Readers of this journal are aware of the huge success that the UK National Health Service Cervical Screening Programme (NHSCSP) has been over the last two decades.1,2 Indeed without it we may have witnessed epidemic rates of cervical cancer.3 Although the actual benefit lies in the successful treatment of screen-positive women identified by the programme the real success of the screening programme, in reducing cervical cancer rates, is due to a number of factors including technical expertise, high coverage rates and rigorous quality assurance in the laboratory, colposcopy clinic and administration offices.4,5
The treatment of squamous cervical intraepithelial neoplasia is to remove or destroy the transformation zone (TZ). It is likely that no method of treatment is superior to another if it is performed properly and the limited available evidence supports this view. The significant advantages of excision (simplicity, cost, outpatient procedure, histological examination of the entire TZ) mean that treatment thresholds may have lowered over the last decade. Long-term pregnancy-related morbidity associated with excision has been reported recently. The evidence would suggest that this increase equates to a genuine increase in serious adverse outcome for cone biopsy but not large loop excision of the transformation zone (LLETZ). The available data also point to an increase in both incomplete excision and premature labour associated with the excision of large endocervical TZs. The clinical implications arising from this are firstly that women with large type 2 and 3 TZs need appropriate counselling before treatment and that the threshold for treating young women with mild abnormalities needs review.
Clinical guidelines for colposcopists
It is remarkable how different clinical practice is across Europe and globally. In the UK women will, in normal circumstances, be referred for a colposcopic examination because of an abnormal cytology test. Apart from abnormal cytology and a small number who have suspicious symptoms or a suspicious looking cervix women will not be referred for a colposcopy examination. Also, colposcopy is practiced exclusively by devoted colposcopists in special colposcopy clinics by accredited and regularly re-certified colposcopists or colposcopists in training. In contrast, in some countries in Europe a colposcopy examination is routinely offered to almost every woman attending a gynaecologist usually in conjunction with cervical cytology. Colposcopy may be used as a screening tool, as a diagnostic aid and to facilitate therapy.
When to treat an individual is not within the scope of this review. In the UK the NHSCSP has produced a clinical guidelines document for colposcopists and trainees. It is a comprehensive and evidence-based list of guidelines, which covers most clinical circumstances that a colposcopist will come across and is widely used.6 The US has a similar consensus document.7 More recently consensus European clinical guidelines have been published in Cytopathology.8,9 Finally, for those interested in clinic management, there is also a joint Royal College of Obstetricians and Gynaecologists/British Society for Colposcopy and Cervical Pathology (RCOG/BSCCP) standards document that describes the requirements necessary to set up and run a quality assured colposcopy clinic service.10
Principles of treatment and type of transformation zone
The treatment of squamous cervical intraepithelial neoplasia (CIN) is to eradicate the transformation zone (TZ). This may be achieved by destruction or excision. Excision may be used for any case but when considering a destructive method defined case-specific conditions must be satisfied (Table 1) so as to avoid under-treating unrecognized invasive or glandular disease.9
|The transformation zone must be fully visible|
|A colposcopically directed diagnostic biopsy must be taken from the most dysplastic area in the TZ|
|There must be no suspicion of invasive disease|
|There must be no suspicion of glandular disease|
|There should not be cytological/histological disparity|
|The patient should not have had previous treatment|
Classification of the transformation zone
The precise position and visibility of the TZ is fundamental to optimising treatment and minimising, or predicting, treatment-related morbidity. There has been real confusion in the published literature of treatment methods concerning the TZ site and the type or extent of treatment. For example, traditionally the term cone biopsy has meant excision of a largely endocervical TZ or the treatment of a woman with, for example endocervical squamous or glandular disease, whereby it is considered necessary to excise a significant amount of endocervical canal epithelium.11 However in many recent publications conisation, or simply cone, is used to imply any kind of excision of the TZ.12 Also, when comparing treatment success or morbidity rates, valid comparisons demand similar case selection criteria. In comparisons of excisional versus ablative techniques it has not always been possible to determine the precise selection criteria. Cases that do not satisfy the necessary conditions for ablative therapy (and therefore will have excisional treatment) should not be compared to cases that do. For example, published series of excisional treatment may include cases of suspected microinvasive disease, glandular disease or endocervical disease. Comparing these cases with women treated by ablation is invalid.
