SEARCH

SEARCH BY CITATION

Keywords:

  • extranodal marginal zone lymphoma;
  • incidence;
  • lymphoma;
  • ocular region;
  • subtypes

Abstract

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

With a lifetime risk of 1% and 700 new cases per year, Non-Hodgkin lymphoma (NHL) is the seventh most frequent type of cancer in Denmark. The incidence of NHL has increased considerably in Western countries over the last decades; consequently, NHL is an increasing clinical problem.

Ophthalmic lymphoma, (lymphoma localized in the ocular region, i.e. eyelid, conjunctiva, lacrimal sac, lacrimal gland, orbit, or intraocularly) is relatively uncommon, accounting for 5%–10% of all extranodal lymphomas. It is, however, the most common orbital malignancy.

The purpose of this thesis was to review specimens from all Danish patients with a diagnosis of ophthalmic lymphoma during the period 1980–2005, in order to determine the distribution of lymphoma subtypes, and the incidence- and time trends in incidence for ophthalmic lymphoma. Furthermore, an extended analysis of the most frequent subtype, extranodal marginal zone lymphoma (MALT lymphoma), was done to analyse clinical factors and cytogenetic changes with influence on prognosis.

A total of 228 Danish patients with a biopsy-reviewed verified diagnosis of ocular adnexal-, orbital-, or intraocular lymphoma were identified. We found that more than 50% of orbital- and ocular adnexal lymphomas were of the MALT lymphoma subtype, whereas diffuse large B-cell lymphoma (DLBCL) predominated intraocularly (Sjo et al. 2008a). Furthermore, lymphoma arising in the lacrimal sac was surprisingly predominantly DLBCL (Sjo et al. 2006). Incidence rates were highly dependent on patient age. There was an increase in incidence rates for the whole population from 1980 to 2005, corresponding to an annual average increase of 3.4% (Sjo et al. 2008a).

MALT lymphoma arising in the ocular region was found in 116 patients (Sjo et al. 2008b). One third of patients had a relapse or progression of disease after initial therapy and relapses were frequently found at extra-ocular sites. Overall survival, however, was not significantly poorer for patients with relapse.

Furthermore, we found that the frequency of translocations involving the MALT1- and IGH-gene loci is low in ocular region MALT lymphoma (2 of 42, 5%), but may predict increased risk of relapse (Sjo et al. 2008b).

In conclusion the incidence of ophthalmic lymphoma is increasing at a high rate in Denmark. Ophthalmic lymphoma consists primarily of MALT lymphoma. The molecular pathogenesis of MALT lymphoma arising in the ocular region rarely involves translocations in the MALT1- and IGH-gene loci.

Acta Ophthalmologica Thesis http://www.blackwellpublishing.com/aos Ophthalmic Lymphoma: Epidemiology and Pathogenesis Lene Dissing Sjö, MD

This thesis is based on the following papers referred to in the text by their Roman numerals:

  • I. 
    Sjo, L.D., Ralfkiaer, E., Prause, J.U., Petersen, J.H., Madsen, J., Pedersen, N.T. and Heegaard, S. (2008). Increasing Incidence of Ophthalmic Lymphoma in Denmark from 1980 to 2005. Invest. Ophthalmol. Vis. Sci; 49(8):3283–3288.
  • II. 
    Sjo, L.D., Ralfkiaer, E., Juhl, B.R., Prause, J.U., Kivela, T., Auw-Haedrich, C., Bacin, F., Carrera, M., Coupland, S.E., Delbosc, B., Ducrey, N., Kantelip, B., Kemeny, J.L., Meyer, P., Sjo, N.C. and Heegaard, S. (2006). Primary lymphoma of the lacrimal sac: an EORTC ophthalmic oncology task force study. Br. J. Ophthalmol. 90, 1004–1009.
  • III. 
    Sjo, L.D., Heegaard, S., Prause, J.U., Petersen, B.L., Pedersen, S. and Ralfkiaer, E. (2008). Extranodal Marginal Zone Lymphoma in the Ocular region – clinical, immunophenotypical and cytogenetical characteristics. Epub ahead of print. Invest. Ophthalmol. Vis. Sci Aug. 2008.

Supervisors

Steffen Heegaard, Associate Professor, MD, DMSc Eye Pathology Institute, Dept. of Neuroscience and Pharmacology, University of Copenhagen

Elisabeth Ralfkiær, Professor, MD, DMSc, Dept. of Pathology, Copenhagen University Hospital

Jan Ulrik Prause, Professor, MD, DMSc, Eye Pathology Institute, Dept. of Neuroscience and Pharmacology, University of Copenhagen

This thesis is available online at: http://www.blackwell-synergy.com/toc/ao/87/thesis1

Contents
Acknowledgementsiv
Abstract1
Introduction1
Background2
 Non-Hodgkin lymphoma2
 Site-specificity of NHL2
 Immunology of the eye and its surrounding tissues2
 Immunology of the orbit and ocular adnexa2
 Immunology of the eye3
 Ophthalmic lymphoma4
 Lymphoma subtypes4
 Lymphoma in the lacrimal sac4
 Incidence of ophthalmic lymphoma4
  MALT lymphoma4
 Structural chromosomal changes in MALT lymphoma5
 NF-κB activation – the unifying concept for MALT lymphomagenesis5
Hypotheses and aims of the study6
 Hypotheses6
 Aims6
Material and methods7
 Material7
 Methods7
 Clinical data7
 Histopathology and immunohistochemistry8
 Detection of translocations by fluorescence in situ hybridisation (FISH)8
 Statistics9
Results10
 Ophthalmic lymphoma (I)10
 Ophthalmic lymphoma subtypes and clinical characteristics10
 Incidence of ophthalmic lymphoma10
 Lymphoma arising in the lacrimal sac (II)10
 MALT lymphoma arising in the ocular region (III)10
 Presence of MALT lymphoma specific translocations10
 Factors associated with prognosis of ocular region MALT lymphoma11
General discussion11
 Lymphoma characteristics and aetiology according to localization11
 Incidence of ophthalmic lymphoma14
 Genetic alterations of MALT lymphoma in the ocular region14
 Clinical characteristics15
 Treatment of ophthalmic lymphoma16
 Prognostic factors in MALT lymphoma arising in the ocular region16
Conclusions and perspectives16
Summaries17
 Danish summary17
References17

Acknowledgements

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

The studies included in this thesis were carried out during my employment as a research assistant at the Eye Pathology Institute, Department of Neuroscience and Pharmacology, Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark 2005–2008.

First of all I would like to thank my supervisors, Steffen Heegaard, Elisabeth Ralfkiær and Jan Ulrik Prause for guiding, supporting and inspiring me professionally and personally through all steps in the project period.

I am greatfull to O.A. Jensen for always offering his linguistic expertise.

My thanks go also to Sanni Pedersen for conducting FISH analyses and to Bodil Laub Petersen for her collaboration in planning and interpreting these analyses. For guidance in statistics, Jørgen Holm Petersen is to be thanked.

My special thanks go to Lise Mette Gjerdrum, Vera Timmermans Wielenga, Eric Santoni- Rugiu, Ben Vainer, Christian Bjørn Poulsen and other colleagues from the Department of Pathology, Rigshospitalet for scientifically fruitful and socially very pleasant meetings. They have all made research so much more fun, and I feel privileged to be their future colleague. My friendship with colleagues from the Eye Pathology Institute and from Topotarget and Experimental Pathology Unit at the Biocenter has also been invaluable.

Colleagues from pathology departments throughout Denmark and Europe are to be thanked for very friendly collaboration in providing specimens for the study. Also colleagues from haematology and ophthalmology departments have provided important clinical information.

Last, but not least my thanks go to all my family, but in particular to Nicolai, Astrid, Frederik and Villiam, for making the world go round.

This study was made possible by grants from The Danish Cancer Society, The Velux Foundation, The Danish Foundation for Cancer Research, The Gangsted Foundation, Direktør Jens Aage Sørensen og hustru Ingeborg Sørensens Mindefond, Øjenfonden, Øjenforeningen Værn om Synet, The Augustinus Foundation, Johs. Clemmesens Forskningsfond, Ingeniør August Frederik Wedel Erichsens Legat and Købmand Kristjan Kjær og hustru Margrethe Kjær født la Cour-Holmens Fond.

Introduction

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

With a lifetime risk of 1% and 700 new cases per year, non-Hodgkin lymphoma (NHL) is the seventh most frequent type of cancer in Denmark (Statistics Denmark 2007; The Danish Cancer Society 2008). The incidence of NHL has increased considerably in western countries over the last decades (Groves et al. 2000; Howe et al. 2001; Clarke & Glaser 2002); consequently, NHL is an increasing clinical problem.

Ophthalmic lymphoma, (lymphoma localized in the ocular region, i.e. eyelid, conjunctiva, lacrimal sac, lacrimal gland, orbit or intraocularly) is relatively uncommon, accounting for 5–10% of all extranodal lymphomas (Freeman et al. 1972). It is however, one of the most common orbital malignancies (Johansen et al. 2000; Shields et al. 2004), and the extranodal marginal zone lymphoma [mucosa-associated lymphoid tissue (MALT) lymphoma] is the dominant lymphoma subtype in the orbit and ocular adnexa. Extranodal marginal zone lymphoma is an extranodal lymphoma considered to be the neoplastic counterpart of the marginal zone cells in reactive follicles. It is a disease of the elderly, characterized by an indolent course.

Recently, several structural chromosomal abnormalities have been demonstrated in MALT lymphoma. They include t(14;18)(q32;q21) involving IGH and MALT1, t(11;18)(q21;q21) involving API2 and MALT1, t(1;14)(p22;q32) involving Bcl-10 and IGH, and t(3;14)(p14;q32) involving FOXP1 and IGH (Inagaki 2007). All of these, except the translocation involving FOXP1, lead to formation or up-regulation of proteins (API2-MALT1, MALT1 and Bcl-10) that ultimately target the same signalling pathway (NF?B) (Lucas et al. 2001), indicating that these translocations are important for the pathogenesis of MALT lymphoma.

The frequency of each of these abnormalities varies with the anatomic site of the lymphoma (Streubel et al. 2004). Furthermore, studies assessing abnormalities in ophthalmic MALT lymphoma have reported varying frequencies of the translocations (Takada et al. 2003; Adachi et al. 2004; Streubel et al. 2004, 2005; Ye et al. 2005). In gastric MALT lymphoma the occurrence of t(11;18) has proven to be of prognostic significance (Liu et al. 2001). In ophthalmic MALT lymphoma the occurrence and prognostic- as well as pathogenic relevance of translocations are still to be determined.

Most of the previous studies investigating clinicopathological characteristics, survival data and prognostic factors in ophthalmic lymphoma have been based on unselected groups of patients with different lymphoma subtypes (Jenkins et al. 2003; Meunier et al. 2004; Sullivan et al. 2005). Because of the great diversity in clinical behaviour and prognosis of different lymphoma subtypes, analyses focusing on each subtype are needed.

