Mitotic count in breast cancer. The influence of training on performance.

Two medical laboratory histotechnologists were trained on mitotic counting for 6 months and 1 year. The training included detailed theoretical education on mitotic counts and microscopy sessions with a motivated pathologist. Most of the training time, however, was used for mitotic counts from a large collection of breast cancer samples. The progress in performance was followed by allowing the technicians to count 20 breast cancer samples several times during the training. The mean value of the mitotic counts of technician 1 varied dramatically during the early phases of training but then stabilized near the levels of technician 2, whose mean values were more constant throughout the training period. Both observers showed changes in the standard deviation and coefficient of variation in the early phases of training but later stabilized at the same level. The proportion of the counting results of the technicians within the 95% confidence limits determined by the performance of already trained, experienced technicians varied considerably at first, but then stabilized at values 0.7 and above. The study showed that training improved the performance in mitotic counting. We suggest that the training of mitotic counts provides the means to achieve economical and reproducible estimates of tumor cell proliferation.