Predicting lymphoma development in patients with Sjögren’s syndrome

ABSTRACT

Introduction: The issue of predicting lymphoma in primary Sjögren’s syndrome (pSS) starts from its clinical and biologic essence, i.e., an autoimmune exocrinopathy with sicca syndrome, inflammation and lymphoproliferation of MALT (mucosa-associated lymphoid tissue) in exocrine glands.

Areas covered: The two major predictors to be firstly focused are persistent salivary gland (SG) swelling and cryoglobulinemic vasculitis with related features as purpura and low C4, or the sole serum cryoglobulinemia repeatedly detected. They are pathogenetically linked and reflect a heavier MALT involvement by histopathology, with the expansion of peculiar rheumatoid factor (RF)-positive clones/idiotypes. Other predictors include lymphadenopathy, splenomegaly, neutropenia, lymphopenia, serum beta2-microglobulin, monoclonal immunoglobulins, light chains, and RF. Composite indexes/scores may also predict lymphoma.

Expert opinion: Prediction at baseline needs amelioration, and must be repeated in the follow-up. Careful clinical characterization, with harmonization and stratification of large cohorts, is a relevant preliminary step. Validated and new biomarkers are needed in biologic fluids and tissues. SG echography with automatic scoring could represent a future imaging biomarker, still lacking. Scoring MALT involvement in pSS, as an additional tool to evaluate disease activity and possibly to predict lymphoma, is welcomed. All these efforts are now ongoing within the HarmonicSS project and in other research initiatives in pSS.

KEYWORDS:

• Sjögren’s
• lymphoma
• cryoglobulinemia
• rheumatoid factor
• HCV
• MALT
• ESSDAI
• activity
• swelling
• autoimmunity

1. Introduction

Primary Sjögren’s syndrome (pSS) is an autoimmune and lymphoproliferative connective tissue disease (CTD) [1], and its pathobiological hallmark is the acquisition of mucosa-associated lymphoid tissue (MALT) in exocrine glands [2–4]. Also for this reason pSS was called ‘autoimmune epithelitis’ or ‘autoimmune exocrinopathy’ [5]. Severe manifestations of CTDs are much less frequent [4,6–16] than sicca syndrome in pSS. By contrast, patients with pSS have a greatly increased risk to develop a B cell non-Hodgkin’s lymphoma (NHL), which has a prevalence of about 5% [17–25], and which represents the major cause of patient decreased survival [26].

Although the clinical phenotype may be different in subsets of patients with pSS, the autoimmune and lymphoproliferative involvement of MALT sites, and in particular of the salivary and the lachrymal glands, is a common denominator in disease pathobiology, and leads to the key symptoms of dryness [2–4,27]. This MALT background is typical of pSS, and not of other CTDs. Thus, it is not surprising that the MALT histotype or, in general, the marginal zone histotype, is detected in the vast majority of B cell NHLs arising in pSS. The second, but rarer lymphoma histotype is diffuse large B cell lymphoma, which may also derive from the transformation of a MALT lymphoma [24,28,29]. The very high prevalence of marginal zone lymphoma (84.2%; De Vita S., personal observation) was recently confirmed in a large tri-centric Italian and Greek cohort from Udine, Pisa and Athens (UPA), involving 140 pSS-NHLs [30]. Strikingly, lymphoma localization is often salivary, i.e., a crucial localization of pSS itself.

In pSS MALT lesions, B lymphocytes are continuously stimulated, and epithelial cells play an active pathogenetic role [3–5]. Chronic triggering, immune cells mediating both the innate and adaptive response, cytokines and growth factors, are all present in the local microenvironment [4], and support the proliferation by B cells with different antibody specificity [31,32]. The emergence of rheumatoid factor (RF)-positive clonal B cell expansion is later facilitated, and these cells in turn become more prone to additional events favouring their expansion, with the final possibility of overtly malignant transformation, antigen-selected [2,33–41].

The issue of lymphoma prediction is of extreme importance in pSS, and not only due to the severity of this possible complication, but also to understand the pathophysiology of pSS itself with the available history (and pathologic tissues) from the fully benign to the overtly malignant stages of B cell expansion. Despite several years of efforts, however, the possibility to predict the outcome of pSS at the beginning of the disease, and also during the subsequent follow-up, is still limited. Lymphoma, as a possible outcome of pSS, makes no exception, although important steps to highlight some clinical and laboratory lymphoma predictors have been done over the years.

