Int. J. Oral Maxillofac. Surg.

2009; 38: 339–345

doi:10.1016/j.ijom.2009.01.009, available online at http://www.sciencedirect.com

Clinical Paper
Reconstructive Surgery

Analyses of speech intelligibility Y. Matsui1,2, T. Shirota2,

Y. Yamashita3, K. Ohno2
1
Department of Oral and Maxillofacial

in patients after glossectomy

Surgery, Graduate School of Medicine,
Yokohama City University, Yokohama
236-0004, Japan; 2Department of Oral and
Maxillofacial Surgery, School of Dentistry,

and reconstruction with Showa University, Tokyo 145-8515, Japan;

3
Department of Oral Rehabilitation, Showa
Dental Hospital, Showa University, Tokyo 145-
8515, Japan

fasciocutaneous/myocutaneous
flaps
Y. Matsui, T. Shirota, Y. Yamashita, K. Ohno: Analyses of speech intelligibility in
patients after glossectomy and reconstruction with fasciocutaneous/myocutaneous
flaps. Int. J. Oral Maxillofac. Surg. 2009; 38: 339–345. # 2009 Published by Elsevier
Ltd on behalf of International Association of Oral and Maxillofacial Surgeons.

Abstract. This study analyzed the results of speech intelligibility tests in patients with
tongue cancer who had undergone resection with the aim of making surgical
recommendations for flap design and inset, to improve speech function. A total of
126 patients, enrolled from 13 Japanese institutions, were classified into 3 groups
according to the resected site: lateral, anterior, and combined. The lateral group was
further divided into 3 subgroups and the anterior group into 2 subgroups according
to the size of resection. The speech intelligibility score was analyzed based on
articulatory site and mode: 5 articulatory sites (linguodentoalveolar, linguopalatal,
linguovelar, and their intermediates); and 7 articulatory modes (plosives, fricatives,
affricatives, grids, nasals, vowels and semivowels). Low speech intelligibility
scores were recorded at sites where flaps contribute directly to the pronunciation in
the lateral and combined groups and at the anterior part of the reconstructed tongue
Keywords: speech function; reconstruction; tu-
in the anterior group. Plosives and glides displayed low values in general. A radial mor surgery; tongue.
forearm flap had higher function in the lateral group than other flaps. The type of
flap had no effect in the anterior and combined groups. Surgical techniques and flap Accepted for publication 16 January 2009
selection to improve functional status in each type of resection are discussed. Available online 24 February 2009

Postoperative quality of life is increas-
ingly emphasized in the evaluation
of outcomes for patients who have
undergone surgical ablation of oral
cancer11–13. The resection of tongue can-
cer often causes postoperative speech
impairment, as the tongue is a primary
active articulator. Recent progress in
reconstructive surgery has offered new
therapeutic options suitable for the com-
plex anatomy and function of the ton-
gue2,3,6,9,10,14.
In a previous multicenter study7, the
authors investigated the postoperative
speech function of tongue cancer patients

