Introduction
The knowledge of professionals generally has a half-time of about 5 years. So half
of everything you will read in this review will be outdated 5 years from now. That
may be not so bad, but it also means that everything you have learned more than 5 years
ago is only half true, and do you know which half? Keeping up with the literature
is an almost impossible task and never complete. With the introduction of the new
WHO classification at the EAHP/SH workshop in Bordeaux last year, several major changes
are introduced into our daily work. This will be the subject of review in the next
issue of the Journal of Hematopathology. For now, there is also quite some interesting
data to digest:
Biology of lymphoma
Hodgkin lymphoma
The successful meeting of the EAHP in Bordeaux focused on tumor microenvironment,
an exiting but difficult area. We know how important the stroma and its cells are
for tumor growth, but the complexity of the interplays between cells, reporter molecules,
and matrix is so large that we have only limited understanding. We may have to await
further large steps in systems biology to get a grasp of this by modeling. Nevertheless,
interesting work is done in especially Hodgkin lymphoma (HL), a tumor in which the
environment is remarkably large. Chetaille et al. [1] studied 63 cases of classical
HL(cHL) by microarray and found that EBV-positive cases differed in profile from EBV-negative
cases by the presence of a Th1 response gene signature. Furthermore, they found that
when an overexpression of genes related to B cells and apoptosis were present, patients
had a favorable outcome. They were able to confirm this by immunohistochemistry: poorer
prognosis in the case of high percentage of either TIA-1-positive reactive cells or
topoisomerase-2-positive tumor cells, whereas high numbers of BCL11A, FOXP3, or CD20-positive
reactive cells had a favorable influence. Immunodetection of BCL11A, a marker of both
B cells and plasmacytoid dendritic cells, had the strongest predictive value. EBV-positive
cHL (n = 108) was also studied by Asano et al. [2], and in comparison with age-related
EBV-positive B-cell proliferations (n = 34), since also the latter often has an inflammatory
background similar to cHL. cHL had a lower ratio of geographical necrosis, lower number
of cytotoxic T cells, and a better prognosis. Many more of such studies will be needed
before there will be a reliable marker set that predicts who of the relatively few
patients with cHL will relapse.
B-cell lymphomas
Also in B-cell lymphomas, the microenvironment is subject of a study: de Jong et al.
[3] studied with immunohistochemistry 61 pretreatment biopsies from patients treated
in a clinical trial and found that the prognostic impact was different for the two
treatment arms. CD69 expression on tumor cells was a poor prognostic sign, and an
interfollicular infiltrate of FoxP3-positive T cells was a good prognostic sign irrespective
of the treatment arm, and the authors suggest that a dense infiltrate of FoxP3-positive
T cells, dense and interfollicular infiltrate of CD68-positive macrophages, and complete
follicular dendritic meshworks were associated with a favorable time to progression
in CVP-treated patients, while being a poor prognostic sign in fludarabine-treated
patients. This finding confirms the importance of moving away from prognostic marker
studies and go to predictive markers (van Krieken [4]). Ai et al. [5] did in vivo
experiments that partially explain the findings of de Jong et al.: follicular lymphoma
B cells in contrast to normal B cells could induce conventional T cells to express
FoxP3 and to acquire regulatory function.
The cell of origin in multicentric Castleman disease (MCD) was hypothesized to be
the HHV-8-positive B cell by Chadburn et al. [6]. The phenotype of this cell was similar
to that of primary effusion lymphomas (PAX5, Bcl-6-; PRDM1/BLIMP1, IRF4/MUM1+; Ki67+)
but varied in maturation, since in MCD, cells expressed also Oct2 and cytoplasmic
immunoglobulin. Mack and Sugden [7] showed that EBV and HHV8 are both needed for sustained
cell proliferation in primary effusion lymphomas.
In vitro studies are used to predict efficacy of drugs. Even though there is often
limited correlation with in vivo studies, these experiments do give insight in the
biology of tumors. Everton et al. [8] studied the effect of rituximab on the proteasome
of follicular lymphoma cells and showed a change in expression for a multitude of
proteins, including molecules involved in migration, adhesion, signaling, ubiquitination,
and others. Lapalombella et al. [9] combined two drugs, rituximab and lenolidomide,
both as single agent effective in CLL, and showed that lenolidomide results in downregulation
of CD20 on CLL cells, resulting in lower efficacy. This results predict that this
combination is not effective in clinical studies. In cell lines of MCL, Jones et al.
[10] showed that a new agent, MI-63, induces p53 and HDM-2 accumulation and inhibition
of proliferation which was synergistic with m-TOR inhibitors.
Roue et al. [11] were able to show that bendamustine, a multifunctional alkylating
agent, has cytotoxic activity in cell lines of MCL and CLL and that in p53 mutated
cell lines, there was a synergistic effect with nucleoside analogues.
