J Jpn Prosthodont Soc 52543-549, 2008

ORIGINAL ARTICLE

Investigation of the Factors Influencing the Outcome of Prostheses on Speech Rehabilitation of Mandibulectomy Patients
Ayako Hagino, DDS, Ken Inohara, DDS, Yuka I. Sumita, DDS, PhD, and Hisashi Taniguchi, DDS, PhD
Department of Maxillofacial Prosthetics, Graduate School, Tokyo Medical and Dental University, Tokyo, Japan

Clinical significance

Speech rehabilitation with prosthodontic treatment af ter mandibulectomy is difficult and hard to predict. Our study revealed that the ease of tongue movement, no soft tissue grafting, and mandibular continuity were related to the recovery of speech ability with prosthodon tic treatment, and we can predict the recovery before prosthodontic treatment.

sue grafting, and the continuity of mandibular bone contribute to the recovery of speech ability with prosth odontic treatment.

Key words: maxillofacial prosthesis, mandibulectomy,

speech rehabilitation, continuity of mandible, tongue movement

Abstract

Purpose: The aim of this study was to investigate the

Introduction
As a result of a combination of surgery and neoad juvant therapy, the survival rate of head and neck cancer patients has markedly improved.1 Howev er, resection of the maxillofacial region can cause functional disorders in speech, mastication, and swallowing.2,3 Speech is very important as a means of communication, and accordingly, speech disor ders caused by resection impair the quality of life of maxillofacial patients.4,5 Therefore, the rehabili tation of speech is one of the most important aims of maxillofacial prosthetic treatment. It has previ ously been reported that the speech ability of the maxillectomy patient can be improved with a max illofacial prosthesis.6-8 On the other hand, there are no detailed reports about speech rehabilita tion of mandibulectomy patients with prosthodon tic treatment. We found in the clinic that the speech ability of some of the mandibulectomy patients improved with prosthodontic treatment, remained the same in some, and deteriorated in others. The aim of this study was to ascertain the factors that effect the recovery of speech ability with prosth odontic treatment. The monosyllable Speech In telligibility (SI) test, which is one of the most wide spread methods of evaluating the speech ability of maxillofacial patients in Japan,9 was used to ex amine the change in speech ability after prosth odontic treatment. It is surmised that factors such as tongue movement and the stability of dentures have an effect on speech ability. Thus the type of resection (marginal, segmental, or hemimandibu lectomy), whether soft tissue grafting was under
543

factors influencing the outcome of prostheses on speech rehabilitation of mandibulectomy patients. Methods: Eleven patients (6 males and 5 females) who underwent mandibulectomy without glossectomy be cause of a tumor participated in the study. A Speech Intelligibility (SI) test was applied without and with a prosthesis to evaluate their speech ability. The type of resection, whether soft tissue grafting was undertaken or not, the continuity of mandibular bone, and the num ber of remaining teeth related to the stability of the prosthesis were determined from the medical records. The some of acoustic features, Formant 1 and Formant 2 range, were investigated to evaluate objectively the limitation of tongue movement. Five questionnaires were sent out to evaluate subjectively the difference in sense of discomfort while speaking with and without the prosthesis. These eleven items were entered into stepwise multiple regression models to determine the predictors of the differences in SI score without and with a prosthesis. Results: Three variables, the ease of tongue movements, whether soft tissue grafting was undertaken or not, and whether the mandibular bone was continuous or not, contributed to the recovery of speech ability with prosth odontic treatment. Conclusion: The ease of tongue movement, no soft tis
 Corresponding to: Dr Ayako Hagino Department of Maxillofacial Prosthetics, Graduate School, Tokyo Medical and Dental University 1-5-45 Yushima, Bunkyo-ku, Tokyo113-8549, Japan Tel: +81-3-5803-5557, Fax: +81-3-5803-5557 E-mail: ayako.mfp@tmd.ac.jp Received on January 25, 2008 /Accepted on September 9, 2008

544

Hagino et al., J Jpn Prosthodont Soc 52 : 543-549, 2008

Table 1 The patients clinical profiles.

means without reconstruction;

means with reconstruction.

taken or not, the continuity of mandibular bone, the number of remaining teeth related the stabil ity of the prosthesis, and Formant 1 (F1) and For mant 2 (F2) range which are some of the acoustic parameters to evaluate the limitation of tongue movement10 were investigated. In order to deter mine whether the patient could speak without dif ficulty, questionnaires were sent out to evaluate the differences in sense of discomfort while speak ing with and without the mandibular prosthesis.

