J Clin Pathol 1982;35:1-13

Review article

Cervical intraepithelial neoplasiaCH BUCKLEY, EB BUTLER, H FOX

From the Departments ofPathology, University of Manchester and St Marv's Hospital, Manchester

SUMMARY The theoretical and practical reasons for replacing the terms "cervical dysplasia" and"cervical carcinoma in situ" by the single diagnostic entity of "cervical intraepithelial neoplasia" arereviewed and the advantages and drawbacks of this newer terminology discussed. The histologicalcharacteristics and cytological features of the various grades of cervical intraepithelial neoplasia aredescribed and the differential diagnosis of this lesion is considered.

In 1969 Govan and his colleagues gave a detailedaccount of the classification, nomenclature, histo-logical features, and cytological characteristics ofthose various abnormalities of cervical squamousepithelium which fall short of a frankly invasivecarcinoma.' This paper has served well as a guideline,and reference text, for many pathologists andcytologists but in the intervening years our know-ledge of cervical pathology has expanded and ourunderstanding and interpretation of cervical epithe-lial abnormalities has altered. One result of thischanging appreciation of cervical lesions has beenthe introduction of a new terminology: this changehas been welcomed by some, but resisted by, andindeed proved unacceptable to, others, with theresult that whilst some pathologists and cytologistsare currently couching their reports in terms of thenew nomenclature others are still using the olderand better established terminology. The concurrentuse of two systems of nomenclature for cervicallesions is unsatisfactory and prone to cause con-fusion and misunderstanding.As advocates of the new system of terminology it is

our aim in this paper to detail the conceptual andpractical reasons for adopting a new nomenclature,to consider the possible objections to its use, and toredefine the histological and cytological features ofabnormalities of the cervical squamous epitheliumin terms of this nomenclature.

Nomenclature of cervical epithelial abnormalities

A fundamental division of cervical squamous epithe-lial abnormalities can be made between those whichlack any potential for evolving into an invasivesquamous cell carcinoma and those in which there isa significant risk of progression to an invasive

Accepted for publication 20 July 1981

neoplasm. The first group of banal changes includessuch entities as basal cell hyperplasia, reserve cellhyperplasia, immature squamous metaplasia, andmature squamous metaplasia, all of which arebenign, indeed usually physiological, conditionsunaccompanied by any increased risk of invasivecarcinoma. Epithelial abnormalities that are poten-tially capable of progression into an invasiveneoplasm have traditionally been categorised eitheras dysplasia or as carcinoma in situ, dysplasticchanges within the epithelium being graded as ofmild, moderate, or severe degree.The new nomenclature applies only to those

cervical epithelial abnormalities associated with anincreased risk of invasive carcinoma, all of which arenow put into the single diagnostic category ofcervical intraepithelial neoplasia (CIN).2, 3 Threegrades of abnormality are recognised: CIN I whichcorresponds to mild dysplasia; CIN II which isequivalent to moderate dysplasia; and CIN IItwhich encompasses both severe dysplasia and car-cinoma in situ.

REASONS FOR THE CHANGE IN

NOMENCLATUREFirstly, it has to be recognised that there has been,and still is, considerable disagreement as to thedefinition of both dysplasia and carcinoma in situ.Thus in 1961, an International Committee onHistological Terminology defined a carcinoma in situas "a lesion of the epithelium in which, throughoutits thickness, no differentiation takes place."4 Thiswas also the view taken by Govan et al who insistedupon complete loss of stratification and of cellulardifferentiation as defining criteria.' There is nodoubt that most histopathologists accept, and rely on,this definition but Burghardt5 has maintained thatthere can be no theoretical objection to the conceptof a differentiated carcinoma in situ, defining this

11 *

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Buckley, Butler, Fox

condition as one "in which cells showing thehistological features of malignancy occupy the fullthickness of the epithelium:" he did, however,remark that many would regard the condition recog-nised by him as a differentiated carcinoma in situ as aform of dysplasia. Burghardt's views are shared byKoss6 who simply defines carcinoma in situ as "anintraepithelial lesion histologically resembling in-vasive cervical cancer:" in this definition it isaccepted that there may or may not be evidence ofdifferentiation. The WHO definition of a carcinomain situ is "a lesion in which all or most of theepithelium shows the cellular features of car-cinoma":7 this is a definition which many patholo-gists would take to include both carcinoma in situand severe dysplasia.

There have been similar discrepancies in thedefinition of dysplasia and to some extent this hasbeen due to a lack of agreement as to the exactmeaning of the term "dysplasia." The strict sense ofthe word implies only an abnormality of growth andit has been applied not only to "premalignant"changes in the cervical squamous epithelium but alsoto examples of epithelial atypia in the colon,8 toclearly benign conditions of the breast,9 and todevelopmental abnormalities.6 Dysplasia has, infact, never been defined as a pathological processand it is not clear whether those who wish to retainthis term consider that a cell in dysplastic cervicalepithelium is an abnormal non-neoplastic cell, a cellwhich has undergone malignant change or one which,in some ill undefined manner, occupies a hypotheticalmiddle ground between neoplastic and non-neoplastic cells. In view of this vagueness as to whatis meant by dysplasia it is hardly surprising thatdefinitions of cervical dysplasia have varied and oftenbeen couched in purely negative terms. Thus theInternational Committee on Histological Termino-logy defined dysplasia as "all other disturbances ofdifferentiation of the squamous epithelium of lesserdegree than carcinoma in situ :"4 as remarked byFerenczy'0 this definition is sufficiently broad andimprecise as to encompass anything apart from anormal cervical epithelium. Others have considereddysplasia to be "those forms of atypical epitheliumwhich cannot yet be classed as carcinoma in situthough cellular abnormalities are present5 whilst theonly positive definition is that employed by WHO,"dysplasia is a lesion in which part of the thicknessof the epithelium is replaced by cells showing varyingdegrees of atypia."7 The grading of a dysplasticepithelium into mild, moderate, or severe dysplasiarests upon even more tenuous grounds and, indeed,no agreed criteria have ever been established for sucha grading. Some grade dysplasia in terms of thedegree of cellular atypia whilst others base their