CyclinD1 overexpression is the hallmark of MCL, and Dal Col and Colcetti [12] tried
to downregulate cyclinD1 in vivo. They show that inhibition of PI3-K/Akt induces a
40% decrease of cyclin D1 half-life as a result of accumulation of the dephosphorylated/active
form of GSK-3beta within the nucleus, where this kinase can phosphorylate cyclin D1.
Translocation of cyclin D1 into the cytoplasm is mediated by the nuclear exportin
CRM1, whose association with cyclin D1 increases following PI3-K/Akt inhibition. Notably,
rapamycin downregulated GSK-3beta Ser9 phosphorylation with concurrent nuclear export
of cyclin D1 only in MCL cells in which GSK-3beta is under the control of mTOR. These
findings suggest that the ability to downregulate cyclin D1 through GSK-3beta may
identify subsets of MCL patients who may benefit from the treatment with mTOR inhibitors.
Two very similar studies show the complexity of the genetic changes in mantle cell
lymphoma (MCL). Both Schraders et al. and Bea et al. performed combined genomic and
expression profiling. Schraders et al. [13] used a restricted group of classical mantle
cell lymphomas, whereas Bea et al. [14] used both cell lines and tumors with different
morphologies. Not unexpectedly, this approach resulted in higher percentages of specific
changes in the Schraders study, pointing to mitochondrial metabolism, DNA repair,
and chromosomal stability, whereas the Bea paper indicates a great variation of genes
in many pathways, but often affected in only a minority of the cases. In a more direct
approach, Gelebart et al. [15] investigated the Wnt pathway in MCL and found it affected
in both cell lines (three out of three) and tumor samples (33 out of 46).
A very different approach for treating MCL patients might be immunotherapy directed
against cyclinD1; this might be possible since Kondo et al. [16] found in the blood
of MCL patients cytotoxic T cells directed against the cyclinD1 positive MCL-cells.
Espinosa et al. [17] studied the NF-kappaB signaling pathway in diffuse large B-cell
lymphoma (DLBL) using immunohistochemistry for the involved proteins. They were able
to show that NF-kappaB activation is seen in a minority of cases, that it is generally
mediated through the classical pathway, not associated with the activated B- or germinal
center type and that it predicts good prognosis.
Uddin et al. [18] studied S-phase kinase protein 2 (SKP2), an F-box protein, that
targets cycle-dependent kinase inhibitor p27KiP1 via ubiquitin-mediated degradation.
Using a large number of tissue samples from DLBL and cell lines, they showed that
about 40% of the DLBL expressed SKP2, which was inversely related to p27Kip1 expression.
Treatment of the cell lines with bortezomib resulted in downregulation of SKP2 and
increased expression of p27Kip1 resulting in increased apoptosis.
Deregulated Myc results in aggressive lymphomas. Mason et al. [19] investigated in
a mouse model the efficacy of ABT-737, a small molecule that mimics antagonists of
bcl-2 proteins. This drug was effective in double hit tumors (bcl2 and myc translocation),
but not in myc lymphomas only.
T-cell lymphoma
The knowledge on the biology of T-cell lymphomas is less than that of B-cell lymphomas,
due to the fact that these tumors are rarer, but also due to the larger complexity
of the T-cell system compared to the B cells. The importance of the WNT signaling
pathway in lymphocyte development and in many forms of cancer prompted Groen et al.
[20] to investigate the nuclear expression of beta-catenin in a large panel of lymphomas
(n = 162). In contrast to Gelebart (see above), they did not find MCL affected, but
found nuclear expression mainly in precursor T-cell malignancy (nine out of 27) and
a few mature T-cell lymphomas, and they showed gain of function mutation in the gene
in a few of them. In the majority of lesions, however, the WNT pathway was still functional
and thus a potential target for therapy.
Anaplastic large-cell lymphoma (ALCL) is relatively well understood, since it has
often a translocation involving the ALK gene. The remarkable morphology of the cells
now seems to be due to the function of the NPM-ALK fusion gene. Ambrogio et al. [21]
showed that the kinase activity of the nucleophosmin (NPM)-ALK fusion regulated the
shape of ALCL cells. NPM-ALK formed a complex with the guanine exchange factor VAV1,
enhancing its activation through phosphorylation. VAV1 increased Cdc42 activity, and
in turn, Cdc42 regulated the shape and migration of ALCL cells.
Marzec et al. [22] report that the NPM/ALK-carrying T-cell lymphoma cells strongly
express the immunosuppressive cell-surface protein CD274, as determined on the mRNA
and protein level. Using the small molecule ALK inhibitor CEP-14083, they were able
to inhibit NPM/ALK function resulting in the downregulation of CD274 expression. These
findings identify an additional cell-transforming property of NPM/ALK and describe
a direct link between an oncoprotein and an immunosuppressive cell-surface protein.
