chapter 3. Squamous intraepithelial lesions: cytology– CHAPTER 3 CHAPTER correlation CHAPTER 1 CHAPTER

This chapter discusses the nat- 3.2 Historical context but were confined to the epithelium. ural history of cervical precancer, The term “” was coined HPV and oncogenesis, cytology no- The precursor phase of the natural about 20 years later by Reagan and menclature, and the cytological and history of is char- Hicks (1953), and dysplasia was cat- histological recognition of cervical acterized by cellular changes within egorized as being mild, moderate, precancer. the epithelial lining of the ; in or severe depending on the propor- other words, the abnormality is en- tion of the epithelial layers involved 3.1 Current understanding of tirely intraepithelial. John Williams in the dysplastic process. Carcino- the natural history of cervical first described intraepithelial cellular ma in situ was considered to have precancer changes in tissue adjacent to inva- a greater degree of abnormality and sive cancer more than 125 years to be the final precancerous state. Cervical cancer has a long precursor ago (Williams, 1888). During the The term “” (halo or vac- stage. The cervix is accessible and early decades of the 20th centu- uolated cytoplasm or empty space sheds exfoliated cells easily, and cy- ry, the concept of intraepithelial cytoplasm) was coined by Koss and tological examination of these cells dysplasia gained acceptance (Cul- Durfee (1956). Meisels and Fortin reveals precancerous changes that len, 1900; Rubin, 1910). It implied (1976) first recognized these cells as are easily eradicated. The essential cancerous-looking cells confined being infected with HPV. causative agent of cervical cancer to the epithelium above the base- Richart (1968) introduced the is the presence of high-risk HPV, ment membrane and led to the term concept of a continuum and subdi- which is easily detectable. Cervical “carcinoma in situ” (Broders, 1932), vided the spectrum of abnormality cancer is a completely preventable which was defined as full-thickness into three categories, called CIN . This is quite apart from the cellular changes that looked mor- grades 1 (mild dysplasia), 2 (moder- availability of an effective vaccina- phologically similar to undifferenti- ate dysplasia), and 3 (severe dyspla- tion. The disease should not exist. ated invasive carcinomatous cells sia). In this classification, carcinoma

Chapter 3. Squamous intraepithelial lesions: cytology–histology correlation 23 in situ was combined with severe Fig. 3.1. Changing terminology and treatment trends for cervical precancer dysplasia. The cytological classifi- over the past century. CKC, cold-knife conization; Cryo, cryotherapy; LEEP, cation was similar in that mild, mod- loop electrosurgical excision procedure; Rx, treatment. erate, and severe were suggestive (but not diagnostic) of CIN1, 2, and 3, respectively. The relative ease of treatment afforded by outpatient therapy, which had begun to replace hysterectomy and cold-knife conization in the 1970s and 1980s, lowered the threshold for treatment of cervical lesions. In an attempt to simplify the classifi- cation and because it had become clear that minor-grade lesions did not often progress to cancer, Richart (1990) proposed a two-tier classifi- cation system. High-grade lesions were thought to be much more likely to be genuinely precancerous. Low- grade lesions were considered to be transient and rarely precancerous. Many low-grade lesions were asso- These lesions have limited, if any, and major numerical and structural ciated with koilocytosis and recog- precancerous potential for progres- alterations of the host chromo- nized as being HPV-related. Howev- sion to cancer. This type of somes. This leads to uneven distri- er, this classification system was not is called a productive infection. The bution of the overall DNA content universally used. Also, moderate ab- key step in the of HPV- (aneuploidy) and is reflected by shifts normalities, some of which were un- linked cancers is the activation of the of the nuclear staining pattern (the doubtedly low-grade in nature, were viral E6 and E7 in the bas- staining intensity). This type of infec- included in the high-grade category al and parabasal cells of the infect- tion is more readily recognized cyto- and perhaps treated too readily. The ed epithelium (Bergeron et al., 2015; logically, colposcopically, and histo- different classifications are repre- Doorbar et al., 2012; Duensing and logically and is called a transforming sented in the diagram in Fig. 3.1, with Münger, 2004). If these viral genes are infection (see Chapter 4). treatment patterns included below. expressed in basal or parabasal cells, Sometimes moderate dyskary- The traditional “screen, diagnose, they trigger chromosomal instability osis (at cytology) or moderate and treat” pathway (Fig. 3.2) worked reasonably well when the threshold Fig. 3.2. Traditional process of screening test, colposcopic assessment, his- for referral to was set tological diagnosis, and treatment. high. The concept of a continuum persisted until relatively recently. A greater understanding of the biology of oncogenic HPV and its different effects in squamous epithelium of the lower genital tract has led to a differ- ent concept. It now seems clear that there are two different types of HPV infection. The first type is an innocent and transient infection, which may produce mild or low-grade lesions that are recognizable cytological- ly, colposcopically, or histologically.

