Scheme 1                                       Scheme 2                                    Bethesda scheme

                                                     HPV changes                                Low-grade lesions

Mild dysplasia                                  CIN I

Moderate dysplasin                          CIN II

Severe dysplasia                               CIN III                                     High-grade lesions

Carcinoma-in-situ

HPV=Human papillomavirus; CIN=cervical intraepithelial neoplasia
PATHOLOGY OF CERVICAL PREMALIGNANCY

        The diagnosis of CIN is based upon the architectural and cytological appcarances of the cervical epithelium.Architectural features include differentiation, stratification and maturation, terms which are closely related but not synonymos. The proportion of the thickness of the epithelium showing differentiation is a useful feature to be taken into account when deciding the severity of a CIN. It is the not most important criterion despite the fact that it is one of the easiest to asscss. In CIN I at least the upper half of the epithelium usually shows good differentiation and stratification whereas in CIN III differentiation may be very slight or event abseint. Nuclear abnormalities are the, Most important combination of features to be taken into, account-when assessing CIN. The nuclei, are cxamined using similar criteria to those employed by the cytologist in assessing a cervical smear, nuclear cytoplasmic rario, hyperchromasia, nuclear pleomor phism and variation in size of nuclei. Both the overall number of mitotie figures and their height in the cpithelium are assessed. The more superficially the. mitotic figures are found, the more severe the CIN is likely to CIN may affect the gland crypts as well as the surface cpithelium.

       Anderson & Hartley (1980) showed that the mean depth of crypt involve ment in women with CIN III was 1.25mm and that the mean plus 3 stnadard deviations (taking in 99.7% of the popularion) was 3.8mm. These figures suggested that treatment to a depth of 5mm into the stroma would be sufficient to eradicate most CIN; however, practical experience has shown that treatment to 10 mm gives much better results without Inereasing morbidity (Soutter et al 1986).

        This underdingnosed lesion is characterized by columnar cells with large hyperchromaric nuclei 'and prominent nucleoli. The nuclei may be stratified and show abnormal mitotic figures. There is often gland budding and a 'back-to-back' arrangement. In some cases, the whole of a gland may be involved but often the lesion occurs as a sharply demarcared area in the deep portions of the glands.

        In the great majority of cases, the lesion lies in the transformation zone, close to the squamocolunmar junction (SCJ). Isolated AIS high in the endocervical canal some distance from the SCJ is tare. It follows that a cone biopsy with a good margin of excision in the endocervical canal will be adequate treatment for most cases of AIS.

        Traditionally, the various intraepithelial lesions have been considered part of a continum of change rather than morphologically or biologically distinet steps that progress inevitably from one to the next (Richart, 1973). This view has been modified somewhat by the discovery that specific HPV; types are generally associated with higher-grade lesions. Moreover, with techniques such as morphometry and DNA microspectrophotometry, lowergrade lesions have been found to have a principally diploid or polyploid DNA content, which in turn correlates with their tendency to regress (Bibbo et al, 1989; Fu et al, 1981; Fu et. al, 1983; Fu el al, 1988). In contrast, higher-grade lesions are frequently biologically aneuploid, demonstrate greater degrees of cytologic atypia,and are more likely (as a group) to persist or progress (Bibbo et al, 1989; Fu et al, 1981; Fu et al, 1983; Fu et al, 1988; wilbanks et al, 1967). Notwithstanding the value of the techniques mentioned above in segregating lesions into general groups (Taylor et al, 1987), it is accepted that one cannot consistently distingnish true precancerous lesion from cytologically and histologically similar lesions that are benign. The variability in natural history of histologically (and biologically) similar lesions may well be influenced by a multitude of facrors.

        Nasiell et al (1983, 1986) found that approximately two-thirds of mildly dysplastic lesions and one third of inocerately dysplastic lesions regressed during followup. Problems cncountered in most studies, howerer,inchide length of followup, arbirtrary criteria for determining the grade of an individual lesion, and the potential alreration of the natural history of a lesion as a result of biopsy (Nasiell et al, 1983; Nasiell et al, 1986; Richart and Barron, 1969).