grading on the proportion of the epithelial thicknessoccupied by undifferentiated cells. It has, however,become a widespread practice, rightly or wrongly, todiagnose mild dysplasia if undifferentiated cells donot occupy more than the basal third of the epi-thelium, to regard moderate dysplasia as a conditionin which undifferentiated cells occupy between onethird and two thirds of the epithelial thickness andto diagnose as severe dysplasia any condition inwhich undifferentiated cells occupy more than twothirds, but not the full thickness, of the epithelium.The degree of individual opinion implicit in thesediagnostic criteria is clearly considerable whilst theoften difficult distinction between a severe dysplasiaand a carcinoma in situ is frequently based solelyupon a highly subjective interpretation as to whethertwo or three layers of flattened cells on the surface ofthe epithelium are differentiated or not. It is thereforenot surprising that experienced pathologists maydiffer markedly from one other in their assessmentand diagnosis of cervical epithelial abnormalities'and that even the same observers looking at the sameset of slides at two different sessions may attainmajor differences between their own diagnoses onthe first and second viewing.'2

It is thus clear that dysplasia is an ill-understoodand imprecisely defined abnormality, that grading ofdysplasia is a highly subjective and arbitrary exercise,that the distinction between dysplasia and carcinomain situ is often only a matter of opinion and thatpathologists can achieve little consistency either withtheir colleagues diagnoses or in their own opinions.These problems would be of little account were it notfor the fact that the use of the terms "dysplasia" and"carcinoma in situ" has implanted into the minds ofmany pathologists and gynaecologists the impressionthat cervical dysplasia is, in an undefined fashion, adifferent disease process to carcinoma in situ. It is, ofcourse, accepted that dysplasia may progress tocarcinoma in situ but it is often assumed that such aprogression involves a basic change from a relativelybenign to a potentially dangerous condition. Thisconcept of a two-disease process has, inevitably, ledto the belief that dysplasia can not progress directlyto an invasive carcinoma without first passingthrough a stage of carcinoma in situ and has resultedin dysplasia, especially of mild or moderate degree,being often regarded as a relatively innocuous con-dition which does not merit the therapeutic measuresdeemed necessary for carcinoma in situ; indeed,dysplasia of mild or moderate grade is often nottreated and is sometimes ignored. It is, however,becoming increasingly clear that the evolution of anintraepithelial neoplasm is not a two-stage processand that the conditions known as "dysplasia" and"carcinoma in situ" simply represent arbitrarily

2

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Cervical intraepithelial neoplasia

defined and artificially delineated stages in theevolution of a single continuous process. Thus it hasbeen the unanimous conclusion of ultrastructural,13cytogenetic,14-16 tissue culture,17 autoradiographic'8and microspectophotometric'9 studies that cells in adysplastic epithelium are virtually identical to thosein a carcinoma in situ and that the pattern ofabnormality in these cells remains constant through-out the whole spectrum of dysplasia and carcinomain situ: particularly striking has been the demon-stration that cells in even a mild dysplasia show thesame degree of chromosomal abnormality, in termsof aneuploidy, that characterise a carcinoma insitu.10

This unity of basic cellular abnormality is furtheremphasised by studies of the biological behaviour ofdysplasia. It is widely recognised that between 45 and70% of cases of carcinoma in situ will eventuallyprogress to an invasive carcinoma20 21 and it is alsorecognised that carcinoma in situ is a condition whichrarely, if ever, spontaneously regresses: dysplasia, onthe other hand, has been thought to undergo spon-taneous regression quite frequently and is oftenregarded as a very infrequent precursor of carcinomain situ with little possibility of evolving into aninvasive carcinoma. To a considerable extent theseimpressions are based on early prospective studies inwhich it was not realised that a cervical punchbiopsy not infrequently removes all the abnormalepithelium and is therefore curative: even in cases inwhich complete excision is not achieved the healingprocess which follows the trauma of biopsy mayeradicate the epithelial lesion. Under these circum-stances it was not fully appreciated that post-biopsydisappearance of the cervical lesion usually representsa cure and not necessarily, or even usually, spon-taneous regression, this leading to a serious mis-interpretation of the biological behaviour of un-biopsied dysplasia: the magnitude of this error isshown by Koss's study6 in which nearly 40% of casesof dysplasia diagnosed by punch biopsy of the cervixsubsequently disappeared. It is noteworthy, how-ever, that in this study 42-3% of cases of dysplasiaprogressed to carcinoma in situ despite the inter-vention of a biopsy whilst nearly 4% evolved into aninvasive carcinoma. The natural history of dysplasiais best demonstrated by a combination of cytologyand colposcopy with the avoidance of biopsy: insuch a study Richart and Barron22 showed that 50%of cases of dysplasia eventually progress to a car-cinoma in situ, that 28 % of cases advance to a higherdegree of dysplasia and 22% of cases of dysplasiapersist unchanged; spontaneous regression of dys-plasia was found to be extremely rare and occurredonly in a very small proportion (6 %) of cases of milddysplasia. If it is accepted that approximately 500%

