Real‐world experience: Introduction of T cell replete haploidentical transplantations in a single center

Abstract Objectives The aim of this study was to describe real‐world data on outcomes of T cell replete haploidentical hematopoietic stem cell transplantation (HSCT) after the introduction of this modality in a single center and to compare them with different donor types. Method Outcomes of 30 consecutive patients with hematological malignancies that received T cell replete haploidentical HSCT with posttransplantation cyclophosphamide (PTCY) from 2016 to 2018 in our center were analyzed and compared to the outcome of human leukocyte antigen (HLA)‐related and unrelated matched donor HSCT (n = 97) and to a historical cohort of T cell depleted haploidentical HSCT (n = 11). Results One year graft‐versus‐host‐free, relapse‐free survival in haploidentical HSCT was comparable with other donor types (haplo 40%, matched related donor [MRD] 33%, matched unrelated donor [MUD] 25%, p = 0.55). Non relapse mortality was high in haploidentical HSCT (50%), mostly due to infectious complications. However, relapse rates were only 3%, and OS and progression‐free survival after 1 year were 47% and thereby also similar to HLA‐matched HSCT in our center (MRD 53%, MUD 48%). Conclusion Our data show that T cell replete haploidentical HSCT has similar outcomes to HLA identical HSCT after introduction in our center. More strict adaptation on infection prevention was a crucial aspect of our learning curve. Overall, this type of transplantation is a feasible option when lacking an HLA‐identical donor. This option has advantages over an unrelated donor as it brings less logistical challenges than MUD transplantations.


INTRODUCTION
that TCD resulted in an extremely low incidence of GVHD [1][2][3]. Non relapse mortality (NRM) with this strategy remained extremely high and resulted mainly from lethal infections due to the prolonged T cell deficient state caused by the extensive TCD by the in vitro CD34+ cell selection. We, and others performed this type of TCD haploidentical HSCT in the past and were not able to duplicate the results from Perugia. We observed an even higher rate of NRM and accordingly extremely low survival rates and have abandoned this type of transplantation [4,5]. Interesting strategies have been developed to overcome this drawback of TCD haploidentical HSCT, for instance by only removing specific cell types in vitro [6,7]. Different types of TCD are explained more in detail by Or-Geva and Reisner [8].
Since the development of in vivo T cell depletion with posttransplantation cyclophosphamide (PTCY), the use of haploidentical donor cells is universally increasing [9,10]. The biological concept of this technology is to functionally impair alloreactive donor T cells, activated by day 3-5 after transplant, and spare non-alloreactive T cells [11]. These unmanipulated T cell replete haploidentical HSCTs are characterized by an encouraging overall survival (OS) and low incidences of GVHD and relapse [9,[12][13][14][15].
Even a large meta-analysis comparing haploidentical HSCT (n = 1410 patients) to MRD HSCT (n = 6396 patients) showed no significant differences between both treatment groups with regard to OS, relapse, and GVHD-free, relapse-free survival (GRFS) [23]. T cell replete haploidentical HSCT has several other benefits: limited costs for the donor transplant (compared to MUD or UCB), rapid availability for almost all patients, and the possibility to collect additional cells for cellular therapy at the time of transplant or thereafter [24].
Given this opportunity, we decided to re-introduce haploidentical HSCT in our hospital in 2016 for patients without HLA-identical donors, however this time with PTCY. We decided to retrospectively evaluate the outcomes and compare them with outcomes from HLAidentical HSCT in the same time frame and with our historical cohort of TCD haploidentical HSCT, as a quality control for our institute to see if our transplant data are comparable to those published.

Patients
This is a retrospective study of 127 consecutive patients (30 T cell   replete

Donor selection criteria
In our institute the following hierarchy for donor choice applied: first donor choice for every patient was an HLA-identical donor (MRD or 10/10 MUD) but when not available a haploidentical family donor was chosen. Both patient and donor HLA typing were performed using sequence-based typing for HLA-A, -B, -C, -DRB1, and -DQB1 loci. In case a haploidentical donor was available, patients were tested for the presence of donor-specific antibodies (DSAs). When DSAs were present in high levels (mean fluorescence intensity > 4000), the donor was excluded.

