Post-Transplant Cyclophosphamide-Based Graft-Versus-Host Disease Prophylaxis After Mismatched Unrelated Donor Peripheral Blood Stem Cell Transplantation

Abstract

PURPOSE

Allogeneic hematopoietic stem cell transplantation (HSCT) is a curative treatment for advanced hematologic malignancies. HSCT using human leukocyte antigen (HLA)-mismatched donors is historically associated with inferior survival. Patients from underrepresented racial and ethnic groups more frequently rely on HLA-mismatched donors. We hypothesized that post-transplant cyclophosphamide (PTCy) based graft versus host disease (GVHD) prophylaxis would improve outcomes for HSCT recipients using peripheral blood stem cells (PBSCs) from HLA-mismatched unrelated donors (MMUDs) by reducing the risk of GVHD.

METHODS

This phase II, nonrandomized, multicenter trial assessed PBSCs in the setting of a GVHD prophylaxis regimen of cyclophosphamide, tacrolimus, and mycophenolate mofetil in two adult strata: myeloablative conditioning (MAC) and reduced-intensity or nonmyeloablative (RIC/NMA) conditioning before HSCT from a MMUD. The primary objective was to estimate 1 year overall survival (OS) for each stratum. Key secondary end points included incidences of acute and chronic GVHD.

RESULTS

A total of 145 patients enrolled, with 59% self-identifying within an underrepresented group. The 1 year OS was 83.8% (95% CI, 73.1% to 90.4%) for MAC and 78.6% (95% CI, 67% to 86.5%) for RIC/NMA. Incidences of grades III to IV acute GVHD at 6 months were 8% (95% CI, 3.2 to 15.6) for MAC and 10% (95% CI, 4.4 to 18.4) for RIC/NMA. Moderate/severe chronic GVHD at 1 year was 10.3% (95% CI, 4.4 to 18.9) for MAC and 8.6% (95% CI, 3.5 to 16.6) for RIC/NMA. 32% of patients whose donors matched at fewer than seven of eight HLA alleles had similar OS compared with those with donor matched at seven of eight alleles.

CONCLUSION

PTCy-based GVHD prophylaxis after MMUD HSCT with PBSC grafts results in favorable 1 year OS. Using MMUDs expands donor availability to all patients regardless of ancestry (ACCESS; ClinicalTrials.gov identifier: NCT04904588).

INTRODUCTION

Pharmacologic suppression of graft-versus-host disease (GVHD) is critical to the success of hematopoietic stem cell transplantation (HSCT).¹ The use of cyclophosphamide for GVHD prevention has markedly reduced the risk of GVHD after HSCT.^2-6 Traditionally administered on the third and fourth day after HSCT, post-transplantation cyclophosphamide (PTCy) has enabled transplants from donors once considered unsuitable because of human leukocyte antigen (HLA) mismatches that posed unacceptable risks of GVHD and graft failure.^1,7-9 Initially developed for prevention of GVHD after HLA haploidentical-related HSCT, later studies demonstrated that PTCy prevents GVHD after both matched and mismatched unrelated donors (MMUDs).^5,6,8,10,11

CONTEXT

• Key Objective

• Can successful outcomes be achieved after hematopoietic stem cell transplantation (HSCT) from mismatched unrelated donors (MMUDs) using peripheral blood stem cells (PBSCs) as the graft source and post-transplant cyclophosphamide-based graft-versus-host disease (GVHD) prophylaxis?

• Knowledge Generated

• Excellent overall survival (OS) was achieved at 1 year after HSCT using PBSCs from MMUDs, with low rates of GVHD. There was no difference in OS on the basis of degree of human leukocyte antigen match, expanding access to HSCT for patients regardless of their ancestry.

• Relevance (C. Craddock)

• In patients transplanted using a MMUD, post-transplant cyclophosphamide represents a major advance in GVHD prophylaxis with impressive transplant outcomes. By increasing donor availability, this advance will significantly extend the curative potential of allogeneic HSCT, most notably in patients from specific ethnic groups, in whom at present, donor identification can be extremely challenging.*

• *Relevance section written by JCO Associate Editor Charles Craddock, MD.

