sialography in diagnosisanteroposterior and true lateral film3 u,ually give all the required...

521

SIALOGRAPHY IN DIAGNOSISBy S. S. ROSE, M.B., F.R.C.S.

First Assistant, Department of Surgery, University of Manchzster

HistoryThe injection of radiopaque media to supple-

ment a clinical opinion has long been regarded asa valuable aid to diagnosis. The importance ofsialography, however, although emphasized re-peatedly by Payne, seems to have escaped generalattention and comparativelv few papers have beenpublished on this subject.The first account of a sialogram carried out

as a diagnostic measure appeared in 1925 whenBarsony described a method of outlining theparotid duct, using 20 per cent. potassium iodide.This medium, however, proved to be so irritantthat he abandoned the technique. A year laterthe introduction of lipiodol provided a furtherstimulus to this form of inve3tigation, but, inspite of its widespread adoption, it was not untilI931 that R. T. Payne published the first of hisclassical series of papers on the subject. Usinga glass pipette of the type shown in Fig. i, heinjected 2 to i ml. of lipiodol very slowly into theappropriate duct orifice until discomfort was com-plained of. An X-ray was taken immediatelywith the instrument in situ. The main points inthe technique were that the apparatus was lightand so less likely to slip out of position during thesubsequent manipulations of the patient's head,and that there was no detachable needle to beforced off the end of the syringe. The latterdifficulty had been previously experienced onaccount of the force necessary to inject the ex-tremely viscid medium. This method of sialo-graphy gave a clear picture of the main duct sys-tem, and often extended into the finer ductsand acini when the injection was complete. Theuse of neohydriol fluid, however, has obviatedmost of the early technical difficulties, as its lowviscosity allows greater ease and speed of injectionwith the minimum of effort.

It has recently been argued that water-solubleradiopaque media should be used on account oftheir low viscosity and because they show no ten-dency to break up into globules. It has beenfound, however, that such media as viskiosol runquickly out of the gland and even when the sialo-gram is carried out with the needle in position,the resulting-picture is poor.

TechniqueThe present technique has been arrived at by

consideration of all the factors involved and hasproved satisfactory in a series of 93 sialograms,without any untoward effects. Routine straightX-rays are taken before sialography.An ordinary 5 ml. record syringe is used with

a blunt number i or 2 intravenous needle angledslightly in the upper third of its shaft to facilitateits introduction into the duct orifice. A lock-onneedle is sometimes advocated (Barsky and Silber-man, 1932) but has not been found necessarywhen the less viscous media are used.

There is usually no difficulty in finding theopening of Stensen's duct, opposite the site of thesecond upper molar tooth, by eversion of the cheekbetween thumb and forefinger. If it is not im-mediately obvious, massage over the gland andduct will produce a rapid flow of saliva from anorifice which dilates to two or three times itsnormal size. The opening of Wharton's duct,however, is situated on a mobile papilla whichmust be stabilized before inserttion of the bluntcannula. This may be achieveUd by the. u-se of afine non-toothed dissecting forceps whi'ch seizesthe papilla immediately above the opening andgently renders taut the line of the duct. Thetongue is pressed upwards and forwards againstthe hard palate. The use of I per cent. hydro-chloric acid solution may sometimes be necessaryto produce a flow of saliva and a supply is keptin readiness on the sialography tray. It is oftennecessary to dilate the duct orifice before insertionof the cannula and, for this purpose, modifiedlacrimal duct dilators are used (Fig. 2). In casesof duct stenosis, preliminary duct dilatation isessential before sialography, but over-dilatationwill give a false picture on X-ray and may allowtoo rapid emptying of the gland. The degree ofdilatation, therefore, should only be that necessaryto allow easy introduction of the cannula.The volume of fluid injected in this series has

been rather more than that advocated by otherauthors, namely I 2 to 2 ml. This does not increasethe discomfort, providing that the injection is notmade too quickly, and it is found that, althoughdiscomfort may appear after i ml. has been in-

by copyright. on January 29, 2021 by guest. P

rotectedhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

522 POSTGRADUATE MEDICAL JOURNAL October I950

I',)I

FIG. i.-Glass pipette of type recommended by Payne( 93 ').

