Comparison of cytology and histology results in English cervical screening laboratories before and after liquid‐based cytology conversion: do the data provide evidence for a single category of high‐grade dyskaryosis?

R. G. Blanks and R. S. Kelly
Comparison of cytology and histology results in English cervical screening laboratories before and after liquid‐based cytology conversion: do the data provide evidence for a single category of high‐grade dyskaryosis? Objective: To determine whether the difference between the positive predictive value (PPV) for cervical intraepithelial neoplasia (CIN) grade 2 or worse (CIN2+) of referral from moderate dyskaryosis and from severe dyskaryosis was reduced after laboratories converted from conventional to liquid‐based cytology (LBC). Furthermore, to explore the cytology/histology agreement after LBC conversion, and to determine post‐LBC whether there was increased support for the use of one single category of high‐grade dyskaryosis (equivalent to high‐grade squamous intraepithelial lesion). Methods: The association between cytology and histology has been examined using annual Korner return data (KC61 returns) collected by laboratories from the English National Health Service cervical screening programme. The study compares return data before and after LBC conversion. Results: The study examined data from 102 laboratories that converted from conventional cytology to LBC. Before conversion the PPV for CIN2+ of severe dyskaryosis was 88% and after increased to 90% (P = 0.003). For moderate dyskaryosis the PPV for CIN2+ increased from 70% to 72% (P = 0.06). The absolute difference of 18% between severe and moderate dyskaryosis was therefore the same pre‐ and post‐LBC conversion. The PPV of mild dyskaryosis for CIN2+ before and after conversion reduced from 23% to 19% (P < 0.001). The agreement between cytology and histology measured using a weighted Kappa statistic increased from 0.52 to 0.60 after conversion to LBC because of small increases in the proportions of severe dyskaryosis or worse with CIN3+ outcomes and mild dyskaryosis with CIN1 or less outcomes. Conclusions: Following LBC conversion there was evidence of a modest increase in the agreement between cytology and histology but no evidence of a change in the absolute difference in PPV for CIN2+ between moderate and severe dyskaryosis. The data support the conclusion that women referred with moderate dyskaryosis will on average have a lower risk of progression to invasive cancer than women referred with severe dyskaryosis. If the data were considered to support the categories of high‐grade dyskaryosis (moderate) and high‐grade dyskaryosis (severe) before LBC conversion then it can be strongly argued that they also support these categories after conversion.     

Fair shares in health care? Ethnic and socioeconomic influences on recording of preventive care in selected inner London general practices. Healthy Eastenders Project.

OBJECTIVE--To describe the association of ethnic and socioeconomic status with recording of preventive care information by selected general practitioners. DESIGN--Random selection of people aged 20-64 registered with 43 general practitioners. Ethnic and social characteristics of stratified samples were determined at interview in the subject''s home. Recording of preventive information was ascertained from general practitioners'' medical records. SETTING--Inner London borough of Tower Hamlets. SUBJECTS--505 ut of 739 people confirmed as residents at their home address (190 white, 86 black, 112 Bangladeshi, 105 Chinese or Vietnamese, 12 other). MAIN OUTCOME MEASURES--Socioeconomic characteristics, consultation with general practitioner, and recorded preventive activities for ethnic groups. RESULTS--Minority ethnic groups were considerably more disadvantaged than white people and five times more likely to be overcrowded (31% v 6%), three times less likely to own their own home(11% v 37%), twice as likely to be in social classes IV and V (54% v 28%) and less likely to be employed (34% v 63%). There were no significant differences between white, black, Bangladeshi, and Chinese or Vietnamese subjects in recording smoking, blood pressure, alcohol consumption, weight, and height in the general practitioners'' medical records. White women were more likely to have a record of mammography (46% v 20%; P=0.03) and of cervical smears than women in minority ethnic groups. CONCLUSION--Despite major socioeconomic inequity, equitable recording of preventive activity for the major causes of death for white, black and Bangladeshi populations is possible. Chinese and Vietnamese people had lower levels of recording and consultation. Mammography and, to a lesser extent, cervical cytology are inequitably recorded and require additional support at practice level.     

Mild and moderate dyskaryosis: can women be selected for colposcopy on the basis of social criteria?