We will discuss the possible prediction of lymphoma development in pSS in clinical practice, at present. The currently used and easy-to-perform methods to evaluate the risk of lymphoma progression at the onset and during the course of pSS will be presented (Table 1). Secondly, we will focus to some recent observations which could be useful, in our mind, to improve the study of lymphoma predictors. Finally, the most recent proposals and collaborative ongoing research approaches to this topic will be briefly updated, mainly in the HarmonicSS project (HarmonicSS, European Union Grant 731944; https://harmonicss.eu) [42].

Table 1. Lymphoma predictors in pSS

LYMPHOMA PREDICTORS in pSS
1. Major predictors Persistent salivary gland swelling Cryoglobulinemic vasculitis and serum cryoglobulinemia Strictly related predictors Manifestations of cryoglobulinemic vasculitis: purpura, low C4, others (glomerulonephritis, peripheral neuropathy) Heavier MALT involvement by tissue salivary histopathology B-cell clonal expansion with peculiar rheumatoid factor-related genes/idiotypes, and ongoing expansion of the same B-cell clone in metachronous tissue biopsies
2. Other possible predictors
Clinical
Younger age at disease onset Lymphadenopathy Splenomegaly Raynaud’s phenomenon Blood Neutropenia Lymphopenia/CD4^+ T lymphocytopenia Anti-SSA/SSB positivity Rheumatoid factor and its titre Increased serum beta2 microglobulin High levels of immunoglobulin light chains Increased serum levels of Flt3-L Increased serum levels CXCL13 Increased serum levels of TSLP Genetic and tissue Chromosomal translocations and oncogene deregulation TNFAIP3 polymorphisms BAFF polymorphisms BAFF receptor His159Tyr mutation High Interferon IFNγ/IFNα ratio in minor SG biopsy Decreased miR200b-5p in minor SG biopsy Tissue TSLP
3. Composite indexes Lymphoma prediction in all pSS patients [10] Lymphoma prediction in pSS patients with persistent SG swelling [9]
4. Investigational HarmonicSS ○ Novel score for lymphoma prediction ○ Blood, saliva and tissue biomarkers ○ Imaging biomarkers: salivary gland echography ○ Composite index of MALT involvement Other research projects BIG DATA SjÖGREN PROJECT PRECISESADS NECESSITY

2. Predictors of lymphoma risk in pSS

Many lymphoma predictors (clinical, laboratory, and pathologic) have been searched in pSS since the 1970s. The more clinically relevant, as supported by the literature, are persistent salivary gland (SG) swelling and mixed cryoglobulinemia repeatedly detected and/or cryoglobulinemic vasculitis (CV; i.e. the presence of clinical vasculitis in addition to positive serum cryoglobulins) [2,6–13,15,27,43], as well as low C4 and purpura, which may be cryoglobulinemia-related (Table 1). These two main clinical predictors are, in any case, strictly linked to additional clinical, pathologic and laboratory proposed predictors of lymphoma in pSS [6–14]. These include other CV-associated features (glomerulonephritis, peripheral neuropathy) [15], a heavier involvement of MALT by SG biopsy [2–9,43,44], and laboratory features as monoclonal gammopathy and rheumatoid factor and its peculiar idiotypes, which also determine the pathophysiology of cryoglobulinemia [15] (Table 1).

Many other predictors have been also suggested, such as lymphadenopathy, splenomegaly, neutropenia, lymphopenia, free immunoglobulin light chains, increased serum beta2 microglobulin, positive rheumatoid factor in general, genetic abnormalities, oncogenetic events, cytokines and growth factors, chemokines, ongoing monoclonal B-cell expansion in metachronous tissue biopsies, and recently the European League Against Rheumatism (EULAR) Sjögren’s syndrome disease activity index (ESSDAI) [2,6–14,26,44–46] (Table 1).

Of note, it is crucial to know exactly whether a so called ‘predictor’ of lymphoma in pSS may rather relate to the progression to B-cell lymphoma and/or to fully deregulating oncogenetic events. If so, it is not a true predictor, even if very useful to monitor the disease and to catch the progression to a malignant lymphoproliferation, True predictors, by contrast, may be present many years before the lymphoma diagnosis. In addition, predictors may be different in different subgroups of pSS patients: the stratification of pSS is, thus, of crucial importance.

