Available via license: CC BY-NC-ND
Content may be subject to copyright.
REVIEW
Follicular lymphoma research: an open dialogue for a
collaborative roadmap
M
elanie Collin,
1
Guillemette Gagey,
1
Vignesh Shanmugam,
2,3
Abner Louissaint Jr,
4,5
Jessica Okosun,
6
Clementine Sarkozy
7
& Bertrand Nadel
1
1
Aix-Marseille University, CNRS, INSERM, Centre d’Immunologie de Marseille-Luminy, Marseille, France,
2
Department of Pathology, Brigham and Women’s Hospital, Boston,
3
Cancer Program, Broad Institute of MIT and
Harvard, Cambridge,
4
Department of Pathology,
5
Krantz Family Center for Cancer Research, Massachusetts General
Hospital, Boston, MA, USA,
6
Barts Cancer Institute, Queen Mary University of London, London, UK and
7
Hematology
Department, Institut Curie, Saint Cloud, France and LITO, U1288, Universit
e Versailles Saint Quentin en Yveline,
Saint Quentin en Yveline, France
Collin M, Gagey G, Shanmugam V, Louissaint A Jr, Okosun J, Sarkozy C & Nadel B
(2025) Histopathology 86, 79–93. https://doi.org/10.1111/his.15344
Follicular lymphoma research: an open dialogue for a collaborative roadmap
Follicular lymphoma (FL) is the second most common
type of lymphoma (20% of all non-Hodgkin lympho-
mas), derived from germinal centre (GC) B cells, and is
characterised by its significant clinical, prognostic and
biological heterogeneity, leading to complexity in man-
agement. Despite significant biological investigation
and indisputable clinical progress since the advent of
the immunotherapy era more than 20 years ago,
much remains to be done to understand and cure this
lymphoma. Today, FL is metaphorically a giant puzzle
on the table with patches of sky, landscape and foliage
clearly appearing. However, many of the remaining
pieces are held by various stakeholders (e.g. clinicians,
pathologists, researchers, drug developers) without
global agreement on what the gaps are, or any clear
blueprint on how to solve the puzzle of understanding
the heterogeneity of this disease and create curative
and tailored therapies. With the advent of new investi-
gation and drug technologies, together with recent
advances in our capacity to manage big data, the time
seems ripe for a change of scale. More than ever, this
will require collaboration between and within all stake-
holders to overcome the current bottlenecks in the
field. As for every investigator, we acknowledge that
this first draft is necessarily biased, incomplete and
some FL expert readers might recognise some
Address for correspondence: A Louissaint Jr, Department of Pathology, Massachusetts General Hospital, Boston, MA, USA. e-mail:
alouissaint@mgb.org
J Okosun, Barts Cancer Institute, Queen Mary University of London, London, UK. e-mail: j.okosun@qmul.ac.uk
C Sarkozy, Hematology Department, Institut Curie, Saint Cloud, France and LITO, U1288, Universit
e Versailles Saint Quentin en Yveline,
Saint Quentin en Yveline, France. e-mail: clementine.sarkozy@curie.fr
B Nadel, Aix-Marseille University, CNRS, INSERM, Centre d’Immunologie de Marseille-Luminy, Marseille, France. e-mail: nadel@ciml.univ-mrs.fr
Abbreviations: Ag, antigen; AI, artificial intelligence; BCL, B cell lymphoma; BCR, B cell receptor; BM, bone marrow; CAR-T, chimeric
antigen receptor T; COO, cell of origin; cFL, constrained FL; CMG, chromatin modifying genes; CPC, cancer progenitor cells; CRR, complete
response rate; ctDNA, circulating cell-free tumour DNA; DZ, dark zone; dFL, DLBCL-like FL; DLBCL, diffuselarge B cell lymphoma; FFPE,
Formalin-fixedparaffin-embedded; FL, follicularlymphoma; FLIPI, FL international prognosticindex; GC, germinalcenters; GOF, gain-of-
function; ICT, immunochemotherapy; Ig, immunoglobulin; LOF, loss of function; LZ, light zone; MRD, minimal residual disease; NHL, non-
Hodgkin lymphoma; ORR, overall response rate; OS, overall survival; PFS, progression-free survival; POD24, progression of disease within
24months of treatment; PRC2, polycomb repressive complex 2; RISC, relapse-initiating subclones; RT-PCR, reversetranscription–polymerase
chain reaction; RR, relapse/refractory; SHM, somatic hypermutation; SLO, secondary lymphoid organs; SOC, standard of care; TFH, T
follicular helper cell; tFL, transformed FL; TME, tumour microenvironment; TNF, tumour necrosis factor; WHO, World Health Organisation;
ZO, zanubrutinib in combinationwith obinutuzumab.
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and
distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
Histopathology 2025, 86, 79–93. DOI: 10.1111/his.15344
remaining gaps not addressed. We hope they will reply
to make this effort a collaborative one to assemble all
the pieces in the most ideal fashion. As such, this
review intends to be a first step and an interactive plat-
form to a collaborative roadmap towards better under-
standing and care of FL.
Introduction: unmet clinical needs as a
starting point
The management of follicular lymphoma (FL) begins
with histological diagnosis by a surgical pathologist,
which is relatively straightforward, and based on the
presence of centrocytes and admixed centroblasts
with an immunophenotype consistent with a GC cell
of origin (COO). FL is, at least partly, usually associ-
ated with a follicular pattern and the presence of
BCL2 and/or BCL6 gene rearrangements, although
subtypes that deviate from this have been recently
characterised.
1
The majority of FL patients present
with lymphadenopathy in multiple sites with variable
systemic distribution, tumour burden and association
with symptoms (e.g. localised or systemic lymphade-
nopathy with no symptoms versus extensive lymph-
adenopathy with B symptoms or symptoms related to
site of involvement). FL is characterised by a clinical
heterogeneity at time of diagnosis that relies upon
biological heterogeneity, which is still challenging to
capture fully. This heterogeneity results in the appli-
cation of a variety of first-line treatment strategies,
including ‘watch and wait’ for patients without symp-
toms to immunochemotherapy (ICT) with rituximab
maintenance for patients with a clinically symptom-
atic disease, associated with a median progression-
free survival (PFS) of 10.5 years, with a 10-year
overall survival (OS) estimate of 80%.
2,3
However,
the absence of a plateau on all progression-free sur-
vival curves after this and other available first- and
second-line therapies reflects the frequent occurrence
of relapses, leading to FL being considered as a
mainly incurable disease. Each sequential relapse
tends to occur more quickly and with increased
aggressivity and refractoriness to subsequent thera-
peutic options.
The clinical prognosis of FL patients after therapy
is also heterogeneous and relatively unpredictable. A
significant subset of treated FL patients (15–20%)
harbour a chemo refractory disease from the outset
or experience an early relapse within 24 months of
first-line chemotherapy (POD24), associated with an
increased risk of death from lymphoma.
