Abstract • 180

The study of minimal residual disease has attracted much attention in clinical oncology for its potentials of tailoring treatment and gaining insights into the nature of cure. Several methods have been proposed to detect MDR in leukemias and have been extensively reviewed. In ALL rearrangements where Ig and TCR genes result in unique recombinations of variable (V), diversity (D), and joining segments, the junctional regions between these gene segments can be regarded as "fingerprint-like" sequences due to the deletion and random insertion of nucleotides during the rearrangement process. PCR-based MRD detection via clonespecific junctional regions generally reaches sensitivity of 10/-4 to 10/-5. Several retrospective studies indicated that the detection of MRD in childhood ALL has prognostic value, although the results of these studies are not fully concordant. More recently three prospective studies have demonstrated that molecular or immunological detection of MRD during the first three months gives relevant prognostic information about the in vivo effectiveness of the treatment. Although technically feasible, several questions are still open and will be discussed: 1. How to use MRD data for future clinical studies in childhood ALL? 2. Which patients could potentially benefit? 3. Do we have clinical options: i.e. intensification of treatment in patients with high risk MRD and decrease in those with very low levels? 4. Does MRD differ according to ALL phenotype? 5. Are studies feasible, cost-effective, and evaluable?