4d

FIG. 3.-Parotid gland. Normal sialogram. I ml.Neohydriol.

jected, it does not increase appreciably as theinjection continues. By this modification of tech-nique it has been found possible to fill the glandso adequately that not only is the duct systemoutlined but an outline of the surrounding glandparenchyma is also visualized. The accuratedelineation of the gland is of particular value inthe diagnosis of space occupying lesions of thesalivary gland, for the lesion may be either toosmall or too peripheral to distort the duct anatomy.The needle is inserted with the patient seated

opposite the surgeon while an assistant supportsthe head. Care must be taken in cannulating the

......

....

FIG. 2.-Svringe, needles and dilators.

FIG. 4.-Parotid gland. Normal sialogram. 2 ml.Neohydriol. Showing better filling and de-finition. This is the small largely retromandibulartype of gland.

parotid duct so that the needle is not inserted so farthat it is occluded by the end of the buccal partof the duct, and it is found that a distance of 2 cm.is adequate. A finger is held over the duct orificeduring the injection and this is replaced by a lintswab which is removed immediately before theX-ray is taken. This ensures a minimum of lossinto the mouth with subsequent complication ofthe radiographic appearance. A good picture maybe obtained up to 20 minutes after the injection.There is, therefore, no need for the rather difficultmanipulation required with the cannula in situ.The routine views taken are as follows:


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

October 1950 ROSE: Sialography in Diagnosis 523

ig

....

FIG. 5.-Parotid gland. Normal sialogram. Large type of gland withwell marked socia parotidis. The gland extends well forward on to themasseter.

PAROTID GLANDAnteroposterior and true lateral film3 u,ually

give all the required information. The obliqueview serves only to distort the normal anatomyand is not carried out as a routine procedure. Itmay, however, be used in conjunction with thelateral view to elucidate spatial relationships ifthese are somewhat obscure.

SUBMANDIBULAR GLANDThe most useful views here are the lateral and

oblique lateral, but a basal view is occasionallyhelpful in establishing the extra- or intraglandularposition of a filling defect. The insertion of adental occlusion film in the floor of the mouthgives information about the intraoral part of theduct but is now only used when specially in-dicated, for example in the exact localization ofa stone in this region.

The Normal SialogramNormal sialograms achieved by this technique

are illustrated in Figs. 3, 4, 5 and 6. Figs. 3 and 4show the different degrees of filling obtained withinjections of i ml. and 2 ml. respectively in the

same gland. The main duct is uniformly fine andthere is no evidence of terminal dilatation. Fig. 7ashows a diagrammatic representation of the parotidduct system. Experience with the normal parotidsialogram shows that the duct anatomy is reason-ably constant. Fig. 7b similarly shows a diagram-matic representation of the submandibular glandshowing the right-angled bend where the ductruns down from the floor of the mouth into thegland substance. It is seen that the submandibularducts on the whole are much shorter and widerthan those of the parotid, an anatomical fact whichmay have a bearing on the relative frequency ofrecurrent infection in the latter gland.

Clinical Applicationr"he type of case which is constantly referred

for sialography is the recurrent parotid swelling,usu lly after a straight film has revealed noevidence of a calculus. The incidence of calculiin the parotid is as low as z per cent. of salivarystones and most cases of recurrent parotid swellingfall into one of the following groups:

(i) Recurrent pyogenic parotitis.(2) Simple duct obstruction.(3) Secondarily infected duct obstruction.


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

524 POSTGRADUATE MEDICAL JOURNAL October I950

p9 .p9 ... Al

.99 ,.9GW9

FIG. 7a.-Diagram of parotid sialogram. There areusually three ascending ducts as well as the ductof the socia, if present, and one or two descending-ducts depending on the size of the gland. Severalsmall retromandibular ducts drain the deep, partof the gland.