24 dysplasia (at histology) may contain Fig. 3.3. Different HPV infection stages. both types of infection, and these are difficult to distinguish using cytology or histology. Fortunately, develop- ments in molecular biology have led to specific biomarkers of cell biol- ogy that can discriminate between these types where doubt exists (see Chapter 4).

3.3 HPV and the genesis of cervical cancer

Several different risk factors have CHAPTER 3 CHAPTER been implicated for cervical cancer and precancer. These include smok- ing, early age at first intercourse, nutritional deficiency, chlamydial infection, multiple sexual partners, multiple pregnancies, and long-term use of oral contraceptives (Bosch et the epithelial cell nuclei and chang- the past 25 years. These revisions al., 1995; Franco et al., 1999; IARC, es from a latent to a transforming in- reflect the changing understanding 2007, Schiffman et al., 1996; Wal- fection. It is in those cases that the of risk associated with different cy- boomers et al., 1999). However, the risk of progression is high. It is not tological and histological reporting fundamental and essential causative known what distinguishes those cas- and a greater understanding of the agent is the persistence of oncogen- es in which the becomes inte- role of oncogenic HPV. The United ic HPV in the epithelium of the TZ grated and transforming from those Kingdom classification now reflects and/or adjacent glandular epitheli- in which the infection is transient and the Bethesda two-tier classification. um. The relationship between onco- harmless. The relationship between To help clinicians manage their pa- genic HPV and cervical precancer oncogenic HPV infection and the risk tients with different grades of ab- appears, at first, paradoxical. Cervi- of progression or clearance is dis- normality, the ASCCP developed a cal cancer is always associated with cussed in Chapter 4. The crucial step series of clinical guidelines linked to oncogenic HPV, but oncogenic HPV is that of the HPV infection becoming the Bethesda classification (Wright is a normal and usually transient in- a transforming infection. et al., 2003). In the United Kingdom, fection that most healthy sexually the NHS Cervical Screening Pro- active women will encounter in early 3.4 Cytology nomenclature gramme (NHS, 2010) produced an reproductive life. The current think- evidence-based guidelines docu- ing is that the oncogenic HPV gains To this day, there are several dif- ment, which linked management to entry to the cervical epithelium at the ferent cytological classifications for the degree of cytological abnormal- new SCJ, possibly associated with cervical precancer. The German ity and other relevant case charac- minor abrasions, and that this allows classification is used in Germany, teristics (e.g. HPV test result, age, the virus to access reserve cells un- Austria, and some countries in east- and smoking history). It has recently derneath the single layer of columnar ern Europe. The United Kingdom been updated (NHS, 2016). Fig. 3.1 epithelium (Fig. 3.3). has its own classification, as do Aus- attempts to relate some of the pre- Most women will be infected tralia and New Zealand. Perhaps the vious cytology nomenclatures to with oncogenic HPV, and the great most widely used classification is the current Bethesda classification, majority will clear the infection with- the Bethesda terminology system, which is probably the most widely out any residual harm or increased first introduced in 1988 by the United used system today. risk of cervical cancer. In a small States National Cancer Institute (Sol- There has always been an in- percentage of women, the infection omon, 1989). It embraced the con- terdisciplinary dependency in man- persists, and in a small proportion cept of a two-tier gradation and has agement of cervical precancer. of those, it becomes integrated into undergone several revisions over Traditionally, this has been using