2.1. Persistent salivary gland swelling and cryoglobulinemia: the two major lymphoma predictors, related to tissue inflammation and lymphoproliferation in pSS

The major risk factor for lymphoma evolution in pSS is represented by persistent SG, which has been defined as swelling lasting at least two months [2,27,33]. A number of studies supports this statement [2,6–13,27,43], although how glandular swelling may be better detected still remains elusive. Importantly, SG enlargement is accompanied by gland inflammation, a higher focus score [10,47,48], and a higher degree of lymphoid organization. The organization of the infiltrating lymphoid cells in SG biopsies in pSS may range from unorganized foci to organized foci, with or without germinal centres. The presence of germinal centres may imply a higher risk of lymphoma in pSS, although the issue is still controversial [8,44,49,50]. Furthermore, SG localization is largely overrepresented in pSS-associated lymphomas, consistent with the earliest clinical observations by Talal [1]. In turn, SG swelling is phenotypically and biologically strictly linked to mixed cryoglobulinemia in pSS [15,43].

In pSS, fully benign SG MALT contains the expansion of B cells with different specificity [31], but the B-cell clones heavily overexpanded in pSS-related myoepithelial sialadenitis and MALT lymphoma (i.e., those with a strong selection advantage) often express the same immunoglobulin genes and idiotypes of cryoglobulins, e.g. IGHV1-69, IGKV3-20 and the 17.109 idiotoype, and may originate cryoglobulins [6,33,38–40,51]. In addition, a heavier SG inflammation and lymphoproliferation is significantly related to cryoglobulinemia in pSS [52]. Furthermore, the sole physical excision of a MALT lymphoma expressing these genes (by parotidectomy) was followed by a marked decrease in serum cryoglobulins, and then by their increase concomitantly to lymphoma relapse [43]. Thus, the salivary and, in general, the MALT involvement, i.e., the biologic hallmark of pSS, appears as the substrate of cryoglobulinemia in pSS, and these two main predictors of lymphoma in pSS are definitely linked each other. The MALT involvement appears to be, in any case, the original substrate: consistently, SG swelling, which reflects MALT expansion, is much more frequent than cryoglobulinemia [9,10,27].

One major clinical issue is that while SG swelling is present in about one third of patients with pSS, only a minor part of them will develop lymphoma. The risk of lymphoma was specifically investigated in pSS patients with SG swelling in only one study [9], providing a useful instrument for a further stratification, in clinical practice, of the whole SG swelling subgroup. In this study, analysing the lymphoma risk in a large multicentre Italian cohort in general, patients with the two major lymphoma predictors (CV and SG/parotid enlargement) were investigated in separate groups. All the selected 661 pSS patients, fulfilling the American European Criteria for the classification of SS, were hepatitis C virus (HCV)-negative and repeatedly tested for a careful evaluation of cryoglobulinemia. Patients were categorized into four groups, including pSS with lymphoma, with CV and without lymphoma, with persistent SG swelling without lymphoma (with or without CV), and pSS controls without lymphoma and without CV or SG swelling. In addition, pSS patients with SG swelling were further differentiated, either evolving or not into lymphoma in the follow-up. Within pSS patients with SG swelling, those with at least two additional risk factors among four (cryoglobulinemia, low C4, leukopenia and anti-La/SSB positivity) were at high risk of lymphoma, whereas the positivity of only one or none of the four risk factors provided a negative predictive value for lymphoma of about 90%. Strikingly, cryoglobulinemia and low C4 were selected as additional predictors together with SG swelling [9].