FIG. 6.-Submandibular gland. Normal sialogram.The right-angled bend where the main duct turnsdown into the gland is well marked. A posteriorsublingual duct is seen entering the main duct.

(i) RECURRENT PYOGENIC PAROTITISThis interesting and by no means uncommon

condition was first described by Payne (I933),who demonstrated the ' bronchiectatic' type ofterminal dilatation of the parotid duct. He alsoshowed that the commonest infecting organism inthis condition was the Streptococcus Viridans,a finding repeated in this series. Other organismsreported have been the Streptococcus Pyogenes,

t

4t%7'

FIG. 7b.-Diagram of submandibular sialogram.The subsidiary duct descending from the angle ofthe jaw to join the angle of the main duct is veryconstant.

:l ..a...

hL

LI -Nl -

-l - l#BX}}.

::::: .:.l - -&i3.I - -

- -|I i - - |-- l -| -

..... ; w

FIc 8.-Early recurrent pyogenic parotitis. S.viridans infection. Lateral view showingmoderate sialangiectasis.

B. Proteus, Pneumococcus, organisms of Vincent'sangina and even atypical coliform. bacilli.

Streptococcus Viridans InfectionThis infection characteristically occurs in

women during the fourth and fifth decades. Only3 out of 24 in the present series were men.The patient first notices an aching pain in front

of or just below the tragus and this is normally


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/


FIG. 9.- Advanced recurrent parotitis. S. viridansinfection. Antero-posterior view, showingmarked sialangiectasis.

accompanied by a slight diffuse parotid swelling.The onset may be preceded by a febrile illness,although more often no predisposing cause can befound. The patient may volunteer the informa-tion that she is subject to attacks of dryness of themouth, particularly when undergoing a period ofmental stress. The earlier attacks may last onlya few hours, with remissions up to six months ora year. They may begin in one gland and befollowed by attacks of greater severity in theopposite gland. At first the swelling subsidescompletely between attacks, leaving only a slightthickening. Later it persists and the gland maybecome palpable throughout its whole extent, thesurface becoming smoothly nodular and of arubbery consistency. If there is only a localizedinfection a firm nodule may be produced which isdifficult to distinguish clinically from a salivarytumour. The history and sialogram will serve todifferentiate between these conditions. The salivaduring the attacks has particulate globules ofmucopus containing desquamated cells and bac-teria, giving the characteristic ' snowstorm' effect(Payne, 1938). The constant discharge of infectedsaliva into the mouth causes the patient to com-plain of an unpleasant taste and occasionally theend of an attack may be announced by a suddenheavy mucopurulent discharge.

In recurrent parotitis there is no close relation-ship between the pain and food, but there is oftena history of an attack commencing at the begin-ning of a meal, particularly in the earlier cases.

e4

FIG. Io.-Recurrent parotitis. Pneumococcal infec-tion. Marked dilatation of duct and grosssialangiectasis. Lateral view.

As the disease progresses, food has less and lesseffect and it may be that by the time the patientis referred for treatment, the early relationship tomeals has been forgotten.

The Sialogram. In the early case, the terminaldilatation may be confined to one area alone orthere may be scattered isolated dilatations (Fig. 8).The duct itself shows no abnormality. The glandparenchyma is usually well outlined and noanatomical abnormalities are revealed.

In the late case of S. viridans infection thetypical picture of sialangiectasis is seen (Fig. 9).There is a saccular dilatation of the terminalducts and acini, and the main ducts become' beaded.' This appearance results from alternatedilatation and narrowing of the duct, and is anindication of the duration of the infection. It isalso noteworthy that the shadow of the glandparenchyma is either very faint or non-existent.This is partially due to lack of filling in parts ofthe gland which are occluded by inflammatoryexudate or by fibrosis, and partially due to thereservoir-like action of the dilated duct system.