The proteasome inhibitor bortezomib is now extensively studied, since it has the potential
of being active in many forms of cancer. Shen et al. [23] had found earlier that it
is effective in NK-cell neoplasms, and now report that it is effective in some cell
lines at the same doses that is used for multiple myeloma inducing mitochondrial apoptosis,
whereas in one cell line, a higher concentration was needed, after which mitotic catastrophe
occurred. In the same tumor type, that is invariable carrying the Epstein–Bar Virus
(EBV), Ian et al. [24] were able to inhibit proliferation by silencing the EBNA1 gene
of the virus.
Epidemiology of lymphoma
Even though there is an improving survival for many forms of cancer, history of medicine
tells us that the discovery of the cause of a disease is of great importance for the
final solution. Etiological studies in cancer are relatively limited, and until now,
epidemiological studies on risk factors did use lymphomas as a broad category. Now
we know that lymphomas form a series of different disease entities, and thus, studies
into risk factors should separate the different forms of lymphoma. Morton et al. [25]
did such an analysis on the more common types, DLBL (although this is likely not a
single entity), follicular lymphoma (FL), marginal zone lymphoma (MZL), and CLL. They
found that late birth order and high body mass index (≥35 kg/m2) increased risk for
DLBCL alone. Autoimmune conditions increased risk for marginal zone lymphoma alone.
The tumor necrosis factor G-308A polymorphism increased risks for both DLBCL and marginal
zone lymphoma. Exposure to certain dietary heterocyclic amines from meat consumption
increased risk for CLL/SLL alone. There were no significant risk factors for follicular
lymphoma alone. These data indicate clearly that such studies need to focus on separate
disease entities.
The association of Chlamydia psittaci (CP) with ocular marginal zone lymphoma is well
known. Ponzoni et al. [26] investigated in 205 extraorbital lymphomas and 135 non-neoplastic
controls the presence of CP by multiple methods. CP DNA was equally prevalent in nonocular
nodal and extranodal lymphomas: among the latter, it was more common in diffuse large
B-cell lymphomas of the skin (P = 0.03) and Waldeyer’s ring. This important study
sheds new light on the causative role of this microorganism in the development of
lymphomas.
Improving outcome for lymphoma patients is well described in clinical studies, but
those studies include always a selection of patients. It is therefore important to
study outcome on a population basis. Al-Tourah et al. [27] investigated the data from
the population-based registry of British Colombia (Canada) and analyzed 600 patients
with newly diagnosed FL. With a median follow-up of 109 months, 170 (28%) developed
transformation, 107 (63%) of them based on biopsy confirmation, the others on clinical
parameters. The annual risk of transformation was 3% continuously through 15 years.
A multivariate analysis of clinical factors at diagnosis identified advanced stage
as the only predictor of future transformation. The median post-transformation survival
was 1.7 years. The 5-year survival was superior for patients with limited extent transformation
compared with those with advanced cases (66% vs 19%, P < .0001). Clinically diagnosed
transformation has an equal impact on outcome as biopsy proven transformation.
Many studies are available that describe specific features of certain lymphoma types
in several regions of the world. By comparing these, one gets further insight in the
background of these lymphoma types, although it is often difficult to understand what
the causes of differences are. Nomura et al. [28] describe 21 pediatric high-grade
B-cell lymphomas in Papua New Guinea and found a frequent association with EBV, regardless
of subtype, and even in lymphoblastic lymphoma (LBL). A large series of endemic Burkitt
lymphomas (BL) was described from Uganda by Ogwang et al. [29]. Among 500 patients,
median age was 6 years and male-to-female ratio was 1.8:1. The age-standardized incidence
was 2.4 per 100,000, being 0.6 in 1–4 year olds, 4.1 in 5–9 year olds and 2.8 in 10–14 year
olds. Most children presented with abdominal (56%, M/F 1.4:1) vs. only facial tumors
(35%, M/F 3.0:1). Abdominal tumors occurred in a little older (mean age: 7.0 vs. 6.0 years)
and more frequently in female children (68% vs. 50%). A study on 234 BL cases from
Brazil by Queiroga et al. [30] revealed that EBV was present in 53% of all BL cases,
varying from 29% (12 out of 42) in the South to 76% (13 out of 17) in the North. The
frequency was higher in the pediatric group. Expression of p53 protein was observed
in 16%, and only rare cases showed p63 expression. They conclude that BL in Brazil
is regionally distinct and has a low incidence of p53 overexpression and a higher-than-expected
association with EBV in sporadic cases.