OBJECTIVE--To describe the distribution of cervical intraepithelial neoplasia grades among women with mild and moderate dyskaryosis after a single cervical smear and to determine whether social criteria could help identify women who are at increased risk of grade II or III disease. DESIGN--Cross sectional analysis within a randomised prospective study. Subjects had a repeat smear, a colposcopic examination, and an excision biopsy of the transformation zone. In addition, women were asked to complete a social questionnaire. SETTING--Colposcopy clinic, Aberdeen. SUBJECTS--228 women with a single smear test showing mild or moderate dyskaryosis. MAIN OUTCOME MEASURES--Histology, age, sexual and contraceptive history, cigarette smoking. RESULTS--159 (70%) women had cervical intraepithelial neoplasia grades II or III. Among current smokers the prevalence of grade II and III disease was higher in women who smoked greater than or equal to 20 cigarettes a day (84%) than among those who smoked less (66%; p less than 0.04). Women with more than one sexual partner also had a higher prevalence (75%) than women with only one partner (50%; p = 0.0028). Use of oral contraceptives and younger age were not significantly associated. The prevalence of grade II or III disease was up to 66% in the lower risk groups. CONCLUSIONS--Because of the high prevalence of cervical intraepithelial neoplasia grades II and III in both the high and the low risk groups social factors are not useful for selecting women with mild or moderate dyskaryosis for either early referral to colposcopy or cytological surveillance.     

Regular audit of colposcopic biopsies from women with a mildly dyskaryotic or borderline cervical smear results in fewer cases of CINIII

Two years after introducing mandatory review of cases in which the cervical smear was discrepant with subsequent colposcopic or histological finding, the predictive accuracy of a first abnormal smear and the need for treatment were analysed. The results were compared with performance figures prior to this form of audit policy. Over 12 months 415 women referred for colposcopy were studied. Three per cent of patients with a single borderline smear and 6% with mild dyskaryosis had cervical intraepithelial neoplasia grade III (CINIII) revealed in histopathological examinations after colposcopy. Only 25% with a borderline smear and 33% with mild dyskaryosis required treatment. Of women with moderate dyskaryosis, 18% had a biopsy showing CINIII and 46% were treated. Of women with severe dyskaryosis in their cervical smear, 61% were shown to have CINIII or invasive cancer on biopsy and 90% were treated. Regular audit improved cytological prediction of grade of epithelial abnormality found on biopsy, allowing accurate, safe surveillance for minor smear abnormalities.     

Regular audit of colposcopic biopsies from women with a mildly dyskaryotic or borderline cervical smear results in fewer cases of CINIII

Two years after introducing mandatory review of cases in which the cervical smear was discrepant with subsequent colposcopic or histological finding, the predictive accuracy of a first abnormal smear and the need for treatment were analysed. The results were compared with performance figures prior to this form of audit policy. Over 12 months 415 women referred for colposcopy were studied. Three per cent of patients with a single borderline smear and 6% with mild dyskaryosis had cervical intraepithelial neoplasia grade III (CINIII) revealed in histopathological examinations after colposcopy. Only 25% with a borderline smear and 33% with mild dyskaryosis required treatment. Of women with moderate dyskaryosis, 18% had a biopsy showing CINIII and 46% were treated. Of women with severe dyskaryosis in their cervical smear, 61% were shown to have CINIII or invasive cancer on biopsy and 90% were treated. Regular audit improved cytological prediction of grade of epithelial abnormality found on biopsy, allowing accurate, safe surveillance for minor smear abnormalities.     

Fertility and legal abortion in England and Wales: performance indicators for family planning services.

The relation between fertility rates and legal abortion rates was investigated in a sample of health authorities in England and Wales to see how these varied. Total period fertility rates and total period legal abortion rates were derived from the average number of live births or legal abortions that would be experienced per woman if women experienced the age specific rates of the year in question throughout their childbearing years. The sample of 30 health authorities was selected by taking the districts with the highest and lowest fertility rates in each English region and in Wales in 1986. Total period fertility rates varied from 1.37 in Riverside to 2.42 in Tower Hamlets, while abortion rates varied from 0.25 in East Yorkshire to 0.99 in Riverside. When the two rates were added to provide a potential fertility rate it became clear that some districts with similar potential fertility rates had very different underlying component rates. Such comparisons can be used for service monitoring, indicating the need for better abortion and family planning services in districts with high fertility rates and for better family planning services in those with high abortion rates.     

Management of women with mild and moderate cervical dyskaryosis.