Background

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

Non-Hodgkin lymphoma

Non-Hodgkin lymphoma is a malignant neoplasm derived from a clonal proliferation of B- or T-lymphocytes. It is a heterogenous group of more than 30 different subtypes that can arise in extranodal tissue and in lymph nodes.

Classification of lymphoma has changed several times during the last decades. Until 1994, where the Revised European American Lymphoma (REAL) classification was published (Harris et al. 1994), there was no global consensus on which classification to use. The REAL classification and more recently the World Health Organization (WHO) classification (Jaffe et al. 2001) are anchored in the perception that NHL subtypes mimic normal stages of B- (and T-) cell differentiation. Thus, distinctive morphological- and immunophenotypical characteristics along with molecular- and clinical features, allow them to be classified according to their postulated cell of origin (Fig. 1).

image

Figure 1.  Schematic illustration of the normal differentiation of B cells in a lymphoid follicle, constituted by the germinal centre, the mantle zone and the surrounding marginal zone. Based on morphological- and immunophenotypic characteristics along with molecular- and clinical features, B-cell NHL can be classified according to their postulated cell of origin.

Download figure to PowerPoint

Site-specificity of NHL

Lymphocytes recirculate continuously, from blood to secondary lymphoid tissue and back via efferent lymphatics. The exit-point from blood to lymphoid tissue of a given lymphocyte is determined by a tissue-specific homing signature on its cell surface, enabling the cell to recognize and leave the blood at specialized endothelial sites expressing the relevant ligands (Roos 1991). When lymphocytes are exposed to antigen in the lymphoid tissue, their exit from the lymphoid tissue is blocked and a clonal expansion with differentiation into a specialized effector- or memory B-cell of the antigenspecific lymphocyte begins. Additionally, the homing signature of the lymphocyte is revised to a unique combination of adhesion- and chemokine receptors, allowing preferential exit from the blood into the type of lymphoid tissue where the initial activation took place (Pals et al. 2007).

As lymphoma cells represent malignant counterparts of lymphocytes with preservation of most of their physiological behaviour, the same molecular mechanisms as described for lymphocytes guide the tissue specific homing and dissemination of lymphoma (Pals et al. 2007). Consequently, lymphomagenesis is site specific, effectuated through the underlying biology and function of the resident lymphoid tissue that generated the specific lymphocyte homing-signature. In this respect the eye and its surrounding tissues differ immensely and will be considered separately in the following section.

Immunology of the eye and its surrounding tissues

Immunology of the orbit and ocular adnexa

Immunity to protect the eye from invading pathogens is prerequisite for the preservation of vision. The cornea is the most important structure of the ocular surface for maintenance of visual function. It consists mainly of dense connective tissue without lymphoid cells. Therefore, for protection from invading pathogens the cornea depends on the conjunctiva as well as soluble factors provided through the tear film by the lacrimal gland (McClellan 1997). The conjunctiva is composed of a loose connective tissue layer covered by a thin layer of epithelium forming a continuous mucosal surface from the eyelid margin. It covers the posterior eyelid surfaces and the external surface of the sclera, inserting at the corneal limbus (Fig. 2).

image

Figure 2.  Schematic topography of the conjunctival zones. The mucosa of the ocular surface is formed by the conjunctiva and cornea. The conjunctiva is the major support tissue for the preservation of the optical function of the cornea. It has different topographical zones, with a distinct morphology (i.e. from the lid margin: marginal, tarsal, orbital, fornical, bulbar and limbal). For immune protection it is equipped (unlike the cornea) with resident lymphoid cells belonging to the mucosal immune system of the body; the figure shows lymphocytes in black and plasma cells in blue. From (Knop & Knop 2005, with permission from Journal of Anatomy).

Download figure to PowerPoint

The connective tissue of the conjunctival stroma contains numerous lymphocytes, dendritic cells and IgA secreting plasma cells, with the highest density (sometimes forming conjunctival lymphoid follicles) in regions covering the posterior eyelid surfaces (Knop & Knop 2000; Chodosh & Kennedy 2002). Such IgA secreting lymphoid components within a mucosal surface are referred to as MALT. They reside at critical anatomic positions where exposure to pathogens is manifest, such as the respiratory and digestive tract. Consequently, the conjunctival lymphoid tissue is encompassed in the common mucosal immune system.

Lymphoid follicles are also found in the caruncle, the lacrimal gland, and in the lacrimal drainage system (Allansmith et al. 1987; Knop & Knop 2001). Hence, the conjunctiva together with the lacrimal gland and the lacrimal drainage system form a functional MALT unit (termed eye-associated lymphoid tissue). This appears to have special importance for corneal immunity (Knop & Knop 2005) (Fig. 3).

image

Figure 3.  Mucosa associated lymphoid tissue of the eye (EALT) at the ocular surface and lacrimal drainage system. Eye-associated lymphoid tissue consists of lymphoid tissue that is continuous (blue tissue outline) from the lacrimal gland throughout the conjunctiva into the lacrimal drainage system. It is functionally connected by the flow of tears over the surfaces (yellow arrows over conjunctiva and lacrimal drainage system). Follicles in EALT allow the detection of ocular surface antigens and generate the formation of specific effector cells (e.g. plasma cells). (From Knop & Knop 2005, with permission from Journal of Anatomy).

Download figure to PowerPoint

The connective tissue and the other components of the orbit are under normal circumstances devoid of lymphocytes (van der Gaag 1988). The presence of lymphatic capillaries in the orbit has been questioned until recently, where evidence of lymphatic vessels has been found, although only in the outer layers of dura surrounding the optic nerve (Dickinson & Gausas 2006). In addition, a part of the lacrimal gland (the orbital lobe), containing MALT and numerous lymphatics is located in the orbit.

Immunology of the eye

The uveal tract, comprising the iris, ciliary body and choroid, represents the vascular organ of the eye. In addition to providing most of the blood supply to the intraocular structures, lymphocytes enter the eye from the uvea. The consequences of inflammation in the eye might be critical, because a slight loss of clarity or micro-anatomical deformation along the visual axis may be sight threatening (Chang et al. 2006). However, the eye is an immune privileged organ, i.e. an organ in which certain immune mechanisms are inhibited, and thus some hostile inflammatory reactions are repressed.

Immune privilege is characterised by dynamic interactions between the immune system and specialized tissues and is comprised of several features (Streilein 1999). The function of the continuous layer of retinal pigment epithelium (RPE), as an immunologic blood–retina barrier, is the most important factor for understanding lymphomagenesis of the eye. The RPE support the photoreceptor cell layer of the retina. Their intercellular connections via tight junctions form a shield that prevents blood-borne cells and molecules from entering the subretinal space (Streilein 1993, 2003). Furthermore, the expression of Fas ligand (FasL) may play a role in constituting an immunologic barrier through Fas–FasL interaction, inducing apoptosis in activated T-cells (Jorgensen et al. 1998).

The absence of lymphatic vessels in the eye is also of importance. The vast majority of aqueous humour is leaving the eye through the trabecular meshwork and is transported into the venous drainage of the conjunctiva. If antigenic material is present in the aqueous humour, this transport pathway ensures that the antigen bypasses lymph nodes and ends in the spleen instead (Streilein 1999).

Ophthalmic lymphoma

Lymphoma subtypes

In light of the immunological aspects of the eye and its surrounding tissues it is to be expected that lymphoma subtypes differ in these diverse regions. Several studies indicate that the most common lymphoma arising from tissues surrounding the eye is the low-grade B-cell lymphoma: so-called extranodal marginal zone lymphoma (MALT lymphoma) (Coupland et al. 2002, 2004b; Cho et al. 2003; Bardenstein 2005; Rosado et al. 2006). This finding is not only applicable to the areas recognized as being part of the MALT system, i.e. orbital lymphoma is also primarily of the MALT lymphoma subtype.

In contrast, as the retina is a part of the CNS and the RPE layer serves as a blood–retina barrier, lymphoma arising in the RPE, sensory retina, and optic nerve can be considered as primary CNS lymphoma, whereas uveal lymphoma usually evolves secondarily to systemic lymphoma. Accordingly, the distribution of primary intraocular lymphoma subtypes is similar to the distribution of lymphoma subtypes in the CNS (Chan et al. 2002; Jahnke et al. 2006).

It is therefore my hypothesis that:

A: MALT lymphoma is the predominant orbital- and ocular adnexal lymphoma subtype in the Danish population, whereas intraocular lymphoma is primarily diffuse large B-cell lymphoma (DLBCL).

Lymphoma in the lacrimal sac

Lymphoma arising in the lacrimal sac is very rare. Data on lymphoma subtypes have been reported in case reports only and interestingly, MALT lymphoma and the high-grade DLBCL each accounted for approximately half of all cases (Saccogna et al. 1994; Nakamura et al. 1997; McKelvie et al. 2001; Mori et al. 2001; O’Connor et al. 2002; De Palma et al. 2003; Parmar & Rose 2003; Schefler et al. 2003; Takada et al. 2003).

Based on the knowledge that the lacrimal sac contains MALT tissue, it is my hypothesis, although challenged by the results from several case reports, that:

B: lymphoma arising in the lacrimal sac is predominantly MALT lymphoma.

Incidence of ophthalmic lymphoma

Incidence rates of NHL in general has risen in the Western countries over the last decades (Groves et al. 2000; Howe et al. 2001; Clarke & Glaser 2002), with a doubling from the mid-1970s to the mid-1990s in the Scandinavian countries (Adamson et al. 2007). Immune deficiency is the strongest and best described risk factor for the development of NHL and some of the increase in incidence of NHL from 1970 to 1990s reflects a rising incidence of HIV and AIDS. The fact that the incidence of NHL has stabilized since active anti-retroviral therapy was taken in to use in the mid-1990s supports this explanation.

HIV/AIDS infection cannot fully explain the magnitude of the changes in incidence of NHL (Hartge & Devesa 1992), neither can changes in classification systems nor improved diagnostic capabilities, and so the reason for the increase in incidence still remains unsolved (Alexander et al. 2007).

Data on incidence of ophthalmic lymphoma is sparse. Two reports indicate that the incidence of ophthalmic lymphoma in the United States of America is increasing at an even higher rate than NHL in general (Margo & Mulla 1998; Moslehi et al. 2006). These reports are based on register data. Verification of lymphoma subtype or even of a lymphoma diagnosis has not been performed. However, these are the only studies assessing the incidence of ophthalmic lymphoma. Thus, the incidence and time trends in incidence have never been evaluated in Scandinavian or other European countries.

I hypothesize that:

C: the incidence of ophthalmic lymphoma in the Danish population is comparable with that found in the American population, and that:

D: the incidence of ophthalmic lymphoma in the Danish population has increased during the last decades.

