A Novel Approach for Minimal Residual Disease Detection in Childhood T-Cell Lymphoblastic Lymphoma (T-LL): A Children’s Oncology Group Report.

Contemporary treatment of childhood T-LL produces 2 yr EFS of ≥ 85% (Reiter et al., Blood, 2000). To further improve cure rates without increasing treatment-related toxicities it is critically important to better understand the biology of the disease and identify prognostically important subgroups (Cairo et al., PBC, 2005). Approximately 15% of children with T-LL have disseminated disease at diagnosis as demonstrated by morphologic examination of bone marrow aspirates and biopsies. We developed a flow cytometric method that can detect 1 malignant T-cell among 10000 or more normal bone marrow and peripheral blood cells. The method relies on expression of CD3 and TdT, combined with the absence of B-cell and myeloid markers, to identify malignant T cells. We applied this method to study bone marrow samples collected at diagnosis from 70 patients enrolled in the Children’s Oncology Group A5971 study “Randomized Phase III Study for the Treatment of Newly Diagnosed Disseminated Lymphoblastic Lymphoma or Localized Lymphoblastic Lymphoma.” Of the 70 patients, 48 (68.6%) had CD3⁺/TdT⁺ lymphoma cells detectable in the bone marrow (median, 0.25%; range 0.01%–28%), a proportion considerably higher than estimates based on morphology. In 27 patients, we studied paired left and right iliac crest aspirates; in all cases, results were concordant (20 positive and 7 negative). The difference in percentages of CD3⁺/TdT⁺ cells measured among positive sample pairs was <1 log, suggesting homogeneous disease dissemination. Among the 70 patients, relapse has occurred to date in 11 patients, 3 in the primary site only, and 8 in other sites (with or without bone marrow involvement). All the 8 latter relapses occurred among the 32 patients with ≥ 0.1% CD3⁺/TdT⁺ cells at diagnosis in bone marrow; none was observed in the 38 patients with lower levels or no detectable disease (P = 0.001 by Fisher’s exact test). Because of findings in patients with T-cell acute lymphoblastic leukemia indicating that levels of minimal residual disease (MRD) in peripheral blood resemble those in bone marrow (Coustan-Smith et al., Blood, 2002), we hypothesized that blood could also be used for detecting disseminated T-LL. Paired marrow / blood samples from 61 patients were available for study. In 41 patients, similar levels of CD3⁺/TdT⁺ cells were detected in marrow and blood; in the remaining 20, CD3⁺/TdT⁺ cells were undetectable (<0.01%) in marrow, but detectable in 7 of the corresponding blood samples (range, 0.01%–0.09%). These results suggested that blood could be used to monitor early treatment response in patients with T-LL. Hence, Study A5971 was amended to determine the prevalence of MRD in blood during remission induction therapy. To date, 146 samples from 42 patients with T-LL have been studied; tests on days 7 and 28 are emerging as the most informative. Our findings provide new insights about disease dissemination in T-LL, and ways to measure early response to therapy and monitoring MRD in these patients. This may have potential implications in the future for alternative therapeutic strategies in the subset of children with T-LL with persistent MRD during induction therapy.