Detection of abnormal cells in white cell differentials: comparison of the HEMATRAK automated system with manual methods.

Automated differential systems can rapidly count larger numbers of cells compared with the standard manual procedure. When a fixed number of abnormal cells are interspersed randomly with a large number of normal cells, it can be shown mathematically that counting more cells increases the chances of detecting at least one abnormal cell. To test this hypothesis in a clinical setting, the authors compared 200-cell and 400-cell automated differentials obtained via the HEMATRAK Model 360 system with results of 100-cell differentials performed either manually or automatically for a group of 141 blood smears. Manual 100-cell differentials also were performed in a reference laboratory for comparison. In close agreement with theoretical expectation, both 200-cell and 400-cell differentials detected significantly more abnormal cells than did either the manual or automated 100-cell differential. Results of the latter two were not significantly different. Eighty-seven per cent of the slides that, according to the 100-cell manual differential, were without abnormal cells were found to have such cells on the 400-cell automated differential. Atypical lymphocytes and nucleated red blood cells were the abnormal cells most frequently identified.