        The time required for a lesion to evolve from a low grade to a high grade or to progress eventually to invasive canncer is not known and impos sible to determine by direct observation. Obviously, it would not be ethically permissible not pragmatically possible to take a large group of women who have what is believed to be a potentially dangerous and progressive epithelial lesion, allow the lesion to run its course without hiopsy or treat ment, and follow them over many years to see what will happen, In order to try to understand what might happen, mathematical models have been developed based on data from several studies involving women of different ages who have different epidemiologic risk factors. Barron and Richart (1968) have calculared that mean time it takes for intraepithelial lesions to progress from one grade to the next is approximately 5 years. 

        The mean time required for high-grade intraepithehal lesions WIN 111) to progress to invasive disease is also uncertain but has been stimated to range from 1 to 30 years, with a reasonable average figure being 10 to 13 years, (Barron etal, 1978; Copielson and Brown, 1975; Gustafsson and Adami, 1989). But some speciall cases meparted in Taiwan, heprogtesing were faster.

 MALIGNANT POTENTIAL OF CIN
 Progression from CIN Ill to invasion

        The malignant potential of CIN III is amply demonstrated by Melndoe et al (1984) in a crucial paper. A group of 131 patients who had been treated for CIN III and who continued to produce abnormal cytology for more than 2 years after the initial treatment were followed for 4-24 years. After 20 years' follow-up, 36% of these women had developed invasive disease.

        Studies of progression from CIN I to CIN III are blighted by the difficulty in accurately determining the grade of the initial lesion. Those reports that relied upon cytology only to determine the initial diagnosis and to document progression or regression are invalidated by the poor correla tion between the grade of cytologic abnormality and that of the histology al, (Soutter et al 1986) Basing the diagnosis of CIN I upon a colposcopic asses ment without punch biopsy, progression to CIN III was observed in 26016 of women within 2 years (Campion et al 1986).

        Given the clear malignant potential of untreated CIN III, the difficulty in identifying CIN I accurately and the high progression rate in women whose CIN I was not treated, it seems to be prudent to treat all wonen with CIN regardless of the grade of abnormality. There is no justification for treating subclinical HPV lesions.

        The principal techniques for removing cervical lesions include cryosurgery, laser therapy, and cone biopsy. In gencral, cryotherapy and laser are considered compar ably effective for removing most cervical lesions. In fact, from a large number of studies cryotherapy appears slightly superior. However, extensive lesions, including those involving the pottio, are usually treared by laser ablation, and lesions of uncertain biology or extending into the endocervical canal are most effectively managed by cone biopsy.

        This technique has evolved as an alternative to electrocautery, the latter being an effective but painful approach to ablating the T zone (Chanen, and Rome, 1983). A cryoprobe of variable size is applied to the cervix, and freezing is achieved by a nitrous oxide refrigerant forced through a small orifice at the base of the probe tip. Tissue destruction is achieved by tissue and cellular dehydration; the formation of ice erystals, with cell membrane rupture ; cellular protein denaturation; thermal shock; and vascular stasis (Charles and savage, 1980).The extent of freezing is determined by the size of the lesion, and the goal is to prouduce an ice ball that extends at Icast 4mm beyond the lesion. Currently, two consecutive deapplications with an interim period of partial thawing are recommended. For larger lesions, multiple applications may be necessary. The success of the technique is linked to several factors. General causes of failure include inadequate freezing (produced by poor contact, short duration of freeze, and low tank pressure) and the freezing of lesions that are, too large for complete coverage (Charles and Savage, 1980). The size of the lesion is of considerable portance, with cure rates dropping substantially for lesions greater than I centimeter in diamcter (Town send,1974). In addition, deep gland (or crypt) involvement may influence thera peutic success. The efficacy of conventional cryosurgery in destroying lesions within cervical clefts ("glands") has been disputed. 

