ATG and Post-Transplant Cyclophosphamide Do Not Abrogate the Inferior Outcome Risk Conferred By HLA-Α and HLA-B Mismatched Unrelated Donors

Background

Donor selection for allogeneic stem cell transplant is a complex process, where several factors are considered for potential impact on transplant outcome. Several publications have suggested that 9/10 mismatched unrelated donors (MMUD) may be equivalent to 10/10 MUDs. However, there is debate as to whether all 9/10 mismatches (MM) are equivalent. We sought to examine this in a single centre retrospective study using ATG followed by post transplant cyclophosphamide (ATG PTCy) as standard GvHD prophylaxis and correlated outcomes in terms of donor MM, donor age and CMV serostatus.

Patients and methods

A total of 414 patients who received HCT from unrelated donors between Jan 2015 and Dec 2019, at Princess Margaret Cancer Centre, Canada, were enrolled in the retrospective study. Patients and disease characteristics, and outcomes are detailed in Tables 1-3.

The probability of overall survival (OS) was calculated using the Kaplan Meier product limit method and heterogeneity of time to event distribution functions was compared by the log rank test. The cumulative incidences of chronic and acute GvHD, relapse, and transplant related mortality (TRM) were estimated using the cumulative incidence method considering competing risk, and groups were compared using Gray’s test. Death was considered a competing event for relapse, acute and cGvHD, and relapse was considered a competing event for NRM, acute and cGvHD. GvHD outcomes shown in Table 4.

Results

Median OS at 5y for the whole cohort (n=414) was 45.8% (38.2-53.0), and relapse free survival was 40.2% (32.9-47.4), with TRM at 1y 23.7% (19.6-28.0), Table 5.

Compared to the whole cohort (median OS 38.8mo [21.6-NA]), patients who received grafts from MMUD (n=86) had median OS 9.2mo (6.0-17.3), p<0.001, in particular median OS for HLA‑A MM was 8.1mo (5.3-18.4), HLA-B MM 5.9 (2.7-11.8) compared to other 9/10 MMUD (n=32) 18.8mo (5.7-NA), p<0.001. This was due to higher TRM at 1y, MMUD 39.1% (24.0-49.4) vs MUD (n=327) 19.2% (15.0-23.8) p<0.001 primarily due to HLA-A MM 39.6% (24.0-54.9) and HLA‑B MM 35.7% (11.7-61.1), p=0.001.

GVHD-Free Relapse free Survival (GRFS) at 5y for the whole cohort was 36.8% (29.7-43.9). GRFS At 2y for 9/10 MMUD was 25.2% (16.1-35.4), HLA-A MM 21.6% (9.9-36.2) HLA-B MM 15.4% (2.5-38.8) vs MUD 51.4% (45.3-57.2), p<0.001.

Median OS for patients who received ATG PTCY (n=298) was 49.2mo (49.2-NA), for 9/10 (n=58) 10.2mo (6.1-19.6), with HLA-A MMUD (n=26) 9.6 mo (5.8-27.4), HLA-B MMUD (n=9) 4.6mo (1‑11.8) (p<0.001), Figure.

A number of donor factors significant by univariate analysis (incl. donor age, recipient/donor CMV, CD34 dose, presence of HLA MM, type of MM, female donor to male recipient), were selected to determine correlates with all mortality by multivariable analysis (MVA). The only significant factor was 9/10 MMUD HR 2.27 (1.65-3.14), p<0.0001, specifically HLA-A MM HR 2.69 (1.77-4.08), p<0.00001, HLA-B MM HR 3.14 (1.71-5.78) p=0.0002, Table 6.

The same donor factors significant by univariate analysis were examined for correlation with TRM and significance was found for HLA MM HR 2.36 (1.57-3.55) p=0.00004, specifically HLA-A MM HR 2.44 (1.42-4.17), p=0.0012, HLA-B MM HR 2.39 (1.03-5.57) p=0.0034.

Graft failure (GF) occurred in 4.6% of patients. Significant donor factors for GF were CD34 dose <4.0x10⁶/kg, HR 5.28 (1.11-24.98) p=0.038; and HLA MM 2.78 (1.27-6.05) p=0.01. Of 414 patients, 395 patients (95.4%) received fresh stem cells and 19 (4.6%) received cryopreserved stem cells. Cryopreserved MUD grafts were associated with GF on MVA, HR 5.56 (1.75-17.4, p=0.0034), which may be due to several possible factors including the travel time prior to cryopreservation.

Among AML patients, CMV negative patients in receipt of grafts from CMV negative donors (CMV-/-) (n=64) had worse survival (HR 2.05, 1.03-4.09, p=0.04) compared to recipient/donor pairs where any were CMV positive. There was a significantly higher relapse rate in the CMV-/- (15/64, 23.4%), compared to CMV positive (30/236, 12.2%, any combination), p=0.01 by Chi-squared test.

Conclusions

These data suggest that HLA-A and HLA-B mismatched donors confer a significantly inferior outcome despite the use of ATG PTCy. Cryopreserved stem cells from unrelated donors were associated with graft failure, and CMV-/- were associated with higher incidence of relapse. Further work is required to develop novel conditioning regimens and GvHD prophylaxis to mitigate these risks.