MALT lymphoma

Extranodal marginal zone lymphoma is the most frequent lymphoma subtype found in the orbit and ocular adnexa (Coupland et al. 2002; Cho et al. 2003). It is an extranodal lymphoma considered to be the neoplastic counterpart of the marginal zone cells in reactive follicles. Morphologically the lymphoma cells infiltrate around reactive B-cell follicles and spread to form confluent areas of heterogeneous small B-cells (centrocyte-like cells, monocytoid cells and small mature lymphocytes) with scattered immunoblast- and centroblast-like cells (Jaffe et al. 2001). Clinically, MALT lymphoma is a disease of the elderly, characterized by an indolent course.

Extranodal marginal zone lymphoma occurs most commonly in the stomach but may affect any organ of the human body. It only rarely arises from native organized MALT as seen in Peyer’s patches of the ileum. Usually the lymphoma arises from MALT developed as a result of a chronic inflammation by persistent infections or autoimmune disorders, even in sites that do not typically contain significant epithelial structures (Pelstring et al. 1991).

In general, MALT lymphoma at different extranodal sites shares some common morphological, phenotypical and molecular features, yet the type of infectious agent differs with the primary site of involvement. In gastric MALT lymphoma Helicobacter pylori has been shown to be the causative agent in almost all cases (Wotherspoon et al. 1991). A possible connection between ocular region MALT lymphoma and Chlamydia psittaci has been suggested (Ferreri et al. 2004). However, results from various studies differ, possibly because of geographical differences (Daibata et al. 2006; Mulder et al. 2006; Rosado et al. 2006; Vargas et al. 2006).

There are currently no generally accepted prognostic factors for MALT lymphoma arising in the ocular region. This is probably because most of the previous studies investigating prognostic factors in ocular region lymphoma have been based on groups of patients with up to 10 different lymphoma subtypes (Jenkins et al. 2003; Meunier et al. 2004; Sullivan et al. 2005; Plaisier et al. 2007). Because of the great difference in clinical behaviour and prognosis according to lymphoma subtype, there is a need for studies focusing on ocular region MALT lymphoma.

It is my hypothesis that:

E: clinical factors are related to prognosis of MALT lymphoma arising in the ocular region.

Structural chromosomal changes in MALT lymphoma

Recently, several structural chromosomal abnormalities have been demonstrated in MALT lymphoma. They include t(11;18)(q21;q21) involving API2-and MALT1-genes, t(14;18) (q32;q21) involving IGH- and MALT1-genes, t(1;14)(p22;q32) involving Bcl-10- and IGH-genes, and t(3;14) (p14;q32) involving FOXP1- and IGH-genes (Auer et al. 1997; Willis et al. 1999; Streubel et al. 2003, 2005).

t(11;18)(q21;q21) – API2/MALT1.  The t(11;18) is the most common structural chromosomal abnormality in gastric- and pulmonary MALT lymphoma (Streubel et al. 2004). It has not been described in other B-cell lymphomas. Interestingly, gastric MALT lymphoma with this translocation is associated with advanced stage disease, which does not respond to H. pylori eradication. However, they only rarely undergo transformation to high-grade lymphoma (Liu et al. 2001, 2002; Streubel et al. 2004).

The t(11;18) fuses the API2-gene to the MALT1-gene, thereby generating the fusion protein API2-MALT1 (Dierlamm et al. 1999). The API2-MALT1 fusion protein is capable of activating NF-κB, although neither wild-type API2 nor wild-type MALT1 has this activity (Uren et al. 2000).

t(1;14)(p22;q32) – IGH/Bcl-10.  t(1;14) is seen exclusively in MALT lymphoma, however, less frequently than the t(11;18) (Willis et al. 1999). It juxtaposes the Bcl-10 gene to an IGH gene cluster locus, resulting in deregulation of the Bcl-10 gene and ultimately an overexpression of the Bcl-10 protein (Lucas et al. 2001). Wild-type Bcl-10 is involved in development and function of B-cells by activating NF-κB. This effect, however, is markedly enhanced when MALT1 is co-expressed in the cells (Lucas et al. 2001).

In mature B-cells, Bcl-10 protein is expressed exclusively in the cytoplasm. However, in MALT lymphoma, the Bcl-10 expression is observed in the nucleus in up to 60% of cases (Ye et al. 2000). Nuclear Bcl-10 expression has been linked to the presence of t(1;14)(p22;q32) or t(11;18)(q21;q21) (Ye et al. 2005; Sagaert et al. 2006b), but the mechanisms of the aberrant nuclear localization of Bcl-10 and its clinical impact remain unclear (Liu et al. 2001; Franco et al. 2006).

t(14;18)(q32;q21) – IGH/MALT1.  The third MALT lymphoma-specific translocation involves IGH and MALT1 and leads to overexpression of the MALT1 gene (Streubel et al. 2003). It is primarily found in MALT lymphoma outside the gastrointestinal- or pulmonary tract (Streubel et al. 2003, 2004). MALT1 works synergistically with Bcl-10 to enhance NF-κB activation (Lucas et al. 2001).

t(3;14)(p14.1;q32) – FOXP1/IGH.  The most recently described translocation t(3;14) brings the forkhead box protein P1 (FOXP1) gene on chromosome 3 under the control of IGH, which deregulates the expression of FOXP1 (Streubel et al. 2005). t(3;14) is not specific to MALT lymphoma, unlike the other three MALT lymphoma-associated translocations; it has been found also in DLBCL (Barrans et al. 2004). Its role in MALT lymphoma pathogenesis is still debated.

FOXP1 is expressed broadly in both germinal centre and pregerminal centre B-cells. The physiological role of FOXP1 in lymphoid tissue is still unclear. However, it seems to be essential for early B cell development (Hu et al. 2006), and overexpression of FOXP1 is associated with a poor clinical course both in DLBCL and MALT lymphoma (Barrans et al. 2004; Sagaert et al. 2006a).

NF-κB activation – the unifying concept for MALT lymphomagenesis

There is evidence that the three MALT lymphoma-specific translocations involving API2/MALT1, IGH/MALT1 and Bcl-10/IGH lead to formation or up-regulation of proteins (API2-MALT1, MALT1 and Bcl-10) that ultimately target the same signalling pathway (NF-κB) (Lucas et al. 2001). This indicates that these rearrangements are important for the pathogenesis of MALT lymphoma. NF-κB is a transcription factor with effect on a number of proliferation-related genes in B cells (Jost & Ruland 2007). Antigen receptor-stimulated NF-κB activation is dependent on three key molecules: CARMA1, Bcl-10 and MALT1 that form a complex capable of triggering the IKK complex. The Inhibitor of nuclear factor κB kinase (IKK) complex phosphorylates IκB and targets it for degradation. In this way NF-κB is released and translocates into the nucleus, where it induces expression of genes essential for proliferation and function of B- and T-cells (Fig. 4). In MALT lymphoma this NF-κB activation seems to be driven in an uncontrolled manner, and in the absence of antigen stimulation.

image

Figure 4.  Hypothetical mechanism for the effect of the chromosomal translocations involving Bcl-10 and MALT1: stimulation of the receptor by antigen promotes the recruitment and oligomerization of CARMA1, BCL10 and MALT1, forming a complex capable of triggering the IKK complex. The IKK complex phosphorylates IκB and targets it for degradation. In this way NF-κB is released and translocates into the nucleus. CARMA1: caspase recruitment domain 1, IKK: inhibitor of nuclear factor (NF)-κB kinase.

Download figure to PowerPoint

The up-regulation of wild-type BCL 10 induced by t(1;14) causes NF-κB activation in a dose dependent manner. The API2-MALT1 fusion product may possess a mechanism for self-oligomerization and thus the t(11;18) product bypasses normal BCL10 to activate IKK directly and similarly results in increased nuclear translocation of NF-κB. Wild-type MALT1 appears to depend on interaction with BCL10 for auto-activation. Excess MALT1 caused by t(14;18), however, is capable of activating NF-κB, presumably because MALT1 binds normal Bcl-10 thereby stabilizing Bcl-10 and causing its accumulation in the cytoplasm with uncontrolled NF-κB activation as a result.

Hence, there is good evidence that NF-κB activation because of Bcl-10, MALT1 or API2-MALT1 deregulation is important in the pathogenesis of MALT lymphoma. However, the frequency of the structural alterations leading to the formation or up-regulation of these proteins, seems to vary depending on the anatomic site at which MALT lymphoma arises (Streubel et al. 2004). Furthermore, only a few studies have analysed the occurrence of the translocations in ocular region MALT lymphoma (Takada et al. 2003; Streubel et al. 2004, 2005; Ye et al. 2005) with varying results, and the true incidence of the translocations in this region is not yet clarified.

It is my hypothesis that:

F: MALT lymphoma in the ocular region arises because of NF-κB activation via structural chromosomal changes leading to the formation or up-regulation of Bcl-10, MALT1 or API2-MALT1,

and that:

G: t(14;18)(q32;q21) is the most frequent translocation in ophthalmic MALT lymphoma, whereas t(11;18)(q21;q21) is rare.

Furthermore I hypothesize that:

H: the presence of translocations involving Bcl-10, MALT1 or API2-MALT1 in ophthalmic MALT lymphoma is associated with an adverse prognosis.

Hypotheses and Aims of the Study

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

Hypotheses

In summary, the hypotheses that will be examined in the thesis are:

  •  A:
    MALT lymphoma is the predominant orbital- and ocular adnexal lymphoma subtype in the Danish population, whereas intraocular lymphoma is primarily DLBCL.
  •  B:
    Lymphoma arising in the lacrimal sac is primarily MALT lymphoma.
  •  C:
    The incidence of ophthalmic lymphoma in the Danish population is comparable with that found in the American population.
  •  D:
    The incidence of ophthalmic lymphoma in the Danish population has increased during the last decades.
  •  E:
    Clinical factors are related to prognosis of MALT lymphoma arising in the ocular region.
  •  F:
    MALT lymphoma in the ocular region arises because of NF-κB activation via structural chromosomal changes leading to the formation or up-regulation of Bcl-10, MALT1 or API2-MALT1.
  •  G:
    t(14;18)(q32;q21) is the most frequent translocation in ophthalmic MALT lymphoma, whereas t(11;18)(q21;q21) is rare.
  •  H:
    The presence of translocations involving Bcl-10, MALT1 or API2-MALT1 in ophthalmic MALT lymphoma is associated with an adverse prognosis.

Aims

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

To approach answers to the hypotheses presented above, this thesis is based on three papers addressing the following aims:

  • 1
    To determine the distribution of subtypes of lymphoma in the eye, orbit and ocular adnexa.
  • 2
    To investigate the distribution of lymphoma subtypes in the lacrimal sac and compare the findings with lymphoma subtypes in the ocular adnexa.
  • 3
    To determine the incidence- and time trends in incidence of ophthalmic lymphoma in Denmark from 1980 to 2005.
  • 4
    To investigate the frequency of MALT lymphoma specific translocations in MALT lymphoma presenting in the ocular region.
  • 5
    To identify clinical-, and cytogenetical factors associated with prognosis of MALT lymphoma presenting in the ocular region.

