Peripheral Blood CD34 Donor Chimerism has Greater Clinical Utility Than CD3 for Detecting Relapse after Allogeneic Stem Cell Transplantation for Acute Myeloid Leukemia or Myelodysplastic Syndrome

Monitoring of donor chimerism (DC) may detect early relapse following allogeneic hematopoietic stem cell transplantation (allo-SCT) for acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS). Most centers use unfractionated peripheral blood or T-cells to monitor DC, although CD34⁺ DC may be more predictive. The limited adoption of CD34⁺ DC may be due to the lack of detailed, comparative studies. To address this knowledge gap, we compared peripheral blood CD34⁺ and CD3⁺ DC in 134 patients who underwent allo-SCT for AML or MDS. In July 2011, the Alfred Hospital Bone Marrow Transplantation Service adopted routine monitoring of DC in the lineage-specific CD34⁺ and CD3⁺ cell subsets from peripheral blood at 1, 2, 3, 4, 6, 9, and 12 months post-transplantation for AML or MDS. Immunologic interventions, including rapid withdrawal of immunosuppression, azacitidine, and donor lymphocyte infusion, were prespecified for CD34⁺ DC ≤80%. Overall, CD34⁺ DC ≤80% detected 32 of 40 relapses (positive predictive value [PPV], 68%; negative predictive value [NPV], 91%), compared with 13 of 40 relapses for CD3⁺ DC ≤80% (PPV, 52%; NPV, 75%). Receiver operating characteristic analysis showed the superiority of CD34⁺ DC, with the greatest value at day 120 post-transplantation. CD3⁺ DC provided additional value in only 3 cases, preceding CD34⁺ DC ≤80% by 1 month. We further show that the CD34⁺ DC sample can be used to detect NPM1^mut, with the combination of CD34⁺ DC ≤80% and NPM1^mut identifying the highest risk of relapse. Among the 24 patients in morphologic remission at the time of CD34⁺ DC ≤80%, 15 (62.5%) responded to immunologic interventions (rapid withdrawal of immunosuppression, azacitidine, or donor lymphocyte infusion) with recovery of CD34⁺ DC >80%, and 11 of these patients remained in complete remission for a median of 34 months (range, 28 to 97 months). In contrast, the other 9 patients did not respond to the clinical intervention and relapsed within a median of 59 days after detecting CD34⁺ DC ≤80%. The CD34⁺ DC was significantly higher in responders than in nonresponders (median, 72% versus 56%; P = .015, Mann-Whitney U test). Overall, monitoring of CD34⁺ DC was considered clinically useful (early diagnosis of relapse enabling preemptive therapy or predicting low risk of relapse) in 107 of 125 evaluable patients (86%). Our findings show that peripheral blood CD34⁺ DC is feasible and superior to CD3⁺ DC for predicting relapse. It also provides a source of DNA for measurable residual disease testing, which may further stratify the risk of relapse. If validated by an independent cohort, our results suggest that CD34⁺ should be used in preference to CD3⁺ DC for detecting early relapse and guiding immunologic interventions following allo-SCT for AML or MDS.