The abnormal glandular smear: cytologic prediction, colposcopic correlation and clinical management

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.

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].