The second major predictor of lymphoma in pSS, besides persistent SG swelling, is the presence of serum cryoglobulins (cryoglobulinemia) and, more consistently, a frank cryoglobulinemic vasculitis – CV, both predictors of lymphoma in pSS [2,7,26,53,54]. The association between mixed monoclonal (type II) cryoglobulinemia and the risk of lymphoma in pSS was first documented by Tzioufas et al. [6], who also reported an increased risk related to the presence of specific monoclonal RF-associated cross-reactive idiotypes (CRI), such as 17109 and G-6. Serum mixed cryoglobulinemia, occurs in about 7–15% patients with pSS, while CV is less frequent (3–7%), though it greatly affects the pSS-related morbidity [7,52,55]. Of note, the current international classification criteria for CV [56] have been further validated both in HCV-positive patients (i.e., the large majority of CV, i.e. about 70–90%, lacking pSS) [57] and in patients with pSS [58], which resulted the most common condition of HCV-unrelated CV [56]. This is consistent with the clinical phenotype showing many similarities [15]. Anyway, clinical differences also exist and, overall, two distinct entities (i.e., HCV-related and pSS-related/HCV-unrelated CV) are associated with mixed cryoglobulinemia and both predispose to B-cell NHL [15,43,59,60]. Of note, phenotypic similarities and differences are also observed between lymphomas complicating the course of HCV-related and pSS-related CV in several studies [51,61]. Furthermore, the similarities and differences are also biologic, since HCV-related cryoglobulinemia is mainly a bone marrow and hepatic lymphoproliferative disorder, while pSS-related cryoglobulinemia depends on lymphoproliferation of MALT [43]. Both in HCV-related and in pSS-related CV, however, there is an expansion of RF-positive B cell clones which employ the same immunoglobulin heavy and light genes [51], indicating common pathogenetic pathways. As a conclusion, HCV infection therefore represents a very good model to study the link between infection, autoimmunity and lymphoproliferation in pSS, where the triggering exogenous antigen(s) and/or autoantigen(s) remain unknown [59,60,62]. Finally, while cryoglobulinemia is a predictor of lymphoma in pSS, there are very limited data about additional predictors in patients who already have cryoglobulinemia [9,63]. Further oncogenetic events definitely occur after the development of cryoglobulinemia itself in the course of HCV infection, since only a fraction of these patients evolves into lymphoma: in turn, this also may occur in pSS.

After the first report on cryoglobulinemia as lymphoma predictor in pSS [6] many other studies focused on this issue. Skoupouli and co-workers [53] reported that the development of pSS lymphoma was associated with the presence of mixed monoclonal cryoglobulins (relative risk, 7.9; p = 0.0012), low levels of C4 complement (relative risk, 7.5; p = 0.0016), and purpura (relative risk, 3.9; p = 0.037), the two latter manifestations being related to cryoglobulinemia. Low levels of C4 were indicated as the strongest predictor for mortality in pSS (relative risk, 6.5; p = 0.0041). Of interest, the Authors observed that the initial presentation of pSS with these manifestations could determine the outcome and mortality of pSS itself [53].

Hypocomplementemia as a risk factor of lymphoma in pSS is herein discussed together with cryoglobulinemia and CV, since these are pathogenetically related to complement consumption [15]. Hypocomplementemia is closely associated with the two main adverse outcome scenarios in pSS, i.e., lymphoma development and death [53]. In addition, low complement, as a lymphoma predictor, may be caused by an immune complex-mediated complement consumption also without detectable cryoglobulins. In fact, rheumatoid factor positivity itself, not necessarily linked to cryoprecipitability, may be as well related to the formation of immune complex and to complement consumption, and pSS-related lymphomas express B-cell receptor sequences with RF reactivity [33,36,40]. Thus, both peculiar rheumatoid factor specificities/idiotypes and low C4 deserve additional evaluation as relevant predictors of lymphoma in pSS, even without cryoglobulinemia.

2.2. Other risk factors

There are less data in the literature as concerns additional proposed risk factors for lymphoma in pSS.

Lymphopenia was reported as predictor of lymphoma in a cohort study by Theander et al. [64] who identified CD4⁺/CD8⁺ T cell ratio ≤0.8 (HR = 10.92, 95% CI 2.80 to 41.83) as the strongest risk factor for developing lymphoma in pSS. Other predictors in this study were CD4⁺ T lymphocytopenia (HR = 8.14, 95% CI 2.10 to 31.53), low C4 (HR = 9.49, 95% CI 1.94 to 46.54), purpura/skin vasculitis (HR = 4.64, 95% CI 1.13 to 16.45) and decreased C3 (HR = 6.18, 95% CI 1.57 to 24.22). Of note, a diffuse large cell lymphoma histotype was more common than a marginal zone histotype in this cohort [64]. Consistently, Baimpa et al. reported that CD4⁺ lymphocytopenia was the only independent variable predicting a lymphoma histotype different from marginal zone lymphoma in pSS, the major histotype being again diffuse large B cell NHL [7].

Among the other possible hematologic manifestations of pSS, Baimpa et al. reported neutropenia as a lymphoma risk factor in pSS (HR 8.9), correlating with splenomegaly [7]. In general, neutropenia appears as one laboratory factor able to predict the transformation of IgM disorders into lymphoid malignancy [65]. The role of anaemia has been sporadically suggested [10,12].