        Although it is generally accepted that cryotherapy will produce tissue necrosis to a depth that will destroy all crypt involvement, Stafl et al (1977) noted that gland crypts may remain in up to 25% of cases, and Savage et al (1982) found that 18% of lesions involving glands recurred following cryotherapy. The chance that invasive cancer may develop as a reault of inadequate destruction of disease within clefts and may be "covered over" by regenerating epithelium appers remote. Rather, these cases appear to be related to improper colposcopie triage G.c., unsuspected intracanal involvement) or may simply represent lesions that were missed owing to their subtle appearance.

         As with any other destructive or ablative modality, success after cryosurgery is defined by three negative Pap smears over one year of followup. An abnormal smear during the first year is presumed to result from treatment failure (Richart et al, 1981).

        A laser, an acronym for "light amplified by stimulated emission of radiartion," produces electromagnetic radiation that is coherent, collimated, and monochromatic. These characteristics produce a beam of light that can be focused on tissue to produce vaporization and necrosis on contact. The wave length of the light depends on the gas being used, which in turn deter mines the effects of the light beam on the tissue. The energy beam produced by the carbon dioxide laser is absorbed by water, and its contact with tissue produces steam and thermal necrosis. The heat produced will seal small blood vessels, producing hemostasis. At high-power densities there is little thermal injury, and epithelial repair has been shown to cover the wound within 10 days Waggish, 1983, Wright, 1988).

        The two approaches to cervical disease include laser vaporization and laser excision. The former is recommended for uncomplicated lesions when invasive cancer has been clearly ruled out. The entire T zone is ablated to a depth of 6 to 8 mm periphcrally and 4 to 5 mm centrally (Baggish, 1983;Baggish, 1980). When invasive cancer has not been completely excluded,or when the lesion can be seen to extend deeper into the canal, the laser may be used to remove a cylinder of intact tissue for pathologic analysiss (Melndoc et al, 1981). A combination of vaporization and excision may be feasible when the lesion extends noto the portio and into the canal. In this case, a cylinder can be removed centrally for pathologic examination, and the well -vistralized peripheral lesion can be ablated.

        A large number of abiative methods are available. The chief advantage of these (with the exception of radical electrodiathermy) is that general anaesthesia is not required. Cryotherapyis the one method most ofteh associated with unsatisfactory results but this has usually been due to inappropriate case selection. A disadvantage common to all these techniques is that they depend heavily upon the exclusion of invasion by colposcopy and directed biopsy. In addition, they are not applicable to all patients some will always require an excisional treatment. The indications for excisional treatment of CIN are listed below:

1.Any suspicion of invasive disease.
2.Any suspicion of a glandular abnormality.
3.SCJ not clearly visible.
4.History of any previous cervical surgery.
5.Elective method of treating any case of CIN?

        Knife cone biopsy was supplanted as the standard treatment by ablative techniques partly because of the complications and partly because of the complications and partly because of the need for general anaesthesia. Complications are listed below:
1.Intraoperative haemorrhage.
2.Secondary haemorrhage.
3.Pelvic infection.
4.Cervical stenosis.
5.Cervical incompetence.

         The complications of laser cone biopsy are fewer than those of knife cone biopsy; the technique is far more precise (Larsson et al 1983) and the distortion of the cervix that results is much less, suggesting that there will be fewer problems in any subsequent pregnancies. In addition, laser cone biopsy can very ofteh be performed under local anaesthesia (Partington et al 1987).

        The ability to perform laser cone biopsy under local anaesthesia,the observation that the complications were no greater than in laser vaporization (Partington et al 1989)and anxieties about invasive cancer being missed led to a widening of the indications for excisional therapy and to the suggestion that laser excision should replace vaporization for most patients. The advent of large loop electrodiathermy excision of the transformation zone (LLETZ) made excisional treatment quicker and reduced the cost of the equipment required (Prendiville et al 1989). This technique employs a blended diathermy current and a loop of very thin stainless-steel wire.