Haploidentical transplantations T cell replete
Chemotherapy-based myeloablative conditioning regimens included thiotepa, busulphan, and fludarabine (TBF) in most of these haploidentical stem cell transplantations according to the myeloablative conditioning regimen used in Genua, Italy [25]. In some indications, a radiotherapy-based conditioning was used with total body irradiation (TBI) consisting of 10 Gray (Gy) in fractionated doses, combined with fludarabine (Flu-TBI) ± cyclophosphamide before transplantation.

TCD
In all cases a myeloablative conditioning regimen was used that included thiotepa, fludarabine, and TBI (8 Gy).

MRD and MUD
A non-myeloablative conditioning regimen with a combination of low dose TBI (2 Gy) and fludarabine was most often used in HLA-identical SCT. Other utilized regimens can be found in Table 1.

Stem cell source
The preferred source of stem cells in T cell replete haploidentical HSCT was bone marrow (BM). Peripheral blood (PB) was preferably used in HLA-identical transplantations. In these transplantations, unmanipulated BM and PB stem cells were given on day 0. Only in cases with ABO major or minor incompatibilities, red blood cell or plasma depletion of the harvested product was performed.
In TCD haploidentical HSCT, PB stem cells were used after ex vivo CD34 selection and cryopreservation.

GVHD prophylaxis
In recipients of T cell replete haploidentical grafts PTCY 50 mg/kg intravenously on days +3 and +5 was given to create in vivo T cell depletion, and they received cyclosporine (CyA) from day 0 until day +180 and mycophenolate mofetil (MMF) from day +1 to day +28 [25].

Most recipients of HLA-identical stem cell transplantations received
CyA from day −3 to day +180 and MMF from day +1 to day +85.
Only patients diagnosed with myelofibrosis also received anti thymocyte globulin (ATG).
Recipients of TCD haploidentical grafts received ATG on days −6 to −2, but no GVHD prophylaxis after stem cell transplantation.

Supportive care
During neutropenia ciprofloxacin and fluconazole were given as selec- tive digestive tract decontamination. Anti-microbial prophylaxis furthermore consisted of valacyclovir, and after 1 year experience on T cell replete haploidentical HSCT cotrimoxazole was added to prevent pneumocystis infections.

Endpoints
The primary endpoint was GRFS at 1 and 2 year. This was defined as time from transplantation until grade III-IV acute GVHD (aGVHD), chronic GVHD (cGVHD) requiring systemic immunosuppressive treatment, disease relapse or death, whichever occurred first [26]. Secondary endpoints were OS, progression-free survival (PFS), relapse rate, NRM, incidence and severity of aGVHD and cGVHD and time to engraftment.

Statistical method
Categorical variables are expressed as number and proportion, and continuous variables as median and range. The Kaplan-Meier method was used for the OS, PFS, and 1-year GRFS analyses.
The cumulative incidences of acute and chronic GVHD were estimated considering death not related to GVHD as a competing event.
For the calculation of NRM, disease relapse or progression was treated as a competing event, and for the calculation of relapse, NRM was treated as a competing event. Outcomes were calculated from the day of transplantation. Comparisons between all groups were made using log-rank and Gray tests, and p-values of these comparisons were given.
Analyses were performed using SPSS, version 25, and R software.

Patients
Patient characteristics are summarized in Table 1.

