Introduction

As a result of recent advances in medical care, the number of patients who receive various immunosuppressive agents has been increasing. These patients are known to be at risk for the development of immunodeficiency-associated lymphoproliferative disorders (LPD) under an iatrogenic immunocompromised status [1,2,3]. Although aberrant infection of Epstein-Barr virus (EBV) in immunosuppressed lymphocytes has been suggested to play a key role in the pathogenesis of these LPD [4,5,6,7], the overall context of this unique disease entity is not yet completely understood.

This insufficient understanding of iatrogenic immunodeficiency-associated LPD is partly attributed to their clinical, histopathological, and genetic heterogeneity [1, 8,9,10]. They include post-transplant lymphoproliferative disorders (PTLD) that arise in patients after solid organ transplantations or hematopoietic stem cell transplantation (HSCT) and other iatrogenic immunodeficiency-associated LPD that arise in patients treated with various immunosuppressive agents for any reason. They include various pathological subtypes including non-destructive hyperplasia of lymphocytes, polymorphic LPDs, and several aggressive types of malignant lymphomas. Their genetic landscapes have not yet been completely revealed. In recent studies, PTLD has been considered to consist of genetically distinct populations: EBV-related or others [6, 11,12,13] and germinal center B cell-like (GCB) or non-GCB subtype [14]. There was also a hypothesis that some subtypes of PTLD might actually be a coincidental occurrence of lymphoid neoplasms among post-HSCT patients [11, 15], although this view has not yet reached a consensus. The genetic backgrounds of other iatrogenic immunodeficiency-associated LPD such as MTX-associated LPD have been scarcely examined.

Regardless of this heterogeneity, LPD growing with an immunocompromised background are known to present worse clinical outcomes than those without such a background [1, 16, 17]. In the analysis of PTLD, the International Prognostic Index (IPI) score, hypoalbuminemia, and the treatment response for rituximab have been suggested as prognostic factors for survival [1, 16, 18, 19]. However, these prognostic factors have been less discussed in other iatrogenic immunodeficiency-associated LPDs [20, 21]. Moreover, although all these iatrogenic immunodeficiency-associated LPDs share immunocompromised backgrounds [22] with some pathological features derived from aberrant viral infection [7, 23, 24], the impact of the immunosuppressive status in each patient on clinical outcomes has rarely been assessed [25]. This should be more carefully examined since it reflects not only the patient’s morbidity but also the anti-viral or anti-tumor effects of lymphocytes.

In this study, we analyzed the impact of lymphopenia on survival in patients with iatrogenic immunodeficiency-associated LPDs. We chose total lymphocyte count as a clinical factor to evaluate patients’ immunosuppressive status since it is easy to obtain and is always examined as an index for immune reconstitution in routine practice.

Methods

Data collection

Clinical data of patients who were pathologically and clinically diagnosed with PTLD or other iatrogenic immunodeficiency-associated LPD over the past 20 years were collected from electronic medical records in Kyoto University Hospital.

Diagnosis was based on the WHO classification at the time and also reviewed according to the WHO classification of 2017 (revised 4th edition) when analyzed. Details of lymphomas and the results of blood examinations such as total lymphocyte count and serum albumin value at diagnosis were also collected from the records. Those associated with primary immune disorders or human immunodeficiency virus (HIV) infections were excluded. All patients gave their informed consent prior to their inclusion in the study. The Institutional Review Board of Kyoto University Hospital, where this study was organized, approved this study.

Statistics

The primary endpoint of this study was overall survival (OS) and the secondary endpoint was progression-free survival (PFS). OS was examined by calculating deaths from any cause; survivors at the last follow-up were censored. PFS was examined by calculating progression/relapse of LPD/lymphomas or death from any cause. Descriptive statistics were used to summarize variables related to patient characteristics. OS and PFS were evaluated by Kaplan-Meier methods and the Cox regression hazards model was used in univariate and multivariate analyses to assess the prognostic significance of the total lymphocyte count at diagnosis. Multivariate analysis was performed using covariates that were selected by preceding stepwise selection in the Cox model with a P value threshold of under 0.2. Covariates assessed were recipients’ sex, clinical background (PTLD, other iatrogenic immunodeficiency-associated LPDs), histological characteristics (monomorphic, polymorphic, or early-phase diseases), International Prognostic Index (IPI) value, primary treatment (rituximab-containing chemotherapies, other chemotherapies or focal radiation, no treatments or reduction in immunosuppressive agents), EBER positivity, serum albumin value at diagnosis, and year at diagnosis (1998–2013, 2013–2017).

