Minimal residual disease (MRD) monitoring has a strong prognostic value in childhood lymphoblastic leukemia (ALL) and is currently utilized in all major pediatric ALL protocols. MRD monitoring is done by multiparameter flow cytometry, IG/TCR quantitative PCR or reverse transcriptase quantitative PCR of leukemic fusion transcripts providing a reliable measurement of treatment response. However, occasionally bone marrow (BM) aspirates may not yield representative material or be misinterpreted due to treatment-induced changes in MRD marker profile, undetected subclones at diagnosis, contamination with peripheral blood or cell adhesion and stroma cell interactions posing a risk for underestimating MRD levels and misclassifying resistant disease that may be detected by traditional BM morphology methods, immunohistochemistry, karyotyping and FISH. We present four cases with high MRD levels where MRD monitoring failed to provide the correct stratification information. Through these cases, we discuss the continued need to consider all available information including BM smears, touch imprints and trephine biopsy preparations not only at diagnosis but throughout remission monitoring in pediatric ALL.