Materials and Methods

1. Subjects and prostheses Eleven patients (6 male, 5 female), who had under gone partial mandibulectomy without glossectomy for a tumor, participated in this study. All were na tive speakers of Japanese, with normal hearing abilities. The clinical profiles of each patient in cluding age, gender, time after surgery, primary tumor, the type of resection (marginal, segmental, or hemimandibulectomy), the type of reconstruc tion, the remaining teeth, and the continuity of mandibular bone are summarized in Table 1. We had found in a prior study that there were no sig

nificant differences in the SI score before and after prosthodontic treatment (before 85.9 12.6, after 83.3 11.4, p = 0.253 ). All had prostheses that had been used for at least one month without prob lems. All subjects used a conventional prosthesis, not an implant-supported one. Three of the eight sub jects in the marginal and segmental groups had no teeth in the mandible and a resin-based, remov able complete denture (RCD) was given to these patients. Four of the subjects in these groups had missing teeth on the non-surgical side of the man dible. A resin-based bilateral free-end saddle re moval partial denture (RPD) with clasps was giv en to these four subjects. The last subject in these groups had no missing teeth on the non-surgical side of the mandible. A resin-based unilateral freeend saddle RPD with clasps was given to the sub ject. One of three hemimandibulectomy subjects had no missing teeth on the non-surgical side of the mandible. A resin-based unilateral free-end saddle RPD with clasps was given to these sub jects. The other subject had no teeth in the man dible and a resin-based RCD was given. Two of the subjects in the marginal and segmen tal groups had no teeth in the maxilla and wore a

Factors Influencing on Speech Rehabilitation with Prosthodontic Treatment after Mandibulectomy

545

resin-based RCD, one wore a resin-based bilateral free-end saddle RPD with clasps, and two wore a resin-based unilateral free-end saddle RPD in the maxilla. All the hemimandibulectomy subjects used prostheses with a palatal ramp in the maxilla, as they had mandibular deviation. Two of three hemi mandibulectomy subjects who had no missing teeth in the maxilla wore a resin plate with a palatal ramp, and one of the hemimandibulectomy sub jects who had no teeth in the maxilla wore a resinbased RCD with a palatal ramp. Informed consent was obtained from each par ticipant verbally and in writing. This research had the approval of the Ethical Committee of Tokyo Medical and Dental University (no. 297).

ences in F1 and F2 ranges between without and with a mandibular prosthesis were calculated.

2. SI test A one hundred monosyllable SI test was applied in the conventional way. One hundred Japanese monosyllables from each patient were recorded on a digital analog tape recorder (TCD-D100, Sony, Tokyo, Japan) using a high quality dynamic micro phone (SM58, Shure, Niles, IL, USA). The sam pling rate was at 44 kHz. The recorded utterances were played back in a quiet room to five listeners who were linguistically untrained and had no im pairment in hearing. The SI score was calculated as the mean percentage of correct responses from three of the five listeners after the maximum and minimum scores had been excluded. In this study, the SI score of each patient under the conditions without and with a mandibular prosthesis was measured, and also the scores of each patient with out prosthesis were subtracted from the score with prosthesis and computed as the difference between the two SI scores. 3. Acoustic analyses The F1 range and F2 range were examined using Sumitas methods.10 The five Japanese vowels, /a/, /e/, /i/, /o/, and /u/, were used as the sample vowel sounds. The speech recording was performed in a sound-insulated room using CSL4400 (KayPentax, Lincoln Park, NJ, USA). Each subject was seated and a high quality dynamic microphone (SM48, Shure, Niles, IL, USA) was positioned at a con stant distance (20 cm) from the lips. Each subject was asked to take a brief breath and to sustain the vowel sound for 34 sec. Acoustic analysis was per formed using a personal computerbased analysis. A segment of 1 sec duration from the middle por tion was extracted for further analysis. The differ