Pneumococcal Recurrent Parotitis (Payne, 1940)This condition deserves special mention on

account of several features which distinguish itboth clinically and sialographically from the S.viridans infection. The attacks of pain and swel-ling are severe and the degree of general dis-turbance of health is considerable. Abscess for-mation may occur by the fusion of several pockets


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

526 POSTGRADUATE MEDICAL JOURNAL October 1950

FIG. II.-Parotid absce3s. Pneumococcal infection.Antero-posterior view.

of pus which may be discharged into the parotidduct, thus terminating the infection. The salivais purulent, often with a characteristic greenishtinge, and on culture a heavy growth of pneumo-cocci is readily obtained.The sex incidence of pneumococcal parotitis is

curious in that in the present series 8 cases outof 9 have been males.

The Sialogram. Marked disorganization of theterminal duct system is evidenced by the grossdegree of dilatation seen, with no evidence of a

FIG. 14.-Secondarily infected duct obstruction. Seetext.

gland outline (Fig. io). The main duct in thiscase is much wider than normal with periodicconstrictions along its length. Occasionally thisis so marked that a 'sausage-string' appearancehas been noted (Ollerenshaw and Rose, 1950) asseen in Fig. I4. It is evident that the reaction ofthe duct to infection is one of alternate strictureand dilatation and that this is the same whetherthe primary cause is obstruction at the orifice ora long-standing recurrent pyogenic parotitis.

Parotid Abscess. The late case of pneumococcalparotitis may give rise to the picture seen in Fig.II, that is, of a localized parotid abscess. Fivecases in the present series showed this appearanceand all were pneumococcal in origin. The historyin all these cases was of long-standing, recurrent,comparatively severe attacks of pain and swelling,each attack resembling the acute post-operative typeof parotitis. They differed from the latter only inthe absence of the extreme general prostrationencountered in the acute post-operative infection.In all these cases the abscess discharged spon-taneously into the duct, and external drainage wasnot required.

SIMPLE DUCT OBSTRUCTIONThis condition is more common than recurrent

pyogenic parotitis and in its uncomplicated formis a separate entity. However, when complicatedby infection, the obstructive type bears a closeclinical resemblance to the inflammatory le3ionand may only be differentiated by a careful historyand by sialography.

It has been shown above that in recurrent paro-titis there is no predisposing obstructive cause inthe main duct. In the early case the abnormalfindings are confined to the terminal ducts andacini, the main duct being of normal size. Suchdilatation as occurs in the later stages is behindmultiple inflammatory strictures. It will be shownbelow that quite a different state of affairs existsin simple duct obstruction.

EtiologyThe two main types of duct obstruction are

produced by similar mechanisms and are depen-dent on an abnormality of mastication following oraccompanying a dental derangement. They maybe divided anatomically into papillary and buccal,according to the portion of the duct involved.The papillary duct obstruction is caused by

trauma to the papilla of Stensen's duct and,though it usually occurs in the presence of ill-fitting dentures, it may occur in young adults inrelation to the eruption of the upper molar teeth.

In the younger age groups an inflamed andtender gum will produce a deviation from thenormal chewing habit. This may cause injury to


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/


*

FIG. 12.-Simple duct obstruction. Papillary type.Oblique lateral view.

Stensen's papilla with subsequent oedema andulceration. When the original inflammation hassettled down, fibrosis may develop at the ductorifice and stenosis ensue. The cases in thisseries occurred between the ages of twenty andtwenty-five.

In the older age group a similar repeated traumato the duct orifice is seen, but here the cause is anill-fitting denture. It is often found that the pro-gressive alveolar absorption occurring with ad-vancing age results in the upper denture becomingloose and undergoing an abnormal degree ofupward and downward movement with mastica-tion. The papilla is repeatedly traumatized asthe denture rides up in the alveobuccal sulcusand ulceration may occur. This again may resultin the papillary type of obstruction. More rarely,where the whole of the inner side of the cheekover the line of the duct is involved, the fibrosismay spread backwards and the buccal type ofobstruction develops. The latter always occursin the older age groups in the presence of den-tures, although even here the papillary ductobstruction is the more common.

Clinical PictureThe patient presents with a diffuse parotid

swelling accentuated by meals, and complains ofdiscomfort rather than pain. The history isusually short, as the persistent nature of theswelling causes advice to be sought during thefirst attack.

