Endocervical glandular involvement is associated with an increased detection rate of high-grade squamous intraepithelial lesions on the Papanicolaou test

Although The Bethesda System for Reporting Cervical Cytopathology does not mandate reporting of endocervical glandular involvement (EGI) in Papanicolaou test specimens with high-grade squamous intraepithelial lesions (HSIL), several studies have suggested that EGI diagnosed on surgical specimens is associated with higher rates of residual or recurrent dysplasia. When suspected, EGI is reported for Papanicolaou test specimens at our institution, but the performance of this diagnosis has not been assessed. The archives were queried for Papanicolaou test specimens with a diagnosis of HSIL-EGI (2006-2017). All follow-up surgical pathology specimens within a year of the Papanicolaou test diagnosis were evaluated for cytologic–histologic correlation. This same query was repeated for all surgical pathology specimens with a diagnosis of HSIL-EGI. All preceding Papanicolaou test diagnoses within a year were assessed for cytologic–histologic correlation. Twenty Papanicolaou test specimen glass slides were reviewed by 6 observers to assess for interobserver variability. Patients with HSIL-EGI on surgical specimens were more likely to have a preceding Papanicolaou diagnosis of HSIL and atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesion (ASC-H) (32.3% versus 25.5%, P = 0.03, and 16.7% versus 11.8%, P = 0.04, respectively). Patients with an HSIL-EGI diagnosis on a Papanicolaou test were significantly more likely to have HSIL-EGI detected on a follow-up histology (41.6% versus 24.0%, P < 0.001). Interobserver concordance was poor for the assignment of EGI in Papanicolaou test specimens. Overall, the diagnosis of HSIL-EGI on Papanicolaou test specimens is complicated by poor sensitivity and interobserver concordance. •The authors conducted a retrospective review of histologic and Papanicolaou test specimens in which a diagnosis of high-grade squamous intraepithelial lesions with endocervical gland involvement (HSIL-EGI) was made.•Compared with a control cohort, the presence of HSIL-EGI on histology was more likely to be preceded by a diagnosis of HSIL or atypical squamous cells, cannot exclude high-grade squamous intraepithelial lesion (ASC-H).•When made on a Papanicolaou test specimen, an HSIL-EGI diagnosis was associated with a significantly increased rate of HSIL-EGI on follow up histology with a modest positive predictive value (especially on larger surgical specimens), but overall was found to have poor sensitivity and interobserver co