Material and Methods

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

Material

The study is based on a review of pathology reports filed in the Eye Pathology Institute, University of Copenhagen. A 26-year-period (1980–2005) was chosen to determine the incidence- and subtypes of ophthalmic lymphoma in Denmark. Additionally, The Danish Registry of Pathology and the Danish Lymphoma Group Registry were examined for ophthalmic lymphoid lesions diagnosed during the period 1980–2005. From the registries all patients diagnosed with a lymphoid lesion in any ophthalmic site (i.e. orbit, lacrimal gland, lacrimal sac, intraocular structures, conjunctiva or eyelid) were collected. Furthermore, pathology institutes affiliated with the Ophthalmic Oncology Task Force of the European Organization for Research and Treatment of Cancer and European Ophthalmic Pathology Society were contacted to collect additional cases of lymphoma in the lacrimal sac.

Histopathological reports from a total of 462 Danish patients with an ophthalmic lymphoid lesion during the 26-year-period were retrieved and reviewed. Specimens with evident reactive lymphoid hyperplasia (n = 199) were identified and not analysed further. The remaining 263 specimens with any suspicion of lymphoid malignancy were reviewed by histology and immunohistochemistry. Thirty-five cases were originally diagnosed as opthalmic lymphoma, but were excluded from the study for technical reasons (18 cases) or because an opthalmic location or a lymphoma diagnosis could not be confirmed (17 cases). Thus, the collected material covered all cases of primary and secondary lymphoma in the ocular region diagnosed in Denmark during the period 1980–2005 plus cases of lacrimal sac lymphoma originating from all over Europe back to 1948.

Discussion

The material investigated is unique as it covers a whole population during a time period of 26 years.

The Eye Pathology Institute, University of Copenhagen serves as a nationwide ophthalmic pathological institute, receiving specimens of tissue from the eye region from both private ophthalmologists and departments of ophthalmology throughout Denmark. Since the early 1940s all tissue specimens and histopathological descriptions have been thoroughly filed.

Some ophthalmological departments send their extra-ocular specimens to the local departments of general pathology. Hence, The Danish Registry of Pathology was included in retrieving material as well. The Danish Registry of Pathology was founded in 1960 with the purpose of registering all pathology diagnoses in Denmark. With the introduction of systemized nomenclature of medicine codes (SNOMED codes) in the late 1970s the registration progressed, covering all Danish departments of pathology. However, some departments registered only a subset of their specimens until the late 1980s. Thus, a few cases not registered in the Danish Registry of Pathology in the early 1980s may have been overlooked. Additionally, the Danish Lymphoma Group Registry founded in 1982 was explored to crosscheck for cases not found in the upper mentioned registries. This search led to no additional cases. However, until late 1990s this registry covered only Western Denmark, and therefore the risk of missing a few cases during the 1980s and 1990s still exists.

The use of three different population based registries, and the fact that all specimens were reviewed gives a reliable representation of the incidence of ocular region lymphoma in Denmark. However, lymphoma classification systems have changed during the 26-year-period and MALT lymphoma was not a described entity until 1983 (Isaacson & Wright 1983). Thus, some MALT lymphomas before that time might have been classified as reactive lymphoid hyperplasia, leading to an underestimation of incidence rates during the 1980s in our study. However, we reviewed all histopathological reports of lymphoid lesions and cases originally diagnosed as reactive lymphoid hyperplasia, but with a description suspicious of lymphoma were re-evaluated. Furthermore, the MALT lymphoma diagnosis went into practical use in the early 1990s and thus the increase found from mid-1990s to 2005 can be considered reliable.

Methods

Clinical data

Clinical files from ophthalmic- and haematological departments involved in the diagnostic process and treatment of the patients were collected and reviewed with particular reference to sex, age, prior history of lymphoma, sites of involvement, stage of disease at diagnosis according to the Ann Arbor staging system (Carbone et al. 1971), symptoms and signs, treatment, International Prognostic Index (IPI) (The International Non-Hodgkin’s Lymphoma Prognostic Factors Project 1993; Armitage 2005), site and date of relapse and date and cause of death.

Anatomical sites within the ocular region were divided into eyelid, conjunctiva, anterior uvea, posterior uvea, retina, orbit, lacrimal gland or lacrimal sac based on clinical, imaging [descriptions of computer tomography (CT) scans or magnetic resonance (MR) scans] and histopathological findings.

Follow-up data concerning relapses were updated to February 2007 via SNOMED codes registered in the Danish Registry of Pathology.

Furthermore, data from the Danish Cancer Registry and the Registry of Cause of Death were collected to achieve additional information concerning date- and cause of death. All death certificates were reviewed to validate the cause of death.

Lymphoma related death was registered for patients, dead in hospital because of lymphoma progression or as a complication to lymphoma treatment. For patients who died out of hospital, with a death certificate stating lymphoma as the primary cause of death, death because of lymphoma was only coded, if the patient was not in complete remission at last follow-up. If lymphoma was not reported as the primary cause of death in the death certificate, (or if clinical files described death because of other illnesses) the patient was coded dead because of causes other than lymphoma.

Overall survival (OS) was calculated from time of diagnosis to time of death from any cause or time of last follow-up. Progression free survival (PFS) was defined as time period from time of diagnosis until the date of first relapse/progression or last follow-up or death.

Discussion.  The Central Population Registry (CPR) was established in Denmark in 1968 providing all persons in Denmark with an individual identification number. The CPR number makes linking of patient data possible from clinical files, the Eye Pathology Institute Registry, the Danish Registry of Pathology, the Danish Cancer Registry and the Registry of Cause of Death. We were thus able to minimize the number of patients lost to follow-up.

Because of the retrospective character of this study, evaluation of clinical parameters was based on the original data recorded by local physicians. Furthermore, some clinical files were not obtainable. Consequently, clinical data, such as Ann Arbor stage and IPI score at diagnosis, were not applicable to all patients included in the study.

The occurrence of relapses of primary ocular region MALT lymphoma was based on information in the clinical files, the Danish Eye Pathology Institute Registry and the Danish Registry of Pathology. Possibly some recurrences may not have been histologically verified and therefore the incidence of relapses may be higher than reported in our study. Also, the question of transformation from low-grade to high-grade lymphoma could not be approached.

In retrospective studies the risk of miscoding date and cause of death is present. However, the cross-linking of information in the Danish Cancer Registry and the Danish registry of cause of death warrants a high validity of registration. It may be safely assumed that persons not reported as dead were still alive (Jensen et al. 1985). Furthermore, we optimised the validity of the coded cause of death by reviewing all death certificates and compared the statements with information in the clinical files and the Danish Cancer Registry. The possibility of miscoding the cause of death was thus minimized.

Histopathology and immunohistochemistry

For re-evaluation of the specimens, formalin fixed paraffin embedded tissue was collected from the Eye Pathology Institute and from all Danish pathology departments (I). In the lacrimal sac lymphoma study (II), 20 unstained slides from each case were obtained from pathology institutes throughout Europe. Morphology and localization in the ocular region were evaluated on haematoxylin and eosin (H & E) stained sections.

A primary immunohistochemical panel was applied with antibodies directed against CD3, CD5, CD20 and CD79α to differentiate between B- and T-cell lymphomas. The B-cell neoplasms were additionally immuno-phenotyped with antibodies directed against Bcl-2, Bcl-6, CD10, CD23, CD30, Cyclin D-1, MUM-1 and κ- and λ light-chains. T-cell lymphomas were immuno-phenotyped with CD4, CD8, CD30, CD56, ALK-1, TIA and Granzyme B. For all cases the mitotic rate was investigated with MIB-1.

The sections were reviewed in consensus by two of the authors (LD Sjö and E Ralfkiaer) to reclassify the lymphoma according to the WHO classification. Several of the specimens included in the lacrimal sac study (II) were discussed in a plenum of five pathologists (S Heegaard, JU Prause, BR Juhl, E Ralfkiaer and LD Sjö).

Discussion.  Classification of NHL is complex and requires integration of morphological, immunohistochemical, clinical and, in some cases, molecular features. Two pathologists at least, evaluated the specimens. In case of doubt a consensus was obtained within all authors. The inter-observer reproducibility was not assessed in our study.

Immunohistochemistry performed on formalin fixed paraffin embedded tissue is a multistep procedure, and the final result is influenced by several of the steps. Thus, duration of fixation in formalin, the completeness of intiltration of paraffin, and the success of epitope retrieval after deparaffination and rehydration are all very important (Arnold et al. 1996; Shi et al. 2001). Furthermore, because of the use of archived material going back to 1980 (and in the lacrimal sac lymphoma study as far as 1948), the preservation of antigenesity of paraffin embedded specimens is highly relevant. However, the preservation of antigenesity is good, when material is fixed in formalin and kept in paraffin blocks, whereas slide monitored sections must be kept cold (max. 4°C or better −20°C/−80°C) and in air-deprived boxes (Jacobs et al. 1996; Shi et al. 1997; Bertheau et al. 1998). Thus, some tumours may have shown false negativity for some of the antibodies used in the reclassification leading to a wrong diagnosis. We collected paraffin blocks whenever possible, and the sections were cut just prior to staining. All antibodies used in the standard panel for lymphoma classification were well-known monoclonal antibodies with high specificity. Negative and positive controls consisting of lymph nodes with benign hyperplasia were included in all stainings. Furthermore, internal controls were evaluated in all specimens and new stainings performed if no internal controls were present.

Detection of translocations by fluorescence in situ hybridization (FISH)

Fluorescence in situ hybridization (FISH) allows specific localization of genetic sequences in morphologically preserved cells or tissue sections. In short, DNA is denaturated by heating and the fluorescence-labeled probe is added to the denatured sample mixture and hybridizes with the sample DNA at the target site as it reanneals back into a double helix. Excess probe is washed out and the specifically annealed probe can be detected by fluorescence microscopy (Fig. 5).

image

Figure 5.  Fluorescence in situ hybridization technique: DNA is denaturated by heating and the fluorescence-labeled probe is added. The probe hybridizes with the sample DNA at the target site as it reanneals back into a double helix. The specifically annealed probe can be detected by fluorescence microscopy.

Download figure to PowerPoint

For detection of translocations we used two different types of FISH probe: split signal probes for detection of the presence of breakpoints in the IGH (14q32), MALT1 (18q21), Bcl-2 (18q21), Bcl-6 (3q27) or Bcl-10 (1p22) genes, and dual colour dual fusion translocation probes for confirmation of the presence of the translocation product (t14;18)(q32;q21) between the IGH- and MALT1 genes.