Patients within racial and ethnic marginalized groups (hereafter referred to as underrepresented) are far less likely than non-Hispanic White patients to identify HLA-matched donors in worldwide donor registries, limiting their access to HSCT.^12-16 To address this disparity, the NMDP, Center for International Blood and Marrow Transplant Research (CIBMTR), and others have evaluated PTCy in MMUD transplants through prospective clinical trials, showing low rates of GVHD and favorable survival outcomes.^8,10,17,18 A prior NMDP/CIBMTR study (ClinicalTrials.gov identifier: NCT02793544) evaluated the use of bone marrow (BM) as the graft source from MMUDs.⁸ However, BM is associated with prolonged cytopenias in recipients and introduces more procedural and anesthesia risk for donors, making mobilized peripheral blood stem cells (PBSCs) the preferred graft source for many adults receiving HSCT.^19-22 The use of PBSCs is associated with more acute and chronic GVHD.^23,24 To determine whether the use of PTCy can effectively control GVHD after a PBSC graft from MMUDs, we conducted a phase II, multicenter study in adults with advanced hematologic malignancies.

METHODS

Trial Design

This prospective, phase II, nonrandomized, multicenter trial evaluated a GVHD prophylaxis regimen consisting of cyclophosphamide, tacrolimus, and mycophenolate mofetil in two adult strata: one for patients receiving myeloablative conditioning (MAC) HSCT and another for patients receiving reduced-intensity or nonmyeloablative (RIC/NMA) HSCT. All patients received PBSC grafts. The primary objective was to estimate 1 year overall survival (OS) separately for each stratum, defined as the time from HSCT to death from any cause.

The protocol (available with the full text of this article at ascopubs.org/journal/jco) was approved by the institutional review board of the NMDP. The trial was performed in accordance with the principles of the Declaration of Helsinki, and all participants provided informed consent. It was funded by the NMDP and conducted by the CIBMTR. The authors vouch for the completeness and accuracy of the data and for the fidelity of the trial to the protocol.

Patients and Donors

The trial planned to enroll 140 eligible patients (70 in each stratum) who were at least age 18 years and undergoing their first allogeneic HSCT from MMUDs. Recipients of MAC were required to be age younger than 66 years and have a hematopoietic cell transplant-comorbidity index (HCT-CI) <5.²⁵ Eligible diagnoses included acute or chronic leukemia without circulating blasts and <5% BM blasts, myelodysplastic syndrome, or lymphoma in remission. Patients with myelofibrosis were not included. Patients were excluded if a suitable HLA-identical sibling or MUD was available. MMUDs were required to have high-resolution HLA matching at 4/8, 5/8, 6/8, or 7/8 alleles (HLA-A, HLA-B, HLA-C, and HLA-DRB1), be considered eligible for donation, and age 18-35 years. Donors were excluded if recipients had donor-specific HLA antibody levels exceeding 3,000 mean fluorescence intensity against the HLA-mismatched alleles.²⁶

Study Treatment

All patients underwent standard pretransplant evaluations and staging procedures. Donor searches were performed by the NMDP together with the transplant centers, which were required to receive search strategy advice from the NMDP. Patients on the MAC stratum received either fludarabine (Flu) and busulfan or Flu and total body irradiation (TBI), and those on the RIC/NMA stratum received one of the following regimens: Flu and busulfan, Flu and melphalan, or Flu, cyclophosphamide, and low-dose TBI. The target CD34+ cell dose was 4 × 10⁶ per kilogram of actual body weight, but no minimum or maximum dose was specified. Grafts were administered either fresh or thawed after cryopreservation.

The GVHD prophylaxis regimen used in all patients (cyclophosphamide 50 mg/kg once daily on days 3 and 4 after HSCT, mycophenolate mofetil, and tacrolimus) has been previously reported and is available in the protocol document.⁶ Recommended target blood levels of tacrolimus ranged from 5 to 12 ng/mL. Tacrolimus taper was recommended to be initiated at 90-100 days after HSCT, if there was no evidence of active GVHD.

Evaluation and Toxicity Assessments

The schedule of patient assessments is provided in the protocol. Clinician assessments of toxicity (National Cancer Institute's Common Terminology Criteria for Adverse Events version 5.0) were performed, and all expected, related grade 5 and unexpected grade 3 to 5 toxicities regardless of relatedness were reported. Monitoring of four key safety end points (primary graft failure [PGF] by day +28, grade 3 to 4 acute GVHD [aGVHD] by day +100, nonrelapse mortality [NRM] by day +100, and grade 3 to 4 cytokine release syndrome [CRS] by day +14) was conducted weekly, with protocol-defined triggers for study pause and data safety monitoring board review. Study assessments continued for 1 year after HSCT in surviving patients.

