. Pre-operative voice assessment was not evaluated in this study, but is suggested for future studies.
Theoretically, regardless of the treatment, voice quality of patients with glottic tumors will never be equal to their voice before the disease. Even the smallest cordectomy will involve resection of part of the mucosa, reducing the ability for vibration at that point.
Although considered a more conservative strategy, radiotherapy also has negative effects on voice quality compared to controls 40  Videolaryngostroboscopy found that 75% (n=6) of patients undergoing type I and II cordectomy presented complete glottic closure. In this same group glottic mucosal vibration was always present; only 1 patient (individual number 8) had supraglottic vibration. This same patient was the only one in this group (cordectomy types I and II) that had laryngeal vestibule hyperconstriction. All subjective and objective parameters in this patient, except for the fundamental frequency, were worse compared to other patients in the same group. A further interesting point is that even in type II cordectomy, glottic closure was complete in 4 of the 5 operated patients. This is relevant because voice quality is intimately related to glottic closure. Frequently in situ carcinoma is treated with type I cordectomy, and the surgeon may find himself surprised with the pathological examination showing compromised surgical resection margins. The next step would be to widen the surgical margins, which means the patient has to undergo a second operation or be referred to radiotherapy. Knowing that the post-operative functional result is similar in type I and II cordectomies, a safer and wider margin procedure for Tis could be advised.
Videolaryngostroboscopy of type III and IV cordectomies showed that 57.1% (n=4) of patients presented incomplete glottic closure, an increased trend to have laryngeal hyperconstriction as a compensation mechanism and no mucosal vibration. Stroboscopic changes correlated with voice quality data in study group patients, probably due to the amount of tissue removed, decreased mucosal vibration and compensation mechanisms.
Although results suggest that patients undergoing CO 2 laser cordectomy have abnormal voice parameters, the possibility of preserving communication is itself reason for satisfaction. Frequently patients lose motivation when receiving a diagnosis of cancer of the larynx, believing that treatment is necessarily related to voice loss, which is so important for social interactions. Expectations of voice professionals may not be reached, but this is not what happens to most patients; they frequently surprise us by being satisfied with their resulting voice quality.

CONCLUSION
Voice perceptive-auditory analysis indicates that most patients undergoing CO 2 laser cordectomy presented some degree of dysphonia, usually hoarseness and soprosity.
Both the general voice assessment made by patients and the quality of life protocol scores suggest that CO 2 laser cordectomy produces acceptable functional results, with minor repercussion on the patient's lives.
Acoustic voice analysis showed a minor increase in the fundamental frequency of operated patients compared to controls. This difference was not statistically significant. Worse jitter, shimmer and the harmonics-to-noise ratio results were seen in patients undergoing CO 2 laser cordectomy.