As a result of these concerns the International Federation for Colposcopy and Cervical Pathology endorsed a classification of TZ types (Figure 1) which categorised them according to site, visibility and size.13,14 Each type may be large or small. Types 2 and 3 may be partially endocervical and large or small, or totally endocervical.
Ablative or destructive methods
The unifying principle behind ablative treatment of cervical pre-cancer is that the entire TZ is destroyed to a depth of approximately 7 mm. This depth derives from the 1981 paper by Anderson and Hartley,15 which first reported that CIN can exist to at least 4 mm in gland crypts and 7 mm is considered a suitable safety margin.
Cold coagulation is somewhat of a misnomer as the tissue is destroyed using a probe tip temperature of between 50 and 120 °C. The method was introduced by Kurt Semm in the late 1960s and was popularised in the UK by Ian Duncan.16–18 Cold or thermal coagulation is simple, effective and outpatient based. It is associated with very low complication rates and may be performed with or without local anaesthetic infiltration. Likewise cryocautery is simple and relatively painless with few treatment related complications. The technique was, and is, more popular in the USA than in the UK.19,20 Laser may be used either to excise or ablate the TZ depending on how the spot size and power density settings are set. Laser vaporisation was an extremely popular technique in the 1980s and 1990s.21 It allows the operator to demarcate the area to be ablated very precisely and has few complications. The area of ablation may be precisely gauged and there is minimal thermal damage distal to the laser spot.22 It is particularly appropriate where dysplastic epithelium extends on to the vaginal walls and where minimal ablation depth of 2–3 mm is appropriate because no glands are present. The technique is usually performed in the outpatient or office setting using local anaesthetic infiltration. The biggest disadvantage of laser ablation is the large capital outlay and high maintenance costs. Its biggest advantage is that laser machines are almost always purchased by large and devoted colposcopy centres, where relative expertise is usually assured.
Finally radical diathermy was used extensively in Melbourne but not many treatment series were published from elsewhere.23 The technique was usually performed under general anaesthesia. Radical diathermy had the ability to destroy more than 7 mm and high success rates were reported but there has been convincing evidence that subsequent pregnancy related morbidity is increased.24
There are essentially three methods of excision, cold knife cone biopsy, laser excision and large loop excision of the transformation zone (LLETZ).25–27 The latter technique also later became known as loop electrosurgical excision procedure (LEEP) in the USA and elsewhere. The scalpel or cold knife has been the traditional method of performing cone biopsy and continues to be popular in many parts of Europe. It is still favoured by some when treating women with glandular or microinvasive disease, because the scalpel inflicts virtually no artefactual damage at the margins of excision. The method is usually performed in theatre and under general anaesthesia. Traditionally the post-cone wound has been managed by using Sturmdorf sutures, which have the effect of inverting the new squamo-columnar junction. This deals effectively with the open wound. However there is an increased risk of cervical morphological damage and of subsequent stenosis-related dysmenorrhoea and incompetence-related premature labour.24
Laser excision, like LLETZ, may be used to treat women in almost any circumstance. The two methods are very similar. They are both capable of excising the type 1 or type 2 TZs as a simple outpatient procedure under local anaesthetic infiltration. Equally when the circumstances demand either technique may be employed in order to perform what was previously known as a cone biopsy in the UK; that is a large/long endocervical excision. This type of excision is perhaps now better described as a type 3 TZ excision given the confusion surrounding the term cone biopsy in the literature. Again both techniques have the advantage of facilitating comprehensive histological evaluation of the entire specimen whereby the true diagnosis may be recognised, the margins of excision established and the possibility of microinvasive or glandular disease revealed or ruled out. Because of this, there is usually no need to perform a colposcopically directed biopsy at a separate or preliminary examination. For this reason a policy of see and treat has prevailed in many clinics for women with cytologically and colposcopically suspected high-grade CIN.27