OBJECTIVE--To compare the outcomes in women with mild and moderate dyskaryosis after increasing periods of surveillance and thereby to define a rational protocol for managing such women. DESIGN--Prospective study with randomisation of women to one of four treatment groups, each with a different period of surveillance; one group in which the women were given immediate treatment and three other groups in which the women were under surveillance for six, 12, and 24 months. SETTING--A dedicated colposcopy clinic in Aberdeen, Scotland. SUBJECTS--902 women who presented with a mildly or moderately dyskaryotic smear for the first time. INTERVENTIONS--Cytological and colposcopic examinations at intervals of six months until the allocated period of surveillance was completed, at which time biopsy was performed. Women with severe dyskaryosis were withdrawn from surveillance and a biopsy was performed. MAIN OUTCOME MEASURES--The histological findings after punch biopsy or large loop excision of the transformation zone, and the trends in cytological appearances of serial cervical smears. RESULTS--793 women completed the study. In all, 769 women had an adequate final smear, of which 197 were normal cytologically, 328 were still mildly or moderately dyskaryotic, and 244 were severely dyskaryotic. Seventeen of the 67 (25%) women with one repeat smear showing non-dyskaryosis had cervical intraepithelial neoplasia grade III compared with only one of the 31 (3%) women with no dyskaryosis in four repeat cervical smears (P < 0.0001). None of the women had invasive cancer. Of 158 women whose index smear showed mild dyskaryosis and who were allocated to the group under surveillance for two years, only 40 had not defaulted or still had dyskaryotic smears by the end of the two years. CONCLUSION--Cytological surveillance, although safe, is not an efficient strategy for managing women with mildly abnormal smears. Women with any degree of dyskaryosis in a smear should be referred for colposcopy.     

Long term follow up of women with borderline cervical smear test results: effects of age and viral infection on progression to high grade dyskaryosis.

OBJECTIVE--To follow up and assess the significance of borderline change in cervical smears. DESIGN--Retrospective study of women undergoing routine cervical cytological screening in 1981. SETTING--Avon Cervical Screening Programme, covering 250,000 women in Bristol and Weston super Mare. SUBJECTS--437 women showing borderline cervical changes in 1981 and 437 age matched controls with normal results in 1981. MAIN OUTCOME MEASURES--Cytological progression to high grade dyskaryosis (cervical intraepithelial neoplasia grade III or invasive carcinoma). RESULTS--During follow up ranging from 13 to 106 months 98 of the 437 women (22.4%) with borderline cytological changes on routine cervical cytology screening had a subsequent smear test showing high grade dyskaryosis compared with three of the 437 women (0.9%) in the control group. The risk of progression was greater in women aged 20 to 39 than in those aged 40 and over. Human papillomavirus infection had initially been diagnosed cytologically in 101 of the 437 (23%) women with borderline results. Significantly fewer of these women developed high grade dyskaryosis (13/98 (13%) v 88/339 (26%), p less than 0.05). CONCLUSIONS--Women with borderline smear test results are at increased risk of developing high grade dyskaryosis, particularly if the borderline changes occur without cytological features of human papillomavirus infection. Progression occurs within three years in 50% of cases, although a linearly increasing risk was sustained over the nine years of follow up and was greatest in women aged 20 to 39. Careful follow up of these women is indicated.     

The Cytological Prediction of Cervical Intraepithelial Neoplasia In Colposcopically Directed Biopsies

The tissue sections and preceding cervical smears of 1262 women who had colposcopic cervical biopsies were reviewed and the reports correlated. Close correlation between the cytological and histological findings, to within one histological grade of cervical intraepithelial neoplasia (CIN), was noted in 86% of cases. However, the biopsy was negative, or contained evidence of wart virus infection only, in 24% of cases where dyskaryotic cells had been observed in the cervical smear. Of particular concern was the fact that negative histological findings were recorded in 13% of cases where the smear contained cells showing a moderate dyskaryosis and in 1.26% of cases where the smear showed severe dyskaryosis. This suggests that colposcopically directed biopsies do not always reflect the underlying pathological changes in the cervix. Management of these cases is discussed. In 45 women with a normal cervical smear prior to biopsy, histology revealed seven cases of CIN 3 and one case of invasive squamous carcinoma. This indicates that referral for colposcopy is advisable whenever there is clinical suspicion of cancer, even if the cervical smear report is normal.     