0901-5027/040339 + 07 $36.00/0 # 2009 Published by Elsevier Ltd on behalf of International Association of Oral and Maxillofacial Surgeons.
340 Matsui et al.
following reconstruction with fasciocuta-
neous/myocutaneous flaps and reported
factors that influenced the functional
outcome of speech. Better results were
obtained for lateral resections, smaller
excisions, greater tongue mobility, youn-
ger patients, and where there was a longer
interval to examination. Mandibulectomy
and radiotherapy were associated with
poorer outcomes, while type of flap
reconstruction had no effect on functio-
nal outcomes including overall speech
intelligibility. That study also revealed
that postoperative functional level varies
widely among patients who have under-
gone the same type and size of resection.
This could be attributable to the surgical
intervention, such as size of the trans-
planted flap, and the suturing technique
used between the remaining stump and
flap.
The total intelligibility score can objec-
tively evaluate overall postoperative
speech function. Speech intelligibility
tests for syllables can also analyze sources
of speech disorders and provide informa-
tion for better speech function through
analyses based on articulatory mode and
site, regardless of the languages pro-
nounced8. Analyses of various articulatory
sites and modes have seldom been
addressed in the maxillofacial literature.
The present study aimed to identify
surgical recommendations for flap design
and inset, to improve speech function in
tongue cancer patients receiving each type
and size of resection, by analyzing the
results of speech intelligibility tests based
on articulatory mode and site.
Patients and methods
Subjects
A total of 126 patients (43 women, 83
men; mean age at examination, 58.5
years; range, 16–83 years) were enrolled
in the present study. Most subjects had
been included in the previous study, but
some were excluded owing to insufficient
data and new subjects were included in
this study. These patients had undergone
reconstructive surgery with fasciocuta-
neous/myocutaneous flaps after ablative
surgery for proven malignant tumor of the
tongue and/or floor of the mouth at 13
Japanese institutions. General informa-
tion about participants was collected by
the responsible surgeon, under a standard
protocol including: age; T stage; extent
of resection; radiotherapy; and type of
flap reconstruction. Staging was T1 in 6
patients, T2 in 74 patients, T3 in 27
patients and T4 in 19 patients.
Subjects were classified into 3 main
anatomical groups according to the
resected site: lateral (n = 95); anterior
(n = 13); and combined (n = 18). The lat-
eral group was further classified into 3
subgroups according to the size of resec-
tion: partial excisions of the anterior two-
thirds of the lateral border of the tongue
(group L1; n = 38); half resections of the
anterior two-thirds of the lateral border of
the tongue (group L2; n = 35); and com-
plete hemiglossectomies (group L3;
n = 22). All patients in the lateral sub-
groups underwent resection of the floor
of the mouth to some extent, but resection
never exceeded the midline. The anterior
group all displayed some excision of the
bilateral floor of the mouth with resection
of the tongue. This group was subdivided
according to depth of resection into: resec-
tion of the anterior floor of the mouth no
deeper than one-third of the full thickness
of the midline up to the mylohyoid muscle
plane (group A1; n = 5); and resection of
the anterior floor of the mouth deeper than
one-third and up to two-thirds of the full
thickness of the midline (group A2; n = 8).
Patients who underwent subtotal hemi-
glossectomy and/or resection of the ante-
rior floor of the mouth deeper than two-
thirds of the full thickness of the midline
were categorized as the combined group
(group C; n = 18). None of the 126
patients had undergone resection of the
soft palate or tonsillar region.
A total of 64 patients had received
radiotherapy, 53 preoperatively and 9
postoperatively. Mean irradiated dose
was 38.5 Gy (range, 18–95 Gy). Mandi-
bulectomy had been performed in 60
patients, with marginal mandibulectomy
in 49 patients and segmental mandibulect-
omy in 11 patients. After segmental
osteotomy, the mandible had been recon-
structed with a titanium plate at ablative
surgery in all except 1 patient, who under-
went simultaneous reconstruction with a
scapula flap. Of these 10 patients who
received reconstruction with a titanium
plate, 7 subsequently underwent mandib-
ular reconstruction, using bone harvested
from the ilium in 4 and the scapula in 3. In
group L1, all except 1 patient underwent
reconstruction with a free radial forearm
flap (RFF). In the remaining patient, a
vascularized pectoralis major myocuta-
neous flap (PMMCF) was used. In group
L2, the most popular flap type was RFF,
although some patients underwent recon-
struction with a free rectus abdominis
myocutaneous flap (RAMCF). In group
L3, reconstruction involved an RFF in
10 patients, a PMMCF in 6, and an ante-
rolateral thigh flap (ALTF) and RAMCF
in 3 each. In group A1, reconstruction
involved an RFF in 3 patients, a PMMCF
in 1, and a RAMCF in 1. In group A2,
surgeons used an RFF in 4 patients, and an
RAMCF and a PMMCF in 2 each. In
group C, flaps used were an RAMCF in
9 patients, an RFF in 6, a PMMCF and an
ALTF in 1 each.
Patients displayed no significant differ-
ences in mandibulectomy, irradiation, age
or time interval between subgroups.
Methods for evaluation
A speech intelligibility test (SIT)8,9 was
used for evaluation of postoperative
speech function. This is a monosyllable
test standardized by the Japanese Hearing
and Speech Association and the Japanese
Society of Logopedics and Phonatrics.
Japanese pronunciation uses 5 vowels.
A consonant is usually followed by a
vowel or semivowel to make a syllable,
which is the smallest vocal unit in Japa-
nese. The SIT was conducted in Japanese.
Each subject was instructed to pronounce
100 listed syllables, and pronunciations
were recorded. Ten untrained volunteers
with normal hearing, who did not know
the patient, then listened to the tape and
transcribed the sounds as Japanese sylla-
bles that they believed they had heard.
Intelligibility score was expressed as the
percentage of correct responses for all
listeners7. Duration between surgical
treatment and completion of the examina-
tion ranged from 6 to 177 months (mean
(standard deviation (SD)), 35.3  29.9
months; mode, 6 months; median, 28
months).
Seventy-nine of the 100 test syllables
were categorized as glossal sounds accord-
ing to palatolingual contact during articu-
lation by a normal speaker. These were
clasified into 5 subgroups based on articu-
latory site (linguodentoalveolar (LDA), lin-
guopalatal (LP), linguovelar (LV) and
intermediates LDA-LP and LP-LV). The
articulatory mode of glossal sounds
comprised 7 groups (plosives, fricatives,
affricatives, grids, nasals, vowels and semi-
vowels). As well as overall score, intellig-
ibility was analyzed based on articulatory
sites and mode (Table 1; Fig. 1).
The influence of the type of flap recon-
struction on speech intelligibility was sta-
tistically analyzed by comparing scores for
articulatory sites and manners for glossal
sounds between two kinds of flaps, as fol-
lows: in subgroup L2, RFF vs. RAMCF; in
subgroup L3, RFF vs. PMMCF, RFF vs.
RAMCF, and PMMCF vs. RAMCF; in
subgroup A2, RFF vs. PMMCF; and in
group C, RFF vs. RAMCF.
 Speech intelligibility after glossectomy 341

