Post-transplant lymphoproliferative disorders (PTLDs) are a heterogeneous group of lymphoid or plasmocytic proliferations occurring after transplantation. The new World Health Organization classification of hematolymphoid tumors categorizes them as any hyperplasia or lymphoma arising in the immune deficiency/dysregulation setting post-transplantation (1). They are associated with iatrogenic immunosuppression and Epstein-Barr virus (EBV) in recipients of solid-organ transplant (SOT) or hematopoietic stem cell transplantation (allo-HSCT) (2). EBV-association is particularly pronounced in PTLD after allo-HSCT (~83%) as compared to cases after SOT (~33%) (3), often coincides with rapidly rising EBV DNAemia and mostly occurs within 100 days after allo-HSCT (4). While PTLDs are rare with an incidence of 1%-3.2% in the allo-HSCT setting, their incidence has been increasing alongside the number of transplantations, unrelated or HLA-mismatched related donors, selective T-cell depletion, antithymocyte globulin use and donor as well as recipient age (4–7). The risk of PTLD increases in patients with two or more risk factors (8). Thus, prospective monitoring of EBV activation and early treatment intervention seems to be reasonable in patients with elevated risk of PTLD after allo-HSCT. Clinically, extra-nodal disease is common, including 10%-15% presenting with central nervous system (CNS) disease (9). Despite improvements in treatment, PTLD remains one of the most life-threatening complications of allo-HSCT with an overall survival (OS) rate of ~50% after 6 years and a PTLD-related mortality rate of ~30% in the first year after diagnosis (3, 4). Standard risk-stratified therapy in patients with PTLDs after SOT consists of rituximab with or without chemotherapy (10, 11). Non-destructive PTLDs, also termed early lesions, tend to regress with reduction of immunosuppression (RIS) (12). Their prognosis seems to be excellent, if immune suppression reduction can be achieved without graft rejection. However, elevated lactate dehydrogenase, organ dysfunction, multi-organ involvement, advanced stage, bulky disease, and older age have been reported to be associated with a lack of response to decreased immune suppression (13, 14).

To address the need for standardization in patients with PTLDs after allo-HSCT, the Sixth European Conference on Infections in Leukemia (ECIL-6) published guidelines for the management of these patients (15). However, randomized controlled trials are sparse and treatment strategies remain to be standardized.

The aim of this retrospective study from three academic centers was to evaluate the characteristics and outcomes of PTLD in adult patients after allo-HSCT in comparison to PTLD after SOT.

The focus of our study was to characterize adult patients with PTLD after allo-HSCT as well as SOT in a cohort study from three University hospitals and to compare differential outcomes.

Latency from transplantation to the onset of PTLD was shorter in the allo-HSCT group, similar to what is reported in the literature (11, 18). This is likely due to the strong immunosuppressive therapy after allo-HSCT. Both, the extent and duration of the immunosuppressive therapy are main risk factors for development of PTLD (2). This is corroborated by the fact that almost all patients of the allo-HSCT group were still on immunosuppressive therapy at the time of PTLD diagnosis. Tailored maintenance immunosuppressive regimens, however, may help to reduce the risk of PTLD onset (19).

In contrast to other reports (3), we provide comprehensive clinical characteristics including ECOG-PS, stage, and IPI at the time of PTLD diagnosis for allo-HSCT patients. Patient characteristics at time of PTLD diagnosis were comparable between both groups. Main differences were the frequency of malignancies as the indication for the transplantation and, in conjunction with that, the fraction of patients that previously received anthracycline-containing treatments. Surprisingly, this higher burden of pretreatment with chemotherapy and lower latency in the allo-HSCT group did not translate into a lower median ECOG-PS. There was, however, a larger fraction of patients with ECOG-PS >2. An ECOG-PS ≤2 was a prerequisite for inclusion in the phase II PTLD-1 trial (NCT01458548) and the phase II PTLD-2 trial (NCT02042391). These trials exclusively recruited patients with PTLD after SOT and demonstrated that a risk-stratified sequential treatment with rituximab +/- CHOP is safe and effective (10, 11). However, the higher fraction of patients with ECOG-PS >2, the high susceptibility to cytopenias after chemotherapy, and the common previous exposure to anthracyclines warrants caution when transferring the results of the safety of the CHOP regimen from these trials to patients with PTLD after allo-HSCT.

The reported treatment strategies of PTLD in our series involved RIS or IS-switch and rituximab in almost all cases in both groups. RIS is an established approach in PTLD after SOT and has been shown to be an independent prognostic factor in PTLD after allo-HSCT (4). However, RIS after allo-HSCT bears the risk of promoting GvHD and can be prohibitive in cases of coexisting extensive active GvHD, such as in one of our cases (HSCT7). How to balance the need for both decreased IS for the treatment of PTLD and increased IS for treatment of GvHD in such cases remains to be elucidated. The combined immunosuppressive and potential anti-lymphoma effects by mTOR inhibitors (i.e. Sirolimus, Everolimus) are hypothesized, but not systematically validated, rationales.

Fewer patients in the allo-HSCT group received CHOP regimens as compared to the SOT group. The relatively low percentage reflects the high prevalence of prior anthracycline-containing therapies for malignancies in the allo-HSCT group (60%). For example, concerns about the level of cumulative anthracycline doses are documented as reasons to omit them in the treatment of PTLD in HSCT9. Therefore, prior treatment with chemotherapy and cumulative toxicities are a major consideration when choosing PTLD treatments in this group.

