A clinical review of borderline glandular cells on cervical cytology

Cervical glandular neoplasia referrals and the diagnosis of adenocarcinoma in situ: Correlating cytology, colposcopy findings, and clinical outcomes

OBJECTIVES

To determine the clinical outcomes of women with possible glandular neoplasia of endocervical type on cervical cytology, and review all diagnoses of cervical adenocarcinoma in situ (AIS) over a 5 year period at our institution.

STUDY DESIGN

A retrospective case-note review was conducted of all women referred to colposcopy with possible glandular neoplasia of endocervical type on cervical cytology or diagnosed with cervical AIS after biopsy or excision, from January 2014 until December 2018 in a London district hospital.

RESULTS

Of 55 women referred with possible glandular neoplasia of endocervical type, 47 (85.4%) had a significant pathology on histopathological analysis: AIS (n = 22); invasive cancer (n = 7); high-grade cervical intraepithelial neoplasia (n = 18). Women with a history of borderline abnormality on cervical cytology within the last 5 years were significantly more likely to be diagnosed with AIS or invasive cancer (P < .05). For the same period 49 women had histologically proven AIS. Among these 22 (44.8%) were referred as possible cervical glandular intraepithelial neoplasia. Other reasons for referral were the following indications: borderline dyskaryosis (n = 13); high-grade dyskaryosis (n = 8); low-grade dyskaryosis (n = 4); postcoital bleeding (n = 2).

CONCLUSION

Due to the raised risk of significant gynaecological pathology in women with possible glandular neoplasia of endocervical type on cervical cytology, excisional biopsy is essential. Colposcopic impression varies significantly and complete excision of the abnormal lesions should be achieved. AIS is a histological diagnosis and should always be considered during colposcopical and cytopathological assessment.

An audit of liquid-based cytology samples reported as high-risk human papillomavirus and borderline nuclear change in endocervical cells

INTRODUCTION

Primary human papillomavirus (HPV) screening, testing for the virus responsible for 99% of cervical cancers, was introduced in 2018-2020 in the UK. This was preceded by HPV triage of low-grade cytology from 2012. Much of the evidence incorporated into current National Health Service (NHS) colposcopy guidance assessed outcomes prior to this change in screening. The aim of this paper is to assess adherence to NHS cervical screening programme standards, determine the incidence of cases reported as high-risk HPV plus borderline nuclear change in endocervical cells, to calculate colposcopic accuracy and assess histological outcomes in this cohort.

METHOD

A retrospective audit of women referred to a colposcopy clinic in one NHS trust from 2016 to 2018. Data relating to histological outcomes, cytological follow-up and demographics were collected.

RESULTS

Of 2001 referrals, 22 data sets identifying HPV-positive borderline endocervical change were eligible for analysis (1.2% incidence). Median age was 29.5. Two-thirds (68.2%, n = 15) had high-grade dysplasia at diagnostic biopsy. Those women with reassuring histology had normal cytological follow-up. Colposcopic accuracy was moderate (positive predictive value 43.8%, negative predictive value 100%).

CONCLUSIONS

Borderline nuclear change in endocervical cells is an uncommon condition but should be treated as a high-grade referral. All women should be offered a diagnostic biopsy at the initial colposcopy; if no histopathological abnormality is identified, alternative sources of pathology should be considered. Excisional treatment should be recommended to unreliable attenders, those with a complete family and inadequate colposcopy (TZ3) and considered in younger women with a TZ3.

Management of borderline change in endocervical cells: a more dependable approach

Background:

There are limited data and guidance from the UK on borderline nuclear change in endocervical cells (BNCs). The objective of this study is to determine the clinical outcome of women with BNCs, to determine the accuracy of colposcopy and propose a more robust management algorithm.

Methods:

This is a retrospective review of all BNC referrals between January 2006 and December 2011 at the Northumbria Healthcare Trust. Histological diagnosis was based on high-grade histology (CIN 2 or worse). Any high-grade histology in the first year of follow-up was included in the final diagnosis.

Results:

Of the 9001 new referrals, 167 women had BNCs. Thirty-seven (22%) were diagnosed with high-grade histology on initial assessment. Sixty women had satisfactory and negative colposcopy, out of which 7 (12%) were detected with high-grade histology/cytology in the first year of follow-up. Overall, 50 high-grade histology (30%), including two invasive carcinomas were detected.

Conclusions:

Current follow-up of BNCs relies heavily on colposcopic assessment. A significant proportion of women with negative colposcopy was found to have high-grade histology in the first year of follow-up. We propose a more robust management algorithm to lower the probability of missed high-grade histology in this subgroup of women.

Analysis of outcome in women with borderline glandular change on cervical cytology

OBJECTIVE

To review the outcome of women presenting with borderline glandular smear on cervical cytology and to investigate correlation between cytology, colposcopy and histology and subsequent smear history.

STUDY DESIGN

A 5 year retrospective analysis of outcome in all women reported with borderline glandular changes on cytology between 2001 and 2005 at the John Radcliffe Hospital, Oxford, and Cheltenham General Hospital, Gloucestershire, was performed. Cytology, colposcopy, histology and follow-up cytology data at 6, 12 and 24 months after index smear were collated.