From Argentina, a series of 40 extranodal oral lymphomas was reported by Keszler et
al. [31]. Overall mean age of patients was 49 years, and frequency was greater in
males. The most common location was the gingiva, followed by the palate. Intraosseous
cases were more frequent in mandible (75%) than in upper maxilla. Of the cases, 100%
were B-cell lymphomas, with a high frequency of BLBL. Only 60% of the plasmablastic
lymphomas in the series came from HIV+ patients.
Paul et al. [32] describe 56 cases of primary central nervous system lymphomas from
India out of a 19-year study period. These lymphomas comprised 1% of all intracranial
neoplasms. The patients ranged from 10 to 75 years of age with a median age of 42 years,
which is relatively young. Only one patient was HIV positive. All cases were diffuse
large B-cell lymphomas on histopathology except for a single case of CD 3-positive
T-cell lymphoma.
Defining entities
Hodgkin lymphoma
Our understanding of Hodgkin lymphoma has increased largely but is by far not complete.
Using the new insights, one might forget earlier findings. Benharroch et al. [33]
reviewed their 201 cases of cHL and used strict criteria for subclassification by
which they detected 12 cases of lymphocyte-depletion. These cases were often EBV positive
and also CD15, although the latter was important in the original diagnosis. To increase
the understanding of the relatively rare nodular lymphocyte predominant HL, Brune
et al. [34] performed expression array analysis on microdissected L&H cells and compared
the results to those of normal B cells and B cells from other lymphoma types. There
was a very high similarity to the tumor cells from T-cell-rich B-cell lymphoma as
well as cHL, with partial loss of B-cell phenotype and altered expression of many
genes involved in cancerogenesis.
B-cell lymphomas
Most follicular lymphomas (FL) present in lymph nodes, but extranodal presentation
may occur. Weinberg et al. [35] compared 27 extranodal FL to 44 nodal cases, and in
addition, eight cutaneous follicle center cell lymphomas, known to be a different
entity. Extranodal cases were more often diffuse, lacked CD10 expression and had,
surprisingly, more often t(14;18), especially the minor breakpoint. Follow-up was
quite similar between the groups. This paper points to the difficulty of the use of
many criteria (clinical, morphological, phenotypical, genetic) to make a diagnosis,
and then use the same criteria as prognosticator or as single criterium.
Two studies deal with ocular/orbital lymphomas, Lagoo et al. [36] focusing on pathological
features, Sjo et al. [37] on more clinical aspects. This relatively rare lymphoma
is not always easy to diagnose and classify, since it often has not the typical features
of MZL compared to gastric cases. Only a minority of the cases has a translocation
that involves the MLT gene (15% in the USA study, 5% in the Danish study). About 25%
of the patients have disseminated disease, and 5-year survival is about 75%.
Takino et al. [38] collected cutaneous MZL from Asia, Germany, and the USA; in total,
60 cases and showed that they have fairly common clinical and pathologic features.
In nine out of 25 Asian cases, there was tissue eosinophilia, which was virtually
absent in the Western cases. In no case was Borrelia or the MALT1-API2 fusion found.
Quite some work is done on the detection of a molecular marker for MZL. It is already
well known that several translocations leading to NFkappaB activation occur in extranodal
MZL. Vinatzer et al. [39] studied 29 cases from different sites with several cytogenetic
methods and document the following: balanced translocations were found in 21 cases.
IGH was rearranged in the majority of cases with balanced translocations (n = 17/21);
three cases had t(11;18)/API2-MALT1, and one case had novel t(6;7) (q25;q11), respectively.
IGH partner genes involved MALT1, FOXP1, BCL6, and four new chromosomal regions on
chromosome arms 1p, 1q, 5q, and 9p. LDI-PCR identified three novel partner genes on
1p (CNN3), 5q (ODZ2), and 9p (JMJD2C). FISH assays were established and confirmed
LDI-PCR results. QRT-PCR showed deregulation of the novel genes in the translocation-positive
cases. These findings underline the molecular heterogeneity of MZL and also demonstrate
that presently molecular classification is not possible. Sinn et al. [40] focused
on colorectal and gastric cases and found MLT translocation in two out of nine cases
and CpGisland methylation (CIMP) eight out of 13 cases. The CIMP-positive cases had
a poorer survival. Chanudet et al. [41] used high-resolution chromosome 6 tile-path
array-CGH which identified NF-kappaB inhibitor A20 as the target of 6q23.3 deletion
and TNFA/B/C locus as a putative target of 6p21.2–22.1 gain. Interphase fluorescence
in situ hybridization (FISH) showed that A20 deletion occurred in MALT lymphoma of
the ocular adnexa (8/42 = 19%), salivary gland (2/24 = 8%), thyroid (1/9 = 11%), and
liver (1/2), but not in the lung [26], stomach [45], and skin [13]. A20 deletion and
TNFA/B/C gain were significantly associated and exclusively found in cases without
characteristic translocation. In ocular cases, A20 deletion was associated with concurrent
involvement of different adnexal tissues or extraocular sites at diagnosis, a higher
proportion of relapse (67% versus 37%), and a shorter relapse-free survival.