Table 1. Classification of Japanese glossal sounds LP had the lowest (20  21%) among the
LDA(23) Plosives (6) t & d (a,e,o) articulatory sites.
Fricatives (4) s (a,o,e,v)
Affricatives (5) tsv, dz(a,o,e,v)
Glides (4) l(a,o,e,v) Score based on articulatory mode
Nasal (4) n(a,o,e,v)
LDA-LP (4) Glides (4) l(ja,jo,jv,i) LDA
LP (36) Plosives (8) p & b(ja,jo,jv,i)
Fricatives (8) R & (a,o,v,i) In L groups, plosives exhibited the lowest
Affricatives (8) tR & dz(a,o,v,i) values (mean, 60–65%), followed by glides
Nasals (8) m(a,o,v,i), D(a,o,v,i) (mean, 66–69%). Significant differences
Vowel (1) i existed between plosives and other phona-
LP-LV (8) Plosives (8) k & g(ja,jo,jv,i) tions, excluding glides. Fricatives scored
LV (8) Plosives (8) k & g (a,o,e,v) the highest values (mean, 84–85%), with
Total (79) significant differences between all other
phonations in each group, except nasal in
Statistical analyses (Fig. 4). There were no significant differ- L3 (Fig. 5). No significant differences could
ences between A1 and A2, although A1 be found in the same phonation in each set
Statistical analyses were performed using
showed higher values than A2 in most of the 3 groups.
Stat View 5.0 J software (SAS Institute,
articulatory sites. In Group C, LDA had Plosives (33% and 49%) and glides
Cary, NC, USA). A paired t-test was
the highest value (39  24%) and LDA- (46% and 47%) in both A subgroups
utilized to analyze the effects of articula-
tory sites for glossal sounds in each sub-
group. An independent samples t-test was
used to compare results within groups L
and A with regard to articulation site for
glossal sounds and modes. An indepen-
dent samples t-test was also used for ana-
lyses on the influence of the type of flap
reconstruction on speech intelligibility.
Values of P < 0.05 were considered
statistically significant.