In split signal probes, the two DNA probes are designed to hybridize upstream and downstream of the breakpoint region (Fig. 6). Co-localization of the probes (translocation not present) results in an orange signal (mix of red and green signal), whereas a translocation in the breakpoint region splits one signal into separate green and red signals.

image

Figure 6.  The MALT1 (mucosa-associated lymphoid tissue lymphoma translocation 1) gene at chromosome 18 band q21 consists of 17 exons and spans a region of ∼80 kb. The two DNA probes within the MALT1 FISH DNA Probe, Split Signal, are designed to hybridize upstream and downstream of the breakpoint cluster region. Co-localization of the probes results in a red/green signal, whereas translocation events in the breakpoint cluster region will split one signal into separate green (fluorescein) and red (Texas Red) signals.

Download figure to PowerPoint

Dual colour dual fusion translocation probes contain two differently coloured probes spanning the entire genes (IGH and MALT1) to be investigated. In normal cells lacking a translocation, a two orange (2× MALT1), two green (2× IGH) signal pattern is seen, whereas cells containing the translocation will show one orange (intact MALT1), one green (intact IGH) and two fusion (IGH/MALT1 and MALT1/IGH) signals.

We performed interphase FISH on 50 primary ocular region MALT lymphomas. Initially we screened the specimens for translocations at the immunoglobulin heavy chain gene cluster- (IGH, 14q32) and at the MALT1 (18q21) gene-loci, using dual-colour split-signal probes for MALT1 (DAKO, Glostrup, Denmark) and IGH (DAKO). In cases positive for IGH- and MALT1 gene break the IGH-MALT1 fusion product was confirmed using an IGH-MALT1 dual colour, dual fusion translocation probe (Abbot Laboratories). In cases positive for an IGH gene break only, dual-colour break-apart probes for Bcl-2, Bcl-6 and Bcl-10 (DAKO) were applied.

In each case the hybridization signals for each probe were evaluated in at least 100 nuclei. A reactive tonsil was mounted as negative control on each slide. The threshold for presence of a translocation was determined counting split signals in 100 nuclei in five different reactive tonsils. The highest number of false positive nuclei plus three standard deviations (SD) was taken as the cutoff point of each aberration.

Discussion.  Fluorescence in situ hybridization provides a rapid and relatively easy-to-handle technique for the detection of specific chromosomal abnormalities independent of the cycle status of cells. It has no use, however, in a more genome wide screening for genomic aberrations. Furthermore, several factors (i.e. duration of fixation in formalin, the completeness of infiltration of paraffin, wash temperature etc.) influence the intensity and number of FISH signals (Shi et al. 2001; Petersen et al. 2004), and it is thus necessary that procedures are strictly standardized. Like in immunohistochemistry it is of major importance that sections for FISH are cut immediately before application of the probe, and furthermore the prepared slides must be kept cold (−20°C/−80°C) and in complete darkness, otherwise the fluorescent signals faint.

It is an advantage that FISH can be performed on paraffin-embedded, formalin-fixed tissue sections where histological morphology is preserved, so that the scoring of signals can be restricted to cancer cells. However, determination of the exact location of FISH signals within each nucleus can be problematic because of nuclear slicing and the high density of cells. Therefore, we counted at least 100 nuclei from each specimen and compared it with a threshold estimated as the upper limit 99% confidence interval of false positive signals in 100 negative control nuclei. All cases scored as positive had at least 20% more positive signals compared to this threshold, implying that presence of translocation was not in doubt in any of the cases.

Statistics

Differences in characteristics in patient subgroups were tested using Fisher’s exact test. For analyses of incidence, population numbers stratified by sex, age and time period were obtained from Statistics Denmark (Statistics Denmark 2007). Non-Hodgkin lymphoma incidence rates, expressing the number of new cases per 100 000 person-years were calculated for the whole population and for population subgroups according to lymphoma-subtype, localization, sex, age and time-period. Age was grouped in ten 10-year age groups, and time-period in five 5-year groups. Because of a low number of cases of the NHL subtypes DLBCL, mantle cell lymphoma and follicular lymphoma, it was necessary to pool time-periods (1980–1992 and 1993–2005). To allow estimation of the effects of age, sex and time period on the NHL incidence rates, a Poisson regression model (Frome 1983) was used. The models were checked for interaction terms and prognostic significance was evaluated using goodness-of-fit test based on the deviance.

Univariate analyses of survival (OS and PFS) were estimated by the Kaplan-Meier method and compared by the log-rank test. Factors with effect on survival in univariate analyses were selected for multivariate analysis using the Cox regression method to determine independently predictive variables.

Probability values <0.05 were regarded as significant. All statistical analyses were performed using spss statistical software version 15.0 (SPSS, Chicago, Illinois, USA) and sas statistical software package version 9.1 (SAS, North Carolina, USA).

Discussion.  The composition of the Danish population changes so that the older generation represents an increasing fraction. For the analyses of incidence, exact population numbers stratified by sex, age and time period were obtained from Statistics Denmark. Hence, fluctuations in the population number and composition were taken into account. Furthermore, we checked for interaction between parameters.

The main limitation of the analysis of incidence was the low number of cases. Some lymphoma subgroups were represented by so few cases that calculation of incidence and of increments in incidence was not possible by using the Poisson regression model.

Results

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

Ophthalmic lymphoma (I)

Ophthalmic lymphoma subtypes and clinical characteristics

A total of 228 Danish patients with a biopsy-reviewed verified diagnosis of ophthalmic lymphoma were identified. One lymphoma subtype predominated in the orbit and ocular adnexa, i.e. MALT lymphoma accounting for 55% (Table 1, Fig. 7C). Diffuse large B-cell lymphoma was the largest subgroup of intraocular lymphoma (33%).

Table 1.   Distribution of lymphoma within the ophthalmic region according to lymphoma subtype.
Lymphoma subtypenBilateral n (%, CI)Eyelid n (%, CI)Conj. n (%, CI)Lacr. sac n (%, CI)Lacr. gland n (%, CI)Orbit n (%, CI)Intraocular n (%, CI)
  1. B-ALL = precursor B-cell lymphoblastic lymphoma; BCL,uncl. = B-cell lymphoma, unclassified, BL = Burkitt lymphoma; CI = 95% confidence interval; Conj. = conjunctiva; DLBCL = diffuse large B-cell lymphoma; FL = follicular lymphoma; Lacr. = lacrimal; MALT = extranodal marginal zone B-cell lymphoma (MALT lymphoma); MCL = mantle cell lymphoma; = number; PTL,NOS = peripheral T-cell lymphoma, unspecified; SLL/CLL = small lymphocytic lymphoma; T-ALCL = anaplastic large T-cell lymphoma.

  2. *The frequency of high-grade lymphoma intraocularly was significantly higher as compared to the remaining ocular region, p = 0.0035.

  3. Bilateral affection was significantly more common in MCL compared to MALT lymphoma, p = 0.0004.

  4. From (I) (Sjo et al. 2008a).

All lymphomas22828 (12, 8–17)11 (5, 2–8)57 (25, 19–30)3 (1, 0–3)16 (7, 4–10)129 (57, 50–63)12 (5, 2–8)
Low-grade types14714 (10, 5–14)7 (5, 1–8)44 (30, 23–37)2 (1, 0–3)10 (7, 3–11)83 (56, 48–64)0
MALT12613 (10, 5–17)5 (4, 0–7)40 (31, 23–40)1 (1, 0–2)6 (5, 2–8)74 (58, 50–67)0
FL171 (6,0–17)1 (6, 0–17)4 (24, 3–44)03 (18, 0–36)9 (53, 29–77)0
SLL/CLL301 (33, 0–87)01 (33, 0–87)1 (33, 0–87)00
High-grade types6110 (16, 7–26)4 (7, 0–13)9 (15, 6–24)1 (2, 0–5)4 (7, 0–13)35 (57, 45–70)8 (13, 5–22)*
DLBCL291 (3, 0–10)1 (3, 0–10)2 (7, 0–16)03 (10, 0–21)19 (66, 48–83)4 (14, 1–26)
MCL209 (45, 23–67)07 (35, 14–65)1 (5, 0–15)1 (5, 0–15)11 (55, 33–77)0
PTL,NOS402 (50, 1–99)0001 (25, 0–67)1 (25, 0–67)
BL3000001 (33, 0–87)2 (67, 13–100)
B-ALL3000002 (67, 13–100)1 (33, 0–87)
T-ALCL201 (50, 0–100)0001 (50, 0–100)0
BCL,uncl.214 (20, 2–38)04 (20, 2–38)02 (10, 0–23)11 (55, 33–77)4 (15, 0–31)
image

Figure 7.  (A) Typical conjunctival lymphoma presenting as a painless salmon coloured tumour in the conjunctiva of a 65-year-old man. (B) Magnetic resonance image. A 65-year-old woman with a 2-year-history of proptosis. The tumour spreads from the left lacrimal gland (white arrow) and infiltrates the orbit diffusely (asterisk). Histopathological examination of the tumour showed a MALT lymphoma. (C) Orbital MALT lymphoma characterized by a diffuse pattern of small centrocyte-like cells in an 81-year-old female patient (Haematoxylin–eosin (H & E); ×200). (D) MALT lymphoma (same patient as Fig. 1C) with preserved germinal centres shown by a CD23 staining (brown) of the follicular dendritic cells (CD23; ×200). (E) Diffuse large B-cell lymphoma in the right lower eyelid. The tumour cells have pleomorphic nuclei with irregularly distributed chromatin and up to three nucleoli. Mitoses (black arrow) are frequently seen. The 80-year-old man died 6 months after diagnosis (H & E; ×400). (F) Diffuse large B-cell lymphoma in the conjunctiva of a 73-year-old woman. The mitotic activity was high as shown here by MIB-1 staining (brown nuclear staining) (MIB-1; ×200). (G) Cyklin D1 positivity in an orbital mantle cell lymphoma in a 85-year-old man with no prior history of lymphoma. Subsequent staging revealed involvement of the bone marrow and axillary- and inguinal lymph nodes (Cyklin D1; ×400). (H) Typical nodular pattern in a follicular lymphoma of a 45-year-old woman. A full-body CT-scan exposed a retroperitonal tumour and chemotherapeutic treatment was initiated (H & E ×100). From (I) (Sjo et al. 2008a).

Download figure to PowerPoint

Most patients were elderly with a median age of 69 years (range 5–96 years). Ophthalmic lymphoma was first presenting symptom in the majority of patients (Table 2) and more than half of the patients had involvement of the ocular region only (stage I, Tables 3 and 4). The orbit was the most frequent localization. Bilateral disease was significantly more common in mantle cell lymphoma compared with MALT lymphoma or all other lymphoma subtypes (Table 1).

Table 2.   Clinical data according to lymphoma subtype.
Lymphoma subtype Number (%, CI)Men n (%, CI)Median age both genders, years (range)Patients with ophthalmic lymphoma as first presenting symptom, n (%, CI)
  1. B-ALL = precursor B-cell lymphoblastic lymphoma; BCL,uncl. = B-cell lymphoma, unclassified; BL = Burkitt lymphoma; CI = 95% confidence interval; DLBCL = diffuse large B-cell lymphoma; FL = follicular lymphoma; MALT = extranodal marginal zone B-cell lymphoma (MALT lymphoma); MCL = mantle cell lymphoma; PTL,NOS = peripheral T-cell lymphoma, unspecified; SLL/CLL = small lymphocytic lymphoma; T-ALCL = Anaplastic large T-cell lymphoma.