of cases of carcinoma in situ eventually progress to aninvasive carcinoma and that 50% of cases of dys-plasia evolve into a carcinoma in situ it becomesclear that approximately 25% of cases of dysplasiawill eventually develop into an invasive neoplasm. Itis therefore apparent that dysplasia cannot bedismissed as an innocuous lesion, a point furtheremphasised by the now firmly established fact that aninvasive carcinoma can develop directly from adysplasia, of even mild or moderate grade, withoutany necessary prior transition to a carcinoma insitu.5

In view of these facts it seems reasonable to discardthe terminology suggestive of a two-stage evolutionof intraepithelial carcinoma, to accept that all casesof "dysplasia" are, irrespective of grade, a form ofintraepithelial neoplasm and to introduce the termcervical intraepithelial neoplasia to encompass bothdysplasia, of all grades, and carcinoma in SitU.2 3Cervical intraepithelial neoplasia may be defined as"a spectrum of intraepithelial change which begins asa generally well differentiated neoplasm which hastraditionally been classified as mild dysplasia andends with invasive carcinoma."'10 It will be appre-ciated that this definition hints at dysplasia being awell differentiated carcinoma in situ and this is aview which has been strongly espoused byBurghardt.5

It is of value for descriptive and epidemiologicalpurposes to subdivide CIN into various grades butthe studies cited indicate that a grading of this type isdevoid of long-term prognostic value: it is true that,collectively, cases of CIN III are more likely toprogress to an invasive carcinoma than are cases ofCIN I or II but, on an individual basis, a woman withCIN II may have as great a chance of eventuallydeveloping an invasive squamous cell carcinoma asdoes a patient with CIN III. It is also true that casesof CIN I which progress will generally take muchlonger to evolve into an invasive carcinoma than willcases of CIN III, this allowing for a period ofobservation, albeit under careful surveillance, ofpatients with CIN I. With these caveats, all cases ofCIN should be regarded as a single entity, the onlytruly valid differentiation of cervical epithelialneoplasms being between those which are stillconfined to the epithelium and those which areinvading the stroma.

ADVANTAGES OF THE NEW NOMENCLATUREThese may be summarised thus:(i) It emphasises the biological and clinical unity ofthe two apparently discrete conditions of dysplasiaand carcinoma in situ thus obviating the concept of atwo-stage evolution of intraepithelial neoplasia.(ii) It indicates that dysplasia, previously a process of

3

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

4

undefined nature, is in fact a neoplastic lesion.(iii) It removes from the pathologist the difficult andsubjective task of differentiating between severe

dysplasia and carcinoma in situ.(iv) It indicates that whilst grading of an intra-epithelial lesion is of collective prognostic value suchgrading offers no grounds for assuring an individualpatient that her cervical abnormality either will or

will not develop into an invasive carcinoma.(v) It allows for a unity of therapeutic approach tocervical intraepithelial lesions and prevents the stateof affairs where a diagnosis of carcinoma in situ isregarded as a definite, and often urgent, indicationfor treatment whilst one of dysplasia, often dif-ferentiated from carcinoma in situ on relatively flimsyand uncertain pathological grounds, is either nottreated adequately or is ignored.

OBJECTIONS TO THE NEW NOMENCLATURE

Those who oppose the introduction of the term"cervical intraepithelial neoplasia" usually base theirobjections on one or more of the following grounds:(i) Many cases of CIN are not true examples ofneoplasia but represent a non-specific reaction toinflammation, trauma or chronic irritation.(ii) The use of the word "neoplasia" to describe an

apparently mild abnormality of the cervical epithe-lium may give a false impression to the gynaecologistand thus lead to overtreatment of innocuous lesions.(iii) The use of a term implying the presence ofcancer may cause serious emotional disturbance inwomen who simply have a mild abnormality ofgrowth and differentiation of their cervical epithe-lium.

The first of these objections represents the mostserious argument that can be offered against the use

of the CIN terminology. It is almost certain, how-ever, that many cases considered to be examples of"mild dysplasia" do not merit this diagnosis beingmisdiagnosed instances of immature squamousmetaplasia or basal cell hyperplasia. Strict adherenceto the diagnostic criteria for CIN I (mild dysplasia)will therefore reduce markedly the number of casesof CIN I that appear to be reactive rather thanneoplastic. It must further be borne in mind thatthere is increasing evidence that viral infections maypossibly induce true cervical neoplasia and that themere presence of an inflammatory process in asso-ciation with CIN does not necessarily mean that theepithelial changes are not truly neoplastic. Havingsaid this it has to be admitted that a residue of cases

will remain in which non-specific reactive changesmimic very closely those of true CIN. This residue isprobably very small, as evidenced by the fact thatonly 6% of cases of "mild dysplasia" spontaneously