FIG. 13.-Simple duct obstruction. Buccal type.Antero-posterior view.

The gland is diffusely palpable, but the edgesare imperceptible, and there is none of thenodularity felt in the infective cases. Pressureover the gland will give rise to a sluggish flow ofclear saliva from a rigid-looking orifice situated ona hypertrophied papilla. The saliva is sterile onculture and shows no increase of mucoid orcellular content.

It is these cases of duct obstruction which areoften cured by the sialogram itself, as it is neces-sary to dilate the duct as a preliminary procedure.Recurrent parotitis, however, requires a routineof at least six consecutive daily injections of theduct with one-half to three-quarters of a millionunits of penicillin. This treatment has proved tobe uniformly satisfactory in i 8 cases reviewedover a two-year period.

The SialogramFig. I2 shows a papillary duct obstruction in

a man 4I years of age. The main duct isdilated, narrowing down to a fine point at its oralend. The gland itself is not involved and thereis no evidence of sialangiectasis. The papilla washypertrophied and exuded sterile saliva on pres-sure over the gland. His dentures were I2 yearsold and the upper set had become increasinglymobile.

Fig. I3 shows a buccal duct obstruction in awoman 59 years of age. Here the strictureextends up the duct to the point where it per-forates the buccinator muscle. Beyond the


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

528 POSTGRADUATE MEDICAL JOURNAL October 1950

FIG. I5.-Submandibular calculus. The calculus liesin the deep part of the duct after it turns downinto the gland. Dilatation of the main duct be-hind this, but no sialangiectasis.

FIG. I7.-Mixed parotid tumour. There is a centralfilling defect with upward displacement ofStensen's duct.

obstruction the duct is dilated in a somewhatirregular manner, indicating a degree of secondaryinfection. There is, however, no evidence ofinvolvement of the terminal ducts and the glandpattern appears to be normal. The dentures inthis case had been worn for 21 years. Six years

... ....

FIG. I6.-Efects of terminal obstruction of Whar-ton's duct. Obstructed for two months on andoff. Note dilatation of duct without sialangiec-tasis.

previously the lower set had been discarded, asthey were loose, and the unsupported upper setshowed a wide range of both vertical and lateralmovement. There was ulceration in both alveo-buccal sulci extending on the right side forwardalong the line of the duct. This relationshipcould be readily demonstrated by insertion of afine probe.

(3) DUCT OBSTRUCTION WITH SECONDARYINFECTION

It has been pointed out that on occasion it isextremely difficult to distinguish between a secon-darily infected duct obstruction and a long-stand-ing recurrent parotitis. The buccal type ofobstruction, in particular, is prone to secondaryinfection and in the case shown in Fig. I4 S. viri-dans was obtained on culture.The history of these case3 is longer than that of

the simple types and there tend to be more painand thickening of the gland. The relationship tofood is constant, however, and there is evidenceof trauma to the papilla or the cheek overlying theduct. The dentures are ill-fitting and may havebeen worn for many years. Occasionally thedentures have never fitted properly and the lesionmay be fully developed within six months.

The Sialogram (Fig. I4) shows that the duct hasborne the brunt of the infection and is seen to bedilated and strictured. Some of the minor ductsshare in the dilatation, but the parenchyma of the


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

October 1950 ROSE: Sialography in Diagnosis

ig

FIG. i8.-Submandibular filling defect.

gland is not involved and there is no acinar abnor-mality. The infecting organism in this case wasthe S. haemolyticus.As the majority of duct obstructions are sterile,

it is felt that it is justifiable to place these infectedcases in a separate category.