Data regarding lymphadenopathy and splenomegaly as risk factors of lymphoma in pSS are scanty [7], may be difficult to be assessed if mild and inconstant, and finally a lymphoma evolution, rather than a true lymphoma prediction, may be also hypothesized.

High levels of serum beta2 microglobulin, of serum and urine immunoglobulin light chains ad their kappa/lambda ratio may also correlate with the burden of proliferating B cells [66–68].

More recently, by comparing the clinical and laboratory features in pSS patients positive or negative for the anti-Ro/SSA and/or anti-La/SSB antibodies, it was found that anti-SSA/SSB-negative pSS showed a lower risk of lymphoma evolution [16].

Moreover, a younger age at the pSS onset seems associated with an increased risk for lymphoproliferation [69,70].

Genetic abnormalities, oncogenetic events, the involvement of NF-kB pathway, of cytokines, growth factors and chemokines [8,44,47,71–85] and the ongoing expansion in tissue lesions of the same B-cell clone [33], are also linked to the increased risk of lymphoma development or to multistep lymphomagenesis in pSS (reviewed in [86] and in [26]). They include B cell activating factor (BAFF) and Tumor necrosis factor alpha induced protein 3 (TNFAIP3) variants, BAFF receptor His159Tyr mutation, chromosomal translocations and oncogene deregulation, increased serum levels of Fms-related tyrosine kinase 3 ligand (Flt3-L), of C-X-C motif ligand 13 (CXCL13), and high Interferon (IFN) γ/IFNα ratio in minor SG biopsy. They are not however, at present, easy-to-access tests in current clinical practice, and are not herein discussed.

2.3. Composite indexes and scores

Composite indexes and scores might improve the prediction of lymphoma development in pSS [7,9,10,87], although focusing on single predictors in the single individual patient remains a useful preliminary step in the clinical approach to the patient. Composite indexes might overweight and/or mainly evaluate the most important predictors. Whether/which additional predictors might be added, and how much they should be weighted, is also a question. Notably, SG swelling and cryoglobulinemia (or monoclonal gammopathy related to it) are included in all the currently available composite indexes for lymphoma prediction in pSS.

The attempt to stratify pSS patients on the basis of their lymphoproliferative risk was assessed for the first time by Ioannidis et al. [87], and confirmed the crucial role of cryoglobulinemia. The presence of parotid enlargement (hazard ratio – HR 5.21, 95% CI 1.76–15.4), palpable purpura (HR 4.16, 95% CI 1.65–10.5) and low C4 levels (HR 2.40, 95% CI 0.99–5.83) at the first visit resulted as independent predictors of lymphoproliferation in pSS. The Authors then proposed a stratification of pSS in two distinct risk categories, according to the presence or absence of low C4 levels and/or palpable purpura (both related to cryoglobulin positivity): the type I high-risk pSS, that accounts of about 20% of pSS at diagnosis, and the more frequent (80%) type II low-risk pSS phenotype, without a significant increase in risk of lymphoma evolution [87]. Brito-Zeròn and co-workers [55], in their cohort analysis, confirmed these data and added to the model the additional major risk factors of lymphoma development, i.e., severe parotid involvement, demonstrated by scintigraphy, and cryoglobulinemia. A survival analysis performed in this cohort found that pSS patients with at least two adverse factors at diagnosis among four, i.e. parotid involvement by scintigraphy, cryoglobulinemia, purpura and low C4, had a significantly lower survival than patients without risk factors.

A subsequent retrospective study investigated hematologic manifestations and predictors of lymphoma in 536 pSS patients (40 with lymphoma) [7]. Among 5 risk factors identified by multivariate analysis as predictors (cryoglobulinemia, low C4 levels, neutropenia, splenomegaly and lymphadenopathy) the proportion of pSS patients developing lymphoma by number of risk factors was 3.62%, 11.96%, 34.78%, 80% and 100% in patients with 0, 1, 2, 3, and 4 risk factors, respectively.

Finally, a more recent case control study by Fragkioudaki et al. [10] compared 92 pSS patients developing lymphoma in the follow-up to 381 pSS controls without this evolution. Distinct multivariate analyses disclosed independent clinical, laboratory and histopathological risk factors for lymphoma in pSS, and a final multivariate analysis of all variables found seven independent lymphoma risk factors, the highest significance being for SG enlargement. Other risk factors included not only items which may be cryoglobulinemia-related, i.e., low C4, monoclonal gammopathy and positive rheumatoid factor, but also lymphadenopathy, Raynaud’s phenomenon, anti-Ro/SSA and/or anti-La/SSB positivity. Development of lymphoma was 3.8% for patient with ≤ 2 risk factors, 39.9% for patients with 3 to 6 risk factors, and 100% in the presence of all the 7 risk factors [10].