4. Questionnaire The questionnaire was applied in each condition, with the mandibular prosthesis and without the mandibular prosthesis. The following questions were included in the questionnaires. Q1: When you speak, can you move your tongue as you wish? Q2: When you speak, do you feel your voice muf fled? Q3: When you speak, do you feel air leaks? Q4: When you speak, can you move your lower jaw as you wish? Q5: When you talk to others, does someone say that your voice is not clear? The details of the questionnaires are shown in Fig. 1. A Visual Analog Scale (VAS) was utilized for eval uating the responses. The questionnaires were sent out at the same time as examinations were carried out by an independent dentist not in charge of the patient so as to reduce bias in the answer. The differences in VAS scores of each question un der the conditions without and with a mandibular prosthesis were calculated as the scores of each patient. 5. Statistical analysis Stepwise multi-regression analysis was used to examine the factors that effect the outcome of prosthesis in speech rehabilitation. The candidate explanatory variables included the type of resec tion (marginal, segmental, or hemimandibulecto my), whether soft tissue grafting was undertaken or not, the number of remaining teeth, whether the mandibular bone was continuous or not, the difference in F1 range and F2 range without and with the prostheses, and the score of each ques tion. All variables were entered into the stepwise multiple regression model to identify independent predictors. The three predictors, the type of resec tion, whether soft tissue grafting was undertaken or not, and the continuity of mandibular bone, were applied as the categorical data and the other pre dictors were applied as numerical values. The nu merical data of eleven predictors which were en tered into a multi-regression model are shown in Table 2. The categorical number of the three pre dictors was given as follows. In the case of conti nuity of mandibular bone, 1 was entered into the model if the mandibular bone was continuous, oth erwise it was 0. In the case of soft tissue grafting, 1 was entered into the model if soft tissue graft ing was undertaken, otherwise it was 0. In the

546

Hagino et al., J Jpn Prosthodont Soc 52 : 543-549, 2008

Fig. 1 The details of the questionnaires evaluated by Visual Analog

Scale (VAS).

Table 2 The numerical values that were entered into the stepwise multi-regression model. The differences in SI, F1 range, and

F2 range without and with a prosthesis, and the scores of Q1, Q2, Q3, Q4, and Q5 were applied as numerical values. The type of resection, whether soft tissue grafting was undertaken or not and the continuity of mandibular bone, were applied as the categor ical data. In this table, SI, F1 range, and F2 range mean the difference of each score without and with a prosthesis, and Q1, Q2, Q3, Q4, and Q5 are the scores for each question.
SI score (%) A B C D E F G H I J K 4.7 0 5.3 6.7 16 2.6 5.6 3.7 9.4 8 7.3 Continuity 0 1 1 1 1 1 0 1 1 1 0 Graft 0 1 1 1 1 0 0 1 1 0 0 Marginal Segmental Remaining F1 range teeth (Hz) 0 0 1 0 0 1 0 1 0 1 0 0 0 0 1 1 0 0 0 1 0 1 6 9 9 4 0 4 0 0 0 11 3 160 24 23.26 27.33 22.54 51 65 55.69 43.95 8 17.93 F2 range (Hz) 16 3 102.03 737.64 40.32 150 146 104.03 8.39 126 25.78 Q1 0 18 0 9 23 3 4 12 47 0 23 Q2 0 30 0 30 0 0 0 60 8 12 19 Q3 0 26 20 0 0 0 4 62 0 0 31 Q4 0 28 0 1 23 3 3 29 0 3 15 Q5 0 9 21 1 44 4 2 28 0 0 21

case of the type of resection, there were three cat egories. We categorized patients in two ways: if the type of resection was marginal, 1 was entered into the model and the others were 0; if the type of resection was segmental, 1 was entered into the model and the others were 0. With a combina tion of these two categorized numbers, three types of resection could be represented. Then eleven vari ables, the type of resection (marginal, segmental, or hemimandibulectomy), whether the soft tissue grafting was undertaken or not, the number of re maining teeth, whether the mandibular bone was continuous or not, the difference in the F1 and F2

ranges without and with prostheses, and the score of each question were inserted into the stepwise regression model to determine their relevance. A p value of less than 0.05 was considered to be statis tically significant. Data were analyzed using the statistical package SPSS II for Windows 13 (SPSS Inc., Chicago, IL, USA).

Results
A high Q1 score, no soft tissue grafting, and man dibular bone continuity were significant predic

Factors Influencing on Speech Rehabilitation with Prosthodontic Treatment after Mandibulectomy

547

Table 3 The result of stepwise regression analysis in which

the difference in SI score was inserted as a dependent vari able. R2 (coefficient of determination) in this model was 0.926. B: non-standardized regression coefficient.
Independent variable The score of Q1 Whether the soft tissue graft ing undergone or not The continuity of mandibular bone

B
0.334 13.025 8.662

Statistical error 0.044 2.001 2.184

p
0.000 0.000 0.005

lar bone was continuous or not (X-axis). Y-axis, the difference in SI score as a percentage. The corresponding p-value was 0.05.