Results
Overall score
As in the previous study7, group L1 had
the highest score (69  25%) and group C
had the lowest (37  32%). Lateral sub- Fig. 1. Portions of the tongue dorsum related to lingography (adapted from Michi et al.8).
groups scored higher than both A sub-
groups and group C. Wider resection
resulted in lower intelligibility scores
within groups L and A.

Score based on articulatory site

Linguals showed lower values than other
sites in each group.
In the L subgroups, LDA (mean, 72–
74%) had the highest score, with a sig-
nificant difference between almost all
other sites. LDA-LP (mean, 37–51%)
had the lowest score, followed by LP-
LV (mean, 39–58%) (Fig. 2). Each set
of the 3 L subgroups revealed no signifi-
cant differences in LDA and LP, but L1
showed significantly higher scores than
L2 and/or L3 at LDA-LP, LP-LV and LV
(Fig. 3). Fig. 2. Results on glossal sounds based on the articulatory sites (L3 group). LDA had the highest
score, with significant differences compared with almost all other sites. LDA-LP had the lowest
LV had the highest values (mean, 62–
score, followed by LP-LV. LDA (LDA-LP: dt = 9.67, P < 0.0001; LP(cons): dt = 1.93,
67%) and the superiority of LDA (mean, P < 0.05; LP(v/sv): dt = 3.71, P < 0.001; LP-LV: dt = 6.4, P < 0.0001; LV: dt = 3.39,
50–59%) over other sites was not as clear P < 0.01), LDA-LP (LP cons: dt = 6.93, P < 0.0001; LP v/sv: dt = (5.77, P < 0.0001; LV:
in the A groups. A significant difference dt = (2.01, P < .0.05), LP cons (LP-LV: dt = 4.21, P < 0.001; LV: dt = 2.21, P < 0.05), LP v/sv
between LDA and another site only (LP-LV: dt = 4.57, P < 0.0001; LV: dt = 1.97, P < 0.05), LP-LV (LV: dt = (2.73, P < 0.01)
existed for LDA-LP in both A subgroups (T = 1.72). (Paired t-test).
342 Matsui et al.

Effects of type of flap

In the authors’ previous study7, overall
scores showed no significant differences
in any types of flap reconstruction within
the individual subgroups examined.
Certain types of flap in the L2 and 3
subgroups displayed statistical superiority
compared with other types in some articu-
latory sites and modes. In the L2 subgroup,
plosives of LP had significantly higher
scores (dt = 2.18, P < 0.03) in RFF
(51  25%) than in RAMCF (30  23%).
RFF also exhibited significantly higher
scores than the other two types of flap in
subgroup L3: RFF vs. PMMC, 58  19%
vs. 34  13% in plosives of LDA (dt =
2.41, P < 0.03) and 65  15% vs.
48  14% in plosives of LP-LV (dt =
2.16, P < 0.05), and RFF vs. RAMCF,
65  15% vs. 45  16% in plosives of
LP-LV (dt = 2.15, P < 0.05). Significant
Fig. 3. Results of glossal sounds based on articulatory sites (L groups). L1 showed significantly
higher scores than L2 and/or L3 at LDA-LP, LP-LV and LV. LDA-LP: L1 vs. L2 (T = 1.67; differences also existed between RAMCF
dt = 2.03, P < 0.05); L1 vs. L3 (T = 1.67; dt = 2.47, P < 0.01). LP-LV: L1 vs. L2 (T = 1.67; and PMMCF of the L3 subgroup: RAMCF
dt = 2.95, P < 0.01); L1 vs. L3 (T = 1.67; dt = 2.01, P < 0.05). LV: L1 vs. L3 (T = 1.67; vs. PMMC, 58  16% vs. 34  13% in
dt = 2.18, P < 0.05). (Independent samples t-test). plosives (dt = 2.41, P < 0.03) and 53 
14% vs. 71  7% in affricatives (dt =
2.16, P < 0.05) of LDA.
The type of flap had no effect on intel-
ligibility scores for each articulatory site
and mode in groups A2 and C.

