An important issue with regard to the provision of reliable results suitable for making treatment decisions is the adherence to rigorous internal quality control (QC) and participation in external quality assurance (EQA) programs. Such programs have been long established for morphological assessment, immunophenotyping, and cytogenetic analysis. However, they remain in their infancy for molecular detection of fusion genes. Since test methods are based on RT-PCR, it is important that QC schemes evaluate all steps of the procedure, particularly as the RT step is a considerable source of variability. For acute leukemias, distribution of primary patient material is unrealistic, whereas distribution of cell lines for QC purposes suffers from the problem that the heterogeneity encountered in breakpoint patterns for patients with PML-RARA and CBFB-MYH11 gene fusions, for example, is not represented among the available cell lines. Nevertheless, a number of European initiatives including the BIOMED-1, Europe Against Cancer (EAC) and European LeukemiaNet programs have been launched over the last few years to improve standardization of methods for the detection of leukemia-associated fusion transcripts by conventional and real-time RT-PCR.28,29 Such initiatives are important to ensure reliability of diagnostic molecular screening and are absolutely fundamental if MRD data are to be used to predict outcome and modify treatment approach.
The EAC QC exercises have highlighted variations in performance between laboratories that are ostensibly carrying out RQ-PCR assays according to a standardized protocol. This reinforces the importance of developing and adhering to standard operating procedures for performance of RQ-PCR tests and generating clear guidelines for reporting of RQ-PCR results. In particular, each assay should be performed in triplicate, so that the degree of reproducibility of individual results may be readily appreciated and hence "outlying results" disgarded. The EAC program revealed that in some instances, fusion gene assays yield an amplification signal in one of the triplicate wells of negative control samples. Since in many instances the participating laboratories had not previously amplified the transcript in question, it is likely that such results reflect nonspecific amplification or cDNA carryover between adjacent wells. This is an intrinsic shortcoming of platforms using the 96-well plate format and has led us to report MRD analyses in which amplification is restricted to one of three wells as PCR negative (but recommending earlier repeat testing, should the result reflect possible residual disease at the limits of assay detection). To report assays as positive, we require specific amplification in at least two of three wells, with Ct values <40 (i.e., equivalent to at least one plasmid copy).29 Since PCR amplification of primary patient samples is less efficient than plasmid controls, reproducible amplification at Ct values between 40 and 45 can occasionally be observed in follow-up patient samples, which most likely indicates low-level MRD; such results are of uncertain clinical significance and again would be seen as an indication for earlier repeat MRD assessment.
Was this article helpful?