3,4
Transfor-
mation to an aggressive lymphoma (tFL) is responsi-
ble for this poor outcome in most of these cases.
5,6
Conversely, 50% of treated FL patients will be
long-term responders without relapses after 10 years
of follow-up and for whom the leading cause of death
will not be lymphoma, but rather independent malig-
nancies, cardiovascular disease or other unrelated
causes. This has led to the evolution of the definition
of ‘cure’ in FL, and the emergence of the concept of
‘functional cure’.
3
There are three critical unmet clinical needs in FL:
(1) lack of useful prognostic tools at the time of diag-
nosis and prior to therapy, to guide clinical manage-
ment; (2) the need for novel and more effective
therapeutic targets derived from a comprehensive
understanding of FL biology and tailored for defined
subsets of FL patients; and (3) the lack of effective
biomarkers to effectively follow therapeutic responses.
In this review we will discuss the issues underlying
these bottlenecks, highlight evidence gaps and propo-
sitions to overcome them.
Tools to predict clinical behaviour
For decades, haematopathologists, clinicians and
translational researchers have dedicated much effort
towards identifying clinical, histological, molecular
and/or imaging features of FL that may risk-stratify
FL patients into groups with different survival
outcomes.
CLASSIC MORPHOLOGICAL APPROACH
Traditionally, cases of FL with higher numbers of
large lymphoma cells (centroblasts) were thought to
be more clinically aggressive. Pathologists have
applied variations of the original Mann and Berard
approach
7
of histological grading based on the num-
ber of centroblasts counted upon microscopic review.
Over the years, the reproducibility and prognostic sig-
nificance of grading in FL has been increasingly
debated among pathologists.
8,9
Several clinical trials
performed for targeted therapies have shown identical
outcome for grades 1, 2 and 3a FL.
2,10,11
In addition,
it has been shown that histological transformation of
FL is not correlated with histological grade.
8
For
these reasons, grading has been removed from the
fifth edition of the World Health Organisation (WHO)
classification (WHO-HAEM5) (Box 1).
1
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
80 M Collin et al.
CLINICAL AND CLINICO-MOLECULAR
APPROACHES
Several clinical prognostic scoring systems have been
developed with the primary goal of identifying
high-risk patients in daily practice and for clinical
trial inclusion selection. Most of these systems focus
on ‘at diagnosis’ or pretreatment tools based on clini-
cal parameters (e.g. FLIPI, FLIPI-2, FLEX) to stratify
patients at diagnosis into groups with different sur-
vival outcomes. Some scoring systems, such as
PRIMA-PI, are easier to compute, relying upon fac-
tors such as bone marrow (BM) involvement and
blood b2M levels. However, BM biopsies are not uni-
versally performed in all patients at baseline.
12–15
As
clinical tools are essentially surrogates for FL biology,
more biologically derived prognostic tools have been
developed. One of the first was the m7-FLIPI, com-
puted using the gene mutation status of seven genes
(ARID1A,CREBBP,EP300,EZH2,CARD11,FOXO1,
MEF2B) with the patient’s performance status and
FLIPI score to segregate patients into low- and
high-risk groups.
16
Later, gene-expression-based
prognostic tools (PRIMA 23-gene) were developed
using several clinical trials and population-based
cohorts.
17
While these tools can identify high-risk
groups, they exhibit variable prognostic accuracy,
particularly in identifying specific high-risk groups
posing the greatest clinical challenge, such as
patients with POD24 or transformation. The ‘high-
risk’ groups defined by these scoring systems remain
heterogeneous, encompassing a range of clinical phe-
notypes. Additionally, the disease trajectory of
patients within any given group is often differentially
influenced by the type of treatment regimen they
receive (Box 1).
THE EMERGENCE OF ARTIFICIAL INTELLIGENCE
(AI) FOR PREDICTING CLINICAL BEHAVIOUR
During the last decade there have been major
advances in artificial intelligence (AI)-based deep
learning to increase clinical-grade accuracy in histo-
pathological image-based cancer classification.
18–20
Pathologists and AI experts are currently developing
strategies to harness this technology to assist and
optimise lymphoma diagnostics,
21
including FL grad-
ing. Moreover, machine learning also holds great
promise for enabling the discovery of tissue-based
prognostic biomarkers.
19
Additionally, multimodal
data fusion methods
22,23
that integrate radiological,
pathological, laboratory and clinical data will proba-
bly be even more powerful in enabling the accurate
prediction of these clinical groups at diagnosis. Fur-
ther progress requires inter-institutional collaboration
between pathologists for the analysis of large num-
bers of heterogeneous FL cases and to harness overfit-
ting issues (Box 1).
How to approach therapeutic innovation
in FL
MAJOR PITFALLS
Two major pitfalls in the approach to therapeutic
innovation in FL are related to (1) the targeted popu-
lation and (2) the drugs considered. Most immune or
targeted agents are classically tested in the relapse/
refractory (RR) setting, aiming to achieve an overall
response rate (ORR) in Phase II to justify investments
in Phase III trials against standard of care (SOC) ther-
apies. The ‘winners’ are then subsequently approved
for all FL patients. However, RR FL represents a het-
erogeneous minority of FL cases and cannot be gener-
alised to the majority of FL patients. Furthermore, the
discrepancy between the ‘discovery (RR)’ and the
BOX 1. Key opportunities for advancing
prognostic biomarkers
Identifying high-risk individuals remains a chal-
lenge. Current strategies, including histological
grading and prognostic scoring systems are imper-
fect, requiring improvement. Furthermore, focus-
ing solely upon this group overlooks the chance to
enhance the quality of life and outcomes for the
largest fraction of follicular lymphoma (FL)
patients. We need to:
•Prioritise defining the large population of
low-risk FL patients with indolent behaviour
to identify those who can safely reduce or
avoid therapy
•Evolve the biomarker strategies from being pri-
marily prognostic-based to also encompass
identification of predictive and dynamic bio-
markers for more precise therapy
individualisation
•Leverage advanced machine learning and arti-
ficial intelligence (AI) technologies, combined
with large-scale, readily accessible FL biopsies,
to drive the development of tissue-based
biomarkers
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
A follicular lymphoma roadmap 81
extension cohorts may lead to limited improvements
when therapies are applied in the first-line setting.
Secondly, given the diversity of the mechanism of
action of the different drugs in the various drug
developer pipelines worldwide, one would think that
the challenges of FL heterogeneity could easily be
addressed with novel drugs targeting different intra-
cellular pathways, modulating epigenetic hallmarks
or the immune and tumour cell cross-talk. However,
very few novel FL therapeutics in development repre-
sent new drug classes, and virtually none of these
drugs (to the notable exception of tazemetostat;
24
see
below) are driven by or specifically tailored to new
discoveries in FL biology. Furthermore, most, if not
all of the most recently approved therapeutic agents
in FL do not target a specific vulnerability present in
subpopulations of FL, but rather are applied to all
patients without a reliable biomarker of response
(Box 2).