Discussion
Mobility of the residual tongue, particu-
larly of the tip, has been emphasized for
better function after glossectomy8. LDAs
of all L subgroups showed the highest
intelligibility scores in analyses of articu-
latory site and no significant differences
existed between any sets of the 3 L sub-
groups. The reconstruction of the anterior
portion now in use can be assumed to be
functionally sufficient for patients who
Fig. 4. Results of glossal sounds based on articulatory sites (A1 group). LV had the highest underwent resection for lateral-type
values and superiority of LDA over other sites was not as clear. A significant difference between malignant tumor. The low scores for plo-
LDA and another site was only identified for LDA-LP. LDA (LDA-LP: dt = 2.66, P < 0.05), sives and glides of LDA indicate that
LDA-LP (LP(cons): dt = 2.91, P < 0.05), LP(v/sv); (LV: dt = (2.37, P < 0.05), LP-LV (LV: movement of the reconstructed tongue
dt = 2.66, P < 0.05). (Paired t-test). remains insufficient for pronunciation dur-
ing which tight contact to and quick
release from the dentoalveoli by the ton-
showed lower values than in L groups. A1 significantly higher in the L1 group gue are required. The low scores for LDA-
and A2 showed no significant differences, (58  27%) than in the L2 (39  25%) LP, LP-LV and LV are attributed to the
although A2 displayed lower values than and L3 (42  31%) groups. fact that only plosives or glides are pro-
A1 in all modes. A1 showed higher values than A2, nounced in these sites. More efforts to give
In group C, plosives were the lowest, at except for semivowels. mobility of the reconstructed tongue are
<20%, and fricatives were the highest, In group C, vowels (56  44%) and needed for better speech function.
at >50%. semivowels (42  35%) displa yed The study on the type of flap revealed
higher scores than consonants (mean, that RFF was statistically superior to
20–31%). Plosives of LP (20  21%) PMMCF and RAMCF in some arti-
LP
showed almost the same score as those culatory sites of plosives. No clear ten-
As a result of LDA, plosives were lower of LDA (19  27%). Fricatives of LP dency was evident between PMMCF and
than almost all other phonations in the (31  21%) were lower than those of RAMCF, although statistical differences
3 L subgroups (Fig. 6). Plosives scored LDA (56  32%). existed in some aspects. These results
 Speech intelligibility after glossectomy 343