  2. *Men were more likely to present with high-grade lymphoma than women, p = 0.0022

  3. In MCL a predominance of men was found, p = 0.0017.

  4. From (I) (Sjo et al. 2008a).

All lymphomas228 (100)116 (51, 44–57)69 (5–96)184 (81, 76–86)
Low-grade types147 (64, 58–70)66 (45, 37–55)69 (8–96)127 (88, 80–91)
MALT126 (55, 49–62)58 (46, 37–54)69 (8–92)116 (92, 85–96)
FL17 (7, 4–11)7 (41, 18–65)63 (45–96)11 (65, 42–87)
SLL/CLL3 (1, 0–3)1 (33, 0–87)75 (62–75)0
High-grade types61 (27, 21–33)42 (69, 57–80)*72 (5–95)42 (69, 57–80)
DLBCL29 (13, 8–17)13 (45, 27–63)77 (17–95)23 (79, 59–89)
MCL20 (9, 5–12)17 (85, 69–100)75 (60–90)11 (55, 33–77)
PTL,NOS4 (2, 0–3)3 (75, 33–100)61 (31–77)3 (75, 33–100)
BL3 (1, 0–3)2 (67, 13–100)41 (34–55)1 (33, 0–87)
B-ALL3 (1, 0–3)2 (67, 13–100)9 (5–34)2 (67, 13–100)
T-ALCL2 (1, 0–2)2 (100, 90–100)45 (35–55)2 (100, 90–100)
BCL,uncl.21 (9, 5–12)11 (55, 33–77)63 (18–77)15 (71, 50–90)
Table 3.   Extraocular ophthalmic lymphoma. Stage of disease at diagnosis according to lymphoma subtype and localization in 179 patients.
Lymphoma subtypeStageOrbit, nOcular adnexa, n
  1. B-ALL = precursor B-cell lymphoblastic lymphoma; BCL,uncl.=B-cell lymphoma, unclassified; BL = Burkitt lymphoma; DLBCL = diffuse large B-cell lymphoma; FL = follicular lymphoma; MALT = extranodal marginal zone B-cell lymphoma (MALT lymphoma); MCL = mantle cell lymphoma; ns = not stated; PTL,NOS = peripheral T-cell lymphoma, unspecified; T-ALCL = Anaplastic large T-cell lymphoma.

  2. From (I) (Sjo et al. 2008a).

MALTI4528
II11
III30
IV126
ns911
FLI33
II01
III12
ns10
DLBCLI91
II02
IV30
ns41
MCLII02
IV72
PTL,NOSI01
II10
BLII10
B-ALLIV20
T-ALCLI01
III10
BCL,uncl.I54
II10
IV11
ns11
Table 4.   Intraocular lymphoma. Lymphoma subtype, localization and staging of 12 patients.
SubtypeStageRetina, nUvea, n
  1. B-ALL = precursor B-cell lymphoblastic lymphoma; BCL,uncl. = B-cell lymphoma, unclassified; BL = Burkitt lymphoma; DLBCL = diffuse large B-cell lymphoma; ns = not stated; PTL,NOS = peripheral T-cell lymphoma, unspecified; Sec. intraoc. = Secondary intraocular lymphoma.

  2. *Localized in the iris.

  3. From (I) (Sjo et al. 2008a).

DLBCLI10
III01
ns10
Sec. intraoc.01
PTL,NOSIV01*
BLSec. intraoc.02
B-ALLSec. intraoc.01*
B-cell, uncl.I01
Sec. intraoc.03
Incidence of ophthalmic lymphoma

Incidence rates were highly dependent on patient age (Fig. 8) with a maximum of 2.6 cases per 100 000 for men at age 90 years and 1.0 per 100 000 for women at age 80 years. Difference in incidence between men and women was not statistically significant.

image

Figure 8.  Age-specific ophthalmic lymphoma incidence rates for men and women in Denmark during the period 2001–2005. The rates are calculated based on all data in a Poisson regression model adjusting for period. Incidence rates in the 5-year time periods from 1981–2000 are proportional to the ones depicted here – although lower as shown in Table 5. Incidence rates were highly dependent on patient age (p < 0.0001) with a maximum incidence rate of 2.6 per 100 000 for men and 1.0 per 100 000 for women. Differences in incidence between men and women were not statistically significant (p = 0.766). From (I) (Sjo et al. 2008a).

Download figure to PowerPoint

Incidence rates for the whole population increased from 0.086 in 1981–1985 to 0.249 per 100 000 in 2001–2005, primarily because of a rise in incidence of MALT lymphoma. The estimated changes in incidence corresponded to an annual average increase in the 26-year-period of 3.4% (Table 5, Fig. 9).

Table 5.   Incidence of ophthalmic lymphoma 1981–2005 in Denmark. During the period there was an average increase of 3.4% per year (CI: 1.4%–5.4%, p = 0.0007).
Time periodNumber of patientsIncidence per 100.000 (CI)Age-adjusted rate ratio (CI)*
  1. CI = 95% Confidence interval.

  2. *p = 0.0068.

  3. From (I) (Sjo et al. 2008a).

1981–1985220.086 (0.057–0.131)1.00 (1.00–1.00)
1986–1990440.171 (0.128–0.231)1.58 (0.94–2.70)
1991–1995380.147 (0.107–0.202)1.60 (0.94–2.71)
1996–2000500.189 (0.143–0.249)1.78 (1.06–2.98)
2001–2005690.249 (0.196–0.316)2.41 (1.47–3.93)
image

Figure 9.  Incidence of ophthalmic lymphoma 1981–2005 in Denmark. The increase in incidence was primarily because of a rise in incidence of MALT lymphoma.

Download figure to PowerPoint

Lymphoma arising in the lacrimal sac (II)

Lymphoma arising in the lacrimal sac is very rare with fewer than 50 cases of primary lacrimal sac lymphoma reported in the past 30 years. We investigated 15 cases from all over Europe and found that the distribution of lymphoma subtypes differed compared with the orbit and ocular adnexa, with an overrepresentation of DLBCL (33%). Consequently, MALT lymphoma and DLBCL were equally common in this region.

MALT lymphoma arising in the ocular region (III)

Presence of MALT lymphoma specific translocations

Evidence of IGH-gene breakage was found in two of 42 cases (5%, Fig. 10F). One was proven to be an IGH/MALT1 translocation, whereas the other showed evidence of Bcl-6 gene breakage suggesting the presence of an IGH/Bcl-6 fusion.

image

Figure 10.  (A) 63-year-old woman with conjunctival MALT lymphoma. The patient was in stage I at diagnosis, and received no treatment. She progressed after four years with bilateral affection. (B) Orbital MALT lymphoma characterized by a diffuse pattern of small centrocyte-like cells in a 67-year-old female patient (Haematoxylin–eosin (H & E) ×200). (C) Conjunctival MALT lymphoma with plasmacytoid differentiation and multiple Dutcher bodies (black arrows) (H & E ×400). (D) Monotypic plasmacells in a conjunctival MALT lymphoma showing κ-light chain restriction (κ×400). (E) Nuclear Bcl-10 positivity in the neoplastic cells of a conjunctival MALT lymphoma. (F) Fluorescence in situ hybridisation with a dual-colour split-signal probe in a specimen with a translocation at the IGH-gene locus. In the neoplastic cells (white arrows) there is one fusion product (orange dot) representing the normal chromosome, and two split-signal products (one green and one red) from the chromosome with an IGH-associated translocation. From (III) (Sjo et al. 2008b).

Download figure to PowerPoint

Factors associated with prognosis of ocular region MALT lymphoma

Table 6 shows the main clinical features of 116 patients with MALT lymphoma arising in the ocular region. One third of patients had a relapse or progression of disease after initial therapy and relapses were frequently found at extra-ocular sites. Overall survival, however, was not significantly poorer for patients with relapse (Table 7).

Table 6.   Clinical features of 116 patients with MALT lymphoma presenting primarily in the ocular region.
Sex61 women (53%); 55 men (47%)
Age, median (range)69 years (8–90 years)
Parametern%
  1. IPI = International Prognostic Index.

  2. *Stage of disease known in 95 patients.

  3. IPI-score obtainable in 51 patients.

Localization
 Orbit69 60
 Conjunctiva38 33
 Lacrimal gland5 4
 Eyelid4 3
Bilateral disease12 10
Stage*
 I72 76
 II2 2
 III3 3
 IV18 19
IPI-score
 0–245 88
 3–56 12
Table 7.   Results of univariate analysis for prognosis evaluated by Kaplan–Meier method and tested for statistical significance with log-rank test.
Prognostic factorProgression free survivalOverall survival
5-year (%)p-value5-year (%)p-value
  1. IGH = immunoglobulin heavy chain gene cluster; IPI = International Prognostic Index.

  2. From (III) (Sjo et al. 2008a).

Sex
 Men (= 55)720.775670.040
 Women (= 62)7082
Clinical stage
 I (= 72)660.736740.877
 II or more (= 23)7674
Ocular region localization
 Orbit (= 69)740.162740.464
 Conjunctiva (= 38)6676
 Lacrimal gland (= 5)5080
 Eyelid (= 4)5075
B-symptoms
 Yes (= 6)750.487630.029
 No (= 99)5775
Treatment, all patients
 None/surgery/prednisolone (= 11)460.264740.951
 Radiation only (= 58) 6472
 Chemotherapy ± radiation (= 36)7675
Relapse/progression
 Yes (= 37)  790.578
 No (= 79)  72
Proliferation status
 MIB-1 < 30% (= 108)730.075750.912
 MIB-1 ≥ 30% (= 7)5063
Nuclear Bcl-10 0.816 0.208
 Positive (= 22)7084
 Negative (= 53)7065
IgH-involved translocation
 Positive (= 2)00.013500.390
 Negative (= 40)7475
IPI-score
 Low (0–2, = 46)610.852840.006
 High (3–5, = 6)8344

For all patients, the estimated 5-year PFS rate was 71% and 5-year OS rate was 75% (Fig. 11). Among 56 patients who died during follow-up, 12 (21%) died from lymphoma progression 4 to 149 months after diagnosis (median: 64.5 months).

image

Figure 11.  Overall survival (OS) curves for all 116 patients (black line) and for patients in stage I (n = 72) outlined in grey. Stage of disease at diagnosis had no effect on OS. From (III) (Sjo et al. 2008b).

