First Results from a Collaborative Initiative To Develop an International Scale for the Measurement of BCR-ABL by RQ-PCR Based on Deriving Laboratory-Specific Conversion Factors.

RQ-PCR provides an appropriate method to monitor CML. However methods are not standardized leading to differences in reported BCR-ABL values. An International Scale (IS) was proposed to generate comparable values when tested in any laboratory, Blood,2006,108,28. The scale is fixed to a major molecular response (MMR), a value with established prognostic significance. We distributed BCR-ABL reference standards to 12 laboratories (labs) to establish lab specific conversion factors (CF) for the IS. The IS MMR value is 0.1%. The Adelaide reference laboratory (ref lab) has an established MMR value (0.08%) and the CF is 1.25 (0.1/0.08). Multiplying by 1.25 converts the ref lab values to the IS. The ref lab prepared standards with 3 BCR-ABL levels (L1, L2, L3) by diluting K562 cells in cells of normal volunteers. Trizol stabilized cells were sent frozen to labs in Europe (2), USA (3), Asia (5) and Australia (2) that use various methods and controls (ABL 7, BCR 3, GUS 1, ABL and GUS 1). Each lab determined mean BCR-ABL/control% values for each level, which were correlated with those of the ref lab. CF were calculated from regression equations using the formulae: Log y=(slope×(log 0.1))+intercept, CF=0.1/antilog y, where y=lab equivalent MMR value and 0.1=IS MMR. 11 of 12 labs showed a linear relationship across the standards and were included in the analysis. Prior to IS conversion there was a wide range between the lowest and highest values; L1 22-fold (r0.01–0.31); L2 48-fold (r0.06–3.0); L3 10-fold (r10–101). After IS conversion using lab specific CF the agreement was substantially improved; L1 1.8-fold (r0.07–0.13), L2 2.4-fold (r0.33–0.77), L3 3.9-fold (r19–76). CF have been validated for 5 labs by patient sample exchange. The ref lab tested up to 20 samples per lab and values compared before and after conversion. Conversion failed to align data in 1 lab and the reason is not yet established. In the other 4 labs there was a significant difference in the median values before conversion (p=0.01) but not after (p=0.49). 75% were within 2-fold and 98% within 5-fold after conversion, a significant improvement compared to pre conversion (45% p=0.005, 72% p=0.0001). The preliminary data suggest the approach is a reasonable process to achieve standardization, which is anchored to a critical BCR-ABL value. It allows for differences generated by various RQ-PCR methods and controls. The on-going validity of conversion is reliant on maintaining performance of analysis within a lab. We anticipate the project will lead to preparation of certified reference standards available to all labs and that standardized molecular monitoring will facilitate more consistent adherence to treatment guidelines.