By contrast, there is only one study where the risk of lymphoma was evaluated in pSS patients who already have a major risk factor, represented by SG enlargement [9]. When comparing 22 pSS patients with persistent SG swelling who developed lymphoma in the follow-up with 19 pSS patients with SG swelling who did not develop it after ≥ 5 years of follow-up, at least 2/4 among serum cryoglobulins, low C4, leukopenia and anti-La/SSB positivity provided a 10.2 odds ratio for lymphoma development (12/22 vs. 2/19 patients correctly classified) [9].

3. Recent observations to improve lymphoma prediction studies

3.1. An improved evaluation of disease activity of pSS is needed

Persistent SG swelling and mixed cryoglobulinemia [6–13,43] well reflect the degree of inflammation and lymphoproliferation of MALT in pSS [2–9,43]. Thus, they reveal the activity of pSS itself [1–4]. In turn, increased activity of pSS is conceivably linked to a higher lymphoma risk. SG swelling and cryoglobulinemia were then analysed in detail [27], in the follow-up pSS patients developing lymphoma (30 cases) or not (225 pSS controls), to assess whether it is better to investigate definite lymphoma predictors rather than the ESSDAI, i.e., a putative lymphoma predictor [14] including a number of lymphoma-unrelated or poorly-related items. Of note, SG swelling and/or cryoglobulinemia at baseline were significantly higher (p = 0.0003) in pSS patients evolving into lymphoma if compared to pSS controls, while the ESSDAI score showed no significant difference [27].

That ESSDAI is not a good lymphoma predictor was also shown in two different studies in independent large cohorts [10,11], besides the aforementioned one [27]. Of note, about one third of pSS patients who will develop lymphoma shows a low activity, as calculated by ESSDAI, at baseline [10,11,27].

Therefore, the concept of the priority to consider the key disease manifestations and biomarkers predicting lymphoma in pSS, rather than a generical index of disease activity including lymphoma-unrelated items, clearly emerges. Consistently, the possible increase in ESSDAI in pSS patients developing lymphoma appears to be mainly linked to the concomitant positivity of true lymphoma predictors (i.e., SG enlargement, cryoglobulinemia, purpura and low C4) and/or to possible lymphoma manifestations (lymphadenopathy) [27].

If pSS activity is evaluated with a composite index scoring the extent of MALT involvement, such pSS activity index might possibly prove useful for lymphoma prediction in pSS, where lymphoma usually arises from MALT.

3.2. The risk of lymphoma may change over time

A second important concept recently underlined [27] is that the risk of lymphoma evolution may change over time in pSS, i.e., the risk at different time points in the follow-up at may differ with that at baseline. As an example, SG swelling or cryoglobulinemia sometimes are lacking at baseline and develop ex novo in the follow-up of pSS. This is now under more extensive investigation in larger joined cohorts.

A dynamic approach is therefore recommended at present, requiring the re-evaluation of lymphoma risk factors by the general practitioner and the medical specialist at any follow-up visit of the pSS patient, also with low ESSDAI values (see Section 3).

3.3. Laboratory biomarkers will be crucial to improve lymphoma prediction

Laboratory biomarkers may be investigated either in biologic fluids (blood, saliva, urine) or directly in pathologic tissue biopsies in pSS [88]. Within the HarmonicSS project (European Union Grant 731944; https://harmonicss.eu), the investigation of biomarkers is one major task to improve patient stratification. A shorter list of laboratory biomarkers to be mainly focused within HarmonicSS was agreed by Experts during the last meeting of the EULAR Sjögren’s syndrome experimental and translational investigative alliance study group (ESSENTIAL) held at the 2018 annual EULAR congress, Amsterdam. Most of the enclosed biomarkers may relate to disease activity and to the risk of lymphoma evolution.