Results

Patient characteristics (Table 1)

A total of 71 patients (PTLD, n = 26; other iatrogenic immunodeficiency-associated LPD, n = 45) were included, 65 of whom had data of total lymphocyte counts at the diagnosis of LPD (52–87,412/μL). The median age at transplantation was 63 years (range, 3–83) and the median follow-up period for survivors was 2.5 years. Diffuse large B cell lymphoma was diagnosed in 33 patients (PTLD, n = 12; others, n = 21), Hodgkin lymphoma in 12 (PTLD, n = 3; others, n = 9), monomorphic B cell LPD not specified in 11 (PTLD, n = 3; others, n = 7), and polymorphic LPD or early-phase diseases in 15 (PTLD, n = 7; others, n = 8). As for the initial treatment, immunosuppressive agents were reduced in 34 patients (PTLD, n = 4; others, n = 30), rituximab-containing chemotherapies were given in 23 patients (PTLD, n = 14, others, n = 9), other chemotherapies were given in 7 (PTLD, n = 4; others, n = 3), and radiation or nothing was given in 5 (PTLD, n = 4: others, n = 1). The median value of serum albumin at diagnosis was 3.18 g/dL. We set a threshold value of 800/μL (10 [9]/L) absolute lymphocyte counts as the lymphopenia definition by calculating the optimal threshold value using a receiver-operating characteristic (ROC) curve. A total of 26 patients (PTLD, n = 14; others, n = 12) were diagnosed with lymphopenia (≤ 800/μL) and 39 patients were not (PTLD, n = 11; others, n = 28).

Table 1 Patient characteristics
Full size table

Impact of total lymphocyte count at diagnosis on OS

The impact of total lymphocyte count on OS was illustrated with reference to a lymphopenia group (total lymphocyte count ≤ 800 /μL at diagnosis) and a no-lymphopenia group (total lymphocyte count > 800/μL at diagnosis) (Fig. 1a). Overall, the 2.5-year OS was 74.8% (lymphopenia group, 38.8%; no-lymphopenia group, 93.1%).

Fig. 1
figure 1

Overall survival. Probability of overall survival for total patients (a), post-transplant lymphoproliferative disorders (PTLD) (b), and other iatrogenic lymphoproliferative disorders (c) with reference to a lymphopenia group (total lymphocyte count < = 800/μL at diagnosis) and a no-lymphopenia group (total lymphocyte count > 800/μL at diagnosis)

Full size image

In the multivariate analysis, lymphopenia at diagnosis was associated with inferior OS (HR, 3.72; P = 0.043; Table 2). Serum albumin values (> 3.18 g/dL vs. ≤ 3.18 g/dL; HR, 0.21; P = 0.010) and high IPI (high vs. low to high-intermediate; HR, 4.37; P = 0.003) also affected the OS. A subgroup multivariate analysis to assess the impact of lymphopenia according to the clinical background showed a similar trend (Fig. 1b, c), although statistical significance was observed only in patients with other iatrogenic immunodeficiency-associated LPD (HR, 26.67; P = 0.012). Progression of lymphoma or LPD was the most common cause of death in the lymphopenia group (PTLD, n = 3; others, n = 6), followed by transplant-related mortality (PTLD, n = 5) (Table 3).

Table 2 Multivariate analysis of OS
Full size table
Table 3 Number of lymphocytes at diagnosis and cause of death
Full size table

Impact of lymphopenia at diagnosis on PFS

The impact of the total lymphocyte count on PFS was illustrated with reference to a lymphopenia group and a no-lymphopenia group (Fig. 2a). Overall, the 2.5-year PFS was 67.1% (lymphopenia group, 36.7%; no-lymphopenia group, 86.6%).

Fig. 2
figure 2

Progression-free survival. Probability of progression-free survival for total patients (a), post-transplant lymphoproliferative disorders (PTLD) (b), and other iatrogenic lymphoproliferative disorders (c) with reference to a lymphopenia group (total lymphocyte count < = 800/μL at diagnosis) and a no-lymphopenia group (total lymphocyte count > 800/μL at diagnosis)

Full size image

In a multivariate analysis, lymphopenia was independently associated with inferior PFS (HR, 3.82; P = 0.012; Table 4). Serum albumin values also showed a strong impact (> 3.18 g/dL vs. ≤ 3.18 g/dL; HR, 0.26; P = 0.013). Trends of inferior PFS in patients with high IPI (high vs. low to high-intermediate; HR, 3.04; P = 0.067) and superior PFS in those who received rituximab-containing chemotherapy as primary treatment (rituximab-containing chemotherapies vs. other chemotherapies; HR, 0.16; P = 0.081) were suggested. Although the non-lymphopenia group showed a trend of superior PFS (Fig. 2b, c), its statistical impact was apparent only in patients with other iatrogenic immunodeficiency-associated LPDs (HR, 9.66; P = 0.010), but not in patients with PTLD (HR, 0.66; P = 0.703) (Table 5).

