Introduction:

Relapse remains the leading cause of treatment failure after allogeneic stem cell transplantation (SCT) for acute myeloid leukemia (AML). Early intervention triggered by measurable residual disease (MRD) is a key strategy for improving outcomes. Detection of leukemia specific mutations is ideal but validated assays with sufficient sensitivity are currently limited to less than 50% of transplants. Donor chimerism (DC) and bone marrow flow cytometry have broader applicability but few studies have compared their test characteristics. Thus, we aimed to evaluate and compare peripheral blood CD34+ DC and bone marrow flow cytometry MRD for predicting early relapse of AML after SCT. The working hypothesis was that combining both methods would enhance prediction of relapse, which would facilitate early post-intervention strategies.

Methods:

This retrospective study included 171 AML patients undergoing SCT at Alfred Health from 2015 to 2024. CD34+ DC was routinely measured in peripheral blood between days +30 and +42, and on days +90, +120, +180, +270 and +365. MRD was assessed by 10-color flow cytometry on bone marrow at day+42 and +90 using a combination of leukemia associated immunophenotype and different from normal approach utilizing 3 tubes across both myeloid and aberrant lineage markers. Relapse prediction performance metrics - sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) - were calculated for CD34+ DC <80% donor (mixed chimerism) and flow MRD positivity > 0.1%. Kaplan–Meier and Cox regression analyses were used for survival analyses.

Results:

Morphological relapse occurred in 51 (30%) patients (median 151 days post-SCT). Mixed CD34+ DC occurred in 49/158 (31%) evaluable (morphological relapse occurring within 14 days of DC measurement excluded) patients, with 57% of these relapsing. Flow MRD was positive in 32/144 (22%) evaluable (monitoring BM in remission performed) patients, with a 62.5% relapse rate. The combination of CD34 loss and MRD positivity within 3 weeks of each other (n=13) yielded a relapse rate of 92%.

Performance metrics:

  • CD34: Sensitivity 74%, Specificity 82%, PPV 57%, NPV 91%

  • Flow MRD: Sensitivity 47%, Specificity 88%, PPV 63%, NPV 79%

  • Combined: Sensitivity 46%, Specificity 99%, PPV 92%, NPV 85%

False positives for CD34 were more common in patients with higher CD34% (median 68% vs 42.5%, p=0.0046). Relapses not detected by DC or flow (false negatives) were predominantly beyond the surveillance window or extramedullary disease.

In cases of discordance, relapse occurred in 43% (6/14) with mixed CD34+ chimerism negative flow MRD, compared to only 15% (2/13) when flow MRD was positive but CD34+ chimerism remained >80%, suggesting greater prognostic weight of mixed CD34+ in discordant scenarios.

Time from MRD detection to morphologic relapse were comparable between CD34 and MRD (median ~6 weeks).

Conclusions:

CD34+ chimerism and flow MRD offer distinct and complementary insights into relapse risk. CD34+ is more sensitive and reassuring when negative, while MRD is more specific and reliable when positive. CD34+ donor chimerism is more reliable for monitoring when results diverge. Combined marker positivity offers exceptional specificity (99%) and PPV (92%), and should prompt early intervention. Together, these findings support a unified monitoring strategy combining CD34+ DC and MRD, with potential for improved relapse prediction and targeted pre-emptive therapy. Prospective studies validating these findings are warranted, but this approach may contribute to post-transplant surveillance and improve relapse-prevention strategies.

This content is only available as a PDF.
2025
Sign in via your Institution