Highlight
• In the multicenter BMT CTN 1702 biological assignment trial, patients categorized as Very Unlikely (90%) to find a MUD.
• Among transplanted patients, MUD was used for 94% of the Very Likely group but only 9% of the Very Unlikely group; despite this distribution, relapse, treatment-related mortality, disease-free survival, and acute and chronic GVHD rates were similar across groups after multivariate adjustment.
Background
Allogeneic hematopoietic cell transplantation (allo-HCT) can be curative for a range of malignant and nonmalignant hematologic disorders, but its success depends importantly on donor selection. Human leukocyte antigen (HLA)-matched related donors are preferred when available; otherwise, searches for matched unrelated donors (MUDs) are standard. Donor-search prognosis tools estimate the likelihood of identifying a MUD rapidly and can guide early planning. Patients with a low probability of finding a suitably matched unrelated donor historically faced delays or suboptimal donor selection pathways and may receive haploidentical related donors, mismatched unrelated donors (MMUD), or umbilical cord blood (UCB) grafts as alternatives.
Advances in graft manipulation, conditioning regimens, post-transplant cyclophosphamide (PTCy), and supportive care have improved outcomes with alternative donors, leading to wider adoption. However, there has been limited prospective, pragmatic evidence comparing outcomes when a donor-search prognosis strategy explicitly drives prioritization of alternative donors in clinical practice.
Study Design
BMT CTN 1702 is a multicenter biological assignment trial designed to reflect real-world practice by using broad eligibility criteria and enrolling consecutive patients undergoing a donor search. Patients initiating a donor search were prospectively stratified into three groups based on a donor search prognosis score: Very Likely (>90% chance of finding a MUD), Less Likely (intermediate probability), and Very Unlikely (<10% chance of finding a MUD). The trial’s primary endpoint was 2‑year overall survival (measured from the start of donor search or alternative donor identification). The Very Likely and Very Unlikely groups were compared in the primary analysis; the Less Likely cohort was observed prospectively and included in secondary/translational analyses.
The study enrolled 1,751 evaluable patients across 47 centers. Key pragmatic design elements included broad enrollment to reflect heterogeneity in diagnoses and transplant practices, and capture of real-world donor use patterns: haploidentical related donors, mismatched unrelated donors, umbilical cord blood, and MUDs were permitted according to center practice and the donor-prioritization strategy recommended for each prognosis group.
Key Findings
Population and donor-use distribution: Of the 1,751 evaluable patients, 958 (54.7%) were categorized as Very Likely to find a MUD, 517 (29.5%) as Less Likely, and 276 (15.8%) as Very Unlikely. Among the 1,179 patients who ultimately underwent transplantation, the distribution of donor types was strongly associated with the initial search prognosis: 94% of transplanted patients in the Very Likely group received a MUD, while only 9% of the Very Unlikely group received a MUD. Alternative donors (haploidentical, MMUD, or UCB) comprised the majority of graft sources for the Very Unlikely cohort.
Primary outcome: 2‑year overall survival
Unadjusted survival at 2 years did not differ between the Very Unlikely and Very Likely groups. The univariate hazard ratio (HR) for death comparing Very Unlikely versus Very Likely was 1.00 (95% CI, 0.82 to 1.21; P = .98). Multivariate analyses adjusting for potential confounders (including age, disease risk, comorbidities, and transplant-related factors) also showed no statistically significant difference: HR 1.07 (95% CI, 0.86 to 1.33; P = .56). These results indicate that prioritizing alternative donors in patients unlikely to find a MUD produced comparable 2‑year overall survival to those expected to receive a MUD.
Secondary transplant outcomes
Among transplanted patients (n = 1,179), multivariate analyses demonstrated no significant differences across the three donor-search prognosis groups in several clinically relevant post-transplant endpoints: relapse incidence, treatment-related mortality (TRM), disease-free survival (DFS), and rates of acute and chronic graft-versus-host disease (GVHD). In other words, when accounting for patient and disease factors as well as transplant variables, the type of donor ultimately used (as driven in part by the search prognosis strategy) did not produce measurable differences in these major outcomes through 2 years.