Outcomes

The primary end point in both strata was OS at 1 year after HSCT. Key secondary end points included incidence of CRS by day 14, cumulative incidence of relapse and NRM at 1 year, hematopoietic recovery, PGF by day 28, grade 2 to 3 and grade 3 to 4 aGVHD at 6 months as defined by the Mount Sinai Acute GVHD International Consortium,²⁷ moderate/severe chronic GVHD (cGVHD) according to NIH consensus criteria,²⁸ GVHD-free, relapse-free survival (GRFS), and patient quality of life, all at 1 year after HSCT. GRFS was defined as survival without grade 3 to 4 aGVHD, cGVHD requiring systemic treatment, or relapse.²⁹ PGF was defined as lack of neutrophil recovery to >500 cells/μL by day +28 after HSCT. CRS assessment was based on ASTCT consensus guidelines.³⁰ There was a planned exploratory analysis of OS at 1 year by donor match level (7/8 v 4-6/8). Quality of life was measured using patient self-reported symptoms and function using the Lee Symptom Scale (LSS) and the Patient Reported Outcomes Measurement System.³¹ Social determinants of health was operationalized using the CDC developed social vulnerability index (SVI).³² Additional secondary and exploratory end points are listed in the protocol and defined in the Data Supplement (online only). The results for the expansion cohort of Stratum 2 and the results for Stratum 3 described in the protocol will be reported separately.

Statistical Analysis

Sample size was based on estimation of margins of error of the 95% CI for 1-year OS; a sample size of 70 patients for each stratum (MAC and RIC separately) was expected to have a margin of error ± 10% when the 1-year OS was 75%. An OS probability of 75% was the targeted 1-year OS based benchmarks from prior studies.^8,10 Kaplan-Meier method was used to estimate OS, event free survival, progression-free survival, and GRFS. Cumulative incidence was used to estimate NRM, relapse, neutrophil recovery, platelet recovery, aGVHD, and cGVHD, with relapse as a competing risk of NRM; death, relapse, and second HSCT as competing risk of GVHD end points; death and second HSCT as competing risk of hematologic recovery; and death as a competing risk of all other end points; 95% CIs for survival and cumulative incidence were obtained using complementary log-log transformations. Incidences of CRS and of PGF were summarized using proportions and 95% CI. Analysis of covariance was conducted to compare the 1-year LSS, adjusted for baseline score, between patients with moderate/severe versus none/mild cGVHD (reported at any time before 1 year). Other end points were summarized using descriptive statistics. All analyses were done using the population of the 145 eligible and infused adults. No interim analyses were done for the study. No data imputations were done for any analysis. Analyses were performed with SAS software, version 9.4 (SAS Institute, Cary, NC).

RESULTS

Patients and Donors

Between September 2021 and August 2023, a total of 145 patients were enrolled from 21 centers in the United States. An additional five patients beyond the anticipated 70 were enrolled to the MAC stratum because of timing of consent. Patient and donor characteristics, stratified by conditioning intensity, are presented in Table 1 and the Data Supplement (Table S1). Overall, 59% of enrolled patients self-identified as belonging to an underrepresented group, with a higher proportion in the MAC group (71%) compared with the RIC/NMA group (46%). The median patient age was 46 years (range, 20-65 years) in the MAC stratum and 65 years (range, 24-77 years) in the RIC/NMA stratum. AML was the most common diagnosis in both groups, with a higher proportion of ALL in the MAC group and more MDS in the RIC/NMA group. A majority of the patients with AML/ALL were transplanted in CR1. The majority of patients with myeloid neoplasms had poor risk disease (Table 1).^33,34 In the MAC group, there was an even distribution between those receiving chemotherapy-only regimens and those receiving Flu and TBI. By contrast, Flu and melphalan was the most frequently used regimen in the RIC/NMA stratum. Notably, most of the grafts were cryopreserved (81%), as the study was conducted during the COVID-19 pandemic.³⁵ Median donor age was the same in both strata (25 years; range, 18-35), and a majority were 7/8 HLA-allele matched (MAC: n = 52; 69%; RIC/NMA: n = 47; 67%), followed by 6/8 (MAC: n = 19; 25%; RIC/NMA: n = 19; 27%), 5/8 (MAC: n = 3; 4%; RIC/NMA: n = 4; 6%) and 4/8 (MAC: n = 1; 1%; RIC/NMA: n = 0; 0%).