Salivary CalculiWhen salivary calculi are considered, the sub-

mandibular gland claims all the attention. It isargued that it is unnecessary to employ sialographyin the diagnosis of salivary calculi as the plainX-ray will establish their presence. This argumentis perfectly justifiable, but must be accepted withcertain reservations. Firstly, Feuz (1932) haspointed out that up to zo per cent. of salivarycalculi are radiotranslucent and, secondly, a know-ledge of the exact whereabouts of the stone may beessential before treatment is decided upon. Acalculus lying in the floor of the mouth may beeasily removed transorally, but if it should lie ator behind the point at which the main duct turnsdown into the gland, satisfactory removal is only

achieved by sacrifice of the submandibular glanditself. Fig. 15 is a sialogram performed to ascertainthe exact position of a salivary calculus. It wasimpossible to say from the straight X-ray whetherthe stone was intraglandular or not, and bidigitalpalpation of the floor of the mouth failed to localizethe calculus. Fig. I5 shows the stone to be lyingin the deep part of the duct, and indicates the lineof treatment.

Fig. i6 is a sialogram taken after removal of acalculus from the termination of Wharton's ductand shows the degree of dilatation the duct systemundergoes. The absence of sialangiectasis againindicates the essentially non-obstructive nature ofthe recurrent salivary infections. This is furthersuggested by comparing the scant incidence ofrecurrent infection of the submandibular glandwith the frequent occurrence of calculi.

Salivary TumoursIt is sometimes difficult to decide whether a

swelling in the parotid or submandibular regionis actually in the gland concerned or is superficial

C


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

530 POSTGRADUATE MEDICAL JOURNAL October I950I .,sb , . . ..............

R: .- iE ,.

,.:'::.- - | ,,> ' ........................ : . . " R -> S - .l ;..eX,-;= l - | |-ES: .. ' : * - * -;-£C * - * -w ewg l - l l'.: :: .,, '.i . S- l l - | l |........ :::'';::£,|,* : :!.'. . : C':' .i :. .. * l -.. ':'.'::.:. " '':: - * | -*!!}::i':::: @!':::,.,.::'::. ':<1 |-: :::}.::: :.::.'::-,! - * | -j J.o : ' . , : - -|'X,7 "9°'.U'.'- ........................................... :'^'. . :: :::.':': : : :: .:.i- /-i3 . ... ; . .. .,,.en,<.,, ...... . . R- A'O.* } t : :. :, eo R',', S: ...... : X t;Rt.'20:. S-' ^-' ':

FIG. 2oa.-Submandibular filling defect. A largetumour is displacing the duct upwards.

.......a

?sa£0>7 .. .: .:

.:;: .. Pa}s

FIG. 19.-Parotid filling defects. There are twofilling defects one above and one below the mainduct. These were intra-parotid tuberculouslymph glands.

to it. The sialogram may be valuable in deter-mining the relationships of such a swelling to thegland parenchyma (Leroux, I947) particularlywhen the latter is well outlined by adequate fillingof the gland (Ollerenshaw and Rose, 1950). Thereis no doubt that interpretation of the sialogram in

I.

FIG. 20b.-Diagram of 2oa, to show tumour outline.

.°.......

FIG. 2I.-Malignant parotid tumour. Gross destruc-tion of gland outline and irregularity of ductpattern.

cases of tumour may be difficult and it is only bycombination of a careful analysis of the picturewith a knowledge of the normal appearances thata diagnosis can be made. The various points thatrequire special attention are illustrated by thesialograms described under the following headings:

THE FILLING DEFECTThis is the common finding with tumours and

depends for its presence on a space occupyinglesion unfilled by the radiopaque medium. The


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

October 1950 ROSE: Sialography in Diagnosis 53II

area shows as a comparative translucency corres-ponding with the- soft tissue swelling (Fig. tcI).An extrinsic tumour may produce no effect onthe gland, the !ialogram being normal, or it maycause a filling .lefect in the profile of the gland(Fig. i8). Here the anterosuperior portion of thegland has not outlined and the defect correspondswith the site of the swelling. In the centre of thetranslucency there is a small area of calcification;a diagnosis of tuberculous lymphadenitis wastherefore made and was confirmed at operation.Similarly, lymphatic swellings in the preauricularregion are readily distinguished from parotidtumours, and in one case the presence of twofilling defects in the parotid gland in conjunctionwith an enlarged cervical gland enabled a diagnosisof tuberculosis of the parotid lymph glands to bemade with confidence (Fig. i9).