DISCUSSION
T cell replete haploidentical HSCT with PTCY is increasingly used worldwide, but data shown in literature are mostly from large multicenter analyses and generally from centers with high volumes. Here, we report the introduction of this transplantation method in our relatively small transplantation center, which is the smallest in the Netherlands [27]. On average we perform 40-50 allogeneic HSCT per year, with a staff of only nine hematologists.
In this retrospective analysis for quality control -we find OS and GRFS of T cell replete haploidentical HSCT to be like those of our MRD and MUD HSCT.
Remarkably, NRM rate was much higher than expected from literature. In large retrospective registry studies, NRM is 10%-25%, while in our patients this was 50% [28][29][30][31]. After all these adaptations, NRM decreased from 60% to 40%. This is still quite high but could partly be explained by the fact that most of the patients were of older age (median 60.3 years) and were all still receiving myeloablative conditioning. Recently after this analysis, we adapted our policy and decided to use a reduced intensity conditioning regimen in older patients to see if this could lower the mortality rate.
However, this might be at the cost of a higher risk of relapse.
In other retrospective studies, a prominent cause of death was disease relapse (with 25%-45% of patients experiencing relapse within 2 years) [10,[28][29][30][31]. Only one of our patients (3%) experienced relapse in the first year in the T cell replete haplo group, so our relapse rate was much lower than expected. This could be partly explained by the high NRM that is a competing risk for relapse and could also be linked to the myeloablative conditioning regimen.
Cumulative incidences of acute and chronic GVHD were low in our haploidentical T cell replete HSCT, and comparable to those in literature when BM is used as stem cell source [28][29][30][31].
Due to the low relapse and GVHD rate, OS, PFS, and GRFS were comparable to outcomes found in literature. However, direct comparisons are difficult to be made since our numbers are limited and because these outcomes are dependent of different risk factors like age, co-morbidity, and disease status.
In our center, HLA-identical HSCT at time of the analysis was still performed without PTCY. The use of PTCY in HLA-identical HSCT can potentially improve the outcome. In a recent randomized trial a 1-year GRFS in HLA-identical transplantations (MRD and MUD) of 45% was shown, but this is in the same range as our data with haploidentical HSCT [32]. Another option to improve GRFS in HLA-identical HSCT could be the addition of ATG, as we did not use this in most of our patients. However, in most studies ATG only lowers GVHD and does not impact OS [33,34] and gives similar results to PTCY [35].
In the past, we performed some TCD haploidentical HSCT using the same method as described by the Perugia group. In contrast to their results, we had an extremely high NRM rate that was mainly due to infections, and we noticed that less than a fifth of patients survived.
We hypothesize that the difference in climate in the Netherlands versus Italy, and perhaps also of living conditions, might be the reason for that. Also other centers could not reproduce the good outcomes of the Perugia group so these cannot be the only factors responsible for the worse outcome. Therefore, in our hands, and in those of others, this type of TCD haploidentical HSCT was not seen as a good alternative for patients lacking an HLA-identical donor, and after this experience we stopped performing haploidentical HSCT for several years. Perhaps, these serious infectious problems could have been partly avoided with a different type of TCD. In the CD34 positive cell selection not only all T cells, but also natural killer cells and B cells are removed, while in other types of in vitro TCD-specific subsets are removed and thereby the immune system remains more intact.
For the T-cell replete haploidentical transplantations that were performed more recently, we see better outcomes than in our historical patient group.
In light of this analysis and the logistical problems we experienced last year due to the COVID-19 pandemic, we recently even adapted our search strategy and are preferring a haploidentical donor above an MUD 10 of 10 when there are no DSAs.

CONCLUSION
In summary, this comparison shows that outcomes after T cell replete haploidentical HSCT are comparable to those of HLA-identical HSCT in our hands. They are better than the outcomes of TCD haploidentical HSCT in our center in the past. We performed this study as quality control for our institute to see if our transplant data are comparable to those published. At the start, we experienced a learning curve but could decrease our initially remarkably high TRM due to infectious complications. Even though this study has its limitations due to the small number of patients, the heterogeneity between them and the retrospective nature, we conclude that the use of a haploidentical donor is a valid real-world choice for patients in need for an allogeneic HSCT. A haploidentical donor is usually quickly available for almost all patients, and the choice for this type of donor is logistically easier to arrange than an MUD.