Adult Rhabdomyoma of the Head

and Neck:
A Clinicopathologic and lmmunophenotypic Study

SILL00 B. KAPADIA, CDR, MC, USNR, JEANNE M. MEIS, MD,

DENNIS M. FRISMAN, LCDR, MC, USNR, GARY L. ELLIS, DDS,
DENNIS K. HEFFNER, CAPT, MC, USN,
AND VINCENT J. HYAMS, CAPT, MC, USN (RET)

Twenty-seven cases of adult rhabdomyoma (ARM) of the head and cle diHel-entiation. I-” F.\,erl though the histolocgy of ARM
neck are reported. The 20 male and seven female patients ranged i5 distinctive, it often is mistaken for a variety of other
in age from 33 to 80 years (median age, 60 years). Symptoms included lesions, particularly granular cell tumor (GCT)?’ Pre-
airway obstruction and a mass within the mucosa or soft tissue.
\kms reports describe the inilnunohistoc~~e~iii~~ll fea-
Median tumor size was 3.0 cm (range, 1.5 to 7.5 cm). Seven patients
tures of a relatively small number of ARhis using a lim-
(26%) presented with multinodular tumors and one tumor was mul-
ited panel of antibodirs.x~‘~ Our study documents the
ticentric. Follow-up was available in 19 cases and ranged from 2
months to 18.5 years after diagnosis (median, 6.0 years). Lesions
clirli~opatholo~i~ fwtures of a large series of ARMS of
recurred locally in eight cases (42%) 2 to 11 years after diagnosis the head and neck as wx4 as a broader ininiunoprofile
(median, 6 years). One recurrence was multicentric. Histologically, than has been recognic.ed previously. The utility and
ARM was composed of closely packed, large polygonal cells having potential pitfalls of imnlunohistochenlical stains as well
abundant, eosinophilic, granular, or vacuolated glycogen-rich cy- as their histogrnetic implications are discussed in the
toplasm with focal cross-striations. Immunohistochemical stains cwntesl of the niorpho1ogic features that aid in differ-
confirmed skeletal muscle differentiation; the majority of tumors ential diagnosis.
stained for myoglobin (21 of 21 tumors), muscle-specific actin (21
of 21 tumors), and desmin (19 of 21 tumors). Focal or rare immu-
noreactivity for vimentin (six of 17 cases), alpha-smooth muscle MATERIALS AND METHODS
actin (17 of 20 cases), S-100 protein (14 of 21 cases), and Leu-7 (10
of 20 cases) also was detected. Cytokeratin, epithelial membrane
antigen, glial fibrillary acidic protein, and CD68 antigen (with KPl)
were not found. The characteristic histology and immunophenotype
distinguish ARM from other lesions with which it is frequently con-
fused, including granular cell tumor, hibernoma, oncocytoma, and
paraganglioma. The expression of alpha-smooth muscle actin has
not been reported previously in ARM; its presence could reflect
aberrant expression of smooth muscle actin in skeletal muscle or
possibly be a recapitulation of early skeletal muscle embryogenesis.
HUM PATHOL 24:608-617. This is a US government work. There
are no restrictions on its use.

1~~iii~~~i~otii~~o~ti~~iii~~;lld;lins wcrc twrfornied in 21 cxc.

Adult rhabdon~yomas (ARMS) of the head and neck E‘i\,~-liri~i.c)iii~t~~ bv( riotis front fo~m~itin-fisect. tx~~~iffin-
are rare benign lesions that exhibit mature skeletal nius- tw~lwxlctett rrixterid ~erc stlldied with I 1 colnnicr-ci:ttl\; ;ir;iit-
~ll,k ;uililxdi~a ~lsillg rhc ;ividiii-l,iotiii ~~oiiiplcs (AB(:)
mctt~otl. Ix The so,,r( CJ.I\ tw . xid dilution of the zintigens used
w(‘rt ~15 foltobvs: musc~le-stxc-ific actin (MSA; inouse mono-
c-ton;rt. b:n/o Bioc~tieni. Inc. New York, NY; dill~tetl I :X,000),
dptl;l-sllloo~ll 11111>(It xtin (SblA: II~OUSC niono~toii;~t, Sip2
(‘.txvllic;it (:(I, S1 I,ouis, MO; ttitrlted 1 :X,000). dtwliin (~noi~sc
~~~onoclond; Boctlringer ~l;u~nl~rim Biochenlicals, Indianap-
olis, IN; diluted I :4(J). nlvoglobin (rxhbit lx~t~~~lo~~tl; DAKO.