Table 4 Multivariate analysis of PFS
Full size table
Table 5 Subgroup analysis (PTLD and other iatrogenic LPDs) of OS, PFS
Full size table

Discussion

The results of our study demonstrated that lymphopenia at diagnosis may predict inferior survival in patients with immunodeficiency-associated LPD, despite its histological heterogeneity. Contrary to the expectation that this high mortality among patients with lymphopenia reflects their fragility with respect to various infections or intensive chemotherapy [21], the major cause of death was disease progression. Whereas the possibility of rituximab as a primary treatment might have had some impacts on disease suppression, the impact of lymphopenia was independently associated with a higher risk of mortality.

These results suggest that lymphopenia itself could influence disease progression among immunocompromised patients. Several biological expectations could support this hypothesis. First, tumor pathogenesis of these LPD depends partially on the underlying infection of oncoviruses such as EBV. Immunocompromised status in lymphopenia patients could progress aberrant expansion of these oncoviruses. Second, anti-tumor effects of lymphocytes are thought to be less efficient in patients with fewer lymphocytes. Studies on graft-versus-lymphoma (GVL) effects [26] or programmed cell death 1 (PD1)-programmed cell death 1-ligand 1 (PDL1) inhibition [27] have revealed that the anti-tumor effects of lymphocytes play important roles in suppressing tumor cells. As suggested in several malignant diseases [28], the total number of lymphocytes might reflect their tumor-suppressive efficacy against lymphoma cells as well, with a clear impact especially among immunocompromised patients. These considerations that follow our results could explain why some immunodeficiency-associated LPD shrink after the cessation or reduction of immunosuppressive agents [24, 29, 30]. They also support a previous suggestion that earlier recovery of lymphocytes after the cessation or reduction of immunosuppressive agents can predict a lower frequency of disease progression [31].

Possibility of rituximab application, higher IPI value, and hypoalbuminemia were reconfirmed as strong prognostic factors for overall survival in our analysis. However, similar to the results of the phase 2 PTLD-1 trial, [32] rituximab-containing chemotherapies such as R-CHOP did not dramatically improve overall survival. The investigation of risk-dependent strategies and the results of other regimens examined in ongoing clinical trials are awaited [16, 33]. Based on our hypothesis, the promotion of the anti-tumor effects of lymphocytes might be another potent strategy to improve clinical outcomes of iatrogenic immunodeficiency-associated LPD. Since frequent somatic alterations in genes encoding PD-L1/PD-L2 were suggested to contribute to the tumor pathogenesis of lymphomas associated with prior EBV infection, [34]PD1-PDL1 inhibitors could be considered as a therapeutic option in EBV-related immunodeficiency-associated LPD [35, 36]. Although more detailed investigation is warranted, modulation of tumor microenvironment should be a potent target in the treatment strategy of immunodeficiency-associated LPD including PTLD [37]. Nevertheless, it is often a big issue to improve and balance immunoreactivities of lymphocytes among patients in post-transplant status or with autoimmune diseases [38, 39]. EBV targeted cell therapies using virus-specific T-cells derived from patients’ own lymphocytes or from third party T-cells have been suggested to be an emerging option with favorable outcomes in patients with EBV-related PTLD [40,41,42,43,44]. Since similar efficacy could be expected for EBV-related immunodeficiency-associated LPD other than PTLD, there is a call for clinical trials for refractory/relapsed cases. Off-the-shelf products are awaited to broaden the application of these novel agents.

Our study has several limitations. First, this was a retrospective analysis in a small, heterogeneous population and some factors could not be collected from clinical records. As shown via the indefinite impact of lymphopenia in the subgroup analysis of patients with PTLD, the heterogeneity and small number of cases might have obscured the results of multivariate analysis. A prospective study using a larger cohort is mandatory to confirm the reproducibility of our findings.

Second, although clinically suggestive, our proposed explanation of our results has not yet been proved biologically and a more detailed biological approach is necessary. Third, lymphocyte subsets were not evaluated in this study. To discuss the anti-tumor effects of lymphocytes more in detail, evaluation of T lymphocyte subsets might be of importance.

Conclusion

Lymphopenia at diagnosis may potentially predict inferior OS and PFS in patients with immunodeficiency-associated LPDs. It might reflect the characteristics of the mechanism of disease progression for these distinct lymphoid neoplasms growing under immunocompromised backgrounds. A more detailed analysis in a larger cohort is needed to clarify the tumor pathology of these LPD and to investigate better risk-stratified treatment strategies against them.