Fig. 4 The difference in SI score and whether the mandibu

Fig. 2 The difference in SI score and the difference in VAS

score of Q1. Y-axis, the difference in SI score as a percentage; X-axis, the difference in VAS score of Q1. The corresponding p-value was 0.05.

with prostheses, and the score of Q2, Q3, Q4, and Q5. The final regression equation was as follows: the difference in SI score = 2.861 + 0.334 (the Q1 score)13.025 (whether soft tissue grafting was undertaken or not) + 8.662 (whether the mandib ular bone was continuous or not). R2 (coefficient of determination) in this model was 0.926.

Discussion
As the score of Q1 (Can you move your tongue as you wish?) was found to be a statistically signifi cant predictor of the difference in SI score without and with a prosthesis, it is considered that the re habilitation of speech ability is highly dependent on the ease of tongue movement. On the other hand, the F1 range and F2 range did not contrib ute to the difference in SI score. In this study, the tongue movement was examined by two methods, one was the F1 and F2 ranges which are consid ered to be an index of the limitation of tongue movement and the other was Q1 which evaluates the ease of tongue movement. It is considered that the ease of tongue movement relates to whether the tongue can be quickly moved. However, in anal ysis of the F1 range and F2 range, a segment of steady and continuous utterance of 1 sec duration was used as a voice sample of each patient and it is not thought that this method evaluates quick tongue movement. Therefore it was clarified that instantaneous tongue forming is needed for artic ulation. Glossectomy patients were excluded from this study. However, some of the subjects answered that they had difficulty moving the tongue, and it

grafting was undertaken or not (X-axis). Y-axis, the difference in SI score as a percentage. The corresponding p-value was 0.05.

Fig. 3 The difference in SI score and whether soft tissue

tors of the recovery of speech ability with prosth odontic treatment (Table 3, Figs. 2-4). There was no significant association between the difference in SI score and eight variables, namely the type of resection (marginal, segmental, and hemimandib ulectomy), the number of remaining teeth, the dif ference in F1 range and F2 range without and

548

Hagino et al., J Jpn Prosthodont Soc 52 : 543-549, 2008

was considered that there was every possibility that the difficulty with tongue movement was caused by the operation. The state of disability of tongue movement was variable in each patient and it is necessary to think about the form of the prosthesis that accommodates each patients re quirements without obstructing tongue movement. However, the appropriate form of the prosthesis that supports the patients tongue movement has not been clarified yet. Therefore, in a further study we will examine the factors that influence the ease of tongue movement with a view to fabricating a prosthesis that can support tongue movement. The reason why the difference in SI score of the patient who underwent soft tissue grafting dete riorated significantly and the number of remain ing teeth did not contribute to the difference of SI score is that the condition of the defect affects the stability of the prosthesis more than the number of remaining teeth. Thus, either the flap was too large and there was not enough space for the pros thesis or the flap was too soft and could not retain the prosthesis, making it difficult for the patient to speak smoothly regardless of the number of re maining teeth. However, objective methods to eval uate the condition of the flap in the defect area that affects the stability of the prosthesis, such as the thickness and the stiffness, have not been es tablished yet. Therefore, in this study, the classifi cation whether the graft was undertaken or not, was used as the clearest method for evaluation of the condition of the defect. In future studies, the establishment of a method for evaluation of the condition of the flap in the defect area will be re quired. From the result that the continuity of the man dibular bone contributed to the difference in SI score, it is suggested that the possibility of the re covery of SI score with prosthodontic treatment is not so high when the mandibular bone is not con tinuous. It is considered that there are adverse ef fects on mandibular movements during speech following surgical resection and the mandibular movements in discontinuous mandibulectomy pa tients exhibit characteristic deviation and rotation in the frontal plane.11,12 In addition, it is difficult to make mandibular movements stable with prosth odontic treatment, and this was reflected in the score of Q4 (Can you move your lower jaw as you wish?). The score of Q4 shows that almost all pa tients thought that they could not move their mandible more easily with the prostheses than