TABLE 1.

Subject Demographics and Baseline Characteristics

Characteristic	MAC, n = 75	RIC/NMA, n = 70
Centers, No.	18	13
Age, median (min-max)	46.1 (20.4-65.6)	65.9 (24.3-77.9)
Sex, No. (%)
Male	45 (60.0)	35 (50.0)
Female	30 (40.0)	35 (50.0)
Race, No. (%)
American Indian or Alaska Native	2 (2.7)	1 (1.4)
Asian	6 (8.0)	8 (11.4)
African American	10 (13.3)	9 (12.9)
Native Hawaiian or other Pacific Islander	1 (1.3)	0 (0.0)
White	49 (65.3)	51 (72.9)
More than one race	1 (1.3)	0 (0.0)
Other race	3 (4.0)	0 (0.0)
Race unknown	3 (4.0)	1 (1.4)
Ethnicity, No. (%)
Hispanic or Latino	36 (48.0)	14 (20.0)
Not Hispanic or Latino	38 (50.7)	53 (75.7)
Ethnicity NA (non-US resident)	1 (1.3)	0 (0.0)
Ethnicity unknown	0 (0.0)	3 (4.3)
Karnofsky Performance Score, No. (%)
100	11 (14.7)	3 (4.3)
90	43 (57.3)	39 (55.7)
≤80	21 (28.0)	28 (40.0)
Primary diagnosis, No. (%)
AML	33 (44.0)	37 (52.9)
ALL	29 (38.7)	5 (7.1)
MDS	9 (12.0)	18 (25.7)
Other	4 (5.3)	10 (14.3)
HCT-CI, No. (%)a
0-1	31 (41.3)	30 (42.9)
2-3	32 (42.7)	28 (40.0)
4+	12 (16.0)	12 (17.1)
ELN risk category (2017), AML only, No. (%)
Normal	0 (0.0)	1 (2.7)
Favorable	5 (15.2)	2 (5.4)
Intermediate	9 (27.3)	13 (35.1)
Adverse	19 (57.6)	21 (56.8)
IPSS-R at diagnosis, MDS only, No. (%)
Low	0 (0.0)	2 (11.1)
Intermediate	0 (0.0)	3 (16.7)
High	2 (22.2)	4 (22.2)
Very high	3 (33.3)	5 (27.8)
Not tested/enough info	4 (44.4)	4 (22.2)
Conditioning regimen, No. (%)
Bu/Flu MAC	37 (49.3)	0 (0.0)
Flu/TBI MAC	38 (50.7)	0 (0.0)
Flu/Bu RIC	0 (0.0)	14 (20.0)
Flu/Mel RIC	0 (0.0)	44 (62.9)
Flu/Cy/TBI NMA	0 (0.0)	12 (17.1)
Was the infused product fresh or cryopreserved, No. (%)
Cryopreserved	58 (77.3)	60 (85.7)
Fresh	17 (22.7)	10 (14.3)
Infused CD34+ (×106/kg)
n/missing	75/0	70/0
Median	5.54	5.42
25th-75th pctl	4.91-6.66	4.94-6.95
Min-Max	2.42-18.7	2.30-22.4
Donor age, median (min-max)	24.8 (18.3-34.8)	25.1 (18.7-35.3)
Donor sex, No. (%)
Male	33 (44.0)	31 (44.3)
Female	42 (56.0)	39 (55.7)
Donor/recipient CMV status, No. (%)
+/+	25 (33.3)	21 (30.0)
+/–	11 (14.7)	10 (14.3)
–/+	23 (30.7)	17 (24.3)
–/–	14 (18.7)	19 (27.1)
Unknown/+	1 (1.3)	3 (4.3)
Unknown/–	1 (1.3)	0 (0.0)
Recipient/donor HLA matching, No. (%)
7/8	52 (69.3)	47 (67.1)
6/8	19 (25.3)	19 (27.1)
≤5/8	4 (5.3)	4 (5.7)

Abbreviations: bu, busulfan; CMV, cytomegalovirus; cy, cyclophosphamide; ELN, European Leukemia Net; flu, fludarabine; HCT-CI, hematopoietic cell transplantation - comorbidity index; HLA, human leukocyte antigen; IPSS-R, revised international prognostic scoring system; MAC, myeloablative conditioning; MDS, myelodysplastic syndromes; mel, melphalan; NA, not applicable; NMA, nonmyeloablative conditioning; RIC, reduced-intensity conditioning; TBI, total body irradiation.

