Sensitivity and Specificity of Chimerism Tests in Predicting Leukemia Relapse Using Increasing Mixed Chimerism

Abstract

Chimerism test was evaluated to predict leukemia relapse. Increasing mixed chimerism (IMC), defined as recipient increase ≥0.1% in peripheral blood total cell chimerism, was used as a surrogate of disease activity. Combination of quantitative PCR and short-tandem repeat method was applied to achieve high assay sensitivity. Total of 184 patients received stem cell transplant for acute myeloid leukemia (N = 110), acute lymphocytic leukemia (N = 41), myelodysplastic syndrome (N = 30), and 2389 chimerism tests (median follow-up, 1054 days). Sixty-six patients relapsed, and 118 patients did not. Cumulative incidence of relapse increased after 1 IMC or ≥2 consecutive IMCs (hazard ratios, 9.9 and 44.4, respectively). Predicted percentage relapse by day 30 after IMC was 0% (0 IMC), 10% (1 IMC), and 40% (≥2 IMCs). The last chimerism results before relapse detected IMC in 57 of 66 relapsed patients (sensitivity, 86.4%). Nine patients had no IMC before relapse (false negative) because of rapidly evolving relapse (N = 4) or extramural relapse (N = 5). In 118 patients without relapse, 158 of 1873 tests detected IMC (false positive, 8.4%; specificity, 91.6%). The false-positive rates increased with higher percentage recipient T-cell chimerism levels, indicating T-cell contamination as a cause. Chimerism monitoring predicts relapse. However, caution must be taken for false-positive or false-negative IMCs. T-cell removal can improve chimerism test specificity in patients with mixed T-cell chimerism.