Assessment of the Accuracy of Cytology In Women Referred For Colposcopy and Biopsy: the Results of a 1 Year Audit

The results of weekly colposcopy review meetings have been audited for 1 year and cases where there was a discrepancy between the referral cervical smear and the initial colposcopy biopsy have been analysed. New referrals (n = 476) for colposcopy were studied. In the final outcome 80% of 326 women referred for moderate or severe dyskaryosis were found to have cervical intraepithelial neoplasia (CIN) grade II or III or invasive carcinoma. Three women found to have invasive carcinoma had been referred for severely dyskaryotic smears. Twenty women were referred for smears with cell changes suggesting glandular neoplasia: five were found to have adenocarcinoma in situ, whereas eight had CIN and seven had negative biopsies. The results justify the referral policy and demonstrate the need for further investigation when initial colposcopic biopsies are negative.     

Evaluation of liquid-based cytology in cervical screening of high-risk populations: a split study of colposcopy and genito-urinary medicine populations

A split study evaluated the ThinPrep(R) PapTesttrade mark (TP; Cytyc Corp., Boxborough, MA) compared with current methodologies of cervical cytology in two high-risk cohorts. One thousand, three hundred cases from a colposcopy clinic and a genito-urinary medicine outpatient clinic were examined. The TP reported increased detection of all grades of dyskaryosis (mild, moderate and severe; + 4.5%) and a decrease in borderline and unsuitable cases (- 4.9%). Four cases of high-grade dyskaryosis (moderate or severe) were detected only using the TP, while an additional four cases classified as high-grade dyskaryosis with the TP were reported as borderline by our conventional methods. The split-study finding of increased sensitivity with the TP provides for improved clinical management of patients in our high-risk cohorts.     

Six years' audit of laboratory workload and rates of referral for colposcopy in a cervical screening programme in three districts.

OBJECTIVE--To determine laboratory workload and rates of referral for colposcopy in a three district cervical screening programme during 1983-9 to assess the feasibility of accommodating call up of all women at risk, recall at three year intervals (now five year intervals), and investigation of women with all degrees of abnormality. DESIGN--Analysis of computerised screening histories dating back to 1977 of women screened in the Avon cervical screening programme. SETTING--Three district health authorities covering the population of Bristol and Weston-super-Mare, comprising 800,000 people, of whom 250,000 were female residents aged 20 to 64. SUBJECTS--196,977 Women aged 20 to 64 screened in cervical screening programme since 1983. RESULTS--Laboratory workload devoted to follow up of women with abnormalities increased sharply between 1987-8 and 1988-9, with increases of 54% (from 2075 to 3196) in the number of smears for follow up of severe dyskaryosis and invasive cancer, 40% (from 1925 to 2695) for mild and moderate dyskaryosis, and 49% (from 1793 to 2677) for borderline change. The increases were partly explained by the introduction in April 1988 of protocols for follow up and investigation based on guidance in an intercollegiate working party report. The proportion of women with mild and moderate dyskaryosis who were recommended for referral for colposcopy increased steadily from 9.9% in 1983-4 to 79.9% in 1988-9, and for borderline change the proportions were 3.5% and 13.6% respectively. Of all women tested in 1988-9, referral for colposcopy was recommended in 3%. CONCLUSIONS--The increase in laboratory follow up work identified, if it continued, could result in half of existing laboratory capacity in Avon being devoted to follow up work by 1993, with little prospect of maintaining call, recall, and quality control. Investigation of all women with minor cytological abnormalities is neither justifiable nor sustainable and will undermine the benefits of screening by increasing the rate of false positive results and the financial costs.     

O‐110 YEAR AUDIT OF GLANDULAR CYTOLOGY REPORTS WITH SPECIAL REGARD TO THE EFFECT OF CONVERSION TO LBC

Objectives: A randomised trial in screening to improve cytology (ARTISTIC Trial) was set up to evaluate the effectiveness of human papilloma virus (HPV) testing in primary cervical screening. Design: Women attending the NHS Cervical Screening Programme in Greater Manchester were screened for HPV in addition to routine cytological testing with ThinPrep Liquid‐Based Cytology (LBC). 25 020 women aged 20–64 were screened at entry and 3 years later. Participants were randomised in a 3 : 1 ratio between those whose HPV test result was revealed and acted upon and those in whom it was concealed. Results: A total of 24 510 women (18 386 revealed and 6124 concealed) had adequate cytology and HPV results. Cytology results at entry were: 87% normal, 11% borderline or mild, 1.1% moderate and 0.6% severe dyskaryosis or worse. Prevalence of HPV decreased sharply with age, from 40% at age 20–24 to 12% at 35–39 and 7% or less above age 50, and increased with cytological grade, from 10% (normal cytology) to 31% (borderline), 70% (mild), 86% (moderate), and 96% (severe or worse). Only 60% of women with severe dyskaryosis showed HPV 16 and / or 18 infection. Conclusion: The ARTISTIC Trial has provided the largest collection of LBC samples with HPV typing in Greater Manchester population. It has shown that the HPV testing is practicable in routine screening in women aged over 30. HPV 16 and 18 are more predictive for underlying disease, but other HPV types account for 30% of high‐grade disease.     