Success or cure rates
Published cure rates after treatment for CIN are very high.28 The ultimate test of cure is, of course, whether a woman develops cervical cancer after treatment. But this is not a practical index of comparison given the long natural history of untreated CIN. Surrogate measures include the persistence of cytological abnormality or the recognition of a histologically proven CIN lesion. The most comprehensive assessment of treatment efficacy is the meta-analysis published by Martin-Hirsch et al.29 in the Cochrane Database of Systematic Reviews. Not all the studies in this review were randomised or even quasi randomised controlled trials. The authors concluded that there was no difference in terms of success rates between any of the methods except for cryocautery which was less successful in the management of high-grade CIN.29 Shafi et al. in their review, concluded that, given the other advantages of LLETZ it seemed to be the ideal method of treatment for CIN.30 The advantages of LLETZ, or laser excision, over the ablative techniques are detailed in Table 2. The efficacy of LLETZ has been established in several short-term observational studies, which show rates of residual disease within one year of between 3% and 8%.25,31,32
|Provides comprehensive histological examination of the TZ|
|Microinvasive disease may be recognized or ruled out|
|Glandular disease may be recognized or ruled out|
|Margin status may be determined|
|Accommodates the treatment of any type or size of TZ|
|Allows for a selective see and treat policy|
|Facilitates self audit of image recognition|
Longer term studies31–33 have confirmed low rates of residual disease and helped identify those at most at risk of post-treatment disease. Age (over 50 years) and positive margin status have been shown to be independent risk factors for residual disease.33 It is now accepted practice to retreat all women aged over 50 years with incomplete excision revealed histologically and any where there is a suspicion of invasive disease.
Incomplete excision, assessed histologically, is undoubtedly associated with an increased the risk of residual disease being found at subsequent follow up of women treated by an excisional technique.34,35 Whilst the rates of incomplete excision range very widely in the published literature (5–51%),26 the rate of histologically proven residual disease rates is much lower and covers a much tighter range (3–7%).30 Also, as Gardeil et al.34 have shown, the absence of incomplete excision does not exclude residual disease and therefore a comprehensive follow up protocol should be offered to every woman after treatment. But the difference in residual cytologically abnormal rates is significant. The most comprehensive study undertaken is a meta-analysis of 65 published (and one unpublished) observational studies of various excisional techniques by Ghaem-Maghami et al.35 Their review included 35109 women and found that in women with complete excision the rate of recurrence of low-grade or high-grade abnormality was 4% and of high-grade alone 3% compared to 20% (low-grade or high-grade) and 18% (high-grade) where the margins were not reported as being clear of the dysplastic process.
A recent study by Dimitriou et al.36 examined the relationship between incomplete excision and the size and type of TZ in over a thousand women who had a LLETZ procedure over a 4-year period from 2004 to 2008 in a large Dublin teaching hospital colposcopy service. The authors did not find a relationship between type of TZ and incomplete excision but did find that in large type 2 and large type 3 TZs there was an almost doubling in risk of incomplete excision when compared to small type 1 TZs. The authors concluded that when presented with a large type 2 or 3 TZ that needed excision, larger removals should be performed and the women counselled accordingly.