Going forward, we should keep in mind that one
size does not fit all in FL and that a personalised
strategy should be the goal. We need to define effec-
tive therapeutic targets that consider the heterogene-
ity observed for decades in FL (clinical and biological)
and the complex layers of biology that underlie this
heterogeneity. This probably means defining different
therapeutic targets for different subpopulations of FL
driven by the unique biological features of that
subpopulation. As part of this effort, it is critical that
specific and reliable biomarkers of response are devel-
oped for each therapeutic agent. This will require the
collaboration of pathologists and clinicians in the per-
formance of clinical trials (including Phase III) that
are carefully designed with paired clinical and biologi-
cal endpoints integrated as part of the trial design.
Naturally, the identification of novel high-yield tar-
gets in subsets of FL will depend upon significant
advances in our understanding of the complex layers
of FL biology (Box 2).
FIRST STEPS TOWARDS THERAGNOSTIC
CLASSIFICATION USING MOLECULAR SUBTYPING
Moving away from a blanket treatment approach in
FL requires detailed understanding of the molecular
underpinnings of different groups of FL patients to
support differential therapeutic approaches. Do molec-
ular subtypes exist in FL? Several recent studies have
shed light on this. Using a targeted gene panel of 293
genes, a first study identified three genotypical sub-
groups identified: one associated with a high burden
of aberrant somatic hypermutation (SHM), a second
with frequent STAT6 and CREBBP mutations, and a
third group enriched for KMT2D mutations without
the features of the prior clusters.
25
Although a key
limitation in this study was using targeted gene
sequencing to resolve these subgroups, none of these
three groups were associated with patient risk or pro-
pensity to transformation. In another study, whole
genome sequencing of 423 diagnostic biopsies from
FL (some with or without later transformation),
transformed FL and de-novo diffuse large B cell lym-
phoma (DLBCL) cases led to the proposition of two
genetic FL subtypes associated with significantly dif-
ferent risks of transformation: constrained FL (cFL)
versus DLBCL-like FL (dFL).
1
The cFL cohort was
associated with a reduced risk of transformation and
genetically harboured a lower mutational burden, less
SHM but an enrichment of CREBBP KAT missense
mutations and mutations in genes involved in
mTORC1 signalling (RRAGC,ATP6V1B2,
ATP6AP1).
26,27
By contrast, dFL had a much higher
risk of HT and was associated with more frequent
aberrant SHM and CREBBP nonsense mutations. A
third study initially used bulk transcriptome on the
FL B cells to resolve three transcriptional states
referred to as inflamed, proliferative and chromatin-
modifying.
28
Each of these states correlated with spe-
cific genetic and immune microenvironment features,
although were not demonstrated to have any prog-
nostic impact. Finally, a study by Han and colleagues
BOX 2. Key challenges and opportunities towards
therapeutic innovation
The current drug market remains mostly follicular
lymphoma (FL)-biology agnostic. We need to:
•Change the way the drug market is approached
and align the development of drugs in FL with
its key biological characteristics/mechanism of
action rather than repositioning drugs showing
efficacy in other indications (or even other lym-
phoma subtypes). Together with this, drug
repurposing based on biological and testable
rationales should become more frequent and
used in parallel to our progress on FL biology
•Advocate for a personalised approach with
theragnostic information as part of pathology
reporting
•Provide the community with a routinely avail-
able biomarker of response, or a theragnostic
classification, for each novel agent in
development
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
82 M Collin et al.
analysing the tumour microenvironment (TME) using
single-cell RNA sequencing defined four FL clusters
based on TME composition: a ‘na€
ıve’ cell-rich cluster,
a ‘warm’ cluster, an ‘intermediate’ cluster and a
‘depleted’ cluster, the latter being associated with
poorer survival outcomes.
29
Although the technical
approaches were different in each of these studies,
they refer the underlying biological heterogeneity of
FL being driven by more than just genetic aberrations
(Box 2).
Biological knowledge to build the bridge
for a biology-informed treatment
Why is FL so effective at escaping current therapies?
The answer lies in the natural history of FL
30
(Fig-
ure 1), involving a long and complex multihit
process, starting decades before diagnosis and/or
symptomatic manifestations. The process involves
four key steps setting the stage for malignant trans-
formation and recurrent relapses: early BCL2 activa-
tion, dysregulation of B cell dynamics, co-evolution of
propitious TME/tumour ecosystems and genomic
instability leading to epigenetic dysfunction. This
indolent, Darwinian-like evolution of FL generates
both the peculiar complexity and heterogeneity of this
disease.
CPC (COMMON PRECURSOR CELLS)
The complex parallel evolution of early expanding
precursor clones resulting from iterative visits to the
GC of a large pool of long-lived BCL2
+
memory B-like
cells (Figure 1) probably constitutes the first and
Figure 1. A model of FL oncogenesis: t(14;18) is occurring during pro-B cell development in the bone marrow and leads to ectopic BCL2
expression without preventing further B cell maturation. Upon cognate antigen stimulation, peripheral na€
ıve BCL2+cells are preferentially
activated by TFH and GCs. There, ectopic BCL2 expression uncouples GC check-point selection from affinity maturation, leading to clonal
expansion, differentiation and exit of BCL2+memory-like B cells with heterogeneous, unselected, low-affinity and potentially polyreactive
BCR. The propensity of such large pools of BCL2+clones to disseminate in blood, niche in SLO and iteratively visit GC can be assumed as
the second (immunological) hit, heading to the accumulation of mutations. Although most will be passengers, some will directly and recur-
rently impact further steps of oncogenesis either through providing competitive advantage (e.g. proliferation) or resistance to the host immu-
nity and/or to therapy (e.g. quiescence). FL, follicular lymphoma; BCL, B cell lymphoma, TFH, T follicular helper cell; BCR, B cell receptor;
GC, germinal centres; SLO, secondary lymphoid organs.
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
A follicular lymphoma roadmap 83
potentially one of the most serious barriers to effective
therapy.
31
This sets the stage for the co-occurrence of
constant new generations of evolving FL progenitor
cells (called CPC) concurrently present at any time,
probably in various anatomical locations and at vari-
ous advanced stages of malignant transformation,
waiting in line to emerge as FL.
32–35
As a direct con-
sequence of CPC dissemination, spatially distinct FL
involved sites (lymph node, BM, blood) exhibit genetic
and transcriptional heterogeneity.
36–38
Concurrent
CPCs might globally benefit from protumoural TME
reshaping by the most advanced TME/tumour ecosys-
tems. The early genealogical branching of this ‘CPC
factory’ (i.e. before malignant transformation) is the
source of intratumoural heterogeneity in FL, and is
most probably the source of resistance and
relapse.
39–41
Because CPC or daughter
relapse-initiating subclones (RISC) are rarely detect-
able at diagnosis, anticipating targeted therapy for
relapse based on late oncogenic alterations found at
diagnosis (post-CPC branching) is not a rational pre-
cision medicine option. Understanding the most fre-
quent sequence of alterations paving FL genesis is
thus mandatory to identify and target actionable
early hits present in all or most subclones. Unfortu-
nately, BCL2 has proved to be a weak functional and
therapeutic target in FL.