Download figure to PowerPoint

By univariate analysis the presence of a translocation at the IGH-locus was found to adversely affect PFS. Male sex, presence of B-symptoms and an IPI score >2 were poor prognostic factors for OS (Table 7). In Cox regression multivariate analysis, IGH-translocation was the only factor associated with PFS, whereas male sex, presence of B-symptoms and unfavourable IPI score retained independent adverse prognostic significance considering OS.

General Discussion

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

We reclassified all ophthalmic lymphomas diagnosed in the Danish population during the period 1980–2005 and found 228 cases. Thus, lymphoma in the ocular region is relatively uncommon with an annual incidence of around two per 100,000 for people above 70 years of age.

We found that MALT lymphoma was the most frequent subtype in the orbit and ocular adnexa. This is in agreement with other studies (Cho et al. 2003; Coupland et al. 2004b; Tovilla-Canales et al. 2004; Rosado et al. 2006; Ferry et al. 2007). Intraocularly, most lymphomas were high-grade lymphoma subtypes with DLBCL as the largest subgroup (Chan et al. 2002; Coupland et al. 2005; Levy-Clarke et al. 2005; Jahnke et al. 2006). These findings are in accordance with the characteristics of lymphocytes in the tissue in which they arise.

Lymphoma characteristics and aetiology according to localization

The retina is considered a part of the CNS and as the RPE layer serves as a blood–retina barrier, lymphoma arising in the RPE, sensory retina, and optic nerve can be considered primary CNS lymphoma, whereas uveal lymphoma usually evolves as a secondary event to systemic lymphoma (Chan et al. 2002; Jahnke et al. 2006). In our material only two of 12 intraocular lymphomas were located in the retina and both were primary intraocular DLBCL. The remaining 10 were located in the uvea of which seven had a history of lymphoma and three (one DLBCL, one PTL,NOS and one B-cell lymphoma unclassified) had uveal lymphoma as first presenting symptom. However, two of these (one DLBCL and one peripheral T-cell lymphoma, unspecified) showed to have systemic involvement of lymphoma at time of diagnosis. Thus, it was likely that the uveal lymphoma was actually a secondary event in these cases.

The conjunctiva, lacrimal gland and the lacrimal drainage system are encompassed in the common mucosal immune system (MALT) and contain numerous lymphocytes and IgA secreting plasma cells equivalent to the lymphoid tissue of the gastrointestinal tract (Allansmith et al. 1987; Knop & Knop 2001; Paulsen et al. 2002). Consequently, a high frequency of MALT lymphoma is to be expected.

Unexpectedly, we found the lacrimal sac to differ in distribution of lymphoma subtypes from the remaining ocular region, with DLBCL and MALT lymphoma being equally frequent. Furthermore, three of 15 tumours were morphologically intermediate between MALT lymphoma and DLBCL. The somewhat hidden location of the lacrimal sac may delay time to diagnosis and thereby increase the probability of further genetic instability leading to transformation from MALT lymphoma to DLBCL. Therefore, some of the DLBCL possibly represents transformation/progression from MALT lymphoma (Matolcsy 1999; Sjo et al. 2006). Furthermore, infections in the lacrimal sac are difficult to treat and often become chronic. This chronicity may augment the antigenic load in the lacrimal sac and allow the progression of MALT lymphoma to DLBCL (Matolcsy 1999).

An understanding of the distribution of lymphoma subtypes in the lacrimal sac compared with the remaining ocular region (except intraocular lymphoma) should perhaps be turned the other way around. DLBCL is overall the most frequent lymphoma subtype in humans and in the most well known MALT containing region, the gastrointestinal tract, DLBCL accounts for 50–66%, whereas only 15–40% is MALT lymphoma (Koniaris et al. 2003). Likewise, the sinonasal region, the eustachian tube and the middle ear are closely related to the ocular region in exposure to antigens and in function of lymphocytes (Matsune et al. 1996; Paulsen et al. 2002), however, MALT lymphoma is hardly ever seen in these regions (Cuadra-Garcia et al. 1999; Merkus et al. 2000; Lang et al. 2003). Seen in this perspective, the interesting question is not why DLBCL is so frequent in the lacrimal sac, but rather why MALT lymphoma is so frequent in the remaining ocular region.

One explanation is that a specific infectious antigen particularly prone to induce malignant transformation of the B-cells in the marginal zone, is the causative agent in ocular region MALT lymphoma, similar to the findings of a connection between gastrointestinal MALT lymphoma and H. pylori (Parsonnet et al. 1994). In keeping with this assumption, an Italian study of 40 patients with ocular region lymphoma found C. psittaci DNA in 21/24 MALT lymphomas. Additionally, three of five DLBCL and eight of 11 unspecified ocular lymphomas showed evidence of C. psittaci DNA (Ferreri et al. 2004). However, these results have not been substantiated in subsequent studies (Daibata et al. 2006; Mulder et al. 2006; Rosado et al. 2006; Vargas et al. 2006). Furthermore, in our cohort of patients a preceding infectious disease was only reported in a minority of the clinical files.

Autoimmune disease is another known risk factor for NHL and an association of MALT lymphoma and Sjögrens syndrome in particular has been reported (Voulgarelis et al. 1999; Smedby et al. 2006). In our group of MALT lymphoma patients, five patients (4%) reported known autoimmune disease and only two of these had Sjögrens syndrome. This is in contrast to the study of Wöhrer et al. (2007) reporting data from an Austrian oncology department where all patients diagnosed with MALT lymphoma routinely undergo clinical and serological assessment for autoimmune disease. They found that 45 of 61 patients (74%) with extragastric MALT lymphoma were diagnosed with an underlying autoimmune disease.

Because of the retrospective nature of our study our assessments are based only on reporting of autoimmune disease in the clinical files. It is possible that the true incidence of autoimmune disease in our cohort of patients indeed is higher than the reported 4%. However, if an underlying autoimmune disease was to be the initiating factor in most cases, clinical signs must have been absent, diminutive or easily treated in most of our patients. Yet, in autoimmune disease the increased lymphoma risk is mainly because of a high inflammatory activity and is thus correlated to severity of disease (Ioannidis et al. 2002). Therefore, it is unlikely that the patients would not have reported some symptoms.

Furthermore, recent assessments indicate that the predominance of MALT lymphoma in Sjögrens syndrome may not be as high as previously believed. In two large studies of patients with Sjögren’s syndrome, DLBCL was reported as the most common subtype (Tonami et al. 2003; Theander et al. 2006). Consequently, the ocular region represents a region with particularly high frequency of MALT lymphoma and neither infection nor autoimmune disease has so far been shown to be the cause for this predominance of MALT lymphoma.

Incidence of ophthalmic lymphoma

The age-adjusted incidence rate of ophthalmic lymphoma in this study is comparable with that in the American population in cancer-registry based analyses (Margo & Mulla 1998; Moslehi et al. 2006).The use of three different population based registries and the fact that all specimens were reviewed and reclassified gives a very reliable result in our investigation. Furthermore, we observed an increase in incidence of orbital- and ocular adnexal lymphoma, as also found in the American population (Margo & Mulla 1998; Moslehi et al. 2006). In Denmark the annual age-adjusted incidence rate more than doubled during the period 1980–2005. The increase was consistent for both men and women, and applied for the lymphoma subtypes MALT lymphoma and follicular lymphoma and for all ophthalmic localizations except intraocular lymphoma. For the lymphoma subtypes DLBCL and mantle cell lymphoma, which are more likely to be part of a systemic disease, we found no statistically significant increase. During the same time-interval the incidence rate of NHL in general has risen as well, but at a slower rate and it seems to have stabilized during the late 1990s (Howe et al. 2001; Clarke & Glaser 2002).

During the period 1980–2005 lymphoma classification has changed, and some cases formerly classified as reactive lymphoid hyperplasia would possibly have been classified as MALT lymphoma according to the REAL- or WHO- classification. However, we found an increase in the total number of cases for both reactive lymphoid hyperplasia and MALT lymphoma. In addition, even though MALT lymphoma was not a described entity until 1983, (Isaacson & Wright 1983) the diagnosis went into practical use in the early 1990s, and thus the increase from mid-1990s to 2005 can be considered reliable.

The reason for the increase in incidence of ophthalmic lymphoma is unsolved. An unidentified subclinical infectious agent with improved living conditions or augmented virulence may be implicated. Also an increase in exposure to sunlight, particularly in the Northern hemisphere, is a reasonable, but not proven explanation (Cartwright et al. 1994).

Genetic alterations of MALT lymphoma in the ocular region

The increasing incidence of ophthalmic lymphoma and especially that of ophthalmic MALT lymphoma, calls for studies clarifying its pathogenesis. It is generally believed that acquired genetic alterations are involved (Inagaki 2007; Sagaert et al. 2007), because at least four balanced translocations resulting in API2/MALT1, IGH/MALT1, Bcl-10/IGH, and FOXP1/IGH rearrangements have been demonstrated to occur in 20–40% of gastric and pulmonary MALT lymphoma.

The frequencies of the translocations, however, vary greatly depending on the anatomic site at which MALT lymphoma arises (Streubel et al. 2004; Remstein et al. 2006), and in our material only two (5%) MALT lymphomas were rearranged. One was proven to be an IGH/MALT1 translocation, whereas the other showed evidence of IGH and Bcl-6 gene breakage, suggesting the presence of an IGH/Bcl-6 fusion. Bcl-6 is a DNA binding transcription factor of B-cells in the germinal centre, and Bcl-6 plays important roles in regulating differentiation, survival and genetic stability of B-cells (Allman et al. 1996). Translocations involving Bcl-6 are frequently seen in nodal DLBCL and in grade 3 follicular lymphoma (Lo Coco et al. 1994; Katzenberger et al. 2004-). However, the occurrence in MALT lymphoma is rare. Recently, Ye et al. (2008) found seven out of 392 MALT lymphomas with a Bcl-6 translocation, of which IGH was found to be the translocation partner in four.

Since the initiation of this study, the frequency of translocations involving the IGH or MALT1 loci in ocular region MALT lymphoma has been assessed in a few other studies (Adachi et al. 2004; Remstein et al. 2006; Sagaert et al. 2006b; Tanimoto et al. 2006b; Ruiz et al. 2007; Schiby et al. 2007). All these studies have reported frequencies comparable with our study of 0–10%. However, Streubel et al. have conducted three studies (Streubel et al. 2003, 2004, 2005) reporting frequencies of the IGH/MALT1 and FOXP1/IGH in 24–37% and 20%, respectively. Even the frequency of the API2-MALT1 translocation, though generally accepted to be almost nonexistent in ocular region MALT lymphoma, has been reported in up to 13% in two studies conducted in Europe and Japan (Takada et al. 2003; Ye et al. 2005). The variable frequencies of these aberrations indicate that not only site, but also geographical, environmental or other unknown conditions may play an important role in the type of genetic abnormalities in MALT lymphoma. Additionally, the low frequency of the known specific translocations raises the question if NF-κB activation is only involved in the pathogenesis of a minority of ocular region MALT lymphoma. Alternatively, NF-κB is the unifying factor leading to MALT lymphoma, yet only a few of the underlying rearrangements are known.