without the prostheses. It was considered that the surgical invasion, regardless of the extent of resec tion, had a greater effect than envisaged and caused a sense disorder of the mucous membrane in the defect area. It is suggested that the type of resection, classi fied by surgical technique, does not contribute to the difference in SI score. From the results, if the continuity of the mandibular bone is regained, re gardless of the type of resection classified by surgi cal technique, the possibility of the recovery of speech ability with prosthodontic treatment is high. As no significant relevance was confirmed be tween the difference in SI score and the score of Q3 (Do you feel air leaks?), it cannot, at present, be said that speech ability is affected by breath leakage and is restored if air leakage can be pre vented. It is thought that this is because air leak age does not affect the distinctness although it does affect vocal quality. Although the results indicated that the score of Q2 (Do you feel your voice muffled?) did not con tribute to the difference in SI score, many of the patients thought that the prostheses made their voice more muffled. Thus, it is suggested that wear ing the prosthesis makes the voice indistinct but the person listening in the determination of the SI score was able to hear correctly even if the voice became more muffled. In this study, two of five patients who did not undergo soft tissue grafting recovered speech abil ity with prosthodontic treatment and the other three did not. Only one of six patients who under went soft tissue grafting recovered speech ability, one of the patients had no change, and the other four had worse speech ability. All three patients whose mandibular bones were not continuous had deterioration in speech ability with prosthodontic treatment. Speech ability in three of the eight pa tients whose mandibular bones were continuous showed recovery, while there was no change in one patient and worsening in four patients. In this study, it was difficult to recruit subjects because there were not so many mandibulectomy patients without glassectomy, making it impossible to de termine statistically significant correlations and the formula cannot be used universally. However the above-mentioned results were almost similar to the results of the multi-regression model and showed similar tendencies. From our study, clini cians can inform patients of the prognosis prior to

Factors Influencing on Speech Rehabilitation with Prosthodontic Treatment after Mandibulectomy

549

the start of prosthodontic treatment that speech ability is difficult to rehabilitate with prosthodon tic treatment in patients with a discontinuity de fect or who have undergone soft tissue grafting, but speech ability can be recovered with a prosthe sis that supports tongue movement. In a further study, an increase in the number of subjects is needed to be able to establish the universal for mula and determine whether prosthodontic treat ment is useful or not for speech rehabilitation in patients with differing conditions.

Conclusion
It is considered that the ease of tongue movement, not undergoing soft tissue grafting, and mandibu lar bone continuity are significant predictors for the recovery of speech ability with prosthodontic treat ment.

References
1. Velich N, Vaszilk M, Nmeth Z et al. Overall sur vival of oropharyngeal cancer patients treated with different treatment modalities. J Craniofac Surg 2007; 18: 133136. 2. Pauloski BR, Logemann JA, Rademaker AW et al. Speech and swallowing function after oral and oro pharyngeal resections: one-year follow-up. Head Neck 1994; 16: 313322. 3. Marunick MT, Mathes BE, Klein BB. Masticatory function in hemimandibulectomy patients. J Oral Rehabil 1992; 19: 289295. 4. Meyer TK, Kuhn JC, Campbell BH et al. Speech in

telligibility and quality of life in head and neck can cer survivors. Laryngoscope 2004; 114: 19771981. 5. Karnell LH, Funk GF, Hoffman HT. Assessing head and neck cancer patient outcome domains. Head Neck 2000; 22: 611. 6. Rieger J, Wolfaardt J, Seikaly H et al. Speech out comes in patients rehabilitated with maxillary ob turator prostheses after maxillectomy: a prospec tive study. Int J Prosthodont 2002; 15: 139144. 7. Sullivan M, Gaebler C, Beukelman D et al. Impact of palatal prosthodontic intervention on communi cation performance of patients maxillectomy defects: a multilevel outcome study. Head Neck 2002; 24: 530538. 8. Umino S, Masuda G, Ono S et al. Speech intelligibil ity following maxillectomyAnalysis according to manner and place of articulation. J Jpn Stomatol Soc 1999; 48: 6065. (in Japanese with English ab stract) 9. Kubo Y, Nakagawa H, Kawano F. Assessment of functional rehabilitation of maxillofacial patients Survey of the papers from Journal of Maxillofacial Prosthetics (19952004). J Maxillofacial Prosthet 2005; 28: 97102. (in Japanese with English ab stract) 10. Sumita YI, Ozawa S, Mukohyama H et al. Digital acoustic analysis of five vowels in maxillectomy pa tients. J Oral Rehabil 2002; 29: 649656. 11. Taniguchi H, Ishiwata H, Ohyama T, et al. Stability of the mandibular position in occlusion of mandibu lectomy patients with lateral discontinuity defect. J Oral Rehabil 1997; 24: 849856. 12. Takahashi M, Hideshima M, Park I et al. Study of mandibular movements in mandibulectomy pa tientsborder movements and functional move ments during mastication, deglutition and speech. J Med Dent Sci 1999; 46: 93103.