Buckley, Butler, Fox

regress and the fact that a small number of cases ofreactive change are being wrongly considered asneoplastic is probably less important than the currentlabelling of a large number of neoplastic lesions asbeing of little importance.Much the same argument can be applied to the

possibility of overtreatment of innocuous lesions bythe gynaecologist. The CIN terminology is nowwidely used in the gynaecological literature andmany, probably most, gynaecologists have not onlyaccepted this terminology but are increasingly awareof its implications. It is not the severity of CIN thatposes the greatest threat to a woman but the size ofthe lesion and a patient with a small, sharply localised,focus of CIN III is more easily cured, and has abetter prognosis, than does one with an extensivearea of CIN I or 11.23 It is not therefore the case thatthe new terminology leads to widespread overtreat-ment of cervical lesions but that the old terminologyled to overtreatment of some lesions-for example, ahysterectomy for a small focus of carcinoma insitu, and undertreatment for others for example,extensive areas of CIN II. The use of the newterminology interlocks with present therapeuticstrategies for cervical lesions and promotes alogical approach to the treatment of individualpatients.The third objection, that of the patient's emotional

response, is clearly one that can be avoided by acommonsense attitude and choice of words. Apatient with CIN does not have "cancer" in the truesense of the word but runs the risk of developing thedisease: so does a woman with "dysplasia."

Histological features of CIN

In all cases of CIN the full thickness of the epithe-lium, whether on the surface or in the crypts, isoccupied by neoplastic cells: in many cases, however,there is cytoplasmic differentiation in the upper partof the epithelium and hence CIN is usually graded interms of the proportion of the epithelial thicknessoccupied by undifferentiated neoplastic cells ofbasaloid type. A single criterion of this type is not,however, a fully adequate basis for grading CIN andaccount has to be taken of factors such as stratifica-tion, the level within the epithelium at which mitoticfigures are found, and the presence of abnormalmitotic figures. It is recognised that grading of CINis just as arbitrary and subjective an exercise as isgrading of dysplasia but, in the case of CIN, this is oflittle importance, the gradings being of descriptivevalue only and lacking any prognostic connotationsor therapeutic implications.CIN I The essential feature of CIN I is that whilstthe cells throughout the full thickness of the epithe-

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Cervical intraepithelial neoplasia

lium show nuclear abnormalities, the cells in theupper and middle thirds of the epithelium undergocytoplasmic differentiation (Fig. 1). Thus, at all

levels of the epithelium, the nuclei tend to haveprominent nucleoli and be pleomorphic, of irregularoutline, enlarged and hyperchromatic with a coarsegranular or filamentous chromatin pattern. The cellsin the lower third, or less, of the epithelium show no

evidence of cytoplasmic differentiation or of orderlystratification, lack clearly defined boundaries andhave a high nucleocytoplasmic ratio with nuclearcrowding whilst those in the middle and upperthirds of the epithelium show, to a variable degree,evidence of stratification and of cytoplasmicmaturation with a decreasing nucleocytoplasmicratio. Hence the surface cells may show a normaldegree of maturation and can differ from normalcells only by their possession of abnormal nuclei.Aberrant keratinisation of individual cells is some-times seen whilst mitotic figures are relatively un-

common, are generally confined to the lower third ofthe epithelium and are usually of normal form.CIN II The histological features of CGN II are

similar to those of CIN I except that undifferentiatednon-stratified cells with pleomorphic nuclei, and ahigh nucleocytoplasmic ratio extend beyond thelower third of the epithelium but not into the upperthird (Fig. 2). The cells in the upper third of theepithelium undergo a variable degree of stratificationand of cytoplasmic differentiation. Mitotic figuresare present in the lower two-thirds of the epitheliumand are not uncommonly of abnormal form.CIN III In this condition undifferentiated, non-

stratified, basaloid cells with nuclear crowding,

5

indistinct boundaries and a high nucleocytoplasmicratio occupy more than two thirds, or the fullthickness, of the epithelium (Fig. 3). The degree of

Fig. I Squamous epitheliumshowing CIN IL Haenmatoxylinand eosin x 350

nuclear pleomorphism is often greater than that seenin GIN I or II. Mitotic figures are frequently seen, arecommonly present in the upper third of the epithe-lium and are often of abnormal form. GIN III of thispattern is often described as being of the "small cellundifferentiated type" and is seen most commonly inthe upper, or proximal, part of the transformationzone, which may be within the endocervical canaland therefore not always be visible on colposcopy.

It is recognised however that other forms of CINIII occur24 and in these the reliance upon the pro-portion of the epithelium occupied by undifferen-tiated cells breaks down: in these forms cells whichare not totally undifferentiated, but are of similarappearance throughout, occupy most or all of theepithelial thickness and it is in these instances thatgreater reliance has to be placed on features such aslack of stratification, the site of mitotic figures andthe normality or otherwise of the mitotic figures.Thus, one variant of GIN III is characterised by thepresence of cells, in more than two-thirds and usuallythe full thickness of the epithelium, which althoughhaving nuclei of neoplastic type, showing noevidence of stratification and lacking distinct cellboundaries are larger than undifferentiated basaloidcells and have a lower nucleocytoplasmic ratio (Fig.4): mitotic figures are usually abundant, are found atall levels in the epithelium and are frequently abnor-mal. The cells in this type of lesion, which is foundmost frequently in the mid-transformation zone, arethought to be of parabasaloid type.

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Buckley, Butler, Fox

: .k

Fig. 2 Squamous epitheliiiiishowing CIN 11. Undiffer-en-tiated cells occupy less thatntwo-thirdds of the epitlhelialstructure. Haematoxvilin andz7eosin x 350.