^a Per protocol: presence of prior malignancy will not be used to calculate HCT-CI for this trial to allow for the inclusion of patients with secondary or therapy-related AML or MDS.

Baseline socioeconomic and patient-reported outcome (PRO) data are shown in the Data Supplement (Table S2). Completeness of the baseline PRO data was 80% (60/75) and 66% (46/70) in the MAC and RIC/NMA strata, respectively; 73% of MAC and 52% of RIC patients had SVI scores indicating medium to high vulnerability. Across the population, over half of the patients reported reduced baseline physical function (MAC: 57%, RIC/NMA: 65%).

Follow-Up and Outcomes

The median (range) follow-up of survivors in months by strata is 12.0 (3.3-12.9) months for MAC and 12.1 (11.2-12.9) for RIC/NMA. Patient disposition during the study is depicted in Figure 1. The primary end point of OS at 1 year in MAC was 83.8% (95% CI, 73.1% to 90.4%) and 78.6% in RIC/NMA (95% CI, 67% to 86.5%; Fig 2). Secondary end points are summarized in Table 2. Incidences of grades 2 to 4 aGVHD at 6 months after HSCT were 30.7% (95% CI, 20.6 to 41.3) for MAC and 44.3% (95% CI, 32.4 to 55.5) for RIC/NMA, whereas incidences of grades 3 to 4 aGVHD were 8% (95% CI, 3.2 to 15.6) for MAC and 10% (95% CI, 4.4 to 18.4) for RIC/NMA (Figs 3A and 3B; Table 2). The incidences of NIH moderate-to-severe cGVHD at 1 year were 10.3% (95% CI, 4.4 to 18.9) for MAC and 8.6% (95% CI, 3.5 to 16.6) for RIC/NMA (Fig 3C). Symptom burden at 1 year was significantly higher among patients with moderate-to-severe cGVHD compared with those with none or mild cGVHD (21.0 v 10.9, P < .01, Data Supplement, Table S3).

FIG 1.

Patient disposition from enrollment through 1-year follow-up by conditioning stratum. HSCT, hematopoietic stem cell transplantation; MAC, myeloablative conditioning; NMA, nonmyeloablative; RIC, reduced-intensity/nonmyeloablative conditioning.

FIG 2.

Overall survival at 1 year after transplantation by conditioning stratum. MAC, myeloablative conditioning; RIC, reduced-intensity/nonmyeloablative conditioning.

TABLE 2.

One-Year Estimates and 95% CIs for Key Secondary End Points

Outcome	MAC (n = 75), Prob. (95% CI)	RIC/NMA (n = 70), Prob. (95% CI)
GRFSa	47.6% (35.7%-58.6%)	50.8% (38.5%-61.8%)
NRM	10.8% (5.0%-19.0%)	12.9% (6.3%-21.9%)
Disease relapse/progression	22.9% (13.7%-33.5%)	21.4% (12.7%-31.7%)
Acute GVHD grades 2 to 4b	30.7% (20.6%-41.3%)	44.3% (32.4%-55.5%)
Acute GVHD grades 3 to 4b	8.0% (3.2%-15.6%)	10.0% (4.4%-18.4%)
Chronic GVHD (all grades)	19.6% (11.3%-29.7%)	14.3% (7.3%-23.6%)
Severe chronic GVHD	3.1% (0.6%-9.8%)	4.3% (1.1%-11.0%)
Moderate-severe chronic GVHD	10.3% (4.4%-18.9%)	8.6% (3.5%-16.6%)
CMV infectionc	18.7% (10.8%-28.3%)	21.5% (12.7%-31.7%)
Symptomatic BK hemorrhagic cystitisc	8.1% (3.3%-15.7%)	5.7% (1.8%-13.0%)

Abbreviations: CMV, cytomegalovirus; GRFS, GVHD-free, relapse-free survival; GVHD, graft-versus-host disease; HSCT, hematopoietic stem cell transplantation; MAC, myeloablative conditioning; NMA, nonmyeloablative conditioning; NRM, nonrelapse mortality; RIC, reduced-intensity conditioning.

