Minimal residual disease in childhood acute lymphoblastic leukaemia quantified by aspirate and trephine: is the disease multifocal?

The level of minimal residual disease (MRD) in marrow early in treatment strongly predicts outcome in childhood acute lymphoblastic leukaemia (ALL). Using PCR we studied 30 pairs of aspirates and trephines taken during induction treatment. Consensus PCR primers showed a monoclonal gene rearrangement in eight pairs, polyclonal rearrangement in 18 pairs and a monoclonal rearrangement only in the trephine in four pairs. MRD was quantified by leukaemia-specific primers in 22 pairs. There was a linear relationship between the logarithms of MRD levels of aspirate and trephine, with a residual variance which increased as the level of MRD fell. The mean level of MRD in the trephines was 4.1-fold greater than that in the aspirates, probably due to greater dilution of the aspirates with peripheral blood. The high variance at low levels of MRD could not be explained by measurement variation, which had an MRD-independent value of 0.42 log10 units, and was attributed to sampling variation due to patchiness of disease at low MRD levels. The magnitude of the variation was such that predictions of outcome could well be confounded for many patients. We suggest that MRD sampling variability could be minimized either by taking multiple marrow samples or by measuring MRD in peripheral blood.