Lymphoplasmacytic lymphoma (LPL) is difficult to separate from MZL. Sargent et al.
[42] studied 17 nodal LPL which were divided into classic (n = 5), vaguely nodular
(n = 5), and others (n = 7). They were able to show that among the classic LPL, four
out of four had an IgM paraproteinemia, five out of five had bone marrow involvement
(BM+), and one out of five had +MALT1. One of one vaguely nodular LPL had an IgM paraprotein,
two out of four were IgM+, two out of four IgG+, one out of three had BM+, and one
out of five had an IgH rearrangement. Among the other cases, two out of three had
a paraprotein, two out of seven were IgM+, five out of seven IgG+, and zero out of
three had BM+. Of these cases, one showed +12, one +18, and one IgH/BCL2 translocation
plus +18. None of the 17 cases had a 6q21 deletion or +3. These results are quite
different from those obtained generally in MZL, but the genetic data differ also from
those seen in bone marrow-derived LPL.
Although FL is characterized by t(14;18), not all cases carry the translocation. Gaqyi
et al. [43] collected bcl-2 negative (translocation and protein) FL and studied the
mutational status of the IgH gene, but did not find differences compared to bcl2-positive
cases. Additional genetic alterations are common in FL. d’Amore et al. [44] found
in t(14;18) positive cases in initial and follow-up biopsies of 360 patients +7, del(6q),
+der(18) t(14;18), +18, and +X as the common alterations, multiple clones were common,
and genetic divergence was found especially in follow-up biopsies.
Since MUM1 is overexpressed in the activated type of DLBL, Hunt et al. [45] investigated
whether MUM1 is translocated in these tumors. In one out of 16 cases, a MUM1 break
was detected, and none of 17 germinal center-type DLBL carried the break.
Burkitt lymphoma (BL) is defined by morphology, phenotype, and MYC translocation,
but Leucci et al. [46] describe in a series of bona fide endemic (African) BL that
five out of 35 cases did not have a break in the MYC gene that could be detected by
a split signal probe covering the gene, and using fusion probes, no juxtaposition
to immunoglobulin genes was detected either. In these cases, the miRNA hsa-mir-34b
was downregulated, and in in vivo studies this miRNA could regulate MYC expression,
providing an alternative pathway in MYC-negative BL. Such cases were not included
in the expression and CGH array studies by Salivierra et al. [47]. In the array study,
it was shown that pediatric and adult cases of Burkitt lymphoma have similar genetic
alterations but that DLBL that are similar to BL using expression array have different
genetic backgrounds. These data confirm that expression arraying is not a tool to
be used in the classification of tumors.
T-cell lymphomas
Regulatory T cells are a relatively recent recognized subset of T-lymphocytes, expressing
FoxP3 (although not exclusively). Bonzheim et al. [48] investigated whether subtypes
of T-cell lymphomas were derived from these cells but detected among anaplastic large
cell lymphomas (ALCL), angioimmunoblastic T-cell lymphomas (AILD), and peripheral
T-cell lymphomas (PTCL) only one such case, which was rapidly progressive.
T-cell lymphomas are molecularly less characterized compared to B-cell lymphomas.
Feldman [49] performed an important study in which the multiple myeloma oncogene-1/interferon
regulatory factor-4 (IRF4) locus was analyzed by FISH. IRF4 translocations are known
to exist in myeloma and some B-cell lymphomas, but were not yet studied in PTCLs.
Among 169 PTCLs, 12 cases had IRF4 translocations. Two cases with t(6;14) (p25;q11.2)
had translocations between IRF4 and the T-cell receptor-alpha (TCRA) locus. Both were
cytotoxic PTCLs, involving bone marrow and skin. Eight of the remaining ten cases
were cutaneous anaplastic large-cell lymphomas (ALCLs) without TCRA rearrangements
(57% of cutaneous ALCLs tested). These findings demonstrate that IRF4 translocations
are a recurrent genetic abnormality in PTCLs. Cytotoxic PTCLs involving bone marrow
and skin and containing IRF4/TCRA translocations might even represent a distinct clinicopathologic
entity.
A key feature of enteropathy-associated T-cell lymphoma is the loss of surface CD3
with remaining cytoplasmic expression. Using a cell line derived from such a lymphoma,
Tjon et al. [50] were able to determine the mechanism responsible for this: defective
assembly or association of T-cell receptor chains. The finding was confirmed in two
other patients.