42
By the time that overt dis-
ease manifests, BCL2 has gradually built a global
stage for oncogenic substitutes and probably does not
constitute a tumour addiction, as evidenced by disap-
pointing clinical trials using BCL2 inhibitors. Thus,
despite the emergence of new generations of more
potent BCL2 inhibitors, targeting BCL2 alone is
unlikely to address the relapse conundrum. Next in
BOX 3. Key requirements towards therapeutic innovation based on biological knowledge
Will be achieved by:
•Fully characterising the common precursor cells (CPC) and relapse initiating subclones (RISC) at the
genetic, phenotypical and functional levels and identify their associated niches. Large-scale efforts using
deep-sequencing and single-cell technologies will be required to decipher and understand such rare cell
populations. Collaborative academic and industry efforts will be needed to develop or adapt drugs to
directly target CPCs aiming to delay or prevent relapses. Drug development roadmap should include
mandatory efforts to minimise toxicity and optimise specificity, in order to envision a prevention/
maintenance type of therapeutic scheme
•Clearer understanding of the factors underlying the dynamics and plasticity of follicular lymphoma (FL)
cell populations and determining if specific transcriptional states are associated with drug resistance or
sensitivity and exploit druggable vulnerabilities
•Developing innovative model systems that truly recapitulates the multiple epigenetic alterations seen in
patients that will allow better dissection of the complex intrinsic and extrinsic cell circuitry and identify
vulnerabilities and epigenetic targets. Current epigenetic therapies have perhaps been underwhelming
because we miss this detailed understanding. We should harness such information to guide drug discov-
ery towards novel best-in-class epigenetic drugs with increased specificity and efficacy, together with the
development of appropriate biomarkers
•B cell receptor (BCR) activation and signalling are complex and varied, influenced by diverse ligands, iso-
types and genetic factors. This complexity, together with patient-specific immune history, poses chal-
lenges for current FL models. Targeting receptor–ligand interactions, together with signalling pathways,
requires the development of more reliable ex-vivo/in-vivo preclinical models.
•Exploring the long-term effects of chronic stimulation by existing bispecific T cell engagers is crucial.
Advancing next-generation chimeric antigen receptor T (CAR-T) designs with more tumour-specificity,
enhanced functionality and persistence with reduced toxicities
•Combining different treatment modalities based on biological rationale and potential synergy is essential.
Identifying effective combinations with minimal side effects, whether targeting tumour cells, epigenetic
circuitry or the tumour microenvironment (TME) is key
•Leveraging ongoing and future clinical trials of novel and existing drugs, with a focus on ancillary stud-
ies. Collaborative biobanking, suitable methodologies and strong research and development commitments
from all stakeholders are essential
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
84 M Collin et al.
line, genetic and deep-seq studies have identified
mutations in CREBBP as the earliest hit following
t(14;18), in some cases appearing in healthy individ-
uals up to a decade before FL diagnosis.
35,43
CREBBP
mutations are found in 60–70% FL patients, and
CREBBP loss of function is actionable through the
inhibition of the antagonist NCOR/SMRT/HDAC3/
BCL6 complex.
44–46
This opens the first precision
medicine perspectives specifically aimed at eradicating
(or at least delaying) FL relapses. Several black boxes
nonetheless remain, and further research efforts will
be necessary to fully rationalise future targeted thera-
peutic approaches (Box 3).
FL DYNAMICS AND PLASTICITY
Normal mature B cells display inherent cell dynamics
and plasticity associated with major (reversible)
changes of transcriptional states [differentiation, de-
differentiation and light zone to dark zone (LZ to DZ)
transition].
30,47
The transition between states (e.g.
from DZ to LZ) is a tightly regulated process, with sets
of genes expressed or repressed with high synchronic-
ity throughout a continuum of intermediate cell
states.
48–51
This synchronicity observed for normal
DZ/LZ B cells during the GC reaction is lost in FL,
48
partly due to chromatin-modifying gene (CMG) alter-
ations putting the brakes on differentiation and even-
tually locking FL cells as GC B cells, preferentially
homing in the GC.
52–57
The locking is, however, not
absolute, as it allows egress and dissemination, but it
may prevent differentiation and/or restrain transition
of states. Similarly to the gradual transition from DZ
to LZ in normal B cells, current single-cell data indi-
cate that FL cells in the tumour bulk also display a
continuum of expression signatures. Unlike normal B
cells, however, the continuum is spanning from
GC-like (centroblasts?) to Mem-like (interfollicular?)
states, with intermediate states (centrocytes?) consti-
tuting most of the bulk FL cells.
48,58,59
Each patient
displays a distinct balance of such intermediate states,
some leaning to GC-like, others to Mem-like, and this
constitutes one of the main components of interpati-
ent transcriptional heterogeneity. The skewing of
these states towards Mem-like could increase the pro-
pensity for tFL transformation.
59
From the therapeutic standpoint, the various tran-
scriptional states have shown various sensitivities to
chemotherapy, immunotherapy and (epi)genetic
inhibitors.
60
If components of the tumour can adapt
and swiftly transit from one state to the other, this
might constitute an important escape route to a given
selective pressure (Box 3).
Altogether, emerging single-cell data are thus
drawing a picture where FL cells remain dynamic,
undergoing profound transitions of states, probably as
a result of TME signalling. This plasticity may drive
inter- and intratumoural heterogeneity and constitute
an additional layer of complexity adding to effective
therapy escape. Clearer defining of each grand cate-
gory of state and their underlying dynamics might
accelerate identifying new Achilles’ heels and devising
combined strategies to target all categories at once
and/or strategies to prevent transition of states.
BCR ACTIVATION AND SIGNALLING PATHWAYS
The B cell receptor (BCR) constitutes a key oncogenic
pathway that has been shown to promote cancer cell
growth and survival in various types of non-Hodgkin
lymphomas (NHL).
61
The targeting of BCR signalling
pathways has proved very successful in several indica-
tions, such as chronic lymphocytic leukaemia, and
thus constitutes an important area of drug develop-
ment for tailored therapy in B-NHL. In FL, the BCR
(rarely lost despite an active SHM process) is assumed
to represent one of the tumour’s addictions and has
therefore also been the target of several trials using
kinase inhibitors. As antigen-independent tonic BCR
signalling involving the PI3K/AKT/mTOR pathway
has been proposed to be essential for FL cells
survival,
62
PI3Ki held great promise and gained
approval in RR patients.
63–65
However, poor response
rates and the absence of biomarkers to enable identifi-
cation of patients who most benefit, together with a
challenging safety profile, led to withdrawal of these
molecules. Next, despite the absence of strong evidence
of FL’s dependency towards the BTK/NFjB pathway,
the success of BTKi in other lymphomas drove the
development of several trials in FL. Although initial
BTKi monotherapies (ibrutinib, acalabrutinib) failed to
produce durable responses in FL, the Phase II random-
ised study of zanubrutinib (a new generation and more
selective BTKi) in combination with obinutuzumab
(ZO) met its primary endpoint by greatly increasing
ORR compared to monotherapy,
66
and a Phase III trial
is now ongoing (NCT05100862).