Fig. 3 CIN III: unidiffereni-tiated cells of hasaloid typeoccupy alnost the fillthickness of the epithelilumZ.The constituent cells in thi.sexanmple ar-e somewhatspindle-shaped. Haematoxylinand eosint x 350.

In a third variant of CIN LII the epithelium islargely or wholly occupied by cells of relatively lownucleocytoplasmic ratio having well defined cellboundaries: there may be well marked surfacekeratosis. There is, however, a complete disorganisa-tion of growth pattern and lack of stratification andpolarity (Fig. 5): the nuclei are of neoplastic typeand mitotic figures, commonly of abnormal form, are

found with some frequency in all levels of theepithelium. This form of CIN III, known as the'large cell keratinising type" usually occurs well outon the ectocervix, and is therefore thought by

some to arise from original, rather tlhani metaplastic,squamous epithelium:5 others who maintain thatCIN always arises in metaplastic squLamous epithe-lium would regard such cases as developing in an

unusually extensive transformation zone.

DIFFERENTIAL HISTOLOGICAL DIAGNOSIS

OF CIN

A variety of benign cervical epithelial changes maybe confused histologically with CIN. The lesionwhich most commonly poses this problem in dif-ferential diagnosis is immature squamous meta-

4k. - -.

6

,-Mpg-"4w

v"aw #Apl &-0

IC.: la- ...

AMOP'--so --T- Mgt...

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Cervical intraepithelial neoplasia

Fig. 4 CIN III: cells ofparabasal type form thisintraepithelial neoplasm.Haematoxylin and eosin x350.

Fig. 5 CIN III showing cytoplasmic maturation but a

total lack of stratification; mitoses occur at all levelsin the epithelium. Haematoxylin and eosin x 350.

plasia: this can be distinguished from CIN however,by the absence of nuclear pleomorphism, the reten-tion of a normal nuclear chromatin pattern, thepresence of a low nucleocytoplasmic ratio, the lack ofabnormal mitotic figures and the presence, in many

cases, of altered columnar, mucus-secreting cells atall levels in the epithelium but particularly on thesurface. It must, however, be remembered that animmature squamous epithelium may also show CINand the two conditions are not mutually exclusive.

Reactive changes associated with inflammatorylesions of the cervix are also often confu ed withCIN, but there are several histological features whichare of help in distinguishing reactive and neoplasticlesions. The basal layers of a reactive epithelium mayshow a failure of maturation similar to that seen inCIN but, in contrast to GIN, maturation of thesuperficial cells follows a normal pattern with bothcytoplasmic and nuclear maturation. Inflammatorycells of various types (neutrophil polymorpho-nuclear leucocytes, lymphocytes, plasma cells,eosinophils) are, in an inflammatory lesion, not onlypresent in the underlying stroma but are also oftenfound within the epithelium itself, a feature notcharacteristic of CIN.The most severe reactive changes, and those which

most closely mimic CIN, are associated, in both ourown experience and that of others,25 with the presenceof an intrauterine contraceptive device (IUCD). Inthese cases the epithelial cells may show severecellular and nuclear pleomorphism, an increasednucleocytoplasmic ratio, prominent nucleoli andcytoplasmic vacuolation. There is, however, usually awell marked inflammatory cell reaction and removalof the device is followed by a reversion to normal: itis important that the pathologist or cytologist isaware of the presence of an IUCD in such cases.Atrophic squamous cervical epithelium is also

sometimes confused with CIN. Such an epithelium isthin and the basal cells, which occupy an unduly high

7

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Buckley, Butler, Fox

* ....proportion of the epithelial thickness, often haveprominent nuclei and nucleoli. Mitotic figures are,however, seen only in the basal layers of the epithe-lium whilst the typical nuclear characteristics ofneoplasia are absent.Condylomata of the cervix are not uncommon and

may present a problem in differential diagnosis, in sofar as the koilocytosis, multinucleation, individualcell keratinisation and occasional mild cellularatypia which are characteristic of this viral lesionmay give a false impression of CIN. This diagnosticdifficulty is more likely to arise with the flat condy-loma than with the papillary form of this lesion but,again, the absence of nuclei showing the typicalfeatures of neoplasia and of abnormal mitoticfigures serve to exclude CIN. It has to be accepted,however, that CIN can, and not infrequently does,develop in a condylomatous lesion; condylomatamerit careful study to exclude complicating CIN and,conversely, all areas of CIN should be scrutinised forfeatures suggestive of the lesion having developedin a pre-existing condyloma.

Cytological diagnosis of CIN

Cervical intraepithelial neoplasia is generally asymp-tomatic and naked eye inspection of the cervix mayreveal no abnormality. Consequently the under-standing, interpretation, and diagnosis of theselesions stems from the introduction of exfoliativecytology in cancer diagnosis.26

Cells which are present in the cervical smear reflectthe cells exfoliating from the surface epithelium ofthe cervix and the cell population is also influencedby the rate of exfoliation. It is well known that alldegrees of intraepithelial neoplasia can coexist andcells from these lesions together with normalsquamous and columnar cells will be found in thesmear.

