Commentary: Liquid automation refreshes Dr Papanicolaou

That the Papanicolaou test has been modified so little during the greater part of its existence is surely a tribute to the insight and inventiveness of the physician scientist from whom it takes its name. When change began to be mooted in the 1980s it was because new levels of computing power were making it realistic to think of automating the microscopic examination of cells from the cervix. But to do so reliably depended on first improving the quality of the smears. This prompted the development of liquid based cytology—and with it a need for realistic comparative assessments: of the performance of the new technique as against conventional methods,1 and of automated as against manual microscopy for examining cervical cell preparations.2

Various liquid based cytology systems are now commercially available. In methods and equipment their details differ from one manufacturer to another, but all dispense with a core feature of the conventional cervical smear test. Instead of smearing the cervical material on to a microscope slide, the tester rinses the brush-like sample collection device in a small pot of preservative fluid. At the cytology lab the floating cells are dispersed and variously centrifuged or filtered to remove blood, mucus, and other debris, and then allowed to form a monolayer on a microscope slide.

In principle this approach confers several advantages. Virtually all the cells collected from the cervix should be present in the suspension, allowing the final sample to be more representative of the original population. The cells themselves, besides being more evenly distributed across the slide, should also be better preserved. Consequently, the proportion of specimens rejected as unsuitable for examination should fall, allowing the productivity of cytology labs to rise. Laboratory staff taking part in a UK Department of Health evaluation study generally approved the system.3

Moving on to automation, here too one of the goals is an increase in productivity—coupled, it has been hoped, with improved accuracy. The ThinPrep imaging system used by Davey et al2 is intended not to replace human judgment, but to facilitate it by drawing the attention of the microscopist to the cells of most relevance.4 It uses a stain that gives an appearance similar to that used in the conventional smear test, but in which the intensity of colouration is closely correlated with the DNA content of the nucleus. The underpinning biological principle is that “abnormal cells have abnormal DNA content.”5

The automatic analyser works its way through a cartridge of slides, scanning each for the largest and darkest cells. It records the locations of 20 that are likely to be of interest to the cytologist. It also identifies two more locations based on the cells' clustering characteristics. When the operator comes to examine the slide, the motorised stage of the microscope moves these sites successively into the field of view. If the operator judges any cells to be abnormal, the entire slide is then scanned. The location of suspicious cells can be electronically marked before the slide is passed to the cytologist for further examination. The system aims to focus human attention and expertise where it is most needed.