ALCL is defined by CD30 expression and morphological features, but the latter may
be misleading in some cases. Bovio and Allen [51] determined by flow cytometry the
expression profile of 20 ALCL, nine ALK1 positive, 11 negative. Almost all cases expressed
at least one T-cell-associated antigen. CD13 and CD33 were expressed in all ALK1 positive
cases, but in only 1 Alk1 negative case. Misinterpretetion for myeloid sarcoma should
be avoided in the positive cases.
Kong et al. [52] did an extensive analysis of 22 Asian cases of subcutaneous panniculitis-like
T-cell lymphoma. Unusual features were found in two cases, namely, CD8 negative. CD30
was negative in all cases, and CD56 was focally positive in two cases. EBV was present
in one case. All cases were clonal using the Biomed 2 primer sets. Cases with angioinvasion
were more aggressive.
The immunophenotype of 84 cases of nasal-type extranodal NK/T cell lymphoma was studied
using TMA’s by Schwartz et al. [53]. CD2 was positive in 93%, CD3 in 84%, CD5 in 27%,
CD56 in 58%, granzyme in 71%, perforin in 86%, TIA1 in 90%, CD30 in 35%, and EBV in
87%; CD20 was absent in all cases.
Cutaneous lymphomas
Cutaneous lymphomas remain an extensively studied group, especially compared to its
relatively rarety. Riou-Gotta et al. [54] described 71 cases from a population-based
registry in France, covering the period from 1980 to 2003. Of the cases, 82% were
of T-cell origin (the majority mycosis fungoides; MF), and thus, 18% were B-cell lymphomas.
There was an increase in incidence from 0.2 to 0.7/100.000 and 5 year survival was
about 65% for MF and also for the other types taken together.
Van Maldegem [55] underlined that cutaneous lymphomas are different from lymphomas
from other sites, also molecularly. They studied an extensive panel of primary cutaneous
MZLs and show that these are different compared to other extranodal MZL. Whereas most
noncutaneous MZLs express IgM, cutaneous MZLs in majority express IgG, IgA, and IgE.
Furthermore, these isotype-switched cases lack CXCR3 and seem to arise in a different
inflammatory environment.
Granulomatous slack skin disease (GSS) and granulomatous MF are rare and puzzling.
Kempf et al. [56] collected four cases of the former and 15 of the latter. These cases
were diagnosed according to the WHO classification. There was little epidermotropism,
and treatment was not successful in most cases, and one third of the patients died
of their disease. Why granulomas occur remains unknown.
New entities/subtypes
Like many immunodeficiencies is the Nijmegen breakage syndrome associated with lymphoproliferations.
Gladkowska-Dura et al. [57] described their experience from 105 such patients, of
whom half developed a lymphoproliferation. From 14 patients, the pathology is described.
Most commonly, DLBL and T-LBL/ALL were present, and all had a clonal disease. Using
split signal probes breaks in Ig and TCR genes were commonly detected, indicating
that the repair defect is indeed responsible for the development of lymphoproliferations.
The description of the pathology in IgG4 disease becomes extensive. Sato et al. [58]
collected experience on ocular lesions from 21 patients. In 17 of them, the lacrimal
gland was involved, the conjunctiva in no patient. There was a dense, fibrotic plasmacellular
infiltrate with IgG4 expression. In two out of 17, the proliferation was clonal.
The subtyping of DLBL is of continuous interest. Lenz et al. [59] performed an analyses
of 203 cases combining expression and CGH array. They were able to show that the three
subgroups, germinal center B-cell-like (GCB), activated B-cell-like (ABC), and primary
mediastinal (PMB) had different patterns of genomic changes, but no specific alterations
were found. INK4a/ARF was commonly deleted, and FoxP1 upregulated in the ABC group.
In the GBC group, amplification of mir-17-92 and deletion of PTEN were often seen.
Wong et al. [60] analyzed 32 DLBL of the upper aerodigestive tract using immunohistochemistry
and found a similar distribution for ABC and GBC cases compared to other sites and
EBV in only one patient. They conclude that DLBL from this site is not a specific
entity.
Dunphy et al. [61] detected a case of PMB with t(14;18) and analyzed 24 cases in detail
using FISH, PCR, and immunohistochemistry. Four of the cases had features that are
associated with germinal center origin, and these might form a subset of this entity.
The role of EBV in lymphoproliferations in immune-deficient patients is well known.
Increasingly, EBV is being recognized in lymphomas from patients with no clear immune
deficiency, especially in the elderly. Kleinschmidt et al. [62] describe four patients
with central nervous system lymphomas that were on imaging suggestive for EBV-associated
cases and who had been treated with a variety of immune suppressive drugs. Since the
association is not well known, and pathologists are not always well informed, the
authors suggest that EBV testing should be done more often.