Prediction of the probable histological picturedepends on assessment of the whole smear but forcommunication and standardisation betweencytology laboratories it is necessary to identifyindividual cells using a standard terminology. It issuggested that the recommendations of the WorkingParty of the British Society for Clinical Cytology27are followed together with the finer modificationsused by Spriggs and Boddington.28

CYTOLOGICAL FEATURES OF CIN

Identification of dyskaryotic cells in the smearThe appearance of the nucleus distinguishesdyskaryotic cells (Fig. 6) from normal cells and frommalignant cells, the term "malignant" cell being usedonly when there is confidence that the cell is shed from

Fig. 6 Differences in chromatin pattern are illustratedin (a) a normal intermediate squamous cell (b) parabasalcells, and (c) malignant squamous cells from an invasivesquamous cell carcinoma. Papanicolaou's stain x 3100.

an invasive lesion. The nucleus of a dyskaryotic cellis large and hyperchromatic, the chromatin patternis irregular in the sense that it shows a granularcondensation of chromatin but the arrangement ofthe condensed chromatin is uniform without theirregular areas of clearing seen in the nucleus ofa malignant cell. The nucleus is round, oval or,sometimes, elongated and the nuclear outline can besmooth or finely wrinkled. The degree of differ-entiation of the cytoplasm is important in thefinal assessment of the whole smear. Normal cyto-plasmic differentiation is seen in mature (superficialand intermediate) and parabasal dyskaryotic cells(Fig. 7) and in these the appearance of the cytoplasmis similar to that seen in the equivalent normal cellbut the nucleus is relatively large with the featuresdescribed above. Dyskaryotic nuclei can also berecognised in cells with no cytoplasmic differentiation(Fig. 8). These cells have a high nucleocytoplasmicratio and they can occur singly, in small groups, incoherent masses or as sheets. They will be referred toas undifferentiated dyskaryotic cells.

Interpretation of the cervical smear depends uponan assessment of the relative numbers of the differenttypes of dyskaryotic cells present and it is usual toreport the "worst" histological lesion anticipated

8

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Cervical intraepithelial neoplasiac

Fig. 7 Two parabasal cells andone mature dyskaryotic cells arepresent in this field and arecontrasted with a normalintermediate cell in the uppersection of the picture; from apatient with CIN II.Papanicolaou's stain x 1250.

Fig. 8 This field shows a loose clusterof undifferentiated dyskaryotic cellswith parabasal dyskaryotic cells; from apatient with CIN HII. Papanicolaou'sstain x 1250.

with the understanding that cells indicating a lesserdegree of abnormality will also be present. Con-sequently the following descriptions of the cyto-logical patterns to be expected in CIN I, II and IIIare a simplification of the actual picture.CIN I The surface maturation of the epithelium isreflected in the smear and only well differentiateddyskaryotic cells are found. These will be pre-dominantly mature dyskaryotic cells but thepresence of occasional parabasal dyskaryotic cells isacceptable.CIN ii As the proportion of well differentiatedparabasal dyskaryotic cells increases the cytologicalpicture of CIN I merges into that of CIN II and it isoften necessary to predict both in the assessment ofthe probable histological lesion.CIN III The range of histological lesions which fallinto this category are reflected in the cervical smear.

Differentiated parabasal and undifferentiated dys-karyotic cells predominate and keratinised dyskaryo-tic cells showing abnormal shapes, resembling fibrecells and tadpole cells, can be seen in cases showingmarked surface keratosis. When CIN III is of thesmall cell undifferentiated type, coherent sheets(Fig. 9) and thick masses of undifferentiated dys-karyotic cells are seen together with undifferentiateddyskaryotic cells singly and in loose clusters. Inthese cases few, if any, differentiated dyskaryotic cellsare found.

CAUSES OF CONFUSION IN CYTOLOGICALDIAGNOSISThe worst histological lesion present can be pre-dicted in about 80% of cases from the appearance ofthe cytological smear.29 Smears may be under-estimated because of variable exfoliation or appar-

9

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Buckley, Butler, Fox

Fig. 9 A coherenit sheet ofundifferentiated dyskaryotic cells froma patient with CIN III. Papanicolaou'sstain x 1250.

ently overestimated because of undersampling of thetissue specimen but this difficulty is not consideredhere. Morphological aberrations which cause con-fusion may be due to cell degeneration, but may alsobe a result of reactive changes in cells to infection ortrauma; immature metaplastic cells, includingreserve cells, can also cause difficulty and variantscan be present alone or in combination.

Cell degeneration Coagulative necrosis in the nucleiof degenerate cells may mimic the coarse clumping ofchromatin with irregular clear areas seen in malig-nant cells and when these cells are keratiniseddyskaryotic cells a false prediction of invasivecarcinoma can be given. The distinction is made byrecognising the diffuse outline of the chromatinclumps and the nuclear membrane together with theabsence of sharp angularities, sometimes beading ofchromatin at the nuclear membrane is seen. Cyto-plasmic vacuolation also occurs in degenerate cells.