EBV-specific T-cell therapy (VST) was used exclusively in allo-HSCT patients. However, adoptive transfer of EBV-specific T cells has shown great promise for the treatment of EBV-viremia and EBV⁺ PTLD in SOT as well (20, 21). We suspect that, because VST are an established treatment for viral infections after allo-HSCT, physicians were more experienced and inclined to provide VST to allo-HSCT PTLD patients. Moreover, the greater use of VST in the allo-HSCT group was probably due to the markedly higher EBV-association. Other works that defined EBV-positivity as the histological finding of EBV-encoded small RNAs (EBERs) have also found the frequency of EBV-positivity to be ~83% after allo-HSCT (3) and to be ~60% in a SOT dominated cohort (22). In our SOT cohort, only two of four patients with EBER-positive PTLD also had detectable EBV levels in the peripheral blood (PB). Data on the comparative detection of EBV genetic material in PTLD lesions and PB are lacking. Thus, EBER tests should always be performed so as not to overlook EBV-association and miss out on VST treatment options. However, EBV-positive PTLD can also develop in the absence of high viral loads. In addition, high viral loads are not predictive in all settings, and a decline in viral loads may not always predict treatment response. Nevertheless, a rapid decline in EBV load occurs almost invariably after anti-B cell monoclonal antibody treatment, irrespective of long term PTLD response (23). The lack of standardization of techniques and results for assessment of circulating EBV DNA has been an additional complicating factor. Therefore, an international WHO standard for determining EBV load was published in 2013 (23).

With the approval of tabelecleucel for relapsed or refractory EBV-associated PTLD after SOT or allo-HSCT (20), we expect this form of therapy to become more widespread.

Beyond these options, patients in both groups received a wide variety of treatments. These were usually chosen to address the specific subtype of PTLD: two patients (HSCT6 and SOT4) received local therapy and methotrexate for PTLD, while three patients received regimens that were more akin to those developed for the non-PTLD entity counterparts (HSCT9 with ‘GMALL-B-NHL 2002’ for Burkitt lymphoma, HSCT14 with AVD + RTx for cHL, HSCT15 with BV for T-cell PTLD). This aspect is often underreported in PTLD studies. PTLD subtypes themselves appear to be prognostic factors for OS (24–26). However, the heterogeneity in treatments even within PTLD subtypes hampers the identification of prognostic factors and optimal treatment algorithms from retrospective analyses. In addition, there is a concern that chemotherapy might result in relapses of original hematological malignancies or graft rejection. Therefore, more prospective studies are warranted towards a PTLD-specific prognostic model.

In our cohort, the ORR was lower in the allo-HSCT group. Consequently, the OS trended towards an inferior outcome for this group. However, a substantial fraction of patients in the SOT group eventually experienced progressive PTLD. Nevertheless, our data compare favorably with other published analyses (11, 27, 28). In the allo-HSCT group, time from transplantation ≤150 days was identified as a prognostic factor for lower OS in a univariate analysis. PTLD onset as a prognostic factor was observed to varying degrees in previous works (4, 10). One study by the Infectious Diseases Working Party of the European Group for Blood and Marrow Transplantation found a trend for lower OS for PTLD onset <100 days after allo-HSCT (4) and another found that later onset of PTLD after SOT had a longer time to progression (10). In our analysis on the total PTLD cohort, neither age, ECOG-PS or IPI were associated with lower OS. In the SOT group, however, ECOG-PS > 2 was significantly associated with lower OS. The low patient numbers of our cohort were prohibitive of a multivariable analysis and warrant caution when interpreting the significance of the identified prognosticators.

Only a very small number of studies directly compared cases of PTLD after allo-HSCT with that after SOT (29–32). These found a later onset, a higher ORR to treatment as well as a trend for a longer OS in the SOT group as compared to the allo-HSCT group. In contrast, the relapse rate was higher in the SOT as compared to the allo-HSCT group (24, 25, 27, 28). However, in most of these studies treatment is not comprehensively reported or patients from the pre-rituximab era are included. Particularly, in these series rituximab was given to only 36% (n=10) of PTLD patients after SOT and 38% (n=5) of PTLD patients after allo-HSCT (29, 31). This presumably influenced the rather poor median OS in the allo-HSCT groups of <12 months in the majority of studies. Therefore, our study provides important data on more recently treated PTLD patients.

Limitations of this study included its retrospective nature, the small size of the cohorts, and the low number of patients who received a PET-based response assessment. Recently, novel insights into pathogenesis (33) and treatment strategies for PTLD have emerged, including adoptive immunotherapy and targeted therapeutics, such as tabelecleucel, CAR-T cells, and CD30-based therapy (34), that could substantially improve outcomes.

In the near future, data from the phase-3 ALLELE study (NCT03394365) testing tabelecleucel vs. placebo in patients with PTLD after allo-HSCT or SOT might yield more insight into differential response patterns in a relapsed or refractory setting.

In conclusion, we show that PTLD after allo-HSCT has an earlier onset, a lower ORR, and tends to have a worse outcome when compared to PTLD after SOT in a multicenter cohort. We highlight that GvHD and pre-exposure to anthracyclines are important considerations in the treatment of PTLD after allo-HSCT and describe existing treatment heterogeneities. Refined prognostic models are needed. Finally, novel therapies for PTLD are being approved and new studies are warranted to analyze their safety and efficacy in a real-world setting.

The datasets presented in this article are not readily available because The raw data supporting the conclusions of this article are restricted in accordance with the local legislation. Requests to access the datasets should be directed t\o s.kayser@dkfz-heidelberg.de.

Ethical review and approval was not required for the study on human participants in accordance with the local legislation and institutional requirements. The patients/participants provided their written informed consent to participate in this study.

PL and SK were responsible for the concept of this paper, contributed to the literature search and data collection, analysed and interpreted data, and wrote the manuscript. PL, AR, UH, MH, UP and SK contributed patients and critically revised the manuscript. AM and LK performed research and critically revised the manuscript. All authors contributed to the article and approved the submitted version.