<:q~ititri~Li, (:A; tlilurtd’ I :3.20(J), \iincntin (mouse niono-
ctond; I).-\K<); ttitutrtt I : I .liOO), S-100 pr-ocein (I-;lhbit poty-
clol1al; D.-\tiO; ttilurett I : I ,tiOO). gliat tit~rilt~ir~ xirlic protein
((~F.41’) (ratAt tx~lyc~t~~~~at:I)AKO; diturrtt I :.Gi)O), cylokentin
“cockr:Al” (mouse I11oI1o(.IoI1;lls. AE: t,‘AE3. RMR. diluted I:
100; (:K- I. I)AKO. ttilu~cd I : 100). epitheliat mrrnhl-ane x1-
lip31 IEXl.4; nlowst n10110~10~1;i1. DAKO, diluted I : IOU), (3tiH
(wirh KI’- 1. muse ~11o~~octo11al. DAKO: diluted 1:20). and I,rw
7 (n1otlse tnono~~lo11al;Be~toll-Di~kinsorl, Mouilt:Cn \‘iew, (:,A;
dilurrtl I :X). Set iions (‘01. deslnin, C;FAP. krratil1, EklA, ;1ntl
W-I ass;iy were predigested wirh 0.05% ptx,traxr‘ \‘I11 (Sig111a)

608
 ADULT RHABDOMYOMA OF THE HEAD AND NECK (Kapadia et al)

TABLE 1. Clinical Features in 27 Cases of Adult Rhabdomyoma

in 0. I moljl. phc~sphate buffer (pH 7.X) at 37°C: tar 3 minutes.

E~~dogenous prt-oxidase was blocked with a solution of 39:
h\drogen prroside in tnrthanol (3(Lminutc- incubation). kc-
tionz WCI-e placed cnernight in a solution of 10% normal horse
wnm for the moue n~onc~lonal antibody assays 01. 10% goat
st~rum for the rabbit polyclonnl antibody assays. Slides wcw RESULTS
then incubated \vitlt the primarv antibod; for 30 minutes fol-
lowed 1,~ biotinylated horse anii-mouse imtnunogl(~l)ulin (Ig) Clinical Features
(\‘rc.tor Labor-atories. Burlingdme, (:A; diluted I :400) for the
The clinical features are detailetf in Table 1. The
I~OLIW ~no~~~clo~~al antibodv assays or biotinylated goat anti-
rabbit Ig for tile rabbit polvclonal antibody assays (Vectot- median patient age was 60 years (range, 33 to 80 years).
I.ahoratories; diluted I :400) ‘and ,4BC Elite (Vector I~hor;I- ‘rhere were 20 male and seven female patients (male to
tories). A solution ot O.O24’$ hydrogen peroxide and 0.016%~ female ratio , 2.9:1). Of the 18 patients for who~n in-
cliaminobetl/.idirle tetrahydrochloride (DAB) (Signla) in phos- formation on race was available, 15 here rvhite. two were
phate-buff ered saline served as the chromogen. Srctions werr Hispanic, and one was black. The initial manifestation
com~terstained with Gill’s hematoxylin (Fisher Scientific Co. was a soft tissue mass (10 cases) or 3 nitucosal lesion in
Pittsburgh. PA). nlounted with Permount (Fisher Scientifit the upper aerodigestive tract (18 cases); separate tunlol
< .o), and covet-slipped, Tissue sections with knowrl I-eactivitc nodules involved mucosal and soft tissue sites in one
vith the antibodies assayed were used as positive controls.
case. In four cases the large mass involved cot$~ous
Normal mouse serum (diluted I :200) was substituted for the
mucosal sites, such as the base of the tongue, pharynx,
primary antibodies and served as the negative rontt-ol for the
mouse monoclonal antibodies assayed, whereas normal rabbit and larynx, with extension to the neck in two of these
serum (diluted I :2.000) served as ;he negative control for the cases. In decreasing order of frequent!. the mucosal
r-ahbit polv~lonal antibodies. Immunostaining was graded on sites involved were the pharynx. oral ca\‘tty, and larynx.
a scale of 0 to .I t based on the number and intensity of pos- The neck was the most common soft tissue site: one case

609
 HUMAN PATHOLOGY Volume 24, No. 6 (June 1993)

one other case AKM wxs an incidental finding at autopsy.

