A METHOD FOR DETECTING VARIABILITY ARISING FROM ERRORS IN SAMPLE PROCESSING OF PARAFFIN-EMBEDDED TISSUE FOR DNA CONTENT-ANALYSIS

We present a method for controlling variability that may arise from in consistencies in sample preparation for DNA content analysis of paraff in-embedded tissue. Human tonsil tissue obtained from routine surgical specimens was embedded in paraffin according to standard protocols. F ifty-micrometer sections were cut from the block and analyzed each day for 20 days to establish control ranges. One tonsil tissue section wa s processed in parallel with each run of clinical specimens. In this c ontext, a run was defined as the simultaneous processing of 50-mum tis sue sections for extraction of cell nuclei (dewaxing and rehydrating). If the tonsil G0/G1 peak coefficient of variation (CV) exceeded 2 SDs of the established mean, and optimum instrument performance and stain ing were verified, all samples prepared with the tonsil control were r eprocessed. Instrument performance and staining were assessed by using the appropriate external controls. By using this rejection rule (1(2S )), the frequency of sample reprocessing in our laboratory was approxi mately 6%. When the run was repeated and the tonsil control CV was wit hin acceptable range, the G0/G1 peak CV of the corresponding clinical specimens improved 25% of the time. Because most investigators are wil ling to accept higher CVs for paraffin-embedded tissue than for fresh tissue, it is desirable to have a control to detect decreased peak res olution, resulting from errors in sample processing.