Interpretation of effect sizes and clinical relevance
The null difference in survival and other major endpoints is clinically meaningful. It supports a strategy of early prioritization of readily available alternative donors (particularly for patients with low likelihood of finding a MUD) rather than prolonged searches for a MUD that may delay transplant or lead to disease progression. The approach appears to preserve disease control and transplant safety at least through the first two years, a high-risk period for relapse and transplant-related complications.
Expert Commentary and Context
Implications for practice: BMT CTN 1702 provides prospective evidence to support pragmatic donor-selection strategies that use an upfront donor-search prognosis to guide whether to prioritize alternative donors. For clinicians, this can inform early transplant planning, expedite donor commitment and scheduling, and potentially shorten time-to-transplant for patients lacking a high probability of finding a MUD. Given the similar outcomes observed, broader use of alternative donor platforms (haploidentical grafts with PTCy, well-matched MMUDs, or optimized cord blood platforms) appears justified when a MUD is unlikely to be identified in a timely manner.
Biological and mechanistic considerations
Advancements over the last decade—including improved GVHD prophylaxis (notably PTCy), refined conditioning regimens, better infection prophylaxis, and enhanced supportive care—have narrowed outcome differences between donor sources. Haploidentical transplantation with PTCy has demonstrated acceptable rates of engraftment, GVHD, and survival in multiple contemporary series; similar statements can be made for selected mismatched unrelated donors and optimized cord blood transplantation protocols. The BMT CTN 1702 results are consistent with this evolution in practice.
Limitations and generalizability
Despite its strengths—prospective design, large sample size, and pragmatic enrollment—several limitations warrant discussion. The trial was a biological assignment rather than a randomized comparison between donor types; therefore, residual confounding by indication or unmeasured center- or patient-level factors could influence results. The Less Likely cohort was observational, making some subgroup comparisons exploratory. The analysis focuses on outcomes through 2 years; longer-term follow-up (e.g., late relapse, chronic GVHD effects on quality of life, and late nonrelapse mortality) will be necessary to fully assess equivalence across donor strategies. Finally, the study aggregated alternative donor types; donor-specific differences (e.g., haploidentical vs cord blood) might exist that are not detectable in these group comparisons.
Practical Recommendations
1) Use donor-search prognosis tools early in the evaluation of patients without a readily available matched related donor to inform parallel donor identification strategies and avoid unnecessary delays.
2) For patients with a Very Unlikely probability of finding a MUD, prioritize alternative donors to expedite transplant, given the comparable 2‑year survival and key post-transplant outcomes reported in BMT CTN 1702.
3) Continue to individualize donor selection based on disease characteristics, urgency, comorbidities, HLA mismatching details, donor age/health, cell dose, and center expertise. Where feasible, multidisciplinary discussion (transplant physician, donor search coordinator, and the patient) remains essential.
Conclusion
BMT CTN 1702 provides prospective, multicenter evidence that a donor-search prognosis strategy—prioritizing alternative donors for patients very unlikely to find a MUD—yields comparable 2‑year overall survival and key transplant outcomes to patients likely to receive a MUD. The findings support pragmatic, early planning that incorporates donor-match probability, thereby reducing delays to transplant without compromising short- to intermediate-term clinical outcomes. Longer-term follow-up and donor-specific comparative analyses will further refine donor-selection algorithms.
Funding and clinicaltrials.gov
The trial was conducted by the Blood and Marrow Transplant Clinical Trials Network (BMT CTN). Detailed funding sources and trial registry information are listed in the primary J Clin Oncol publication (Lee SJ et al., J Clin Oncol. 2025). Please refer to the original article for the full funding statement and ClinicalTrials.gov identifier.
References
Lee SJ, Logan B, Horowitz MM, Dehn JG, Pidala J, Grunwald MR, Westervelt P, Farhadfar N, Hogan WJ, Bashey A, Hayes-Lattin B, Hill L, Brunstein C, Arai S, Srour SA, Symons H, Uberti J, Vasu S, Pusic I, Juckett M, Leifer E, He N, Devine S, Shaw BE, Ciurea SO. Primary Results From Blood and Marrow Transplant Clinical Trials Network 1702: Clinical Transplant-Related Long-Term Outcomes of Alternative Donor Allogeneic Transplantation. J Clin Oncol. 2025 Nov;43(31):3369-3380. doi: 10.1200/JCO-25-00206 IF: 41.9 Q1 . Epub 2025 Sep 18. PMID: 40966481 IF: 41.9 Q1 .