^a Events incude relapse or progression of underlying disease, grade 3 to 4 acute GVHD, chronic GVHD requiring systemic immune suppression, and death.

^b Estimates for acute GVHD at are 6 months after HSCT.

^c Estimates for CMV, infection, and symptomatic BK hemorrhagic cystitis are 100 days after HSCT.

FIG 3.

Key secondary end points by conditioning stratum. (A) Cumulative incidence of acute GVHD grade 2 to 4 at 6 months after transplantation, (B) cumulative incidence of acute GVHD grade 3 to 4 at 6 months after transplantation, (C) cumulative incidence of chronic GVHD moderate/severe at 1 year after transplantation, and (D) cumulative incidence of relapse at 1 year after transplantation. GVHD, graft-versus-host disease; MAC, myeloablative conditioning; RIC, reduced-intensity/nonmyeloablative conditioning.

The disease relapse rate at 1 year was 22.9% (95% CI, 13.7 to 33.5) for MAC and 21.4% (95% CI, 12.7 to 31.7) for RIC/NMA (Fig 3D). NRM at 1 year was 10.8% (95% CI, 5 to 19) for MAC and 12.9% (95% CI, 6.3 to 21.9) for RIC/NMA (Data Supplement, Fig S1a). GRFS at 1 year was 47.6% (95% CI, 35.7 to 58.6) for MAC and 50.8% (95% CI, 38.5 to 61.8) for RIC/NMA (Data Supplement, Fig S1b). The incidence of grade 3 to 4 CRS within 14 days was 1.3% (95% CI, 0% to 7.2%) and 1.4% (95% CI, 0% to 7.7%) for MAC and RIC/NMA, respectively (Data Supplement, Table S4). No PGF was observed in the MAC stratum, whereas four cases occurred in the RIC/NMA stratum (Data Supplement, Tables S5 and S6). Secondary GF occurred in one MAC and two RIC/NMA patients. The median time to neutrophil recovery and platelet recovery to 20,000/mm³ were 15 (range, 12-27) and 17 days (range, 12-60) and 22 (range, 16-56) and 28 (range, 1-98) days for MAC and RIC/NMA, respectively (Data Supplement, Table S5; Data Supplement, Figs S2a and S2b).

In an exploratory analysis, OS did not differ significantly by HLA match level. In the MAC stratum, OS at 1 year was 80.6% (95% CI, 67% to 89.1%) with 7/8 HLA-matched donors and 90.9% (95% CI, 68.1% to 97.6%) with donors matched at <7/8. In the RIC/NMA stratum, OS at 1 year was 76.6% (95% CI, 61.7% to 86.3%) with 7/8 HLA-matched donors and 82.6% (95% CI, 60.1% to 93.1%) with donors matched at <7/8 (Data Supplement, Table S7 and Fig S3). Additional secondary and exploratory analyses, including an unplanned post hoc analysis of genotype frequency, are provided in the supplemental appendix (Data Supplement, Tables S8-S15 and Figs S4-S6).

DISCUSSION

The results of this study demonstrate excellent OS at 1 year for recipients of MMUD HSCT enrolled in both conditioning strata. They also demonstrate PTCy effectively controls GVHD despite use of PBSCs, simplifying the graft procurement process and rendering it easier and safer for donors. Importantly, over half of the enrolled patients self-reported as belonging to a racial or ethnic group often underrepresented in HSCT clinical trials, suggesting that MMUD HSCT using PTCy-based GVHD prophylaxis expands access to HSCT for all patients.^36,37