Six of the patients. all over 50 years of age. died of
uiirelated C‘;IIISC’S
(fi)ur of’ cxcinoniii and one of aspi-
ration pneumonia) from 1 to ti years after diagnosis. In
eight of‘ 1!J cases (~‘,Yo) turnor recurred locally L’ to 1 1
years after diagnosis (median inter~~al, 6 \~ears). Tunror
Las kliowll lo be illc-onipletel~~ excised in one of the
eight cases that recurred and in one other case complete
escisioir \vas perfometl at diagnosis and at both recur-
rences. The type of excision ‘~vas not stated in the rc-
maining six &es. Tunmr recurred tbic.e in two eases.
In another patient (case no. 0) a second AKM occurred
in a separate (niultkentlic) site. Of seven cases with a
Iiiultinod~il~il- tum01~ at onset, three recurred after sur-
gery, three were f’iec of tumor at lost follow-up, and
one was lost to follow-up. There were no nietastases.

Pathologic Features
On gross esaniinatioli the tumors were deep tan to
red-brown. cil~c.~inls~l~ihetl, lotmlated. and soft or fleshy
(Figs I ;untl 2). Subnnucosal tumors were sniooth, poly-
paid, and either sessile OJ‘ pedunculated. The niedian
tunio~~ six IV;IS3.0 cni (range, 1.5 to 7.5 CIII) in 20 cases
in which six bxs stated; in seven other cases the tunio~
FIGURE 1. Multinodular rhabdomyoma in buccal sulcus (cour- size ~vas tither unstated or described as “large.”
tesy of Dr P.H.J. Krogh, DDS).
Hi.stologically, all cases displayed the distinctive
features of AKM (Figs 3 to 7). Tu~~~c~r-swere composed
of‘~~elI-dc~n~u-~~lted. u~len~ap~ulatcd lobules (Figs 3 and
invoh4 the face. The duration ofsyniptcmis associated 4) of’ closel!, packed. large polygonal cells (Fig 5) with
with a mass lesion ranged frond 2 weeks to 3 years (me- sc‘ant stroma containing ec.tatic. to collapsed thin-walled
dian, 2 years). The mass w;is xi incidental &ding at vascular c~hannels. me tumor cells had small, round,
autopsy in one case and in two others it was found on centrallv or peripherally located nuclei with prominent
routine physical exmination for an unrelated condition. nucleoli and abundant, cosinophilic’. granular, 01‘ \ a(‘-
At onset the tumor usually was either 3 single mass or uolated cytoplasm (Fig 5). ,4 “spider web” ;q~pearatlce
nodule (19 cases) or multinodular (seven cases) with dis- was imparted to cells in which residual, radially oriented
crete tumor nodules in the sane anatonlic location (Fig strands of c$oplasni were seen around varying-sized,
I); one case was multicentric. (case no. 16) with separate peripheralI) located vacuoles (Figs 5 and 7). The cyto-
tumors in the larynx and suhmandibulal~ region. Ii plasm in all c‘ascs was rich in glycogen (periodic. acid-
Follow-up was a\-ailahfe in 1I) c‘ases and ranged li-oiii Sc~hifl-positive, diastase sensitive) (Fig 8). (:ytoplasnlic
2 months to 18.5 years (median interval, ti years); in c.l’oss-striatioris (Fig 6) and haphazardly arranged cq%al-

FIGURE 2. Gross appearance of tumor

from neck.