Unrecognised glandular or invasive disease
A particular advantage of excision is the ability to recognize or rule out glandular or invasive disease. With the exception of Cecil Wright in Canada, colposcopists have not considered colposcopy to be a reliable method of recognizing glandular disease. Wright’s excellent atlas of glandular disease37 reveals classical features that would be hard to miss but which are not often present or obvious enough for the competent colposcopist to identify and most cases of glandular disease are heralded cytologically and confirmed histologically. Cullimore et al.38 have reviewed the subject comprehensively. It is also unfortunately true that colposcopy is not a reliable method of recognizing early invasive cancer and between 0.6% and 1% of LLETZ procedures will reveal colposcopically unsuspected microinvasive disease.26
Complications and long-term morbidity after treatment of CIN
It has been clear for many years that there is a low rate of short-term complication after LLETZ39 and the ablative techniques, with the exception of radical diathermy.24 It is also established that cold knife cone biopsy is associated with an increased risk of long-term morphological damage, and that this may result in functional complications which are both obstetrical and gynaecological.40,41
Cervical pre-cancer is common. As mentioned at the outset, its recognition and appropriate treatment is associated with a huge reduction in risk of cancer.2 As most women who need treatment for CIN are young, long term functional complications, specifically those related to pregnancy, are important. LLETZ is by far the most popular method of treatment at this time. Despite relative equivalence in terms of success between the different treatment modalities, LLETZ offers a number of advantages over the ablative techniques (see Table 2). The question of premature labour and similar pregnancy-associated treatment complications has been the subject of much published research over the last three years.24,41 The subject is not an easy one to unravel given that no randomised controlled trial has been sufficiently powered to answer questions relating to pregnancy morbidity. However two large recent reviews have attempted to get as close to the truth about the relative risk as is possible using the meta-analytical model. The first, by Kyrgiou et al.,41 examined preterm pregnancy-related outcomes in women treated for CIN by cold knife cone biopsy, laser ablation, laser cone and LLETZ who subsequently conceived. The studies were not randomised. The authors reported an increased risk for preterm labour of 1.7 (95% CI 1.2–1.4) in women who had been treated by LLETZ compared to women who had not had treatment. Similarly increased risks for low birth weight and for premature rupture of membranes were observed. There were greater risks for these indices in women treated by cold knife cone and smaller increased risks for women treated by laser cone excision. The authors concluded from their data that any excisional technique was associated with an increase in risk of premature labour. This paper precipitated a flurry of concern about using excisional techniques, particularly in young women. This concern has not as yet persuaded many colposcopists to change their preferred treatment method, probably because of the genuine advantages detailed in Table 2. However it may well have influenced those with a low threshold for treating low-grade disease in young women, where the risk of progression to cancer is small.
The second and more recent review by Arbyn et al.24 attempted to determine the relative risk of more serious outcome measures like perinatal mortality, severe preterm delivery (<32/34 weeks), extreme preterm delivery (<28/30 weeks) and low birth weight (<2000, 1500 or 1000 g). It was similar in design; that is, a meta-analysis of pooled data derived from cohort studies selected from pubmed-medline and embase databases. Figure 2 represents the increase in risk of perinatal mortality associated with cold knife cone biopsy but not with LLETZ. The data for laser cone is not clear enough to make a conclusion. The data concerning radical diathermy suggests an increased risk (1.54 CI 0.84–2.82) but only one study was included. The review also found no increase in risk of severe or extreme preterm delivery associated with treatment by LLETZ.
The authors estimated that treatment by cold knife cone biopsy, laser cone biopsy or radical diathermy would result in about one perinatal death in every 70 pregnancies. The risk of perinatal death for women treated by LLETZ was as low as 2 per 1000. Similarly the risk of severe and extreme preterm delivery and low birth weight were common after cold knife cone and diathermy but rare after LLETZ.
In trying to interpret the results of these two reviews one may reasonably say that removing or damaging a large part of the cervix affects its function during pregnancy. A secondary analysis of the data in the first review (Figure 2) supports this concept. Only three studies had data available for valid comparison and, again, with the caveat that these were not randomised controlled trials. Nonetheless the figure clearly points to a risk of preterm labour (PTL) associated with excisions greater than 10 mm in length (called depth in the figure) and no increased risk with excisions less than 10 mm. This concept has some biological plausibility and work presented at the BSCCP meeting this year supports it.42 In this study, of 353 women who conceived after LLETZ in the same hospital nearly 10% (35) had a preterm delivery. The authors examined the relationship between volume, length and thickness of the extirpated transformation zone with the risk of preterm delivery. There was a more than threefold increase in the risk of PTL if the excision volume exceeded 6 cm3 (RR = 3.17, 95%CI 1.56–6.43) when compared to small excisions defined as volumes of less than 3 cm. The study also found that the length of excision (from the ectocervical to endocervical margin) was not an independent risk factor for PTL but that excised specimens that were greater than 12 mm thick at any part, were again associated with a threefold increased risk of PTL (RR = 3.05, 95%CI 1.31–7.08). Although it was surprising to find no relationship between length of excision and PTL, the relationship between very thick and very large volume excisions enhances the plausibility the concept that the underlying increased risk for PTL is related to the amount of tissue is removed. Larger studies are needed and whether epithelial or stromal volume loss is more important is at this stage uncertain.