Ketzenberger et al. [63] described a series of 35 FL with predominantly diffuse growth
pattern. Almost all of the cases lacked the t(14;18) but had a typical immunophenotype
(CD10, bcl6, and CD23 expression). The expression profile of four of the cases was
studied and comparable with more classic FL. The patients had often low-stage disease
and 27 out of 29 cases had a 1p36 deletion, all in all suggesting that this may represent
a separate (sub)entity.
The testis and the central nervous system are considered as immune-privileged sites,
and DLBL arising in these sites (IP-DLBL) have some special features. Booman et al.
[64] performed a gene expression and CGH-array study on 15 central nervous system
(CNS) cases, nine testicular cases, and 15 nodal DLBL, which revealed that loss of
6p21.32–p25.3, including the HLA genes was associated with both types of IP-DLBCL,
whereas gain of 2p16.1–p25.3 was associated with nodal DLBCL. Gain of 12q15–q21.1
and 12q24.32–q24.33 was associated with CNS DLBCL and gain of 19q13.12–q13.43 with
testicular DLBCL. Analysis of candidate genes in site-specific regions and minimal
common regions revealed two major groups of genes: one involved in the immune response,
including regulation of HLA expression, and the other involved in apoptosis, including
the p53 pathway. The presence of both shared and site-specific aberrations in CNS
and testicular DLBCLs underlines the concept of IP-DLBCL but also indicates that IP-DLBCLs
of the CNS and testis do not form a single entity.
DLBL from extranodal sites have generally a better outcome than nodal cases. This
was confirmed by Sato et al. [65] who studies oral cases (excluding tonsillar ones),
even though 17 out of 21 cases were of ABC type as determined by immunohistochemistry.
A recently described lesions is atypical lymphocytic panniculitis (ALLP). Magro et
al. [66] describe seven such patients who have waxing and waning placques. Since a
continuum with subcutaneous panniculits like T-cell lymphoma had been suggested, they
performed clonality testing, and all lesions were clonal, and the clones detected
were preserved in follow-up biopsies. This situation, very similar to refractory celiac
disease, and also monoclonal gammopathy of undetermined significance points to the
importance of interpretation of clonality testing in the complete clinicopathologic
setting.
Akiyama et al. [67] describe five cases of CD8+, CD56+ T-cell lymphoma in the small
intestine, while the patients lack enteropathy. There was no EBV, and the TCR was
clonally rearranged, and the clinical course was dismal. These tumors seem closely
related to EATL.
After seeing two cases of ALCL in the breast of women who had silicone implants, de
Jong et al. [68] performed a national case-control study in the Netherlands. In 16 years
11 primary ALCL in the breast had been diagnosed, and five of these were in women
with implants. In women with other primary breast lymphomas, only one had an implant.
These data suggest that breast implants have a small but significant risk for developing
ALCL. Another unusual presentation of ALCL was discovered by Mosunjac et al. [69]
who analyzed five autopsy-detected ALCL cases. These patients had had a rapid clinical
problem that was reminiscent of sepsis or fever of unknown origin and lacked lymphadenopathy.
These cases had extranodal involvement and often hemophagocytosis.
Rodriquez-Pinilla et al. [70] collected 146 cases of T-cell lymphoma to investigate
the cell of origin. First, they looked for PD-1-positive cases which were classified
as AILT, known to be derived from follicular helper T cells. The other cases were
studied by TMA using a wide panel of markers and EBV; of the non-AILT cases almost
30% expressed PD-1 as well and many of these expressed other follicular helper T-cell
markers also (CXCL13, bcl6, CD10). The authors suggest that the spectrum of AILT may
be wider than previously thought but also that other types of T-cell lymphoma may
be derived from follicular T cells.
Pitfalls in lymphoma diagnosis
Pathologists recognize that lymphoma classification requires special expertise. LaCasce
et al. [71] investigated whether the WHO-classification has facilitated the correct
diagnosis of lymphomas by evaluating the results from their referred cases. For this
analysis they focused on the five most common entities and found that the discordancy
rate was 6%, with the most discordancies due to grading FL. The majority of discordant
cases (80%) might have had a different treatment. Since less common lymphomas total
generally up to about 20–25% of all cases, this study favors review of lymphoma even
in this era of WHO classification and many ancillary techniques.
Castleman’s disease remains an enigmatic lesion with a variable morphology. Based
on 11 cases, Kojima et al. [72] describe the different types of follicles one may
encounter in Castleman’s disease and delineate the sometimes difficult differential
diagnosis with reactive lesions on the one side, and HL on the other.
Cancer cells do not behave like normal cells and therefore unusual expression of markers
can occur. Wang et al. [73] report four cases of CD-3 positive B-cell lymphomas and
Rahemtullah et al. [74] nine cases of CD20-positive T-cell lymphomas, stressing the
need for double positive and negative markers for lineage determination in malignant
lymphoproliferations.