67
Most importantly, the unanticipated outcomes of the
various kinase inhibitors so far illustrate the important
gaps in the current knowledge regarding FL signalling
pathways highlighting the disconnection between biol-
ogy rationale and drug development. A clearer under-
standing of FL’s BCR signalling is urgently needed to
avoid future random achievements and disappoint-
ments, but also to consolidate success. Indeed, even if
positive and approved, the ZO will ideally require a
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
A follicular lymphoma roadmap 85
companion biomarker for patient selection, given that
only a fraction of the patients is likely to benefit from
this strategy.
Another area of great interest lies in upstream BCR
activation. During the decades of preclinical FL devel-
opment with a GC check-point selection invalidated
for BCR/antigen (Ag) fitness, an ‘Ag switch’ probably
operates from the initial cognate Ag towards other
BCR stimuli, probably with great interpatient variabil-
ities in an opportunistic fashion. Among these, the
acquisition of N-glycosylation sites in immunoglobu-
lin (Ig) variable regions might provide a recurrent
(~80% FL cases), continuous and low-intensity
Ag-independent BCR activation pathway via interac-
tion with lectins in the TME (e.g. DC-SIGN on
macrophages).
68
Notably, this peculiar activation
would drive better signalling through an IgM than
an IgG. As most patients with early relapses express
IgM,
69
disrupting this signalling
70,71
might open
important new targeting avenues.
Altogether, BCR activation and signalling circuitry
are complex and more varied than initially antici-
pated. Many knowledge gaps remain to reliably guide
tailored drug development of signalling pathway
inhibitors in FL (Box 3).
EPIGENETICS AND CELL INTRINSIC CIRCUITRY
Epigenetic regulators play a prominent role in the
clonal evolution of FL, as 90% of FL cases harbour at
least one somatic mutation in histone modifiers
including KMT2D,CREBBP,EP300 and EZH2.
39,40,72
Mutations in KMT2D, a histone methyltransferase
specific to H3K4, and CREBBP, a histone acetyltrans-
ferase specific to H3K27, commonly lead to a loss of
function (LOF), while mutations in EZH2, a histone
methyltransferase that forms the catalytic component
of the polycomb repressive complex 2 (PRC2) respon-
sible for laying repressive methylation marks on
H3K27, are gain-of-function (GOF). Consequently, all
CMG mutations in FL lead to repression. A
cell-intrinsic circuitry model accounting for the prom-
inent role of epigenetic repression in FL lymphoma-
genesis suggests that such circuitry prevents further
differentiation of GCB-like cells, in line with FL’s
COO.
52–57
Mutations in epigenetic regulators may
also contribute strongly to inducing an immune eva-
sive TME by dampening down different components
of the immune synapse. For example, EZH2 muta-
tions reprogramme T follicular helper cell (TFH) sig-
nalling and their cross-talk with FL cells, while
CREBBP aberrations down-regulate the antigen pre-
sentation machinery.
57,73,74
The key pathogenic role played by these epigenetic
regulators combined with their high frequency makes
them interesting targets. The discovery of GOF EZH2
mutations in ~25% of FL
75
led to the only example of
drug development issued from specific FL discovery.
The first-in-class EZH2 inhibitor, tazemetostat, dem-
onstrated a greater ORR in EZH2-mutated (69%)
compared to EZH2-unmutated patients (35%). The
presence of responses in the EZH2-unmutated patients
is consistent with the fact that, as normal GC-B cells,
FL cells express EZH2, with the mutant form increas-
ing PRC2 repression activity through the enhance-
ment of H3K27me2 to H3K27me3. The activity and
equivalent duration of response in both groups led to
approval of tazemetostat for both mutated and unmu-
tated patients.
24
Eventually, this questions whether
EZH2, the sole actionable mutation in FL, represents
a truly predictive biomarker for tazemetostat, given
that it is found in merely a quarter of FL patients and
that its inhibition does not provide a curative answer,
particularly in monotherapy. Combination treatment
approaches are currently being evaluated. EZH2’s role
in TME reprogramming is also being investigated to
determine how it may enhance the efficacy of immu-
notherapies such as chimeric antigen receptor T
(CAR)-T.
76
Finally, as EZH2 LOF are oncogenic in
leukaemia, secondary cancer concerns ought to be
monitored in the long term, with potential restriction
of EZH2 inhibitors usage in later lines of the thera-
peutic sequence (>3L).
Similar drug development efforts are ongoing for
KMT2D (KDM5 inhibitors) and CREBBP mutations
(HDAC3 inhibitors), although with more challenging
specificity/toxicity issues due partly to the lack of
direct targeting of LOF mutations and structural hur-
dles of the targeted proteins (Box 3).
46,77
TME ADDICTION AND IMMUNE-BASED THERAPIES
The tumour microenvironment (TME) plays a key role
in clonal evolution, tumour cell survival and clinical
evolution of FL patients.
78–80
This cell-extrinsic circuit
is crucially implicated in FL lymphomagenesis. FL TME
mainly consists of immune T cells,
29,81
including Tfh
cells, BCL6+, inducible costimulatory (ICOS+), C-X-C
chemokine receptor type 5 (CXCR5+) and programmed
cell death ligand 1 (PDL1+), involved in lymphomagen-
esis through secretion of cytokines such as interleukin
(IL)-6, IL-21, IL-4 and CD40-L.
82–84
Regulatory T cells
(Treg) CD4
+
, forkhead box protein 3 (FOXP3) and
CD25
+
are involved through immunosuppressive activ-
ity related to inhibition of cytotoxic cells infiltrating the
tumour.
85
Additionally, other T cell populations have
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
86 M Collin et al.
been identified, such as Tfr cells FOXP3+CXCR5+shar-
ing characteristics of Tfh and Treg cells,
86
cytotoxic
CD4
+
T cells and cytotoxic CD8
+
T cells. Composition of
FL TME is linked to clinical outcomes of patients; a low
abundance of intrafollicular memory CD4
+
T cells is
associated with poorer progression-free survival,
87
whereas a rich infiltrate of cytotoxic CD8
+
T cells is
associated with a better prognosis.
88
Insights into the role of TME in FL clinical behaviour
has stimulated novel immune-based approaches,
including immune check-point inhibitors (ICI), bispeci-
fic T cell engagers (anti-CD3/CD20) and CAR-T cells
(targeting CD19). In RR, ICI assessed in combination
with rituximab led to conflicting data,
89–91
and bio-
markers are required to evaluate whether some (even if
few) patients could benefit from this strategy. Regard-
ing T cell engager, several molecules showed impres-
sive ORR,
92,93
and combinations with lenalidomide are
being developed
94,95
with exciting preliminary results.