        Prior to the era of colposcopy, cone biopsy was the standard approach to cervical precursor lesions. With colposcopy, the gynecologist can now visualize the entire cervix, so that fewer patients will require the removal of tissue to rule ort invasive cancer. Cone biopsy is indicated when (a) the endocervical limits of a lesion cannot be determined, (b) the results of endocervical curettage are clearly positive, (c) a high﷓grade lesion on Pap smear cannot be confirmed on biopsy, (d) invasion is suspected based on one parameter, or (e) the colposcopist is concerned about patient compliance or does not feel sufficiently experienced to rule out cancer in a given case (Jones, 1983; Jones and Butler, 1980). In practice, only about 10% of patients referred with abnormal Pap smears will need to undergo conization (Jones and Butler, 1980).

        Conization is a surgical procedure not suited for the office or clinic setting and requires a surgical facility and anesthesia. Complications following cone biopsy vary. The most common is bleeding, which can oceur during or immediately after conization, or secondary hemorrhage, occurring 5 to 14 days after the procedure. In one study, 2% of patients eventually required transfusion (Jones, 1983). The risk for an incompetent cervix varies. Some studies report no difference in cervical competence between patients who have undergone cone biopsy and control subjects, whereas others demonstrate a significantly greater risk for premature delivery, which appears related to the size of the cone specimen (Jones and Butler, 1980). At present, ikii view of the increasing use of the carbon dioxide laser, laser conization may provide results comparable or superior to coldknife conization. However, the latter approach is still required by many practitioners who are not familiar with use of the carbon dioxide laser.

        Two additional techniques for conservative management of cervical precursors include cold coagulation and large-loop excision of the T zone (Carter, 1984; Duncan, 1983; Prendiville et al, 1989). Contrary to its anme, cold coagulation employs a thermocouple that delivers heat via a thermosound in the range of 50 to 120 *C . Tissue destruction is achieved by boiling, in which case the epithelium is "blistered off", with thermal destruction of the underlying crypt epithelium. Failure rates have been reported to be approximately 5% over 12 months (Duncan, 1983). Largeloop excision employs a wire loop of variable size that functions as a diathermy electrode, allowing excision (paring off) of the T zone with minimal tissue damage (Cartier, 1984). Porponents of this technique point out that it is casy to apply and leads to a tissue dignosis whlie removing the lesion. As the instrument containing the wire is drawn across the tissue, the current produce a cutting effect. Loops of different sizes are used to remove larger portions of tissue. Prendiville et al (1989) reported a 2% failure rate with this technique.

        Although recommended principally for exterisive vaginal lesions, intravaginal 5-fluorouracil may be useful for treating multiple exophytic low-grade cervical warts, usually when they are associated with vaginal lesions (Kcebs, 1987). However, the adjunctive value of this approach in treating cervical disease and the risk/benefit ratio inherent in using this compound in young women with uncomplicated warts is unknown.

Pregnancy outcome following LLETZ

                                                                                LLETZ group                                   Control Group

                                                                                    (N=149)                                           (N=298)

Pregnancy outcome

Live birth146                                                            146(97.9%)                                    292(97.9%)

Fetal loss                                                                      3(2.0%)                                          6(2.0%)

Mode of delivery Vaginal Cesarean

Vaginal                                                                    133(89.3%)                                      260(87.3%)

Cesarean                                                                    16(10.7%)                                       38(12.7%)

Mean gestation(weeks)                                                   38.9                                                 39.2

Preterm delivery                                                            4(2.6)                                              3(1.0)

 <28 weeks

Preterm delivery                                                           14(9.4)                                             15(5.0)

 <37 weeks

Mean birthweight                                                    3,380(range                                   3,372(range

                                                                        2,26o-4,430)                                  1,270-4,610)­

<2, 500g                                                                       4(3.1)                                             9(3.2)