These include the so called ‘old biomarkers’, i.e., those proposed by several studies in the past (beta2 microglobulin, free light chains, RF), the so called ‘old-new’ biomarkers, i.e. those more recently proposed by different groups by preliminary studies (BAFF, CXCL13, Flt3-L, TNFAIP3 polymorphisms, IFN signature), and two definitely novel biomarkers proposed by single groups within HarmonicSS, i.e. miRNA200b-5p and thymic stromal lymphopoietin (TSLP) [89,90]. In a very recent paper [91], low miRNA200b-5p was reduced in prelymphoma and lymphoma pSS patients, proving to be a useful predictor. TSLP increases in the serum of pSS patients, with higher levels in the myoepithelial sialadenitis and mixed cryoglobulinemia, and with the highest level in patients with B-cell lymphoma [92]. Importantly, a concomitant increased expression in tissue B-lymphocytes, maximal in lymphomatous ones, is noticed [93].

These biomarkers will be validated in the large harmonized cohort of pSS patients within the HarmonicSS study. Of course, the simultaneous investigation of a much wider panel of possible laboratory biomarkers is equally needed, and is also planned within HarmonicSS itself and other cooperative research projects in pSS (see Section 3).

4. Current novel initiatives

Different ongoing cooperative initiatives in Europe are expected to contribute to the issue of lymphoma prediction, including, besides HarmonicSS, the BIG DATA Sjögren project [94,95], the PRECISESADS study (European Union Grant 115565; https://www.precisesads.eu), and the NECESSITY project (European Union Grant 806975).

Evaluation of lymphoma risk is a major goal of the HarmonicSS project, and the following current initiatives have been taken in it:

1. The development of a novel composite index for lymphoma prediction in pSS, based on harmonized data in the largest expected cohort of cases and controls, is one of the key clinical unmet needs in HarmonicSS. Both traditional statistical methods and novel engineer-based methodologies [42,96] will be used to achieve the best result. Two previous scores for lymphoma prediction, i.e. the more recent one in the largest series [10] and the only one available in patients who already have persistent SG swelling [9], will be also validated.

2. The investigation of very large array of serum and plasma biomarkers for pSS has been planned, including possible novel ones and those which could be more implicated for lymphoma prediction. The use of a unique, customized kit, and also saliva biomarkers, are under investigation.

3. The development of a possible imaging biomarker will be investigated in pSS, by means of SG ultrasound [97]. This will then occur not only for diagnostic purposes, but also to improve pSS follow-up, including lymphoma prediction. To this end, very novel techniques, including image segmentation and artificial intelligence, are being applied to this powerful, imaging tool [97,98].

4. A final task of HarmonicSS is the development of a composite index of MALT involvement as an additional tool to evaluate disease activity in pSS. As said, MALT inflammation and lymphoproliferation represent the biologic essence of pSS. Many pSS patients cannot be recruited in clinical trials, despite having sicca symptoms, since the ESSDAI is low. This does not exclude, however, biologic and tissue disease activity in pSS [27]. The MALT composite index for pSS will be developed based on histopathology as the golden standard for the definition of the extent and the quality of tissue inflammation and lymphoproliferation. One novel concept is to simultaneously investigate, within the large harmonized pSS cohort, additional ‘surrogates’ of MALT histopathology, i.e., the different possible clinical, laboratory and imaging abnormalities significantly correlated. These may include SG swelling and CV (clinical), cryoglobulinemia, rheumatoid factor positivity and titre, low C4, and other tests in biologic fluids (laboratory), and SG abnormalities by ultrasound (imaging). Finally, the composite index, based both on bioptic and ‘surrogate’ items, will be developed and then validated.

5. Expert opinion

Primary Sjögren’s syndrome is an autoimmune and lymphoproliferative connective tissue disease. Its clinical hallmark is the sicca syndrome, and its pathobiological hallmark is the acquisition of mucosa-associated lymphoid tissue (MALT) in exocrine glands. The issue of predicting lymphoma in pSS should start from these clinical and biological cornerstones.

Lymphoma prediction is relevant, both at baseline and in the follow-up assessment of pSS, for two main reasons. A clinical reason, since lymphoma represents the main cause of increased mortality, and a biologic reason, since it gives the rare possibility to investigate prospectively the different multistep pathogenetic events implicated in lymphoproliferation, and to pSS itself.

Many lymphoma predictors (clinical, laboratory and pathologic) have been highlighted in pSS. Only two of them, however, may be currently considered as major predictors, since they are the ones more consistently reported in the literature and also well linked to MALT autoimmunity and lymphoproliferation in pSS: they are persistent salivary gland (SG) swelling, and cryoglobulinemic vasculitis (CV; with related clinical and laboratory features such as purpura and low C4). Repeatedly detected cyoglobulinemia, even without overt clinical features of vasculitis, should be also considered in pSS.