to the palate or a particular airway for each

phonation. The direct contribution of flaps
for phonation decreases speech intellig-
ibility. To reduce the adverse effects of
flaps, reconstruction with symmetry on
coronal section is mandatory because flaps
that are too bulky or too small hinder
movement and/or positioning of the
remaining tongue for adequate articulation
in harmony with the residual stump. The
step of the reconstructed flap and the
remaining tissue should be avoided, as
this interferes with the normal linguo-
palatal contact and has some adverse
effects on speech intelligibility. In the
LV region, direct suture4 between the
margin of the pharyngeal wall and the
tongue base, if not hindering the mobility
Fig. 5. Results of LDA based on articulatory modes (L1 group). Plosives exhibited the lowest
of the tongue remnant, would be helpful
values. Conversely, fricatives scored the highest values. Plosives (fricatives: dt = 6.30, for L2 and L3 groups by facilitating ton-
P < 0.0001; affricatives: dt = (4.69, P < 0.0001; nasals: dt = 5.71, P < 0.0001), fricatives gue contact to the velar. Intraoperative
(affricatives: dt = 3.21, P < 0.001; glides: dt = 4.05, P < 0.001; nasals; dt = 1.73, P < 0.05), care to avoid pulling down the flap for
affricatives (glides: dt = 2.45, P < 0.01), glides (nasals: dt = (3.73, P < 0.001). (Paired t-test). the neck is also necessary to maintain the
posterior part of the flap in the upper
position and facilitate contact between
the velar and base of the tongue.
A jejunum flap9 might be another can-
didate for group L, as it is more pliable and
extensive than a dermal or myocutaneous
flap owing to the many surface folds and
apparently moves together with the stump.
It is difficult to reconstruct the symmetry
on the coronal section and maintain the
posterior part of the flap to the upper
position because it lacks flap volume.
Group A, as in the previous study, was
associated with lower total intelligibility
scores than group L. This study revealed
that the lower scores result from the low
scores at the front region; LDA. Subgroup
A showed lower scores than the L sub-
groups in plosives and glides of LDA. As
plosives pronounced at LDA, /t/ and /d/
require both broad contact for constriction
and subsequent quick motion of the tongue
Fig. 6. Results of LP based on articulatory modes (L2 group). Plosives showed the lowest value for release. Glide /l/ also needs both ton-
among all articulatory modes. Plosives (fricatives: dt = (6.43, P < 0.0001; affricatives: gue volume for complete linguopalatal
dt = (2.85, P < 0.01; nasals: dt = (5.76, P < 0.0001; semivowels: dt = (6.30, P < 0.0001; constriction along the dental arch or on
vowels: dt = (6.97, P < 0.0001), fricatives (affricatives: dt = 2.17, P < 0.05; nasals: the palate, and rapid release for adequate
dt = 2.19, P < 0.05; semivowels: dt = (1.84, P < 0.05; vowels: dt = (2.39, P < 0.05), affrica- pronunciation5. Functional reconstruction
tives (semivowels: dt = (2.94, P < 0.01; vowels: dt = (3.08, P < 0.01), nasals (semivowels: is particularly needed in the anterior part
dt = (2.56, P < 0.01; vowels: dt = (3.11, P < 0.01), semivowels (vowels: dt = (2.67, P < 0.01). of the tongue after ablation of anterior-
(Paired t-test). type malignant tumors. Differing from L
and C groups, the upper surface of the
tongue in group A can be reconstructed
objectively indicate that RFF should At LP, where no significant difference only with remnant tissue. The lack of
represent the first choice for reconstruc- existed among L subgroups, plosives significant difference between the two A
tion in group L, owing to the pliable showed significant differences between subgroups was partially attributed to this
nature of the flap. In the anterior part, L1 and the other 2 groups. Transplanted characteristic. At reconstructive surgery,
the line between tip and anterior border flaps in the L2 and L3 groups play a the remaining anterior muscles of both
of the floor of the mouth should be elon- greater direct role than that of L1 in pro- sides should be sutured directly to