Posters. P11 Positive predictive values for high and low‐grade cytological abnormalities as surrogates for specificity and sensitivity

Introduction Positive predictive value (PPV), measuring the percentage of moderate dyskaryosis or worse confirmed as CIN2 or worse, is used as a measure of accuracy in cervical screening. However, it relates more to specificity than sensitivity because the denominator includes false positives rather than false negatives. Low values reflect over‐reporting of high‐grade dyskaryosis but high values may reflect under‐reporting. Sensitivity is impossible to measure from correlation of cytology with outcome because women with negative cytology are rarely referred for colposcopy. Rates of CIN3 resulting from referrals for low‐grade cytology may be used as a surrogate for sensitivity, as high values may reflect under‐reporting (ref). Study design Outcome of colposcopy referrals was monitored during a period of 4 years, using a fail‐safe database. Results PPV at Guy's & St Thomas rose from 54% in 1998/1999 to 69% in 2001/2002. The former was below the NHSCSP recommended range. During the same period of time CIN1 rates for moderate dyskaryosis fell from 37% to 24%, reflecting the main source of discrepancy. While specificity increased (as reflected by increasing PPV) sensitivity remained constant in that CIN3 rates for mild dyskaryosis and borderline remained below 6%: average rates in England have fallen over the last 3 years and were 7.4% in 2000/2001 (ref). CIN2 rates for mild dyskaryosis also remained constant at 11% to 12%. Conclusion Correlation of biopsy results with high‐ and low‐grade cytological abnormalities is a useful method of monitoring accuracy of cytology reporting, and can be used to measure over‐ and under‐reporting as surrogates for specificity and sensitivity.     

Improving the NHS cervical screening laboratory performance indicators by making allowance for population age, risk and screening interval

OBJECTIVE: One of the key performance measures in the monitoring of the NHS cervical screening programme is the targeting of laboratories with very high or low percentages (outside the 10th-90th percentile) of adequate smears that have moderate dyskaryosis or worse. These laboratories are assumed to include those laboratories that may have extremes of sensitivity and specificity. A clear limitation with this methodology is that laboratories do not examine smears from women with the same underlying risk, age distribution or screening interval and adjustment for these factors should considerably improve the method. METHODS: This paper describes a method that allows for these confounding variables and a new age-risk-interval adjusted moderate dyskaryosis or worse rate (ARI-adjusted mod+ rate) can be calculated. The adjusted rate is the rate of moderate or worse dyskaryotic smears that the laboratory would have detected had it been screening women with an English 'average' age-risk-interval. All laboratories can therefore be compared using this method. RESULTS: The methodology is illustrated using data from the NHSCSP South West Region. The particularly low percentage of moderate or worse smears detected by one or two laboratories can be shown to be due to a local screened population with a very low risk because of a high mean age, relatively short screening interval and census variables associated with a low risk, rather than any under-calling by the associated laboratories. CONCLUSIONS: The ARI-adjusted mod+ rate requires to be calculated for all laboratories in England if it is to be used as a primary performance indicator. Alternatively, it can be used to further examine laboratories that are deemed to be outliers using the current methodology.     

Five-year follow-up of women with borderline and mildly dyskaryotic cervical smears

This study investigated the 5-year follow-up status of women with cervical smears showing borderline nuclear changes (BNC) or mild dyskaryosis and the effect of koilocytosis on the outcome. Thirteen per cent of women with cervical smears showing BNC had high-grade cervical intraepithelial neoplasia (CIN). In contrast, 28% of women with cervical smears showing mild dyskaryosis had high-grade CIN. The presence of koilocytosis (24% for borderline smears and 34% for mild dyskaryotic smears) did not appear to influence the risk of developing high-grade CIN. Our results suggest that the simultaneous implementation of the British Society for Clinical Cytology proposed terminology and the colposcopy guidelines from the British Society for Colposcopy and Cervical Pathology could have an impact on colposcopy services.     