Reactive changes due to infection The nuclei areenlarged and can be as big, relative to the cytoplasm,as the nuclei of mature and parabasal dyskaryoticcells whilst the dispersed pattern shown by thenuclear chromatin may be mistaken for granularity.The distinction is made by recognising that thesenuclei are hypochromatic rather than hyper-chromatic and that the cytoplasmic staining isusually dense and indeterminate rather than clearlyeosinophilic or cyanophilic. This type of reactivechange is non-specific but when condylomata arepresent more specific changes are seen. These includethe presence of mature, often keratinised, dyskaryo-tic cells, koilocytotic cells (Fig. 10), bizarre, multi-

nucleated cells and sheets of smaller keratinised cellswith irregularly shaped degenerate nuclei. These arereferred to as sheets of dyskeratotic cells (Fig. 11)and reflect the parakeratosis seen in the tissue.Koilocytotic cells in the cervical smear are charac-teristic of condylomatous epithelium but they arenot always present as they cannot exfoliate if there isdense parakeratosis at the surface.29 It has to beremembered that CIN can coexist with a condylomaand the presence of more severely dyskaryotic cellsshould not be ignored when cells characteristic of acondyloma are recognised.

Reactive changes due to trauma Cells which reflecthealing cause most problems in these cases and arefound when there is ulceration due to a ring pessaryand in the presence of an intrauterine contraceptivedevice (IUCD). The cells seen in the smear arepleomorphic, and poorly differentiated with a highnucleocytoplasmic ratio: the nucleoli are pro-minent but the nuclear chromatin pattern is normaland hypochromatic rather than hyperchromatic.

Immature metaplastic and reserve cells Immaturemetaplastic cells which show reactive changes due toinfection present particular problems because thecombination of the more granular chromatin patternof an immature cell together with nuclear enlarge-ment and dispersion of chromatin due to infectionapproximates closely to the appearance of a dys-karyotic nucleus. In these cases it is only proper toacknowledge the difficulty in the cytology report.A similar situation can arise when coherent masses

or sheets of undifferentiated epithelial cells arepresent.28 This cytological presentation is seen in

10

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Cervical intraepithelial neoplasia

Fig. 10 A koilocytotic cell iscontrasted with a normal intermediatesquamous cell. Papanicolaou's stainx 1875.

Fig. 11 A sheet of dyskeratoticcells in which the cytoplasm iskeratinised and the nuclei aredyskaryotic in type and often showdegenerative changes, from CIN Iwith hyperkeratosis. Papanicolaou'sstain x 1250.

CIN III of the small cell undifferentiated type but canalso be due to reserve cell hyperplasia, particularlywhen accompanied by infection: it must also beremembered that comparable groups of cells exfoliatein cases of adenocarcinoma of the endocervix.3' Inmany cases assessment of the whole smear willreveal features which lead to a correct interpretationbut when these are absent it is only possible to givethe range of the differential diagnosis.

Atrophy The normal atrophic smear does notusually cause difficulty even when cells from thethinned epithelium shed as syncytial sheets with a

high nucleocytoplasmic ratio. In these cases theinert regressed appearance of the nucleus is charac-teristic and a single small chromocentre is also seen.

An infected atrophic smear from a patient withsenile vaginitis can cause problems because of the

changes due to degeneration and infection inimmature cells. In some cases it can be impossible tobe certain that dyskaryotic cells are not present aswell. However the position can be clarified by theapplication of an oestrogen cream to the vaginalepithelium for two to three weeks; normal epithe-lium will mature but areas of CIN will remain thesame and dyskaryotic cells will be obvious in therepeat smear.

Cervical intraepithelial neoplasia versusmicroinvasive carcinoma

It is perhaps worthwhile to conclude this account ocervical intraepithelial neoplasia by emphasising theself-evident point that this term excludes micro-invasive carcinoma. It is outside the scope of thisreview to consider microinvasive carcinoma in any

11

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

12

detail but in general terms this diagnosis implies thepresence of an invasive lesion in which, however, thedegree of invasion is sufficiently minimal for theneoplasm to be treated as if it were still confined to

the epithelium. Unfortunately there is continuingdisagreement about the precise definition of a micro-invasive carcinoma, argument centring about suchpoints as the depth of invasion, the point from whichthe depth of invasion is measured, the importance ofa confluent pattern of invasion, the significance ofinvolvement of lymphatic channels and the stress tobe placed upon multiplicity of sites of invasion.These controversial topics are fully discussed else-where32-34 and from the standpoint of this review theonly decision that has to be taken is whether a lesionis still confined within the epithelium or is beginningto invade the underlying stroma. Unfortunately,early stromal invasion, though sometimes obvious,may be extremely difficult to assess: useful diagnosticaids include the fact that tongues of invasivecarcinoma often show better squamous differentia-tion than that seen in the intraepithelial lesion fromwhich they have arisen and that the site of invasion isoften stigmatised by a well marked stromal lympho-cytic infiltration and by local stromal oedema. Ifdefinite invasion can be recognised, of even the mosttrivial degree, a diagnosis of cervical intraepithelialneoplasia can no longer be sustained.

References

Govan ADT, Haines RM, Langley FA, Taylor CW,Woodcock AS. The histology and cytology of changes inthe epithelium of the cervix uteri. J Clin Pathol 1969;22:383-95.

2 Richart RM. Natural history of cervical intraepithelialneoplasia. Clin Obstet Gynecol 1967;10:748-84.

3Richart RM. Cervical intraepithelial neoplasia: a review.Pathol Annu 1973 ;8:301-28.

4lnternational Committee on Histological Definitions,1961. Acta Cytol 1962;6:235-6.

Burghardt E. Premalignant conditions of the cervix. ClinObstet Gynecol 1976;3:257-94.

6 Koss LG. Dysplasia: a real concept or a misnomer?Obstet Gynecol 1978;51:374-79.

7Poulsen HE, Taylor CW, Sobin LH. Histological typing offemale genital tract tumours. International HistologicalClassification of Tumours No 13. Geneva: WorldHealth Organization, 1975.