gated, as a bilobe flap1,15, as much as nunciation at LDA-LP, LP and LP-LV, facilitate quick motion to the fullest
possible for higher mobility of the tip of because the reconstructed area of the two extent possible. Excessive, direct suturing
the tongue. subgroups is always used to form contacts induces deformed contours of the dorsum
344 Matsui et al.
at LDA-LP and LP sites. Such deformities
interfere with adequate contact and release
in these regions. Flaps should be com-
bined to avoid hindering mobility and to
sustain the tongue with adequate form in
the upper position for complete palatolin-
gual constriction along the dental arch or
on the palate. If the lift of the dorsum is
insufficient, intelligibility at the posterior
portion would decrease. The type of flap
had no effects on intelligibility scores for
each articulatory site and mode in A2
group. This is because a flap is mainly
used to sustain and reform the contours of
the dorsum not to mobilize the residual
tongue. Poor intelligibility of sounds
requiring the rear of the tongue has been
reported in patients who underwent resec-
tion of the anterior portion of the floor of
the mouth and reconstruction with a jeju-
num flap9. The results obtained here sug-
gest that, although direct investigation has
not yet been performed, such poor results
must be attributed to deficits in flap
volume.
Group C had the lowest score among
subjects at all articulatory sites. These
patients showed similar tendencies to
groups L and A in articulatory modes, as
plosives and glides showed very low scores.
Wide resection of the dorsum caused
patients to pronounce almost all phonations
with flaps. Aviv et al.1 suggested that a
bilobed RFF offers a promising recon-
structive option when trying to achieve
functional speech after near-total tongue
resection. The tip reconstructed using a
flap, whether bilobed or not, cannot be
elevated using the tongue remnant in group
C. The result that the type of flap had no
effect on intelligibility scores for each
articulatory site and mode in group C indi-
cates that the bulk of the reconstructed
tongue, rather than the type of flap utilized,
is essential for higher speech intelligibility.
Acknowledgements. The authors would like
to express sincere gratitude to the follow-
ing doctors: Kiyohide Fujita, Satoshi
Umino, and Ryoichi Kawabe, Department
of Oral and Maxillofacial Surgery, Yoko-
hama City University School of Medicine,
Yokohama Japan; Satoru Ozeki, Section of
Oral Oncology, Department of Oral &
Maxillofacial Surgery, Division of Oral
& Medical Management, Fukuoka Dental
College, Fukuoka, Japan; Kazunari Oobu,
Section of Oral and Maxillofacial Oncol-
ogy, Division of Maxillofacial Diagnostic
and Surgical Sciences, Faculty of Dental
Science, Kyushu University, Fukuoka,
Japan; Sadao Okabe, Department of Oral
Surgery, Saitama Cancer Center, Saitama,
Japan; Yukihiko Kinoshita, Institute for
Frontier Oral Science, Kanagawa Dental
College, Yokosuka, Japan; Masaru Hos-
oda, Department of Oral and Maxillofacial
Surgery, Kawasaki Medical College, Kur-
ashiki, Japan; Mikihiko Kogo, Tomomi
Yamamoto, Seiji Iida, Masaya Ohkura,
First Department of Oral and Maxillofacial
Surgery, Osaka University Graduate
School of Dentistry, Ohsaka University;
Kan-ichi Seto, First Department of Oral
and Maxillofacial Surgery, School of Den-
tal Medicine, Tsurumi University, Yoko-
hama, Japan; Katsunori Ishibashi, Kouichi
Asada, and Toru Sato, Second Department
of Oral and Maxillofacial Surgery, School
of Dental Medicine, Tsurumi University,
Yokohama, Japan; Masaro Matsuura, Sec-
tion of Oral Implantology, Department of
Oral Rehabilitation, Division of Clinical
Dentistry, Fukuoka Dental College,
Fukuoka, Japan; Takaomi Satomi and Hir-
oshige Chiba, Department of Oral and
Maxillofacial Surgery, Tokyo Medical
University, Tokyo, Japan; Teruo Amagasa,
Shoji Yamashiro, and Jyun-ichi Ishii, Max-
illofacial Surgery, Maxillofacial Recon-
struction and Function, Division of
Maxillofacial and Neck Reconstruction,
Graduate School, Tokyo Medical and Den-
tal University, Tokyo, Japan; Hidemi
Yoshimasu, Department of Community
Oral Health Care Science, School of Oral