Chapter 3. Squamous intraepithelial lesions: cytology–histology correlation 25 cytology to screen, using colposco- occasionally seen in the basal layers infection (LSIL). Abnormal nuclei py to assess and direct biopsy, and (Fig. 3.4a). and other cell changes in parabasal using histology to confirm the di- Histological examination of a tis- and basal cells are typical of a trans- agnosis (Fig. 3.2). In this idealized sue biopsy of normal squamous epi- forming infection (HSIL). In the case scenario, the cytology screening test thelium will reveal normally stratified of an LSIL, as in Fig. 3.5a, there is a identified cases that may or may not epithelium with regular maturation productive viral infection, and cytol- have genuine precancer, colposco- and few mitotic figures in the basal ogy will reveal enlarged nuclei with py was able to recognize or rule out layers. As with cytology, there will be vacuolated cytoplasm in superficial the lesion, and a colposcopically normal nuclear–cytoplasmic ratios and intermediate cells. directed biopsy facilitated definitive and the nuclei will be morphological- histological proof of disease before ly normal (Fig. 3.4b and Fig. 2.2). 3.5.2.2 Histology treatment was advised. But all three of these disciplines are subjective in 3.5.2 LSIL (HPV infection; Histological determination of ab- nature. Until recently, histology was CIN1; mild dyskaryosis) normality is essentially recognition considered the gold standard and of abnormal cellular proliferation. It HSIL was considered the threshold 3.5.2.1 Cytology is based on the morphological as- at which treatment was necessary. sessment of cells in the epithelium, It is now clear that morphological The cytological recognition of ab- the architecture of the cellular layers, assessment at histology is also less normality is based on the finding of and the degree of maturation and than perfect, in particular the deter- nuclear enlargement and variation cellular differentiation. The relative mination of disease severity when in the size and shape of abnormal proportion of the epithelium that is in- morphological or histopathological cells. An increased intensity of stain- volved with abnormality, the degree examination reports HSIL-CIN2. ing with irregular chromatin patterns of maturation, and the persistence of A paper from the Lower Anogeni- is another common feature of abnor- mitotic figures throughout the epithe- tal Squamous Terminology (LAST) mality. These abnormalities in the lium are the usual parameters used Project (Darragh et al., 2012) finally superficial and intermediate cells are to grade the abnormality. Histolog- confirmed the relative subjectivity of koilocytosis, typical of a productive ical examination of LSIL will reveal , especially in the mid- dle grade of CIN2. The WHO 2014 Fig. 3.4. (a) Normal cytology preparation; intermediate cells are indicated histology terminology (Kurman et al., with arrows. (b) Normal histological section of squamous epithelium. 2014) proposed a two-tier classifica- tion, HSIL and LSIL, with the help of a b biomarkers to differentiate the diffi- cult or equivocal cases.

3.5 Cytological and histolog- ical recognition of cervical precancer

3.5.1 Normal cervical epithelium Fig. 3.5. (a) Cytology slide of LSIL. (b) Histological section of LSIL. Cytological examination of exfoliat- a b ed cells from the normal ectocervi- cal squamous epithelium will reveal mostly superficial cells; the nuclei are small, are not hyperchromatic, and have normal density and shape with normal chromatin patterns. Cru- cially, the nuclear–cytoplasmic ratio is low, and mitotic figures are only Koilocytosis

26 koilocytosis in the superficial layers cannot distinguish between CIN2 architecture changes in the epithe- and even part of the intermediate and CIN3. The changes seen at cy- lium are relatively unequivocal and layer, but the undifferentiated cells tology will usually include a definite are disordered throughout all cellular will be limited to the lower third of the increase in the nuclear–cytoplasmic layers (Fig. 3.6b). Cytological exam- epithelium (Fig. 3.5b). ratio as well as abnormal nuclear ination of an HSIL cannot be as pre- size and density and altered chro- cise, and a cytologist reporting HSIL 3.5.3 HSIL (CIN2, CIN3; matin patterns of basal or parabasal will probably describe basal cells that moderate dyskaryosis, cells (Fig. 3.6a). have risen to the intermediate or su- severe dyskaryosis) perficial layers, which are abnormal 3.5.3.2 Histology with enlarged nuclei and reduced cy- 3.5.3.1 Cytology toplasm, as in Fig. 3.6b. At histological examination of a However, the histological diagno- With a severely abnormal CIN3 le- clear case of CIN3, the great ma- sis is not robust in the middle grade, sion, cytology will report the diag- jority of pathologists will agree, be- and the category of CIN2 or HSIL- CHAPTER 3 CHAPTER nosis of HSIL. Cytology, by itself, cause the morphological cellular and IN2 contains some cases where the virus is transforming and the risk of Fig. 3.6. (a) Cytology slide of HSIL. The arrows indicate abnormal squa- progression is real and some cas- mous basal cells. (b) Histological section of HSIL-CIN3. Cellular abnormality es where the virus is proliferative prevails throughout the full thickness of the epithelium. There is an and not transforming and the risk of increased nuclear–cytoplasmic ratio, anisocytosis, and a loss of nuclear progression to cancer is very small. polarity. Several mitoses are present throughout the upper two thirds of the Morphological examination of tissue epithelium. biopsies from CIN2 cases is not reli- able, and pathologists will often not a b agree. Some will call the case CIN3, and some will call it CIN1. In this sit- uation, molecular biology tests can resolve the disparity. To appreciate how molecular biology tests can help, it is necessary to understand a little about the biology of oncogenic HPV and its effect on squamous epi- thelium (see Chapter 4).

Key points • Oncogenic (or high-risk) HPV is an extremely common infection in healthy sexually active women of reproductive age.

• Cervical cancer is a very rare outcome of oncogenic HPV infection but does not occur in its absence. Up to 80% of women will harbour oncogenic HPV during their reproductive life, but only 1 in 10 000 or fewer will develop cervical cancer.

• A positive high-risk HPV test does not imply cancer, precancer, or even an active infection.

• Cytological, colposcopic, and histological recognition of cervical cancer precursor states are all imperfect, because of their innate subjectivity.

Chapter 3. Squamous intraepithelial lesions: cytology–histology correlation 27