A novel intercontinental external quality assessment program for flow cytometric measurable residual disease assessment in pediatric AML

Abstract

Abstract Background Assessment of measurable residual disease by multidimensional flow cytometry (MFC-MRD) is critical for early response evaluation and risk-adapted treatment of pediatric acute myeloid leukemia (pedAML) serving as a robust surrogate endpoint of response and outcome in clinical trials. To guarantee consistent MFC-MRD diagnostics across multiple laboratories, an intercontinental initiative by the Pediatric Acute Leukemia (PedAL) and European Pediatric Acute Leukemia (EuPAL) foundations, in cooperation with the I-BFM-FLOW network, has implemented a novel format for external performance evaluation. This unique approach uses near real-world, spiked pedAML MRD test samples designed to mimic post-therapy conditions. The aim is to leverage international cooperation among MFC-MRD reference laboratories to ensure high-quality and reproducible results in clinicaltrials fit for regulatory filing even in a transcontinental setting. Methods Enumerated AML blasts from diagnostic bone marrow (BM) samples of pedAML patients are spiked into MRD-negative, non-AML post-chemotherapy regenerating BM samples to simulate MRD test samples. To preserve cell integrity upon shipping samples were treated with a stabilizer. In contrast to conventional External Quality Assurance (EQA) programs that use peripheral blood, this format includes a full spectrum of regenerating BM cell populations, including CD34⁺ progenitors, thus accurately modeling the complexity of a post-treatment hematopoietic milieu. These samples, containing defined MRD levels, are distributed quarterly to reference laboratories worldwide (‘ring trials’ [RTs]). Each laboratory receives two clinical case sample triplets comprised of a diagnostic sample (with 5–20% blasts) and two test samples, which contain MRD at defined levels from ≥0.05% to 5% or are MRD-negative. Each laboratory conducts MRD assessments using its local routine diagnostic procedures. Results of each quarterly assessment are compared centrally to the expected values, allowing objective evaluation of individual and overall performance. Results Since start of the initiative, 9 pedAML-MRD RTs have been completed from 04/22 to 03/25, with a total of 324 MRD test samples distributed to and analyzed across the 9 reference laboratories worldwide (AT, DE, FR, GB, IT, NL, SE, US and AU). These laboratories include sites participating in transatlantic intent-to-file pediatric leukemia early phase clinical trials for which this EQA program is obligatory to prove consistent MRD assessment capabilities amongst involved laboratories throughout the active trial period. The results of these MRD RTs demonstrate the feasibility of this EQA approach. Notably, reliable results were obtained even with extended intervals between test sample preparation and analysis at the local site (median 4 days, range [2–16 days]). A key outcome has been identification of substantial variation in result accuracy over time and across laboratories. Correct MRD assessment rates ranged from 52% in the first ring trial (RT1) to 94% in the most recent RT9, indicating a clear overall improvement. However, individual laboratory performance still varies considerably with a median correct value rate of 82.4% [range 44-100%], highlighting the need for targeted training. In particular, MRD-negative cases with challenging or less well-defined leukemia-associated immunophenotypes (LAIPs) accounted for the majority of incorrect assessments. Structured, individualized feedback is provided by the coordinating center after each RT, allowing participants to reflect on their performance and adjust their approaches. This consistent feedback loop not only enhances diagnostic reproducibility but also helps refine MRD gating strategies and improve assessment capabilities. Importantly, we found no difference in performance between centers following “different-from-normal” MFC-MRD analysis versus those using multidimensional analysis strategies incorporating LAIP assessment. Conclusions This collaborative initiative implementing near real-world MRD RTs offers a novel, unique and first-in-man/first-in-child model for external performance evaluation of MFC-MRD in pedAML. Successful participation by international reference centers demonstrates its feasibility and educational value. The program builds a solid foundation for MFC-MRD in pedAML to support regulatory qualification of clinical trials and to inform clinical decision-making for therapeutic development.