The immunologic detection of minimal residual disease in acute leukemia.

Certain combinations of differentiation antigens are expressed on leukemia blasts and are absent or extremely rare among normal progenitors in the fetal liver and fetal and regenerating bone marrow. These combinations include cCD3/TdT, a thymic feature retained on thymic-acute lymphoblastic leukemia (T-ALL) blasts outside the thymus, and the coexpression of TdT and myeloid markers (CD13, CD33) on a proportion of ALL and acute myeloid leukemia (AML). Thus, double marker immunofluorescence assays are operationally leukemia-specific and can be applied in 35% of acute leukemias for detecting minimal disease at a less than 10(-4) level; only rare cases, 2 of 35 in our study, switch these relevant features during relapse. The sensitivity and specificity of these assays was tested as follows. First, bone marrow samples taken from patients who had originally presented with blasts expressing the leukemia-associated combinations but were in full morphologic remission were studied, and varying numbers (less than 0.01% to 10% of the mononuclear fraction) of cells with aberrant features were identified in 11.6% of the cases. Second, the outcome of 19 patients with minimal disease identified immunologically while in complete morphologic remission was investigated: all 19 patients have developed systemic relapse within 4 to 25 (median 14.5) weeks. In contrast, 17 of 25 patients also morphologically in complete remission and without residual disease identifiable immunologically after repeated testing are still in morphologic and immunologic remission (follow-up 17 to 114 weeks, median 28 weeks). Only eight patients in this group have relapsed so far: in two patients the relapse was localized in the cerebrospinal fluid, while in six patients a systemic relapse was observed 6 to 51 (median 21.5) weeks after the last negative immunologic bone marrow examination. In conclusion, no false-positive results were detected with these sensitive assays, and the introduction of appropriately planned prospective studies, including the immunologic detection of residual leukemia, is justified on the basis of these observations.