610
 ADULT RHABDOMYOMA OF THE HEAD AND NECK (Kapadia et al)

FIGURE 3. Soft palate mucosa overlying well-cje-

marcated lobules of ARM

lik ~t~~uctt~rw (Fig 7) were identified focalI) in henta- l)ttmot!p (ttt~o~lo~~itt-I)ositi\.e, JIS.4-positi\e. .ittd dcs-
t(~s\~litt-ro~iti-at~titted sections of tttost casc’s. ‘Fl~r mitt-positi\.c). hlttsc.le-specifics actitt (Fir< IO) and tttyo-
I( tr&t. wre better delineated with the ~Iasson tri- globitt bw~t‘ tlifitselv and strottglv posiiiw (3 I0 -I+) in
c.ltt-cmte (Fig 9) and phospltotungstic acid stains, and 2 1 01 L’I ~3st5. r&ttin slaining’ ( I !I of L’I (xY’~) was
l’ut-~hcr a(~ ctttuated with ittmtunostains for MSA (Fig cbf’lrrt p;tt(~tt\~ 01 more’ itttettsc* at tttv ~wri~~ttct-v 01‘ (ttc
IO), destttin, and tttyoglobin. Mitoses were not sc’en. Foci tissiw. proh1tt~l\ rc1atc.d to thr dqwta of fisatiott or ati-
of intlattttnation and necrosis with reactive nuclca~~ f’ca- tigc’tt pr~set~\at’ic~tr. Focal positi\,it\, 1’01’alpha-5M.4 was
tttres wet-c scc‘n in three ntucosal ARMS. prtwtit in 17 of20 c;tsc\. Vat-iahlv: oft(.tt Ivc~th positi\it\
Imtttrtttop”r”xidase staining results, tabulated in was c~t~srrvtd iii 1tmtor wlls for vitttt~ttlitt IIsix of I7 cases)
I able 2. c.ottfirrrted skeletal ntuscle differ-entiatiott in aitd S-t 00 proleitt (I 4 of 2 1 ( ;Iscs). ssc~;ll1tTrcl 1tll11ol‘

ARM. The tumor cells and adjacent r~ornl~~l skeletal c.vlls wcw positi\c with I.eu-i (10 of’20 C~SCX).No wac-
imscle, whert present, had a characteristic inttnttno- ti\itv IV;ISot,stwetl in ttttttor ( 4s li)r (;t*Al’. c.ylr)kc~t-atin.

FIGURE 4. Low-power view of multiple,

unencapsulated, circumscribed tumor lob-
ules.

611
 HUMAN PATHOLOGY Volume 24, No. 6 (June 1993)

FIGURE 5. Uniform closely packed polyg-

onal tumor cells with abundant granular or
vacuolated cytoplasm, well-defined cell
borders, and peripherally located nuclei.

EMA, or CD68 antigen (WI); scattered slronial cells studies of the extracardiac rhabdomvomas are necessary
were positive fi)r Wl and plasma cells fi)r EMA. 10 determine whether these rare l&ions are true neo-
In addition to ARM, the contI-ibutors’ differential 1,1as111s.
diagnoses included, in decreasing O&I- of fiquerq~. The c.linicopatholo~c features of ARM documented
granular cell tumor, hiber-nonia, onc‘or‘ytoma, and pxa- in our series are comparable to those previously re-
gariglioma. In one case a rhal~doIii~os~~~~~o~Ii~I was con- ported. ‘,‘wJ” Although rare cases have been reported
sidered. ill children, “‘Z’S”! all of our cases were in adults, with
a 3: 1 male predominance. An occasional ARM may be
asymptomatic and found on routine physical examma-
DISCUSSION tion or at autopsy. Howecer. the majority of patients
present with airway obstruction or a mass in the head
Rhahdo~n);orr~as are rxe benign mesenchymal tu- and neck region. Similar to the findings in our series,
mors with striated nius~le differentiation rhac are c.las- the most c‘o~iimm mucosal site is the pliaryIix,“.L”i.““-“‘i
sified by location into c.ardiac’!‘~“’ and extracardiat~ followed by the oral cavity (tongue, floor of the mouth,
types. ‘z”~~~~‘i.2’The more conIII~m cardiac rhabdomy- palate. or buccal niucosa)
I .J.,,i.,7.L”l,‘li.3X and lar-
wnas are composed of modified myocardial cells and ynx. ‘.’ ‘J!‘-‘~ The neck is the single most con~nIon soft
may be multiple or diffusely involve the heart. “‘z” The) iissuc site of ARM.‘,“,‘,‘9
are reported to occur in approximately 50% of patients The terms “I~~~l~~ifi,cal”X.li~:“‘~:‘fi.~l~~” and “lllulti_

with tuberous sclerosis, may spontaneously regress, and centric”” have been applied, at times interchangeably,17
are thought to represent haillartomas rather than true to indicate ARM occurring as more than one separate
neoplasms. ’ W’ Extracardiac rhabdomyomas occur much tumor; these occurred either in the same location’ OI
less frequently than cardiac rhabdomyomas and have in separate sites, such as in the submandibular region
no known association with tuberous sclerosis. ‘.’ The, and larvnx. Ii floor of the mouth hilaterally,gx bilateral
have a predilection for the head and neck region (90% lleck,+.G ’I_’
lght neck and retrolaryngeal area,“” floor of the
of cases), especially the upper aerodigestive tract (phar- mouth aild larynx, ” and oropharynx, base of tongue.