Of note, persistent SG swelling and cryoglobulinemia reflect a heavier MALT acquisition in pSS by pathologic studies, which has been also related to an increased lymphoma risk.

Although pSS-related fully benign SG MALT contains B cells with different antigen specificity, the B-cell clones using peculiar immunoglobulin genes linked to rheumatoid factor (RF) specificity have a strong selection advantage, and are those overexpanded in pSS-related myoepithelial sialadenitis and MALT lymphoma. Then, RF-related peculiar clones/idiotypes are implicated in pSS-related lymphomagenesis, possibly also in the lack of RF-related cryoglobulinemic activity. Overall, and consistent with the clinical finding of SG swelling being more frequent than cryoglobulinemia both in pSS and in pSS-related lymphomas, the MALT acquisition, with possible SG swelling, appears as the earliest pathogenetic event. Consequently, the evaluation of MALT in pSS is a starting point to predict lymphoma and to ameliorate this prediction in the future.

At present, all the pSS patients should be carefully evaluated with clinical, laboratory and imaging tools for present and past SG swelling, and CV features or cryoglobulinemia, also by careful laboratory testing. Other reported lymphoma predictors should be then studied, including clinical ones (lymphadenopathy, splenomegaly) and laboratory ones (neutropenia, lymphopenia, increased serum beta2 microglobulin, immunoglobulin monoclonal components and light chains in biologic fluids, and RF, in particular its higher titres).

Two composite indexes/scores currently available to predict lymphoma in the general pSS population or in the pSS patients who already suffer of SG swelling may be finally used. Conversely, the ESSDAI score does not appear useful to predict lymphoma. The evaluation in a reference Centre for pSS would be required if an increased lymphoma risk is suspected, with patients also undergoing thorax computed tomography, abdomen and pelvic ultrasound, the study of gastric MALT, and bone marrow biopsy/aspiration if cryoglobulinemia is present. HCV infection should be always searched.

Recent observations and cooperative initiatives to improve lymphoma prediction in pSS in the future are crucial, and include:

• a dynamic approach requiring the re-evaluation of lymphoma risk factors by the general practitioner and the medical specialist at any follow-up visit, also in pSS with low ESSDAI values;

• the importance of validated and novel laboratory biomarkers in biologic fluids (blood, saliva, urine) and directly in pathologic tissue biopsies in pSS. Their more widespread use will be needed in the future;

• the importance of large multicentric collaborations, preceded by a very careful patient characterization. To this end, the preliminary harmonization and stratification of large pSS cohorts is crucial, and is now in course within the HarmonicSS (European Union Grant 731944; https://harmonicss.eu);

• the need also of a powerful imaging biomarker, which could be easily repeated during the follow-up, and showing a good reproducibility. SG ultrasound with the use of automatic scoring and artificial intelligence is being investigated to this end in HarmonicSS;

• finally, the need of an easy-to-perform composite index of MALT involvement in pSS, as an additional tool to evaluate disease activity and possibly to predict lymphoma in pSS.

Article highlights

• The clinical hallmark of pSS is sicca syndrome, and the pathobiological hallmark is the acquisition of MALT (mucosa-associated lymphoid tissue) in exocrine glands, predisposing to lymphoma.

• Although many lymphoma predictors (clinical, laboratory and pathologic) have been suggested in pSS, only two of them may be currently considered as major predictors: persisting salivary gland swelling and mixed cryoglobulinemia, which are pathogenetically linked.

• B-cell clones linked to rheumatoid factor specificity are implicated in pSS-related lymphomagenesis, and they have a strong selection advantage within a polyclonal, benign B-cell expansion. They may either show or lack a rheumatoid factor-related cryoglobulinemic activity.

• An easy-to-perform composite index of MALT involvement in pSS is needed, as an additional tool to evaluate activity and possibly to better predict lymphoma in pSS. Clinical, laboratory and imaging (salivary gland ultrasonography) tools relate to salivary gland histopathology, and might represent surrogates of it in pSS. By contrast, ESSDAI does not predict lymphoma.

• The preliminary harmonization and stratification of large pSS cohorts, now in course within the HarmonicSS project (European Union Grant 731944; https://harmonicss.eu), is a crucial preliminary step for further clinical and applied research.

Declaration of interest

S. De Vita has acted as Consultant Expert for Roche, Human Genome Science, Glaxo Smith Kline and Novartis. Doctor S. Gandolfo has no conflicts of interest to declare. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.