Atypical Glandular Cells: Interobserver Variability according to Clinical Management

Area of Science:

• Cytopathology
• Gynecologic Oncology
• Diagnostic Accuracy

Background:

• The 2014 Bethesda System provides criteria for classifying atypical glandular cells (AGC) in cervical cytology.
• Limited data exists on the reproducibility and interobserver variability of these AGC criteria.
• Understanding diagnostic variability is crucial for guideline-driven patient management.

Purpose of the Study:

• To assess interobserver variability in the application of the 2014 AGC diagnostic criteria.
• To evaluate how diagnostic categories of AGC correlate with guideline-specified patient management.
• To identify potential impacts of diagnostic uncertainty on clinical practice.

Main Methods:

• Three cytopathologists re-evaluated 51 Papanicolaou tests previously diagnosed as AGC.
• Diagnoses were categorized based on guideline-specified management and lesion type (glandular vs. squamous).
• Interobserver agreement was quantified using the κ statistic.

Main Results:

• Interobserver agreement for AGC, when categorized by management guidelines, showed substantial variability (κ = 0.009–0.530).
• Half of the observer pairings met criteria for fair to moderate agreement.
• Categorization of lesions as glandular, squamous, or mixed revealed poor agreement (κ = 0.015–0.250).

Conclusions:

• Significant interobserver variability exists in the diagnosis of AGC.
• Fair to moderate agreement was observed when AGC cases were grouped by clinical management, indicating diagnostic variability affects patient care.
• Cytologists and clinicians must be aware of this diagnostic uncertainty.