Cytologist should be encouraged to say precisely what abnormality is present in as many cases as possible(1) and should always strive to provide a clear specific diagnosis to the clinician(2). However there are occasions when a specific diagnosis is not possible and the term borderline nuclear changes (BNC) should then be used in those cases where there is genuine doubt as to the dyskaryotic nature of the cells or where the nuclear abnormalities fall short of mild dyskaryosis. The term was devised to be used in cases where it is impossible to decide if the cells are a product of inflammation or if they have neoplastic potential.
It is interesting to note that in the original concept, for which the term borderline nuclear abnormalities was devised by the British Society of Clinical Cytology (BSCC), there would be relatively few smears where the report would remain equivocal (1) . Yet current data shows the BNC report to have an occurrence in 3.8% of smears in England whereas dyskaryosis accounted for 3.3% of smear reports(3) . It would seem that despite much discussion and debate as to what constitutes a cell showing BNC and the circumstances in which such a report would be issued, the term is not used in the spirit of the original recommendations.
As a continuous range of nuclear abnormalities occurs ranging from minor changes that are usually associated with inflammatory conditions to the changes that correlate to cervical intraepithelial neoplasia (CIN) (1), the potential areas in which a borderline report may be appropriate are wide and varied.
Guidelines(4) were introduced some eight years after the original terminology was introduced and the term borderline nuclear changes (BNC) rather than borderline nuclear abnormalities was used. The guidelines set out the occasions when it would be appropriate to use the BNC report and these were reiterated some six years later(5). The BNC report may be used in the presence of human papilloma virus (HPV) infection when the pathognomonic koilocytes show milder nuclear changes than those required for mild dyskaryosis, or when there are small dyskeratotic cells which have variable nuclear enlargement with increased nuclear cytoplasmic ratio and condensed nuclear chromatin. The second situation in which a BNC report is valid covers a diverse group of conditions in which it may be difficult to distinguish between benign, reactive or degenerative changes from higher degrees of dyskaryosis or even invasive cancer. Such instances may arise in the presence of inflammation or metaplasia where there is an increase in the nuclear cytoplasmic ratio in the absence of cellular degeneration or overt dyskaryosis; where there is mild hyperchromasia in enlarged nuclei; or when the chromatin pattern is coarsened slightly. The third occasion when a BNC report is appropriate is when endocervical cells show three dimensional groups with disorderly cell arrangements, irregular or coarse grainy chromatin, hyperchromasia with intercellular variation in the depth of nuclear staining, and or an irregular nuclear membrane not due to distortion by a cytoplasmic vacuole.
Parham et al(6) suggested than BNC represented two distinct groups, those with reversible reactive changes and those with early dysplastic changes that may or may not progress. They stated that the histological term basal abnormality of uncertain significance(7) equated to the BNC seen in cervical smears.
Edwards et al(8) recognized that 10% of smears with BNC would show a histological outcome of high grade abnormality and suggested that there would be advantages in highlighting this subgroup of patients. They suggested a borderline high grade (BL/HG) and a borderline low grade (BL/LG) category. However although the two observers failed to agree on the criteria used to classify the BL/HG category, features such as the nuclear cytoplasmic ratio, irregularity of nucleoli and chromatin characteristics were used for evaluation. They found the positive predictive value of recognizing BL/HG to be 44% compared with the 13.9% for all borderline abnormalities. However, their paper made no mention of BNC in glandular cells.
The smear showing features between the accepted definitions of normality and abnormality have presented classification difficulties in both the BSCC terminology and the Bethesda System(9). The previous Papanicolaou classification(10) made no provision for the smear that was not negative, worse than inflammatory in nature but less than dysplastic. Cytologists acknowledge that in some cases sufficient diagnostic criteria are lacking and that terminology should reflect the diagnostic uncertainty but such use of such terms should be minimized(2) .
The Bethesda term ASCUS (atypical squamous cells of undetermined significance) does not constitute a diagnostic entity but encompasses a spectrum of cellular changes reflecting a variety of pathological processes that cannot be more specifically classified(2) and as such has similarity to the BNC terminology. The term AGUS (atypical glandular cells of undetermined clinical significance) has been applied to glandular cells that demonstrate changes beyond the benign reactive process yet are insufficient for a diagnosis of adenocarcinoma. However, the recent Bethesda classification(9) does not include the AGUS category but uses the term atypical applied in two different ways; firstly to endocervical, endometrial or glandular cells in general with more specific elaboration in the text of the report; and secondly to endocervical or endometrial cell with the report stated as favouring neoplasia.
The BSCC set up a Terminology Conference in 2002 and proposals for change were based on consensus discussion(11) . It was felt that the borderline category provided an essential buffer and should be retained. Modifications to BNC were that the should be retained and subdivided into BNC (NOS), BNC glandular and BNC high-grade dyskaryosis not excluded, however the borderline smear showing koilocytosis would be grouped with mild dyskaryosis.