Prognostic factors in lymphoma
Investigations into prognostic factors remain popular. Quite a series of papers on
this topic were published, and only a small sample of results is given. None of the
studies were confirmatory in nature, and those are the only ones in which clinical
application can be found. Schreck et al. [75] demonstrate that a large number of Th2
cells in HL is related to better survival; Dong et al. [76] show that bcl10 expression
(next to t(11;14)) is related to HP-eradication unresponsiveness; according to Wang
et al. [77], cytoplasmic sox11-expressing MCLs have a poorer survival; mutation nor
polymorphism of the CD20 gene is predicting rituximab response (Sar et al. [78]);
Aktas et al. [79] show that high apoptotic index is related to good prognosis in pediatric
lymphomas; Lenz et al. [80] describes new gene signatures that predict therapy response
in DLBL; Peh et al. [81] found that bcl-6 expression is associated with poor survival
in DLBL and immunohistochemically determined ABC not; according to Johnson et al.
[82], DLBL with low CD20 expression have lower survival rate; Ki67 and Pim1 are independent
indicators of poor survival in MCL (Hsi et al. [83]); low Ki67 is a negative predictor
of survival in DLBL (Hasselblom et al. [84]); SIRT1 expression is associated with
poor prognosis in DLBL (Jang et al. [85]); in FL, Johnson et al. [86] show that secondary
genetic alterations indicate aggressive behavior; Hasselblom et al. [87] show that
the number of CD68-positive cells in DLBL does not predict prognosis; TCL1A is associated
with more aggressive clinical behavior in CLL and MCL according to Aggarwal et al.
[88]; aberrations in the MYC locus are associated with poor outcome in DLBL (Klapper
et al. [89] and Yoon et al. [90]); expression of P-glycoprotein indicates therapy
unresponsiveness in nasal type NK/T cell lymphoma according to Wang et al. [91];
Ancilliary techniques
Flow cytometry is increasingly used for the phenotyping of lymphomas, but HL is often
difficult to recognize due to the scarcity of tumor cells. In a large study using
flow cytometry of 76 HL and 156 reactive lymph nodes, Seegmiller et al. [92] showed
that HL can be suggested by the presence of many CD7- and CD4-positive cells. Using
the same technique, Pierson et al. [93] show that aberrant CD26 expression indicates,
together with other T-cells markers, T-cell lymphoma, but it is not specific.
FISH is now commonly used for the detection of aberrations in specific genes. De Rijk
et al. [94] describe a robust paraffin tissue approach using split signal probes,
tested in eight different laboratories using 16 different probes covering the most
important regions involved in translocation in lymphomas. This method uses tissue
sections, so that even small areas of tumor can be evaluated without microdissection.
Using isolated nuclei paraffin embedded tissue from 20 cases of CLL, Flanagan et al.
[95] were able to determine almost all aberrations detected in blood samples. Huh
et al. [96] used this method for staging on bone marrow aspirates and showed that
FISH is more sensitive than classic cytogenetics, but no correlation with immunohistochemistry
was done.
Clonality testing is nowadays an important tool in the management of lymph node biopsies,
but is prone to mistakes. Groenen et al. [97] use a case-based approach to describe
common pitfalls in the interpretation of TCR clonality testing and reiterate the importance
of interpretation of the molecular results in the context of the histological and
immunohistochemical findings.
Santon et al. [98] performed clonality testing on gastric biopsies from 19 patients
treated for gastric MZL and found persistent clonality even though histology did not
show lymphoma anymore. This was especially the case in tumors with t(11;14). The clinical
meaning of this finding is not yet clear.
Morales et al. [99] tackle the difficult problem of separating reactive (n = 23) form
malignant (n = 26) B-cell proliferations in the skin, using the Biomed set of primers.
In patients with a lymphoma, a clone was detected in 85% of the samples, and in only
4% of the reactive infiltrate was the data very promising.
Tan et al. [100] analyzed the occurrence of B-cell clonality in 74 T-cell lymphomas.
Using the Biomed approach, 86% had a clonal TCR, and 8% a concurrent clonal B-cell
rearrangement. There was no association with the presence of EBV.
Cytology obtained from fine needle aspirates (FNA) is a limited invasive option for
diagnosing lymphomas but has limitations. Pai et al. [101] used FNA-obtained cells
for morphology, phenotyping, and molecular tests to evaluate enlarged lymph nodes
in patients with mycosis fungoides. In seven out of ten patients, they could determine
the presence of tumor cells in the lymph node; three of them had large-cell transformation.
Remarkably, two of the patients had HL in the lymph node. This study shows possibilities
of FNA in an experienced setting.
Also, multicolor FISH is possible on FNA material as shown by Caraway et al. [102].
In 49 out of 50 patients with CLL, they were able to obtain data on genetic aberrations,
with similar prognostic impact as known from traditional studies.