Using a quality control approach to define an 'adequately cellular' liquid-based cervical cytology specimen

Objective: To define a minimum acceptable total squamous cellularity for (ThinPrep((R))) liquid-based cervical cytology (LBC) specimens using quality control techniques. Methods: Two hundred LBC preparations were made containing varying numbers (< 200) of severely dyskaryotic squamous cells and with varying total cellularities. Results: Ninety-eight per cent of the LBC preparations that were missed by one or more of three cytoscreeners had fewer than 16 abnormal objects (single dyskaryotic cells or clumps of cells) and 87 dyskaryotic cells. The minimum ratio of dyskaryotic to total squamous cells that, in a preparation of 5000 squamous cells has a probability of at least 0.98 that 87 or more dyskaryotic cells will be present is 1 : 47. Twenty-three preparations diagnosed as abnormal had ratios of dyskaryotic to total squamous cells of between 1 : 2.5 and 1 : 4596. There is thus no feasible minimum acceptable squamous cellularity that will give an acceptable probability of detection of all specimen vials containing abnormal cells in the observed proportions. Conclusions: It is suggested that the minimum acceptable cellularity for LBC specimens is set pragmatically by the screening programme to give a feasible percentage of repeat tests.