LLETZ                  LASER

Number of             Number of

patients Percent     patients Percent Pvalue

Immediate complIcations

  Bleeding              2         1.1       0       0.0       NS

  Pain                  2         1.1       0       0.0       NS

Bleeding complications

Bleeding               21         11.1      11     10.8       NS

Pain                   6          3.2       3       2.9       NS

Discharge              20          10.6     7       6.9       NS

        Hysterectomy is rarely indicated as treatment for CIN alone. The guidelincs provided by the American College of Obstetricians and Gynecologists (ACOG), published in 1993, state that some patients with high-grade CIN may be candidates for hysterectomy (3). Another indication is the socalled cancerphobie patient. This situation is extremely rare and is usually the case of the patient being cancerphobic due to misinformation provided by her surgeon. Oh the other hand, if the patient has an indication for hysterectomy, shch as uterine prolapse or adnexal mass, then the finding of CIN on a colposcopic biopsy may be an indication for treatment of the CIN by hysterectomy. The patient will usually be required to undorgo a cone biopsy first, unless the elinincian is absolntely confident of the findings of the colposcopic examination and directted biopsies.

        Positive surgical margins alone are not an indication for hysterectomy (20). Traditionally, patients with positive margins were oftehn advised to undergo hysterectoomy. Several studies have demonstrated that only 30% to 35% of patients with CIN involvement of the surgical margin will have residual CIN in the hysterctomy specimen. A more worrisome finding is the presence of CIN III in both the surgical margin and postcone biopsy; several authors have reported high rates of residual CIN and invasive cancers, especially among older women (11,21). If hysterectomy is considered in these patients with CIN III in the surgical margin and postcone ECC, Kobak and colleagues recommend a repeat cone prior to hysterectomy (22). This will avoid inappropriate simple hysterectomy for the patient with an invasive cancer undetected by the initial cone biopsy.

        The hysterectomy for CIN can be done either as a vaginal procedure or as an abdominal procedure. The indications for the hysterectomy will determine whether the surgeon also plans to remove the ovaries; for treatment of CIN there is no reason to add oophorectomy to the procedure. In selected patients, if a vaginal hysterectomy is indicated, this can be done as an outpatient, same-day procedure.

        Abnormal findings on Pap smear after ablative therapy indicate either treatment failure or recurrence. For practical purposes, failure is defined as a lesion recutting in the first 12 months of followup, preferably after three Pap smears have been abtained. After one year cervical abnormalities are presumed to be persistent lesions or lesions that have developed directly from occult virus adjacent to the treatment field (Fcrenczy et al, 1985; Richart et al, 1980; Richart et al, 1981). Recurrences are defined as lesions detected after three negative Pap smears over the first year of followup. Richart et al (1980) found the long﷓term (after one year) recurrence rate to be 0.44%/year, which according to their calculations approached the level of risk of populations known to be at risk. In a study of cone biopsies, Bjerre et al (1976) reported a recurrence rate of 0.5% for 749 women who initially had negative smears after conization. Kolstad and Klem (1976) reported a 2.4% recurrence rate following cone biopsy, but only 0.8% of these lesions developed after 2 years of followup (Baggish, 1983).

        With cone biopsy, a primary concern is the status of the cone margins and their relationship to recurrent disease and the risk of missed cancer. Ostergaard (1980) reported that 30% of 516 cone biopsies were found to have discase in the margins. However, the reliability of this information is unclear. For example, in Ostergaard's study, 84% (21/25) of patients with positive m,arglins with long-term followup did not suffer a recurrence. Similarly, Bjerre et al (1976) found that 60% of women whose cone biopsy specimons had positive margins had negative. Pap smears for at least 5 years. Conversely, Ostergaard found that 16% of 268 women with negative margins on cone biopsy had residual disease at subsequent hysterectomy. 

        Thus, the management of women after cone biopsy is based primarily on the use of repeated cytologie evaluation of the cervical os and canal The actual risk of invasive cancer is not certain. Ostergaard reported microinvasion in the endocervical canal in 2 of 26 women with positive apical cone margins. In practice, determining which women should proceed to hysterectomy to confirm this should be based upon postoperative cytology, careful analysis of the specimen to rule out invasion, and the amount of disease adjacent to the endocervical margin when this margin is decmed positive.