To contextualize these results, a recent randomized study in the RIC/NMA setting conducted by the Blood and Marrow Transplant Clinical Trials Network (BMT CTN 1703) demonstrated an OS rate of 77% at 1 year (after adjusting for multiple covariates) in the arm using the same PTCy-based GVHD prophylaxis as on the ACCESS study following predominantly HLA-matched related and unrelated donor HSCT. Rates of acute GVHD grade 2 to 4 and 3 to 4 in that trial were 52% and 7%, similar to the 44% and 10%, respectively, observed in this study.⁶ Comparisons to contemporary cohorts of HLA-matched patients receiving MAC are less available. The BMT CTN 1301 study was a randomized trial that used PTCy-based GVHD prophylaxis in one of its three arms, but BM was the graft source, and no additional immune suppression was added to the PTCy backbone.⁵ OS at 1 year was 84.6%, similar to the 83.8% observed in this trial. Rates of grade 2 to 4 and 3 to 4 aGVHD were 37.6% and 10.1%, respectively, and moderate/severe cGVHD was 19.4% at 1 year, also close to the rates of acute and chronic GVHD observed on the MAC stratum of this trial. The results are also in line with those recently observed in a phase II study of MMUDs in the MAC setting, although that study was smaller and limited to donors matched at 7/8 alleles.¹⁸ Although the results observed at 1 year after HSCT are encouraging, we acknowledge that longer follow-up is necessary to confirm the durability of engraftment and low risk of chronic GVHD.

We chose a phase II design for the ACCESS study for a variety of reasons. First, we recognized the need to confirm that the low rates of GVHD observed in our previous study using donor BM could also be achieved with donor PBSC.⁸ Second, to provide more data supporting the notion that excellent outcomes could also be achieved in recipients of donor grafts matched at <7/8. Third, when the study was designed in 2020, haploidentical related donor transplantation using PTCy for GVHD prophylaxis had become the standard alternative donor source in the United States and phase II data from ACCESS, if promising, would be required to generate the equipoise necessary to perform a randomized trial comparing MMUDs with haploidentical donor transplantation. The data from ACCESS now provide the support for such a phase III study.

Our focus on improving outcomes after MMUD HSCT was driven by knowledge that this donor type significantly expands access to curative therapy for all patients with blood cancers regardless of ancestry.¹² In addition to enrolling a high proportion of patients from underrepresented groups, we found, using the SVI score, that many of the patients came from socially vulnerable households with significant rates of financial toxicity. Exploratory data (Supplemental Results and Data Supplement, Figs S5 and S6) further demonstrate that most study participants were unlikely ever to identify a MUD because of factors such as rare HLA genotypes and/or haplotypes, complex genetic admixtures, or HLA antibody reactivity. Additionally, the HLA types of most study patients were rare, with genotype frequencies below those of the global pool of more than 40 million volunteer donors. For such patients, prompt consideration of alternative donor options, including MMUDs, haploidentical relatives, or umbilical cord blood grafts, is essential. The NMDP has developed a search prognosis tool that may provide critical decision support for timely donor selection in patients with a low likelihood of identifying a matched donor.³⁸ In fact, the use of this tool has recently been validated in a prospective study conducted by the BMT CTN.³⁹ That study also demonstrated a shift in preference of alternative donor type toward MMUDs over the course of the study, resulting in a similar time to HSCT for all patients regardless of search prognosis.

The expansion of the donor pool through MMUD HSCT not only addresses disparities in access but also enables the prioritization of younger donors, a key factor consistently associated with improved outcomes in allogeneic HSCT (Data Supplement, Table S10).^40,41 Increased donor age correlates with diminished hematopoietic cell clonal diversity, reduced regenerative capacity, and higher relapse rates.^42-44 By broadening the range of potential donors, MMUD strategies make it possible to select younger donors, which is often not feasible when relying solely on 8/8 HLA-matched or even haploidentical donors. Of note, all donors on this study were age 35 years or younger, and the median age was 25 years. Registry data confirm that survival outcomes are optimal with donors younger than 30 years but decline significantly with older donors.^45,46

The low rates of cGVHD were associated with a meaningful reduction in symptom burden experienced by patients, as self-reported on the LSS (Data Supplement, Table S9). The LSS was an exploratory end point with low numbers and is therefore supportive and hypothesis generating, but not expected to be definitive. We can only speculate what, if any, influence the high proportion of cryopreservation had on the low rates of cGVHD observed, but a previous CIBMTR observational study did find an association between cryopreservation and risk of cGVHD.³⁵ Nevertheless, we observed high rates of bacterial and viral infections in both strata (Data Supplement, Table S10 and S11). This is consistent with previous observations in patients receiving PTCy-based GVHD prophylaxis and is likely due to high doses of cyclophosphamide resulting in both prolonged myelosuppression and diminished T-cell–mediated antiviral immunity.^47-52 The rates of organ toxicity observed were in line with previous studies and remain a concern when using PTCy, particularly in older recipients or in those receiving stringent MAC regimens that carry a higher risk of organ toxicity.⁵³ Future studies should focus on mitigating toxicity of PTCy by dose reductions, which have been reported by other groups in small studies to reduce the risk of organ damage and myelosuppression without compromising control of GVHD.^54-56 The ongoing NMDP sponsored OPTIMIZE study (ClinicalTrials.gov identifier: NCT03797196) seeks to assess the impact of up to a 50% PTCy dose reduction on infectious risk, organ toxicity, GVHD, and OS.