8 Riddell RH, Shove DC, Ritchie JK, Lennard-Jones JE,Morsen BC. Precancer in ulcerative colitis. In: MorsonBC, ed. The pathogenesis of colorectal canicer. Philadel-phia, London, Toronto: Saunders, 1978:95-118.

9 Azzopardi JG. Problems in breast pathology. London,Philadelphia, Toronto: Saunders, 1979.

10 Ferenczy A. Cervical intraepithelial neoplasia. In: BlausteinA, ed. Pathology of the female genital tract. New York,Heidelberg, Berlin: Springer-Verlag, 1977:143-65.

Briggs RM. Dysplasia and early neoplasia of the uterinecervix: a review. Obstet Gynecol Surv 1979;34:70-99.

1 Cocker J, Fox H, Langley FA. Consistency in the histo-logical diagnosis of epithelial abnormalities of thecervix uteri. J Clin Pathol 1968;21:67-70.

BuckleY, Butler, Fox

3 Shingleton HM, Richart RM, Wiener J, Spiro D. Humancervical intraepithelial neoplasia: fine structure ofdysplasia and carcinoma in situ. Cancer Res 1968;28:695-706.

" Kirkland JA, Stanley MA, Cellier KM. Comparative studyof histological and chromosomal abnormalities incervical neoplasia. Cancer 1967 ;20:1934-52.

'5 Granberg 1. Chromosomes in preinvasive. microinvasiveand invasive cervical carcinoma. Hereditas 1971 ,68:165-218.

16 Spriggs Al, Bowey CE, Cowdell RH. Chromosomes olprecancerous lesions of the cervix uteri. Cancer 1971 ;27:1239-54.

' Richart RM, Lerch V, Barron BA. A time-lapse cine-matographic study in vitro of mitosis in normal humanicervical epithelium, dysplasia and carcinoma in situ.J Natl Cancer Inst 1967;39:571-7.

18 Richart RM. A radioautographic analysis of cellularproliferation in dysplasia and carcinoma in situ of theuterine cervix. Am J Obstet Gvnecol 1963;86:925-30.

'9 Wilbanks GD, Richart RM, Terner JY. DNA content ofcervical intraepithelial neoplasia studied by tv o-wavelength Feulgen cytophotometry. Ami J ObhstiG lnecol 1967 ;98:792-9.

20 Kottmeier HL. Evolutioni et traitemetnt des epithelionas.Rei' FranC Gv necol 1961 ;56:821-5.

21 Fidler HK, Boyes DA, Worth AJ. Cervical cancer detectioniin British Columbia. Br J Obstet Gv(naecol 1968;75:392-404.

22 Richart RM, Barron BA. A follow-up study of patientswith cervical dysplasia. Am J Obstet GYnecol 1969;105:386-93.

21 Richart RM. The patient with an abnormal Pap smear--screening techniques and management. N Fngl J Medl1980 ;302 :332-4.

21 Patten SF. Diagnostic cjytologj of the utterinie cerOix. Basel:Karger, 1969.

25Gupta PK, Burroughs F, Luff RD, Frost JK, Erozan YS.Epithelial atypias associated with intrauterine con-traceptive devices (1 U D). Ac ta Cl tol 1978 ;22 :286-91.

26 Papanicolaou GN. Survey of actualities and potentialitiesof exfoliative cytology in cancer diagnosis. Anni InternMed 1949;31 :661-74.

27 Recommendations of a Working Party of the BritishSociety for Clinical Cytology. Problems of cell nomen-clature in cervical cytology. J Clini Pat/ol 1978;31 :1226-7.

28 Spriggs Al, Boddington MM. Progression and regressionof cervical lesions: review of smears from womenfollowed without initial biopsy or treatment. J ClinPatihol 1980;33:517-25.

29 Husain OAN, Butler EB, Evans DMD, Macgregor JE.Yule R. Quality control in cervical cytology. J C/litPat/ol 1974;27:935-44.

3 Stanbridge CM, Mather J, Curry A, Butler EB. Demon-stration of papilloma virus particles in cervical andvaginal scrape material: a report of 10 cases. J C/inzPathol 1981 ;34:524-31.

:1 Stanbridge CM, Butler EB, Langley FA. Problems incervicovaginal cytology: fine structure as an aid to

diagnosis. Acta Citol 1980;24:335-43.32 Roche WD, Norris HJ. Microinvasive carcinoma of the

cervix: the significance of lymphatic invasion andconfluent patterns of stromal growth. Cancer 1975;36:180-6.

33 Ferenczy A. Carcinoma and other malignant tumors of thecervix. In: Blaustein A, ed. Pathology of the fenmalegenital tract. New York, Heidelberg, Berlin: Springer-Verlag, 1977:171-205.

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from

Cervical intraepithelial neoplasia 13

34 Nelson JH, Averette HE, Richart RM. Detection, diag- Requests for reprints to: Professor Harold Fox, Depart-nostic evaluation and treatment of dysplasia, carcinoma ment of Pathology, University of Manchester, Stopfordin situ and early invasive cervical carcinoma. Cancer Building, Oxford Road, Manchester M13 9PT, England.1979;29:174-92.

on March 22, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.35.1.1 on 1 January 1982. Dow

nloaded from