Accurate assessment of cell density in low cellular liquid-based cervical cytology

Scant cellularity is the most important source of unsatisfactory liquid-based cytology. Although still being debated, low cellularity is thought to compromise the detection of squamous lesions. Thus, reliable assessment of cellularity is essential. The aim of the present study was to determine the cellularity range for ThinPrep(®) slides of low cellularity and to establish the most accurate cell-counting protocol.

A series of 60 ThinPrep cases representing the full spectrum of adequate, 'satisfactory but limited by' (SBLB) and unsatisfactory reports were included. Two cell-counting protocols with three different magnifications, using ×10, ×20 and ×40 objectives, were evaluated and related to the true cellularity, together with a reassessment of the degree of adequacy originally reported. The cell-counting protocol that showed the highest correlation coefficient was considered the most accurate.

Based on seven (re)assessments a majority score for adequacy was established. There were 42 cases with a majority score 'unsatisfactory' or 'SBLB' (low cellularity) of which 41 contained fewer than 20 000 squamous cells; and 18 cases with a majority score 'satisfactory' of which one had fewer than 20 000 cells. The cell-counting protocol that showed the significantly highest correlation with the reference standard was the Stichting Kwaliteitsbewaking Medische Laboratoriumdiagnostiek (SKML) protocol with a ×10 objective.

ThinPrep slides reported as unsatisfactory or SBLB were shown to contain fewer than 20000 squamous cells. The most accurate protocol for estimating the cellularity of these slides was cell counting in five non-adjacent microscope fields along the horizontal axis and five along the vertical axis of the slide with a ×10 objective and applying a correction factor of 1.24× to correct for underestimation of the true cellularity.