Health Care Sciences, Faculty of Dentistry,
Tokyo Medical and Dental University;
Koji Fujibayashi and Yutaka Imai, Depart-
ment of Oral and Maxillofacial Surgery,
Dokkyo University School of Medicine,
Tochigi, Japan; Kenichi Tomitsuka,
Department of Oral and Maxillo-facial
Surgery, Course of Metabolic and Regen-
erative Medicine, School of Medicine,
Faculty of Medicine, Yamagata Univer-
sity, Yamagata, Japan; Mitsunobu Ono,
Kanchu Tei, and Yasunori Totsuka, Oral
and Maxillofacial Surgery, Department of
Oral Pathobiological Science, Graduated
School of Dental Medicine, Hokkaido Uni-
versity, Sapporo, Japan; and Ken-ichi
Michi, Yukihiro Michiwaki, and Satoko
Imai, Department of Oral and Maxillofa-
cial Surgery, School of Dentistry, Showa
University, Tokyo, Japan.
References
1. Aviv JE, Keen MS, Rodriguez HP,
Stewart C, Gund E, Blitzer A.
Bilobed radial forearm free flap for func-
tional reconstruction of near-total glos-
sectomy defects. Laryngoscope 1994:
104: 893–900.
2. Bodin IK, Lind MG, Arnander C. Free
radial forearm flap reconstruction in sur-
gery of the oral cavity and pharynx:
surgical complications, impairment of
speech and swallowing. Clin Otolaryngol
1994: 19: 28–34.
3. Hara I, Gellrich N-C, Duker J, Schön
R, Fakler O, Smelzeisen O, Honda T,
Satoru O. Swallowing and speech func-
tion after intraoral soft tissue reconstruc-
tion lateral upper arm free flap and radial
forearm free flap. Br J Oral Maxillofac
Surg 2003: 41: 161–169.
4. Hsiao HT, Leu YS, Lin CC. Primary
closure versus radial forearm flap recon-
struction after hemiglossectomy: func-
tional assessment of swallowing and
speech. Ann Plast Surg 2002: 49: 612–
616.
5. Imai S, Michi K. Articulatory function
after resection of the tongue and floor of
the mouth: palatometric and perceptional
evaluation. J Speech Hearing Res 1992:
35: 68–78.
6. Jacobson MC, Franssen E, Fliss DM,
Birt BD, Gilbert RW. Free forearm flap
in oral reconstruction: functional out-
come. Arch Otolaryngol Head Neck Surg
1995: 121: 959–964.
7. Matsui Y, Ohno K, Yamashita Y,
Takahashi K. Factors influencing post-
operative speech function of tongue can-
cer patients following reconstruction with
fasciocutaneous/myocutaneous flaps – a
multicenter study. Int J Oral Maxillofac
Surg 2007: 36: 601–609.
8. Michi K, Suzuki N, Yamashita Y, Imai
S. Visual training and correction of
articulation disorders by use of dynamic
palatography: serial observation in a case
of cleft palate. J Speech Hear Disord
1986: 51: 226–238.
9. Michiwaki Y, Schmelzeisen R, Hacki
T, Michi K. Articulatory function in
glossectomized patients with immediate
reconstruction using a free jejunal flap. J
Craniomaxillofac Surg 1992: 20: 203–
210.
10. Nicoletti G, Soutar DS, Jackson MS,
Wrench AA, Robertson G, Robert-
son C. Objective assessment of speech
after surgical treatment for oral cancer:
experience from 196 selected cases. Plast
Reconstr Surg 2004: 113: 114–125.
11. Rogers SN, Ahad SA, Murphy AP. A
structured review and theme analysis of
papers published on ‘quality of life’ in
head and neck cancer: 2000–2005. Oral
Oncol 2007: 43: 843–868.
12. Rogers SN, Miller RD, Ali K, Minhas
AB, Williams HF, Lowe D. Patients’
perceived health status following primary
surgery for oral and oropharyngeal can-
cer. Int J Oral Maxillofac Surg 2006: 35:
913–919.
13. Schliephake H, Jamil MU. Prospective
evaluation of quality of life after oncolo-
gic surgery for oral cancer. Int J Oral
Maxillofac Surg 2002: 31: 427–433.
14. Schliephake H, Schmelzeisen R,
Schönweiler R, Schneller T, Alten-
bernd C. Speech, deglutition and life qual-
 Speech intelligibility after glossectomy 345

ity after intraoral tumour resection. Int J Address: Kanazawa-ku

Oral Maxillofac Surg 1998: 27: 99–105. Yoshiro Matsui Yokohama 236-0004
15. Urken ML. Customized tongue recon- Department of Oral and Japan
struction with the radial forearm free flap. Maxillofacial Surgery Tel: +81 45 787 2659;
Presented at the Third International Con- Graduate School of Medicine Fax: +81 45 785 8438
ference on Head and Neck Cancer, San Yokohama City University E-mail: ymatsui@yokohama-cu.ac.jp
Francisco, July 1992. 3-9 Fukuura