However, so far there are no data to justify the choice
of a combination versus a single agent use. Keeping in
mind that the leading cause of death in these trials is
infection, limiting combinations and duration of treat-
ment appears required. The field of FL is also moving
towards CAR-T cells, as lisocabtagene maraleucel, axi-
cabtagene ciloleucel and tisagenlecleucel were recently
approved [complete response rate (CRR) of 94, 79 and
68%, respectively].
96–98
While extremely successful
and potentially curative, the financial toxicity and spe-
cific side effects suggest this should be reserved for
high-risk patients. Available translational data report
on the correlation of CAR-T expansion,
97
tumour
immune contexture or pretreatment levels of
Treg-related chemokines and inflammatory markers
IL-2Raand tumour necrosis factor (TNF)-awith
outcome.
99,100
These data are only preliminary, and
more needs to be done to have a reliable biomarker for
a risk-stratified approach for CAR-T selection. Last, but
not least, we need targets that are not universal on
immune cells (i.e. other than CD20, CD19 for exam-
ple), but rather FL-specific, to avoid long-term side
effects, and more particularly infection complications
(Box 3).
SPATIAL TECHNOLOGIES CAN ENABLE DEEPER
TME CHARACTERISATION AND BIOMARKER
DISCOVERY
The past decade has seen explosive growth in the
development of spatial technologies that enable the
systematic dissection of the tumour microenviron-
ment at ever-increasing spatial and molecular
resolution.
101,102
There are two broad classes of
emerging spatial technologies; first, image-based spa-
tial proteomic and transcriptomic technologies offer
excellent spatial resolution but are limited in their
multiplexing capacity. Conversely, sequencing-based
technologies provide the advantage of high multiplex-
ing capacity, including transcriptome-wide measure-
ments, but suffer in their spatial resolution. Some
newer technologies, such as Slide-tags,
103
attempt to
bridge this important gap between multiplexing
capacity and spatial resolution, where one can obtain
single-cell genome-scale measurements at truly
single-cell resolution. This technology also enables
multimodal measurements, a notable limitation of
current spatial genomic methods. High cost and lack
of FFPE tissue compatibility are important limitations
and active focus areas for technology development.
Recent studies
69,104
have applied single-cell and
spatial (multiplexed immunophenotyping) technolo-
gies on clinically annotated cohorts of FL and have
led to intriguing initial insights into the cellular and
architectural features of the microenvironment,
including exhausted T cell subsets, stromal desmopla-
sia and changes to the follicular growth pattern, that
are associated with outcomes. These studies highlight
the power of these technologies and underscore the
need for additional studies in larger well-defined clini-
cal cohorts (Box 3).
The application of these technologies to FL sam-
ples across space (different anatomical sites) and
time (diagnostic and progression/transformation
specimens), and computational methods for inference
of paracrine and niche-specific ligand–receptor inter-
actions by integrative analysis of single cell and spa-
tially resolved data,
105–107
offer great promise in
achieving the goal of developing a holistic under-
standing of interactions within the tumour microen-
vironment that drive tumour growth. Specifically,
these tools and data sets should enable the discovery
of novel microenvironment-derived tumour cell tro-
phic factors and mechanisms of immune evasion.
These discoveries can lay the groundwork for devel-
oping the next generation of immunomodulatory
therapies, next-generation model systems and bio-
markers, which are much needed in the field
(Box 3).
Development of dynamic biomarkers for
the evaluation of therapeutic response
In order to effectively triage novel targeted therapies
to the appropriate subsets of FL patients, it will be
critical to develop therapy-specific biomarkers that
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
A follicular lymphoma roadmap 87
can be used to effectively and reliably evaluate thera-
peutic responses.
BCL2–IGH REARRANGEMENT DETECTION BY
RT-PCR
Post-induction treatment biomarker evaluation has
historically centred on minimal residual disease (MRD)
measurement of a key molecular hallmark, the
t(14;18), BCL2–IGH rearrangement
108
by reverse
transcription–polymerase chain reaction (RT-PCR).
Multiple studies in FL demonstrate that the presence of
conventional MRD in either the peripheral blood or BM
following treatment is an important predictor of
relapse.
109–112
However, this method has limitations.
First, not all FL patients have this molecular marker, as
up to 40% of patients lack a BCL2–IGH target that can
be tracked and are ineligible for monitoring.
110
Sec-
ondly, detection of very low MRD levels requires very
high input DNA and, in turn, difficulty in distinguish-
ing low-level signals that reflect true residual lym-
phoma cells from non-specific amplification of normal
DNA. Thirdly, MRD negativity following chemoimmu-
notherapy is very high (e.g. nearly 90% of assessable
patients were MRD-negative on the GALLIUM and
FOLL12 trials
110–112
), many patients still relapse, indi-
cating that clinical relapse cannot be accurately
predicted by this approach. Lastly, healthy individuals
who do not develop FL have been shown to harbour
the BCL2–IGH rearrangement, and therefore cells bear-
ing this rearrangement may not all represent cells with
the potential to induce relapse (Box 4).
THE PROMISE OF CIRCULATING CELL-FREE
TUMOUR DNA (CTCDNA)
The remarkable breadth and evolution of intratu-
moural heterogeneity that occurs in time and space
(before and after treatment) in FL patients suggests
that this will require dynamic biomarkers that can be
easily and repeatedly monitored over time. Given the
logistical impossibility of performing multiple serial or
longitudinal biopsies, analysis of ctDNA fragments
released into the blood—a means of liquid biopsy—
may represent the best opportunity to capture and
provide a better representation of therapeutic
responses. In DLBCL, ctDNA has been shown to cap-
ture both the mutational landscape and clonal evolu-
tion, while demonstrating prognostic relevance at
various time-points.
113,114
Studies evaluating ctDNA
as a dynamic biomarker in FL are emerging. The pre-
treatment ctDNA levels in FL patients correlated with
prognosis and tumour burden as quantified by
imaging.
115
Due to the much lower ctDNA levels in
FL compared to aggressive lymphomas such as
DLBCL, higher precision assays are needed for disease
monitoring. Approaches ranging from clonotypical
analyses to individual patient-defined amplicon
mini-gene panels to broader non-individualised tar-
geted gene sequencing have been investigated in pilot
cohorts, demonstrating feasibility and varying prog-
nostic accuracy.
116–118
Of added interest is the possi-
bility of utilising ctDNA to predict FL transformation,
especially as some genetic events associated with
transformation can be detected several months
earlier.
119
This minimally invasive modality might
offer the opportunity to capture heterogeneity while
dynamically monitoring disease response to treatment
and the ability to forecast progression, but this still
requires validation in larger cohorts and continued
refinement of the assay precision (Box 4).