ynx, larynx, and oral cavity) or soft tissue, although they and vallecula.” In our series, 26% of tumors were mul-
may occur rarely in rmusual sites, such as the orbit. va- tinodular at onset but only two occurred as multicentric
gina, bladder, esophagus, trunk, and extremities.“.“.“~” tumors; one of these involved the submandibular region
Extracardiac I-habdomyonlas often present as a solitary and larynx at presentation and was reported
mass and may recur if incomplerely excised. Whether preGous1) “; in another case. a laryngeal ,4RM recurred
they represent hamartomas or true neoplasms is unclear. 5 years later as a separate tumor in the soft palate. The
However, a recent case study of a recurrent ARM dem- local recurrence rate was high in our series (42%), pre-
onstrated clonal cytogenetic abnormalities in tumor cells sumably occurring after incomplete resection. Three of
that involved a reciprocal translocation of chromosomes the seven originally multinodular tumors in our study
15 and 1’7 and miscellaneous abnormalities in the long recurred locally, as did five of 13 uninodular masses;
arm of chromosome 10; these findings support the no- the recurrence rates are comparable and suggest that
tion that ARM is a true neoplasm. “’ Further cytogenetic multinodular masses are not associated with a higher

612
 ADULT RHABDOMYOMA OF THE HEAD AND NECK (Kapadia et al)

FlGl IRE 6. Mild variation in tumor c:ell size

and shaDe Note cross-striations.

.f’lte diqqosis of ARM is based on recognition of

its distinctive histologic appearati~e’~“~‘~“.!‘~’ ’ or ewti
c.ltatxteristic. f’eatur-es in cytologic tttatet~ial obtained 1,)
fine-nredle aspir;ttiott.X~’ ‘di.‘H.4!’A variet!; of ncoplasnls
lxw~ abundant eosiuophilic. vacuolated. or gr-m~dar cy-

toplasm and at-e ft-equently confused with AKM: the

ntonl frrqucnt is ( 4.1, ’ ” _ ‘~“.‘~“’ followed 1)~hiberttottia, 01t-

FIGURE 7. Peripheral cytoplasmic vacu-

oles and haphazardly arranged crystalline
inclusions in tumor cells.

613
 HUMAN PATHOLOGY Volume 24, No. 6 (June 1993)

FIGURE 8. Periodic acid-Schiff-positive in-

tracytoplasmic glycogen (left) that is dias-
tase sensitive (right).

to either the Golgi apparatus or pliagolysosoti~es.~~~~‘~ either benign or malignant (oncocytic carcinomas). They
Hibernotnas are benign tutttors of brown fat usually oc- are composed of polyhedral cells having finely granular,
curring in the subcutaneous adipose tissue of the inter- eosinophilic cytoplasnt rich in tttitochortdria, as dent-
scapular region, axilla, or neck. Patients are frequently otistrated on ultrastructural examination. l’aragan-
younger than 40 years of age. Histologically, these tu- gliotnas are composed of polyhedral cells arranged in
mors are cotnposed of polygonal adipocytes with discrete an organoid pattern (Zellballen). The neuroendocrine
cell borders and central small nuclei. The abundant nature of tumor cells in paragangliottta may be con-
granular or finely vacuolated cytoplasm contains nu- firtned with ultrastructural demonstration of neuro-
tnerous small uniform lipid vacuoles and occasional ktitit secretor)’ granules or itnrnunohistochet~ti~~tl stains for
tinges of lipofuscin-like pigtnent. Oticocytotiias are ep- ntwroti-specific enolase. synaptophysiti, and/or cliro-
ithelial neoplasm of salivary gland origin that may be tiiogratiin.

FIGURE 9. Cross-striations accentuated

with the Masson trichrome stain.