Follow up of women with smears showing BNC and subsequent histological outcomes have shown that there is a 23% excess of cases showing dyskaryosis in subsequent repeat smears after a diagnosis of BNC(6). Most of the risk was found to be in the first two years after the initial smear and women were eight times more likely to develop dyskaryosis in the short term than a woman with a negative smear. It was suggested that random sampling error, and the fact that some cases of BNC represent the earliest cytological changes in a dysplastic epithelium, accounted for the findings.
Patients referred for colposcopy for unsatisfactory smears or smears with changes less than dyskaryosis(12) were found to have CIN of all grades on punch biopsies and large loop excision of the transformation zone (LLETZ) in women less than 50 years of age but not in those over 50. These findings would appear to uphold the view(4) that the cytological expression of HPV and dyskaryosis is suppressed in atrophic smears.
Other studies(13) in follow up of women with borderline cervical smears have found subsequent high grade lesions in about 10% of patients two years following the index smear. Correlations with histology(14) have found that cases reported cytologically as BNC were attributed to various grades of CIN up to CIN3. Such cases should not be interpreted as a misdiagnosis as BNC is a holding category for cases in which the changes identified cannot definitely be attributed to neoplasia. Cone biopsies show areas of low grade CIN and it may be that these are the areas sampled by the smear taker. When correlating cytology with histology it is important to remember that biopsies are taken under colposcopic guidance and represent specific sampling whereas smears taken in the GP setting are by naked eye examination and represent random sampling.
As the classification and recognition of BNC has undergone review and modification over the years so has the management of a patient with a borderline smear. The guidance in 1987(15) was that the borderline smear should be repeated in 3 months and then managed according to the result, with persistent abnormalities being an indication for colposcopy. Such guidance suggests that patients with negative repeat smears would then have been returned to normal recall without follow up.
Current National Health Service Cervical Screening Programme (NHSCSP) guidelines(5) state that BNC in squamous cells should be repeated at least once and preferably twice at a 6-12 month interval and referred to colposcopy should the abnormality persist. The guidelines make allowance for referral on the first borderline smear if the cytologist is concerned that there may be an underlying high grade abnormality. The repeat interval after a borderline smear should be at a maximum interval of six months and follow up recommendations are that there should be a least 3 negative smears six months apart before a women with a previous borderline smear is returned to normal recall. The BNC in glandular cells calls for a repeat at 6 months with endocervical brushings followed by colposcopic referral if the abnormality is still present. Follow up of BNC referred for colposcopic assessment is a repeat smear after treatment or after an interval of 3-6 months.
The British Society for Colposcopy and Cervical Pathology (BSCCP) set out draft guidelines(16) regarding the management of BNC and suggested that women should be referred for colposcopy after three smears with a BNC report in squamous cells, in keeping with current NHSCSP guidelines(5). They recommend that women should be referred for colposcopy after one smear report of BNC in endocervical cells. If the new terminology of the BSCC(11) is adopted in conjunction with the BSCCP guidelines, women with koilocytosis alone, that hitherto had been graded as BNC, will now be classified as mild dyskaryosis and will therefore warrant referral for colposcopic assessment without any follow up smears.
In order to further manage the borderline smear HPV testing is being studied in two trails(17). In the first the HPV pilots linked to liquid based cytology (LBC) follow the protocol that if the HPV test is negative, then there will be a repeat smear in 6 months time and a second HPV test. If the apparent abnormality has either persisted as BNC or mild dyskaryosis or progressed to a high-grade smear, or the women is found to be HPV positive, she is then referred for colposcopy. If the abnormality has regressed or is no worse than borderline nuclear change and the woman is HPV negative on both occasions then she will be discharged back to routine screening.
The initial protocol has been revised in two of the pilot centres, and women who test positive for HPV following a low-grade smear are only referred to colposcopy immediately if they are aged 35 or over when their smear is taken. Younger women are asked to return for a repeat smear and HPV test in six months and the referral is made on the basis of this second test.
In the second pilot study the question of the most appropriate way to deal with HPV positive results and associated psychological issues is being addressed. The Medical Research Council has funded the Trial of Management of Borderline and other Low grade Abnormal smears study (TOMBOLA). The seven year multi-centre trial has a number of aims: to determine whether a policy of cytological surveillance or initial colposcopy is the more efficient and effective policy for further investigation and clinical management in women with mild or borderline cervical dyskaryosis; to determine whether, following colposcopic examination of women with mild or borderline dyskaryosis, immediate LLETZ or biopsy and recall for LLETZ, is the more effective and efficient method of treatment; and to evaluate the contribution of HPV testing to the effectiveness and efficiency of the existing procedures for management of women with mild or borderline dyskaryosis.
Liquid based cytology may help in reducing the reporting rate of BNC and may also be of use with the reflex testing of HPV and so contribute to the management of the woman with a borderline smear. However LBC itself is controversial, some workers (18) have found that LBC reduced the ASCUS/AGUS rate by nearly 27%, whereas the LBC trials in England(20) have found no clear evidence of any impact of LBC on the detection of borderline smears.