Concluding remarks
EVOLVE OUR INFRASTRUCTURE
•The need for multicentre collaboration to
create large, collaborative biorepositories with
well-annotated clinical data spanning the full spectrum
of disease phenotypes from real-world cohorts. This
BOX 4. Key requirements for MRD detection and
response-driven strategies
We need to:
•Determine which liquid biopsy-based assays
have sufficiently high sensitivity for minimal
residual disease (MRD) detection. This will
require the creation of biobanking efforts for
sequential biopsy collection from cohorts of
patients with detailed clinical history to facili-
tate multiple assay evaluation and subsequent
validation
•Identify the best time-point for MRD assess-
ment and outcome prediction, which will
probably depend upon baseline clinical/
biological characteristics. Multimodal integra-
tion will be required to develop the most cost-
effective MRD model
•Leverage knowledge obtained from dynamic
response monitoring to design response-
adapted trials with retreatment and/or escala-
tion strategies to minimise toxicity while
improving outcome for slow responders
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
88 M Collin et al.
approach will enable the generation of multi-modal
data, including various omics, imaging, and pathology,
on large patient cohorts. Such comprehensive datasets
will be invaluable for meaningfully addressing
outcome-driven questions at scale, such as AI-based
algorithms for prognosis and response prediction.
•Collaboration is essential from a wide range of
stakeholders, including clinicians, researchers, admin-
istrators, pharma, funders, patients and advocacy
groups. Addressing the regulatory and logistical chal-
lenges—such as governance, data sharing and harmo-
nisation—will be critical to making this collaboration
effective.
EVOLVE OUR RESEARCH APPROACH
•Identify key research priority areas (as
highlighted in this review) and adopt a reverse trans-
lational mind-set by rapidly taking observations
learnt from prior research and clinical challenges
back to the laboratory to model the complexities seen
in patients.
•Rather than have siloed research teams compet-
ing in parallel, which can lead to redundancies,
incentivise team science and multidisciplinary initia-
tives to pool expertise, resources and increase effi-
ciency of research spending.
EVOLVE OUR CLINICAL TRANSLATION
•Biological discovery and rationale should be at
the heart of future clinical trial innovation—moving
towards biology-driven platform trials. To achieve
this, we must include translational research from the
trial conception, dynamic specimen biobanking dur-
ing the trial and funding to support these ancillary
initiatives. This will allow all manner of biomarkers
(prognostic, predictive and dynamic) to be evaluated
in addition to ones associated with drug toxicities.
Ultimately, these need to be translated into clinically
accessible, affordable and tractable biomarkers.
•Prioritise existing dynamic biomarkers of response
to support development of response-adapted trials
that would curtail treatment in patients who are poor
responders and spare toxicity in those that have
already achieved a good response.
Acknowledgements
All authors contributed to the writing and proofread-
ing of the final version of the manuscript.
Conflicts of interest
The authors declare no conflicts of interest.
Data availability statement
Data sharing is not applicable as no new data were
created or analysed in this study.
References
1. Dreval K, Hilton LK, Cruz M et al. Genetic subdivisions of fol-
licular lymphoma defined by distinct coding and noncoding
mutation patterns. Blood 2023; 142; 561–573.
2. Bachy E, Seymour JF, Feugier P et al. Sustained progression-
free survival benefit of rituximab maintenance in patients
with follicular lymphoma: long-term results of the PRIMA
study. J. Clin. Oncol. 2019; 37; 2815–2824.
3. Sarkozy C, Maurer MJ, Link BK et al. Cause of death in follicu-
lar lymphoma in the first decade of the rituximab era: a
pooled analysis of French and US cohorts. J. Clin. Oncol.
2019; 37; 144–152.
4. Casulo C, Dixon JG, Le-Rademacher J et al. Validation of
POD24 as a robust early clinical end point of poor survival in
FL from 5225 patients on 13 clinical trials. Blood 2022; 139;
1684–1693.
5. Freeman CL, Kridel R, Moccia AA et al. Early progression after
bendamustine-rituximab is associated with high risk of trans-
formation in advanced stage follicular lymphoma. Blood
2019; 134; 761–764.
6. Munta~
nola A, Mozas P, Mercadal S et al. Early progression in
follicular lymphoma in the absence of histological transforma-
tion or high-risk follicular lymphoma international prognostic
index still has a favourable outcome. Br. J. Haematol. 2023;
200; 306–314.
7. Mann RB, Berard CW. Criteria for the cytologic subclassifica-
tion of follicular lymphomas: a proposed alternative method.
Hematol. Oncol. 1983; 1; 187–192.
8. Rimsza LM, Li H, Braziel RM et al. Impact of histological grad-
ing on survival in the SWOG S0016 follicular lymphoma
cohort. Haematologica 2018; 103; e151–e153.
9. Kroft SH. Stratification of follicular lymphoma: time for a par-
adigm shift? Am. J. Clin. Pathol. 2019; 151; 539–541.
10. Hiddemann W, Barbui AM, Canales MA et al. Immunochem-
otherapy with obinutuzumab or rituximab for previously
untreated follicular lymphoma in the GALLIUM study: influ-
ence of chemotherapy on efficacy and safety. J. Clin. Oncol.
Off. J. Am. Soc. Clin. Oncol. 2018; 36; 2395–2404.
11. Marcus R, Davies A, Ando K et al. Obinutuzumab for the
first-line treatment of follicular lymphoma. N. Engl. J. Med.
2017; 377; 1331–1344.
12. Solal-C
eligny P, Roy P, Colombat P et al. Follicular lymphoma
international prognostic index. Blood 2004; 104; 1258–
1265.
13. Federico M, Bellei M, Marcheselli L et al. Follicular lymphoma
international prognostic index 2: a new prognostic index for
follicular lymphoma developed by the international follicular
lymphoma prognostic factor project. J. Clin. Oncol. Off. J. Am.
Soc. Clin. Oncol. 2009; 27; 4555–4562.
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
A follicular lymphoma roadmap 89
14. Bachy E, Maurer MJ, Habermann TM et al. A simplified scor-
ing system in de novo follicular lymphoma treated initially
with immunochemotherapy. Blood 2018; 132;49–58.
15. Mir F, Mattiello F, Grigg A et al. Follicular lymphoma evalua-
tion index (FLEX): a new clinical prognostic model that is
superior to existing risk scores for predicting progression-free
survival and early treatment failure after frontline immuno-
chemotherapy. Am. J. Hematol. 2020; 95; 1503–1510.
16. Pastore A, Jurinovic V, Kridel R et al. Integration of gene muta-
tions in risk prognostication for patients receiving first-line
immunochemotherapy for follicular lymphoma: a retrospective
analysis of a prospective clinical trial and validation in a
population-based registry. Lancet Oncol. 2015; 16; 1111–1122.
17. Huet S, Tesson B, Jais J-P et al. A gene-expression profiling
score for prediction of outcome in patients with follicular lym-
phoma: a retrospective training and validation analysis in
three international cohorts. Lancet Oncol. 2018; 19; 549–
561.
18. Lipkova J, Chen TY, Lu MY et al. Deep learning-enabled
assessment of cardiac allograft rejection from endomyocardial
biopsies. Nat. Med. 2022; 28; 575–582.