 ADULT RHABDOMYOMA OF THE HEAD AND NECK (Kapadia et al)

FIGURE 10. Tumor cells demonstrate im-

munoreactivity for MSA. Note promil nent
cross-striations, A similar staining pattern was
observed for myoglobin and desmin.

111cmnfusing or clifidl cases the slwletal muscle IIiimulior-eat-ti\,it~ for viment in was f~ountl in otil!; six of
origin of ARM tna\: he confirmed by the dcmionstra- I 7 (uses in our study and often was l’oc~d alit1 pale. Sev-
tion of cytoplasnii;~ qlvcogcu and cross-striations 1,) twl of the AKILls previously rrporled li:rv~ keri negative
histochemical stains”‘+ or h,, ultrastructur;d ctetec- 101. vinieutiii.
tion of mvofilanieuts, Z-bands, and glycogcn gram The expression of alpl~~c-SR1,4 iti riial)dotiiyoblasts
ules TXII.IL’.~.~.L’~.~I:,: ‘is well iis Worig, diffuse immuno- of AKM has not ken reported prv\k)ual\..“’ Anti-alpha-
rractivity for MS,\, tlesmin, and myoglohin.“~’ ‘.“‘.‘-l Our SMA reacts spm3idlp with alpha-ShlA ~~ntl is riot found
study shows that tumor cells in ARM also ma\ demo~l- ill normal adult striated muscle.” Howt‘wr, in a recent
st rat’& focal or rare ilnl7iunorractivity for vim&tin aid 4tud\ 0I liuuian rhal)doni~osai.coIii~is. likalli vt al ob-
alpha-SM.4 along_ with MSA. clesmin, and mwglobin. wr~~i c.o-vspwssion of alpha-SMA and dpha-sarcom-

TABLE 2. lmmunohistochemical Staining of 2 1 Adult Rhabdomyomas

I :I, + ?i
I>
I- It I + ,’

3 il Ki If
1 4. I -t It I t
.rl I, j
_. It
Ii It It “+
7 it It L’t
8 Id 2t
!I I'
IO I/ It ‘1+
II I t- Ii “f
I3 1+ Ki :)I+
l-l 1+ Kt :<t
Iti 17 2 i- It
I7 It I+ It
I !I It I+ I+
'I'{
_. 1t If
L'-l 4t It
k)r,
_. 11 ITV
"6 1t I t-
27 It
 HUMAN PATHOLOGY Volume 24, NO. 6 (June 1993)

eric actin in one neoplasm.“” Alpha-SMA co-expression,

along with alpha-sarcomeric actin and desmin, also has
been observed by other investigators in chemically in-
duced rhabdomyosarcomas in rats.“’ These findings, as
well as those of our study, raise the possibility that alpha-
SMA may be present at some stage of rhabdonlyoblastic
differentiation in both rhabdomyosarcoma and rhab-
domyoma, and conceivably at some step in the embry-
onic development of skeletal muscle.” Further study of
normal myogenesis may be useful in understanding the
co-expression of alphaSMA and alpha-sarcomeric actin
in striated muscle tumors, such as rhabdoniyosarc.onla
and rhabdomyoma. Whether the expression of actin
isoforms in rhabdomyoma or rhabdomyosar-ct,nla faith-
fully recapitulates the pattern of expression seen during
normal skeletal muscle differentiation is unknown.5’i
S-l 00 protein and Leu-7 immunoreactivities are
potential sources of confusion in distinguishing ARM
from NIT; therefore, these antibodies should be studied
in conjunction with skeletal muscle markers. The tmlror
cells in ARM stain weakly for S-100 protein and only
scattered cells stain with Leu-7, whereas %Ts usually
stain diffusely and strongly for S-l 00 protein.
In summary, our study reaffirms the clinicopatho-
logic features of ARM previously reportecl in the liter-
ature, demonstrates the utility of imnlunohistotl~e~~~ic~~l
stains in distinguishing ARM from other tumors with
which, despite its distinctive morphology, it is frequently
confused, and demonstrates the previously unreported
expression of alpha-SMA along with skeletal muscle
markers in the rhabdomyoblasts of some ARMS. The
significance of alphaSMA in ARMS as well as rhabdo-
myosarcomas and developing striated muscle remains
to be elucidated.

616
ADULT RHABDOMYOMA OF THE HEAD AND NECK (Kapadia et al)

617