ABC3 Part II: a review of the new criteria for evaluating cervical cytology in England

R.G. Blanks
ABC3 Part II: a review of the new criteria for evaluating cervical cytology in England The new Achievable Standards, Benchmarks for Reporting, and Criteria for Evaluating Cervical Cytopathology, 3rd edn (ABC3) includes radical changes in the criteria for evaluating cervical cytology. First, they include a new mission statement ‘the objective of cervical screening is to reduce cervical cancer incidence and mortality by screening with a high sensitivity for the detection of CIN2 or worse, whilst maintaining a high specificity’. Second, the original four performance measurement criteria where laboratories were examined further if they were below the 10th or above the 90th percentile has been changed to three and laboratories are only mandatorily examined if they fall below the 5th or above the 95th percentile. The old criteria related to the percentage of samples that were inadequate, the percentage of all adequate samples reported as moderate dyskaryosis or worse (equivalent to high‐grade squamous intraepithelial lesion or cancer), the percentage of adequate samples reported as mild dyskaryosis or borderline (equivalent to low‐grade squamous intraepithelial lesion or atypical squamous/glandular cells) and the positive predictive value. The new criteria are percentage of inadequate samples, positive predictive value and a new measure termed referral value. These changes mean that far fewer laboratories will require mandatory examination. Third, a raft of optional performance measures have been introduced to help laboratories examine their annual statistical return to the Department of Health in comparison with other laboratories. These measures have been designed to produce a more uniform national programme, and to help laboratories decide whether they are maximizing the benefit of screening while minimizing the harm, which is the goal of all screening programmes. This review examines in detail the new criteria and explains in more detail some of the thinking behind them.     

Requests for late termination of pregnancy: Tower Hamlets, 1983.

The case histories of all women seeking late (more than 20 weeks'' gestation) abortion in the NHS district of Tower Hamlets in 1983 were assessed. Of 12 women requesting late abortion, seven underwent termination of pregnancy. All the women had severe social or psychological problems, or both. The main reasons for late presentation were denial of pregnancy, youth, and mental disorder. In a small group of atypical women late abortion seems to be justified for reasons other than fetal abnormality.     

Management of Women With A Cervical Smear Showing A Mild Degree of Dyskaryosis: A Review of Policy

Three hundred and thirty-seven women who presented for the first time with a cervical smear showing a mild degree of dyskaryosis were followed for a minimum period of 3 years and 9 months. Of the 305 women with complete cytological and histological records, 178 were biopsied and 127 remained on cytological follow up. In the biopsied group one case of microinvasive squamous cell carcinoma was diagnosed within 1 year of the patient's first abnormal smear. A further 24% showed cervical intraepithelial neoplasia (CIN) I, 17% showed CIN II and 29% showed CIN III. The overall regression rate for the group of 305 women was 34%. Our results indicate that cytological surveillance is acceptable provided that biopsy is advised when dyskaryosis persists. No major modifications to laboratory policy are indicated and in approximately 34% of cases an unnecessary hospital referral would be avoided.     

Cervical intraepithelial neoplasia grade III (CIN III) and invasive cervical carcinoma: the yawning gap revisited and the treatment of risk

In a 3-year study of the population of Southampton and south-west Hampshire there were 10 times as many cases of CIN III compared with invasive squamous carcinoma (700 compared with 70). The peak incidence of CIN III per 1000 screened women years was in those aged 25-29 years, which was 20 years earlier than the peak incidence of invasive cervical cancer per 1000 women years at risk. Ninety percent of CIN III was diagnosed in women under 50 years. There were 14 cases of cervical glandular intraepithelial neoplasia grade III (CGIN III), three coexisting with CIN III, all in women aged under 50 years: the gap between intraepithelial and invasive lesions was not seen for glandular neoplasia. Although referral was for at least moderate dyskaryosis in 86.8% of women with CIN III or CGIN III, most had been screened previously, either having had mild abnormalities requiring repeat cytology (39.8%) or negative cytology (34.5%). Only 12 women aged > or = 50 years had previous negative cytology: 21.4% compared with 35.6% of women aged < 50 years (P = 0.034). The results of this study suggest that the best opportunity for preventing invasive squamous cell carcinoma lies in screening women aged 20-39 years when the incidence of CIN III in the screened population is highest and before the peak incidence of invasive disease. The results also indicate the importance of repeated screening and follow up of minor cytological abnormalities in the detection of CIN III. The benefit of screening must be regarded as a treatment of risk, since it is almost certain that a high proportion of CIN III regresses or persists unchanged.