Personalized Desensitization for DSAs in Allogeneic Blood or Marrow Transplantation: Impact on Graft Outcomes and Access to Transplant Free

Introduction: Donor-specific anti-HLA antibodies (DSAs) increase graft rejection risk in allogeneic blood or marrow transplantation (alloBMT), especially as mismatched donor transplants become more common. However, standardized methods for assessing DSA strength and risk are lacking, and institutional approaches to donor selection and desensitization vary widely. We describe our center's DSA assessment and management strategy, its impact on transplant access, and outcomes of desensitization.

Methods: We retrospectively analyzed 408 consecutive patients undergoing their first reduced-intensity alloBMT at Johns Hopkins with post-transplant cyclophosphamide prophylaxis (2014–2022), all of whom screened positive for a DSA against at least one potential donor, as defined as any candidate who underwent confirmatory HLA typing. DSA strength was assessed using multiple solid-phase immunoassays and flow cytometric crossmatch (FCXM), with mean fluorescent intensity (MFI) cutoffs determined by historical correlations with cell-based assays (FCXM and complement-dependent cytotoxicity crossmatch, CDC-XM). DSAs correlating with positive FCXM or CDC-XM were generally avoided. During unrelated donor (URD) searches, DSAs to mismatched HLA alleles (even those not included in formal match criteria) were excluded. Desensitization was offered to FCXM⁺ patients, those with an MFI >5,000 or additional risk factors (e.g., rising or multiple DSAs, child-to-mother direction), and consisted of therapeutic plasma exchange (TPE) with IVIg. Treatment intensity (TPE number, use of tacrolimus/MMF) prior to conditioning, was guided by estimated DSA strength.

Patients were analyzed in two ways:

• Among DSA⁺ transplants, we compared desensitized (n=71) and untreated (n=137) patients for change in MFI, graft failure, donor chimerism, and engraftment (neutrophils by day 30, platelets by day 60).

• All 408 patients were stratified by maximum DSA strength during screening considering all potential donors(Ab^high vs. Ab^low/int) and evaluated for associations with time-to-BMT and donor search outcomes.

Results:Desensitization outcomes: Median pre-desensitization MFI was 7,197 (IQR, 5,006–13,265), which was reduced to 1,331 following treatment—comparable to 1,567 in untreated patients (p=0.24). Desensitized patients received a median of 3 TPE/IVIg sessions (IQR, 1–6). Treatment was tailored by DSA strength: 39 of 71 received tacrolimus/MMF and pre-conditioning TPE/IVIg, whereas 32 received only day -1 TPE/IVIg. Haploidentical related donors were used in 97% of desensitized and 94% of untreated cases. Neutrophil engraftment by day 30 was 92% vs. 91% and platelet engraftment by day 60 was 87% in both. Graft failure occurred in 4% of desensitized vs. 7% of untreated patients (p=0.74). Overall survival was 75% (95% CI, 0.65–0.85) vs. 84% (0.78–0.90); non-relapse mortality was 14% (0.06–0.22) vs. 9% (0.04–0.14).

Impact of DSA strength on access to transplant: Compared to the Ab^low/int group (n=276), Ab^high patients (n=132) were more likely to be female (95% vs. 55%, p < 0.001), have myeloid malignancies (75% vs. 62%, p = 0.04), and receive unrelated donors (47% vs. 8%, p < 0.001). Time-to-BMT was longer (median 120 vs. 90 days; p < 0.001), and more donors were typed (median 4 vs. 2; p < 0.001). In multivariable modeling, non-Hispanic white (NHW) race, unrelated donor (URD) use, and high antibody levels interacted significantly. Minority patients in the Ab^high group paired with URDs had the longest time-to-BMT (median 219 vs. 145 days for NHW; p = 0.027), were more likely to receive mismatched URDs (100% vs. 45% among URD transplants). Minority patients in the Ab^high group were less likely to receive an URD transplant, a disparity not observed in the Ab^low/int group (odds ratio 0.39 vs. 0.75, p = 0.03).

Conclusions: Patient-tailored desensitization can effectively reduce DSA levels and enable successful transplantation with graft outcomes comparable to low-level, untreated DSA⁺ patients. However, high-level DSAs significantly prolong transplant timelines—particularly for minority patients reliant on mismatched URDs—partly due to cross-reactive DSAs that disqualify otherwise viable donors. Given the demonstrated success of desensitization, these data support more liberal use of DSA+ donor-recipient pairs to reduce barriers to transplantation and mitigate disparities in access.