The smear with BNC has proved problematic in diagnosis and management since the time of Papanicolaou and continues to do so today. In the factual debate over BNC, cytological and histological evidence has been gathered in an attempt to provide a more accurate histological diagnosis that has resulted in the various proposed classifications and sub-classifications. The management of patients with BNC has been changed over the years and may move from cytological follow up and colposcopy, to colposcopy and HPV testing as proposed in the trial studies.
Despite the discussions and accumulated evidence relating to the borderline smear and its management there are areas of fantasy regarding BNC. One area of confusion has been the actual terminology that has metamorphosed from borderline nuclear abnormalities to borderline nuclear changes. There is also the undisputed fact that 3.8% of smears are reported as having BNC. However the original concept that the BNC category would be for those few cases where there was genuine doubt to the dyskaryotic nature of the cells suggests that the borderline report is used more than envisaged. In fact the phrase "genuine doubt" has an area of the fantastic about it for although many papers refer to the 1986 BSCC terminology paper as a source for the phrase, it was not used by the working party group.
One question to ask is, are there really 151,000 women in England placed in a holding category because cytological examination, by experienced cytopathologists, has failed to provide a precise diagnosis? Yet such reporting patterns are not unique to the UK as other reporting systems such as Bethesda allow for the smear with features of undetermined significance. Perhaps it is the fear of litigation that allows the BNC report to profligate despite the facts that many BNC reports are fantasies in that no abnormalities are found on follow up. On the other had there are the indisputable facts that a small, but significant, number of BNC smears do result in high grade CIN.
Perhaps by more vigorous application of the various diagnostic criteria, coupled with professional judgement of patient management and bold reporting on behalf of the cytopathologist the rate of BNC may be reduced. If not the use of new techniques such as HPV testing as a spin off from LBC may make the true nature and risk of the BNC smear clearer and allow the facts to rise above the fantasy.
1. Evans D M D, E A Hudson, C L Brown, M M Boddington, H E Huges, E F Mackenzie and T Marshall. Terminology in gynaecological cytopathology: report of the Working Party of the British Society for Clinical Cytology. Journal of Clinical Pathology 1986. 39: 933 -944.
2. Solomon D ASCUS and AGUS Criteria. Acta Cytologica 1998 42 (1) 17-23.
3. Cervical Screening Programme, England 2001 -2002. DoH.
4. Buckley C H, A Herbert, E F Mackenzie, J Johnson, P A Smith, L S Turnbull, J H F Smith, V Wadehra and C M Stanbridge. Borderline nuclear changes in cervical smears: guidelines on their recognition and management. Journal of Clinical Pathology 1994. 47: 481 - 492.
5. Achievable standards, benchmarks for reporting and criteria for evaluating cervical cytopathology 2nd edition. 2000. NHSCSP publication no 1.
6. Parham D M, E K Wiredu and K A Hussain. Significance of borderline nuclear abnormality in cervical smears. Cytopathology 1992. 3: 85 - 93.
7. Anderson MC, C L Brown, C H Buckley, H fox, D Jenkins, D G Lowe, B T B Manners, D H Melcher, A J Robertson and M Wells. Current views on cervical intraepithelial neoplasia. Journal of Clinical Pathology 1991. 44: 969 -978.
8. Edwards J M, P J Hermansen, V E Hillier and A J Howat. Borderline nuclear change: can a sub group be identified which is suspicious of high-grade cervical intraepithelial neoplasia, i.e. CIN2 or worse? Cytopathology 2002. 5: 267 - 272.
9. The Bethesda System 2001 Terminology. http://www.bethesda2001.cancer.gov
10. Papanicolaou G N 1963 Atlas of Exfoliative Cytology Harvard University Press p 20 21.
11. BSCC Terminology 2002 http://www.clinicalcytology.co.uk/
12. Heately M K and J P Bury. The correlation between the grade of dyskaryosis on cervical smear, grade of cervical intraepithelial neoplasia (CIN) on punch biopsy and the final histological diagnosis on cone biopsies of the cervix. Cytopathology 2002. 5: 93 - 99.
13. Usman E and R Hammond. Cervical screening: management of patients referred for colposcopy with smear abnormalities less severe than dyskaryosis. Cytopathology 1998. 2: 100 - 107.
14. Heatley M K Follow up of women with borderline cervical smears as defined by national guidelines. Journal of Clinical Pathology 1999. 52: 787 -788.
15. Evans D M D, E A Hudson, C L Brown, M M Boddington, H E Huges, E F Mackenzie. Management of women with abnormal cervical smears: supplement to Terminology in Gynaecological Cytopathology. Journal of Clinical Pathology 1987. 40: 530 -531
17. HPV pilots http://www.cancerscreening.nhs.uk/cervical/research.html#hpv-pilots
18. Papillo J. L, M A Zarak andn T L St John. Evaluation of ThinPrep Pap test in clinical practice. Acta Cytologica 1998 42 (1) 203-208.
19. Moss S M, A Gray, R Legood and E Henstock Evaluation of HPV?LBC Cervical Pilot Studies. First report to the Department of Health on evaluation of LBC (December 2002).