Clinical characteristics

The presenting symptoms of ophthalmic lymphoma vary according to site of involvement. Most patients with conjunctival, lacrimal gland- and eyelid lymphoma have a visible, characteristic ‘salmon coloured’ tumour, whereas orbital lymphoma most frequently presents with proptosis. Only a few patients, primarily with orbital lymphoma, complained of pain. Proptosis, and especially unilateral proptosis should always lead to a CT- or MR scan on the suspicion on malignancy. A following biopsy will provide the lymphoma diagnosis. For lymphoma arising in the lacrimal sac, epiphora and swelling are the predominant symptoms. These symptoms are non-specific, and might be misdiagnosed as chronic dacryocystitis with nasolacrimal duct obstruction. There is only one patient in the literature with primary lacrimal sac lymphoma with bloody epiphora (Saccogna et al. 1994). This sign strongly recommends further evaluation, as it is never found in nasolacrimal obstruction because of chronic dacryocystitis (Francis & Wilcsek 2006). Furthermore, the extension of the swelling above the medial canthal ligament should raise the suspicion of malignancy (Parmar & Rose 2003).

Intraocular lymphoma remains one of the most difficult diagnoses to establish, as it frequently masquerades as an idiopathic steroid-resistant chronic uveitis. Patients complain of blurred vision, a painless loss of vision and/or ‘floaters’ (Coupland et al. 2004a). In our material visual loss was the most common presenting symptom, occurring in 70% of patients with intraocular lymphoma (data not shown). Even though several other eye diseases can cause visual loss, fundus examination may show creamy yellow deposits characteristic of lymphoma. Hence, clinical presentation (in extraocular lymphoma) and fundus findings (in intraocular lymphoma) should lead to lymphoma suspicion. The final diagnosis and lymphoma subtyping, however, depend entirely on histopathological and immunohistochemical examination.

Treatment of ophthalmic lymphoma

The management of ophthalmic lymphoma follows the same modalities proposed for nodal lymphoma. Yet, several criteria must be considered in the initial assessment of the disease to clearly define optimal treatment, they are: (i) the histopathologic subtype of lymphoma, (ii) extension of the disease (in the ocular region and systemically), (iii) prognostic factors related to the patient – and to the disease and (iv) the impact of the lymphoma on the eye(s) and visual function. Thus, treatment decision is different from patient to patient and uniform guidelines for treatment of ophthalmic lymphoma do not exist.

It is generally accepted that excellent local control is achieved with radiation therapy in localized primary ocular region lymphoma (Tsang et al. 2003; Uno et al. 2003). Chemotherapy or a combination of chemo- and radiotherapy is given to patients with systemic spread. For patients treated with radiotherapy alone, the risk of relapse in distant extranodal sites remains a significant problem and it is tempting to consider an initial role for chemotherapy. In our study, we found that the 5-year relapse rate for stage I MALT lymphoma patients was higher for patients not treated with chemotherapy. However, the difference was not significant compared to radiation therapy only (Table 7). Furthermore, there was no effect on OS, and consequently, chemotherapy does not seem relevant for patients with MALT lymphoma, at least in the early stages.

Another controversial question considering treatment for stage I primary ocular region MALT lymphoma is whether radiation or chemotherapy treatment is necessary (Tanimoto et al. 2006a). In our cohort 11 patients were not treated with radiation- or chemotherapy initially. Relapse/progression rate was higher for this subgroup, but not statistically significant. Furthermore, there was no effect on OS, even though the majority of patients did not receive any treatment during a median follow-up of 65 months (range 5–204 months). Finally, monoclonal anti-CD20 antibody seems to be an efficient, but expensive treatment for primary ocular region MALT lymphoma (Ferreri et al. 2005) and perhaps even for DLBCL localized intraocularly (Rubenstein et al. 2007).

Prognostic factors in MALT lymphoma arising in the ocular region

In our study around one fourth of patients had involvement of other organs at presentation, and one third had relapse or progression of disease after initial therapy. However, with a 5-year OS of 75% and only 10% of patients dying from their lymphoma, our data confirm that MALT lymphoma usually has a quite indolent course even in cases with widespread disease.

There are currently no generally accepted prognostic factors for primary ocular region MALT lymphoma. Most of the previous studies investigating prognostic factors in ocular region lymphoma have been influenced by the wide variety of lymphoma subtypes (Jenkins et al. 2003; Meunier et al. 2004; Sullivan et al. 2005; Plaisier et al. 2007). Recently, Tanimoto et al. (2007) reported data concerning 114 Japanese patients with primary MALT lymphoma, and found like we did that PFS- and OS rates were not related to the initial stage of disease. In our analysis, the IPI-score was the most reliable prognostic factor for OS. Furthermore, we found a statistically significant shorter PFS for patients with IGH translocation-positive tumours. Even though this should be interpreted with caution because of the low number of positive tumours, our results indicate that IGH translocation may be associated with a more aggressive course of disease. In gastric MALT lymphoma, t(11;18) is more common in cases that show spread to regional lymph nodes or distal sites (Liu et al. 2001). However, despite this significant association of t(11;18) with advanced disease, translocation positive MALT lymphoma rarely, if ever, undergoes high grade transformation. An effect on PFS or OS has only been assessed in one previous study of 90 cases with gastric MALT lymphoma (Nakamura et al. 2007) and in this study PFS and OS was not affected by IGH-translocation status.

Conclusions and Perspectives

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

The material studied in this thesis is unique, covering all lymphomas diagnosed through a 26-year period in the entire Danish population. We have obtained knowledge that will be of value for optimising counselling and treatment of patients in the future.

All hypotheses were tested and the answers summarises the results of the study:

Orbital and ocular adnexal lymphoma is predominantly MALT lymphoma, whereas the most frequent intraocular lymphoma subtype is DLBCL (A). However, lymphoma arising in the lacrimal sac was surprisingly predominantly DLBCL (B).

The incidence of ophthalmic lymphoma in the Danish population is comparable with that found in the American population (C), and during the period 1980–2005 the incidence increased corresponding to an average annual increase of 3.4% (D).

When considering the most frequent ocular region lymphoma subtype, MALT lymphoma, about one fourth of patients had widespread systemic disease (stage IV). However, stage of disease at diagnosis was not related to prognosis. The most reliable prognostic clinical factor was the IPI score (E).

Only one of 42 MALT lymphomas had a translocation leading to upregulation of one of the known NF-κB activating proteins (MALT1). Therefore, this study does not confirm the hypothesis that MALT lymphoma arises because of NF-κB activation (F). The t(14;18)(q32,q21) translocation is not frequent in ocular region MALT lymphoma accounting for 2% in our study, whereas t(11;18) (q21;q21) was not found in any of the specimens (G).

Two of 42 specimens had translocations involving IGH and time to relapse was shorter for these two patients compared with the translocation negative patients. However, because of the small number of events this should be interpreted with caution. Furthermore, one of the IGH translocated cases did not involve Bcl-10, MALT1 or API2-MALT1, but Bcl-6 and the relevance of this translocation in the pathogenesis of MALT lymphoma is not established (H).

The increasing incidence of ophthalmic lymphoma and especially that of ophthalmic MALT lymphoma, emphasizes the relevance of studies clarifying its pathogenesis. As the structural aberrations associated with MALT lymphoma seem to be very infrequent it is likely that the spectrum of chromosomal changes associated with MALT lymphoma will include several additional, hitherto unknown changes. Recently, a number of novel chromosomal translocations have been identified in MALT lymphoma of various sites (Streubel et al. 2007). However, the incidence and their pathophysiological impact remain to be investigated.

Despite the identification of further chromosomal translocations, it is highly likely that a high proportion of MALT lymphoma is negative for chromosomal translocations. Comparative genomic hybridization investigations of the stomach and salivary glands have shown a conserved pattern of chromosomal gains (Zhou et al. 2006, 2007). It is important to explore further whether these recurrent chromosomal gains are a common feature of translocation negative MALT lymphoma, and to identify genes targeted by such genomic gains. Furthermore, the search for an underlying causative infectious agent should not be disregarded. Not only to elucidate the pathophysiology, but also because of a major clinical importance, as many patients might be treated with antibiotics instead of radiation- or chemotherapy.

Danish summary

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References

Non-Hodgkin lymfom (NHL) er den syvende hyppigste cancertype i Danmark med en livstidsrisiko på 1% og 700 nye tilfælde pr. år. Incidensen af NHL har været stigende i de vestlige lande gennem de seneste årtier og derfor er NHL et stadigt større klinisk problem.

Lymfom i øjenregionen (lymfom lokaliseret i øjenlåget, konjunktiva, tåresækken, tårekirtlen, øjenhulen eller intraokulært) er en sjælden sygdom som udgør 5–10% af alle extranodale lymfomer. Alligevel er det den hyppigste maligne tumor i orbita.

Formålet med denne afhandling var at lave en oversigt over alle danske patienter med lymfom i øjenregionen i perioden 1980–2005 for at bestemme fordelingen af lymfom undertyper samt incidensen og den tidsmæssige udvikling i incidens. Desuden var det ønsket at undersøge kliniske faktorer, samt kromosomale forandringer med betydning for prognosen for den hyppigst forekommende lymfomtype, extranodalt marginal zone lymfom (MALT lymfom).

Vi verificerede diagnosen lymfom i øjenregionen hos 228 danske patienter efter revurdering og reklassificering af tidligere vævsprøver. Mere end 50% af lymfomerne i området omkring øjet var MALT lymfom, mens de intraokulære lymfomer primært var diffust storcellet B-celle lymfom (DLBCL). Incidensraten var meget aldersafhængig. For hele populationen var der en stigning i incidensraten fra 1980–2005 svarende til en gennemsnitlig årlig procentvis stigning på 3,4%.

Et hundrede og seksten patienter havde MALT lymfom opstået i øjenregionen. En tredjedel af disse patienter fik recidiv eller progredierede efter initial behandling. Recidiv sås ofte andre steder end i øjenregionen. Dødeligheden var dog ikke højere for patientgruppen med recidiv.

Vi fandt desuden, at frekvensen af translokationer involverende MALT1- og IGH generne er lav i MALT lymfomer opstået i øjenregionen (2 af 42, 5%), men tilstedeværelsen af disse øger muligvis risikoen for recidiv.

Konklusionen er at incidensen af lymfom i øjenregionen er markant stigende i Danmark. Den hyppigste lymfom undertype i denne region er MALT lymfom. Den molekylære patogenese af MALT lymfom i øjenregionen involverer kun sjældent translokationer i MALT1- og IGH generne.

References

  1. Top of page
  2. Abstract
  3. Acknowledgements
  4. Introduction
  5. Background
  6. Hypotheses and Aims of the Study
  7. Aims
  8. Material and Methods
  9. Results
  10. General Discussion
  11. Conclusions and Perspectives
  12. Danish summary
  13. References