RESULTS

Of 92 women with borderline glandular smears, 56 were referred for colposcopy. This accounts for 0.9% of the total referral (6293) to the colposcopy units. A significant rate of abnormal histology was noted, with CIN 1, 2, 3, CGIN or worse in 20 women (36%) and benign pathology in 18 women (32%). Colposcopy had a sensitivity of 82%, specificity of 39%, negative predictive value of 80% and positive predictive value of 49% in predicting abnormal (premalignant/malignant) histology. Only one woman with normal histology at presentation had subsequent abnormal cytology (1/31), whereas abnormal histology at presentation was strongly associated with subsequent abnormality on follow-up cytology over 24 months (8/21) with a p value=0.0058.

CONCLUSION

Colposcopic examination and biopsy of colposcopically identified abnormalities is reliable, with negative colposcopy having a high negative predictive value. We also recommend thorough colposcopic examination before any invasive investigation as over treatment may impact on the future reproductive outcome [1].

Risk of invasive cervical cancer in relation to management of abnormal Pap smear results

OBJECTIVE

We sought to evaluate the management of women with abnormal cytology in terms of subsequent risk of invasive cervical cancer.

STUDY DESIGN

The screening histories of all invasive cervical cancer cases diagnosed in Sweden 1999-2001 and of 5 population-based controls per case were reviewed. In all, 159 patients and 258 control subjects aged < 67 years had an abnormal smear result 0.5-6.5 years prior to cancer diagnosis. The cervical cancer risk was estimated in relation to management by calculating odds ratios.

RESULTS

Histologic assessment of low-grade squamous abnormalities strongly reduced the risk compared to repeated cytology (odds ratio, 0.46; 95% confidence interval, 0.24-0.89). Delaying histologic assessment was also associated with a higher risk (odds ratio, 5.65; 95% confidence interval, 1.39-23.05). After high-grade squamous atypia, absence of any cytologic or histologic specimen was a major determinant of cancer risk (odds ratio, 12.52; 95% confidence interval, 1.42-infinitive).

CONCLUSION

For adequate protection against invasive cervical cancer, further assessment with histology must be recommended also for women with low-grade squamous abnormalities.

Glandular prediction: the pride and the prejudice

For the cytologist and clinician alike, glandular lesions pose possibly the greatest challenge in cervical screening. Worldwide, with increasing confidence in cytological prediction, terminology is evolving. In the UK, with the adoption of liquid based methods, the technical aspects of cervical cytology are being addressed, it is now time to standardise our terminology in glandular reporting. Consideration of the cytological complexity, clinical needs and international protocols is essential in this endeavour.

A retrospective clinical audit of cervical smears reported as 'glandular neoplasia'

The aims of this study were to review the diagnostic pathway of women with smears reported as 'glandular neoplasia' and to outline the management, colposcopy findings, treatment and final histological diagnosis in these women. The design was a retrospective review. A total of 114 women were identified over a 5-year period from the cytology database at the Royal Liverpool University Hospital Cytology Department, whose hospital case notes were available for review. Methods included a review of the case notes for the demographic details, indication for smear, colposcopic findings, investigation and/or treatment procedures, histology, final diagnosis and current disease status. Of 114 smears reported as 'glandular neoplasia', 67 were reported as consistent with cervical glandular intra-epithelial neoplasia (CGIN), six with endocervical adenocarcinoma, 36 with endometrial adenocarcinoma and five with other glandular neoplastic abnormalities. The average age was 46.5 years. 79 (69.3%) smears were routine call/recall and 36 (30.7%) women were symptomatic. The positive predictive value (PPV) for a significant histological abnormality in the CGIN smear group was 80.6% (23.9% invasive carcinomas, 43.3% CGIN and 13.4% CIN) and the PPV of an 'endometrial adenocarcinoma' smear was 86.1%. Smears indicating glandular neoplasia are associated with a high probability of clinically significant lesions, the PPV of a CGIN smear being over 80%. Immediate referral for colposcopy and assessment by an experienced colposcopist is recommended.

An audit of cervical smears reported to contain atypical glandular cells

The terminology and classification system for the reporting of cervical smears is currently under review. However, evidence on which to base the reporting of atypical cervical glandular cells is lacking. This audit of glandular atypia reporting in cervical smears provides evidence to support a three-tier sub-classification system for cervical glandular atypia. The three sub-classifications are distinct with respect to their positive predictive values and to the cell type of the abnormality as confirmed by histology and therefore relate to further clinical investigation.

Cytology suggestive of glandular neoplasia: outcomes and suggested management

Eighty-three cases having a cervical smear result showing abnormal glandular cells were identified and matched up with the diagnostic histology result. Thirty-four (41.0%) were associated with malignancy and 26 (31.3%) with a cervical intraepithelial lesion without invasion. Thirty-eight (45.8%) had conditions of the cervix as follows: 12 cases had invasive disease of the cervix; nine (10.8%) adenocarcinoma of cervix and three (3.6%) squamous carcinoma of cervix. Nineteen (22.9%) had cervical intraepithelial neoplasia (CIN/SIL) alone and seven (8.4%) had cervical glandular intraepithelial neoplasia (CGIN) +/- CIN. There were 16 (19.3%) cases with malignancies of the uterine corpus and six (7.2%) had a malignancy arising from another primary site. Twenty-three (27.7%) had no malignant or pre-malignant condition. The risk of malignancy was related to age and ranged from 18.2% in those under 35 years to 67.9% in those 55 years and over. A protocol for the management of these cases is described.