Follow up after treatment for CIN
There is universal agreement of the need for comprehensive follow up of patients after treatment for CIN. Although failure rates are low, they are not insignificant.25,31,32 Until recently, opinion as to which method of follow up is best (cytology versus colposcopy versus HPV testing) has varied. However a recent meta-analysis by Arbyn et al.43 has clearly shown that routine HPV testing using Hybrid Capture 2 technology is the most sensitive test of cure available at this time. Because of the relative rarity of treatment failure, sensitivity is the most important characteristic of a test after treatment. The recent test of cure study44 has confirmed the Arbyn review that HPV testing is the most sensitive test, although it has its own false negative rate, and it is likely that it will be adopted in many countries in the near future.
Long-term risk of cancer after treatment for CIN
Three groups have highlighted the importance of continued monitoring of women after treatment for CIN because of the increased risk of cervical cancer in these women. In 2006 Soutter et al.35,45 in London published their review of studies which had reported treatment of CIN success rates. They selected 25 studies and found that the rate of post-treatment CIN declined steadily to a rate of 190 per 100 000 women by 10 years. In contrast and of great concern the rate of cervical cancer after the first year of follow up remained relatively constant at about 56 per 100 000 women, a rate which is about 2.8 times greater than expected. Soutter et al. postulated that this was due to declining compliance with follow up protocols and advocated annual cytology for at least 10 years. Likewise Kalliala’s group46 in Finland reported the long-term risk of cervical cancer after treatment by drawing on the Finnish national cancer register. They reported the follow up data of 7564 women treated for CIN between 1974 and 2001 who were followed up until 2003. They also found an increased risk of cervical cancer post-treatment (22 of 7564 women treated) and reported that the greatest risk occurred during the second decade after treatment. Finally Strander et al.47 examined the Swedish cancer registry in a prospective cohort study and reported an increased cervical cancer risk of 2.34 (standardised increased risk, 95% CI 2.18–2.50) amongst women previously treated for CIN. They concluded that this risk had increased since the 1960s and persisted 25 years after treatment. These three reports have obvious implications for follow up protocols of women treated for CIN. Previously there was relative complacency about the risk of cancer with temporal distance from treatment for CIN. Clearly the small but definite increased risk of cervical cancer which exists for women after treatment for CIN needs to be considered for at least 20 years and these women need continuing surveillance for at least this long in order to reduce the risk of cervical cancer to below that of the background population risk.
There is no perfect method of treatment for CIN. Electrosurgical excision approaches excellence and is the most popular method of treatment at this time. Within the spectrum of electrosurgical excision techniques LLETZ will accommodate most circumstances. Also whilst both treatment success rates and morbidity risk will be similar for the great majority of routine cases (the type 1 TZ) the operating colposcopist and individual case characteristics are more important than the choice of therapy for the type 1 TZ (where conditions for ablation are satisfied). For cases that do not satisfy the conditions for ablative treatment excision by LLETZ or laser excision will accommodate any TZ type. Because of the increased risk of both incomplete excision and longer term morbidity associated with the treatment of large type 2 and type 3 TZs, particular care is necessary when approaching these cases. Choices of therapy must be made according to the case characteristics of TZ type, size and abnormality in women who will need treatment. As our understanding of treatment morbidity improves so our counselling to patients and threshold for therapy will need to be modified.