OBJECTIVETo investigate the comparability of bcr-abl (P210) transcript levels detected in different hospitals.

METHODSTen hospitals in China took part in the four times of sample exchange and comparisons from April, 2010 to August, 2011. The exchange samples were prepared by Peking University People's Hospital. Firstly, the BCR-ABL (P210)(+) cells from a newly diagnosed chronic myeloid leukemia patient were 10-fold serially diluted by BCR-ABL (P210)(-) cells and they covered 4 magnitudes. Then, TRIzol reagents were thoroughly mixed with cells in each tube. Every 12 samples (three samples per magnitude) were sent to the other 9 hospitals. The cell number of each sample was 8×10(6). The detection of bcr-abl transcript levels by real-time quantitative PCR were performed in every hospital according to their own protocols. Conversion factors (CF) were calculated using regression equation.

RESULTSDifferences in bcr-abl transcript levels did exist among results of 10 hospitals in each comparison. In general, the results of the most of hospitals were in line with the dilutions of cells. CF of every hospital fluctuated. Three hospitals had relatively stable CF, and their ranges were 2.8 - 5.2, 1.2 - 2.8 and 2.2 - 6.8, respectively; two hospitals had unstable CF with ranges 0.76 - 7.0 and 2.1 - 18.7; three hospitals couldn't be calculated CF one or two times because of the significant deviation of the results from the actually bcr-abl transcript levels, and their ranges of CF which could be calculated were 1.9 - 19.2, 3.6 - 7.6 and 0.18 - 14.7; One hospital only had two CF (3.3 and 5.0) because of the replacement of an important reagent during the period of comparisons.

CONCLUSIONSComparability of bcr-abl (P210) transcript levels between different hospitals could be achieved through CF which acquired by sample exchange and comparison. The stable and reliable detection system is the premise to acquire correct CF.