DISTORTION OF THE DUCT ANATOMYIt has been stated that the sialogram shows a

constant duct anatomy, and any gross deviationfrom the normal pattern must be accounted for. Themain duct may be pushed out of position and beseen running over an obvious intraglandulartumour or two adjacent subsidiary branches maybe widely separated by an interposed neoplasm.In most neoplasms of the salivary glands, sialo-graphic localization depends on a combination ofduct distortion and local filling defect (Fig. 2oa).Reference to the accompanying diagram (Fig.2ob) shows how in this case the submandibularduct is displaced upwards and compressed intoa small area, while the gland below it is replacedby a large spherical translucent area. This wasa large 'mixed' salivary tumour which had beenpresent for nine years.Two sialograms in the present series, one of

which is shown in Fig. 21, caused special commenton account of the peculiar distortion seen, thesmaller ducts seeming to fill in an irregular man-ner, leaving many areas completely unfilled.There was also little evidence of glandular outline

and the whole appearance suggested a gland in-volved by an infiltrative process. Clinically, thepain, hardness and fixity of the swelling stronglysuggested a malignant parotid tumour (Fursten-burg, 1948). It has been suggested that a malig-nant tumour will give a characteristic picture, butthe comparative rarity of the lesion has so far givenlittle opportunity for sialographic study. We wereunable, therefore, on the appearances shown to domore than suggest a diagnosis of malignancy.Further experience with correlation of patho-logical findings may enable a more dogmaticdiagnosis to be made.

SummaryThe history and technique of sialography is

discussed and the author's modifications described.The normal parotid and submandibular sialo-

grams are demonstrated and emphasis is laid onthe constancy of the duct anatomy.The appearance of the sialogram in cases of

recurrent pyogenic parotitis, parotid duct obstruc-tion, submandibular calculus and salivary tumoursis described. The sialogram is correlated with'theclinical picture in each case.The duct obstructions are classified anatomically

with reference to the sialogram and their etiologyis briefly discussed.

Sialography is invaluable in the diagnosis of ductobstruction and recurrent infection and may beuseful in the diagnosis of intra- and extraglandularswellings.

AcknowledgmentAcknowledgments are due to Professor A. M.

Boyd, who has stimulated and encouraged this lineof study, to Dr. E. D. Gray and the staff of theDepartment of Radiology of Manchester RoyalInfirmary for their help and willing co-operation,to the Department of Medical Photography forproviding the illustrations, and to the HonorarySurgeons who made all possible by referring theircases of parotid swelling for inve3tigation.

BIBLIOGRAPHY

BARSKY, A. J., and SILBERMAN, H. (I932), Ann. Surg., 95, 46BARSONY, T. (1926), Klin. Wschr., 4, 2500.FEUZ, J. (I932), Schweiz. Med. Wschr., 62, 666.FURSTENBURG, A. C. (1948), Y. Amer. Med. Ass., 136, I.LEROUX, G. F. (I947), J. de Radiol. et D'Elect., 28, 85.OLLERENSHAW, R. G. W., and ROSE, S. S. (5950), Brit. J.

Radiol., in presi.PAYNE, R. T. (I93 I), Brit. J'. Surg., I9, 142.PAYNE, R. T. (1933), Lancet, I, 342.PAYNE, R. T. (1938), P'oC. Roy. Soc. Med., 31, 398.PAYNE, R. T. (I940), Brit. Aled. J7., I, 287.PYRAH, L. N., and ALLISON, P. R. (1932), Brit. Med. J'., 2

1028.ROSE, S. S. (I950), Brit. J7. Surg., in press.SAMUEL, E. (I9;0), Brit. Y. Radiol., 23, 157.SWINBURNE, G. (1940), Brit. Y. Surg., 27, 713.


rotectedhttp://pm

j.bmj.com

/P


j.26.300.521 on 1 October 1950. D

ownloaded from

http://pmj.bmj.com/

sialography in diagnosisanteroposterior and true lateral film3 u,ually give all the required...

Documents