Dr Schetelig on Challenges Assessing Haploidentical vs Mismatched Transplantation in AML/ALL/MDS

“One feature which that should be should be mentioned is that we had to stop the trial prematurely purely due to recruitment, and this highlights a specific challenge for clinical trials.”

Johannes Schetelig, MD, director of medical research at DKMS; as well as an a professor at Technical University Dresden, discussed the operational challenges encountered during the phase 2/3 HAMLET trial (NCT03275636), which aimed to compare survival outcomes following haploidentical related donor hematopoietic stem cell transplantation (HSCT) with those from single human leukocyte antigen (HLA) loci–mismatched unrelated donor transplantation in patients with high-risk acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), and myelodysplastic syndrome (MDS).

The trial was terminated prematurely due to poor accrual, a limitation that Schetelig noted underscoreds a central difficulty in conducting randomized studies in the transplantation setting, Schetelig began. Successful enrollment in such trials requires coordinated support from referring physicians, transplant centers, patients, and families, he noted.

In HAMLET, recruitment was hindered by strong preexisting preferences regarding donor source. Both physicians and patients frequently expressed fixed positions, with —some favoring the perceived genetic advantages and familial connection of a related donor, and others preferring the potentially broader availability and institutional experience with unrelated donors. These preferences, present in both directions, led many eligible patients to decline randomization, thereby therefore constraining the recruitment pool.

Schetelig highlighted that this resistance reflects a deeper challenge when trial arms represent established clinical practices supported by retrospective outcomes data. In these situations, equipoise—the condition in which clinicians and patients view both interventions as reasonably comparable—may be difficult to achieve. Without equipoise, willingness to accept randomization diminishes, particularly in the transplant context where donor choice is often highly individualized and emotionally charged.

Schetelig also noted that this study suggests that future comparative studies in HSCT may benefit from alternative designs that address patient and physician preference bias while maintaining scientific rigor. Potential strategies include adaptive trial frameworks, preference-based randomization, and integration of prospective registry data. These methods may allow for broader participation while still generating high-quality comparative evidence.

Ultimately, Schetelig emphasized concluded that while although the trial could not meet its accrual target, the lessons learned are valuable for shaping future transplantation research.