Approximately one third of the donors selected for recipients on this study were mismatched at more than one HLA allele, with no noticeable differences in survival, similar to our previous study using BM grafts.⁸ The majority of these were recipients of 6/8 donor grafts. This exploratory analysis was underpowered and therefore not definitive. Nevertheless, the results observed in the 46 recipients within the <7/8 donor group are encouraging and support searching for <7/8 MMUDs in recipients who could benefit from HSCT but otherwise lack a suitable 7/8, haploidentical related, or cord blood donor. NMDP registry modeling data suggest the ability to include <7/8 donors in searches increases the likelihood of identifying a potential unrelated donor to virtually 100% of patients, regardless of their ancestry.¹² This would permit prioritization of other donor characteristics associated with better survival, such as younger donor age.^12,57

In conclusion, this study demonstrates that HSCT with MMUD PBSC and PTCy-based GVHD prophylaxis offers adult patients with advanced hematologic malignancies access to HSCT with excellent outcomes. Of note, many participants reported significant social and financial vulnerabilities that might have excluded them from transplant access outside the setting of a clinical trial. Beyond race and ethnicity, SVI scores indicate that MMUD HSCT improves access to curative therapy for socially vulnerable populations. Although future studies are necessary to optimize the safety and effectiveness of HLA-mismatched donor HSCT, the NMDP, transplant programs, and societies dedicated to health equity must work to develop interventions to address barriers to HSCT that extend beyond HLA.^36,58,59

DATA SHARING STATEMENT

A data sharing statement provided by the authors is available with this article at DOI https://doi.org/10.1200/JCO-25-00856.

AUTHOR CONTRIBUTIONS

Conception and design: Monzr M. Al Malki, Brent Logan, Janelle Olson, Erin Leckrone, Jeffery J. Auletta, Craig Malmberg, Medhat Askar, Brian C. Shaffer, Farhad Khimani, Mahasweta Gooptu, Karilyn Larkin, Muna Qayed, Sung Won Choi, Larisa Broglie, Bronwen E. Shaw, Steven Michael Devine, Antonio Martin Jimenez Jimenez

Administrative support: Sarah Smith, Erin Leckrone

Provision of study materials or patients: Dipenkumar Modi, Farhad Khimani, Mahasweta Gooptu, Alison Loren, Karilyn Larkin, Sung Won Choi, Bronwen E. Shaw, Antonio Martin Jimenez Jimenez

Collection and assembly of data: Monzr M. Al Malki, Janelle Olson, Jianqun Kou, Sarah Smith, Erin Leckrone, Juan Wu, Jeffery J. Auletta, Stephen R. Spellman, Medhat Askar, Mahasweta Gooptu, Karen Ballen, Alison Loren, Karilyn Larkin, Sally Arai, Sung Won Choi, Larisa Broglie, Bronwen E. Shaw, Steven Michael Devine

Data analysis and interpretation: Monzr M. Al Malki, Stephanie Bo-Subait, Brent Logan, Janelle Olson, Jianqun Kou, Erin Leckrone, Juan Wu, Heather E. Stefanski, Jeffery J. Auletta, Stephen R. Spellman, Medhat Askar, Rachel Cusatis, Brian C. Shaffer, Dipenkumar Modi, Farhad Khimani, Mahasweta Gooptu, Mehdi Hamadani, Abeer Madbouly, Martin Maiers, Stephanie Fingerson, Rachel Cook, Karen Ballen, Alison Loren, Karilyn Larkin, Sally Arai, Muna Qayed, Sung Won Choi, Larisa Broglie, Bronwen E. Shaw, Steven Michael Devine, Antonio Martin Jimenez Jimenez

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Post-Transplant Cyclophosphamide-Based Graft-Versus-Host Disease Prophylaxis After Mismatched Unrelated Donor Peripheral Blood Stem Cell Transplantation

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jco/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).