19. Lipkova J, Chen RJ, Chen B et al. Artificial intelligence for
multimodal data integration in oncology. Cancer Cell 2022;
40; 1095–1110.
20. Song AH, Jaume G, Williamson DFK et al. Artificial intelli-
gence for digital and computational pathology. Nat. Rev.
Bioeng. 2023; 1; 930–949.
21. Li D, Bledsoe JR, Zeng Y et al. A deep learning diagnostic plat-
form for diffuse large B-cell lymphoma with high accuracy
across multiple hospitals. Nat. Commun. 2020; 11; 6004.
22. Chen RJ, Lu MY, Williamson DFK et al. Pan-cancer integra-
tive histology-genomic analysis via multimodal deep learning.
Cancer Cell 2022; 40; 865–878.e6.
23. Chen RJ, Lu MY, Wang J et al. Pathomic fusion: an integrated
framework for fusing histopathology and genomic features for
cancer diagnosis and prognosis. IEEE Trans. Med. Imaging
2022; 41; 757–770.
24. Morschhauser F, Tilly H, Chaidos A et al. Tazemetostat for
patients with relapsed or refractory follicular lymphoma: an
open-label, single-arm, multicentre, phase 2 trial. Lancet
Oncol. 2020; 21; 1433–1442.
25. Crouch S, Painter D, Barrans SL et al. Molecular subclusters
of follicular lymphoma: a report from the United Kingdom’s
haematological Malignancy Research Network. Blood Adv.
2022; 6; 5716–5731.
26. Okosun J, Wolfson RL, Wang J et al. Recurrent mTORC1-
activating RRAGC mutations in follicular lymphoma. Nat.
Genet. 2016; 48; 183–188.
27. Ying ZX, Jin M, Peterson LF et al. Recurrent mutations in the
MTOR regulator RRAGC in follicular lymphoma. Clin. Cancer
Res. 2016; 22; 5383–5393.
28. Krull JE, Wenzl K, Hopper MA et al. Follicular lymphoma B
cells exhibit heterogeneous transcriptional states with associ-
ated somatic alterations and tumor microenvironments. Cell
Rep. Med. 2024; 5; 101443.
29. Han G, Deng Q, Marques-Piubelli ML et al. Follicular lym-
phoma microenvironment characteristics associated with
tumor cell mutations and MHC class II expression. Blood Can-
cer Discov. 2022; 3; 428–443.
30. Milpied P, Gandhi AK, Cartron G et al. Follicular lymphoma
dynamics. Adv. Immunol. 2021; 150;43–103.
31. Sungalee S, Mamessier E, Morgado E et al. Germinal center
reentries of BCL2-overexpressing B cells drive follicular
lymphoma progression. J. Clin. Invest. 2014; 124; 5337–
5351.
32. Bogn
ar A, Csernus B, B€
od€
or C et al. Clonal selection in the
bone marrow involvement of follicular lymphoma. Leukemia
2005; 19; 1656–1662.
33. Carlotti E, Wrench D, Matthews J et al. Transformation of fol-
licular lymphoma to diffuse large B-cell lymphoma may occur
by divergent evolution from a common progenitor cell or by
direct evolution from the follicular lymphoma clone. Blood
2009; 113; 3553–3557.
34. Haebe S, Keay W, Alig S et al. The molecular ontogeny of fol-
licular lymphoma: gene mutations succeeding the BCL2
translocation define common precursor cells. Br. J. Haematol.
2022; 196; 1381–1387.
35. Schroers-Martin JG, Soo J, Brisou G et al. Tracing founder
mutations in circulating and tissue-resident follicular lym-
phoma precursors. Cancer Discov. 2023; 13; 1310–1323.
36. Araf S, Wang J, Korfi K et al. Genomic profiling reveals spatial
intra-tumor heterogeneity in follicular lymphoma. Leukemia
2018; 32; 1261–1265.
37. Haebe S, Shree T, Sathe A et al. Single-cell analysis can define
distinct evolution of tumor sites in follicular lymphoma. Blood
2021; 137; 2869–2880.
38. Nagy
A, B
atai B, Kiss L et al. Parallel testing of liquid biopsy
(ctDNA) and tissue biopsy samples reveals a higher frequency
of EZH2 mutations in follicular lymphoma. J. Intern. Med.
2023; 294; 295–313.
39. Okosun J, B€
od€
or C, Wang J et al. Integrated genomic analysis
identifies recurrent mutations and evolution patterns driving
the initiation and progression of follicular lymphoma. Nat.
Genet. 2014; 46; 176–181.
40. Pasqualucci L, Khiabanian H, Fangazio M et al. Genetics of
follicular lymphoma transformation. Cell Rep. 2014; 6; 130–
140.
41. Kridel R, Chan FC, Mottok A et al. Histological transformation
and progression in follicular lymphoma: a clonal evolution
study. PLoS Med. 2016; 13; e1002197.
42. Davids MS, Roberts AW, Seymour JF et al. Phase I first-in-
human study of Venetoclax in patients with relapsed or
refractory non-Hodgkin lymphoma. J. Clin. Oncol. 2017; 35;
826–833.
43. Roulland S, Kelly RS, Morgado E et al. t(14;18) translocation:
a predictive blood biomarker for follicular lymphoma. J. Clin.
Oncol. 2014; 32; 1347–1355.
44. Pasqualucci L, Dominguez-Sola D, Chiarenza A et al. Inacti-
vating mutations of acetyltransferase genes in B-cell lym-
phoma. Nature 2011; 471; 189–195.
45. Jiang Y, Ortega-Molina A, Geng H et al. CREBBP inactivation
promotes the development of HDAC3-dependent lymphomas.
Cancer Discov. 2017; 7;38–53.
46. Mondello P, Tadros S, Teater M et al. Selective inhibition of
HDAC3 targets synthetic vulnerabilities and activates immune
surveillance in lymphoma. Cancer Discov. 2020; 10; 440–
459.
47. McHeyzer-Williams LJ, Milpied PJ, Okitsu SL et al. Switched-
memory B cells remodel B cell receptors within secondary ger-
minal centers. Nat. Immunol. 2015; 16; 296–305.
48. Milpied P, Cervera-Marzal I, Mollichella M-L et al. Human ger-
minal center transcriptional programs are de-synchronized in
B cell lymphoma. Nat. Immunol. 2018; 19; 1013–1024.
49. Holmes AB, Corinaldesi C, Shen Q et al. Single-cell analysis of
germinal-center B cells informs on lymphoma cell of origin
and outcome. J. Exp. Med. 2020; 217; e20200483.
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
90 M Collin et al.
50. Kennedy DE, Okoreeh MK, Maienschein-Cline M et al. Novel
specialized cell state and spatial compartments within the ger-
minal center. Nat. Immunol. 2020; 21; 660–670.
51. Attaf N, Baaklini S, Binet L, Milpied P. Heterogeneity of ger-
minal center B cells: new insights from single-cell studies.
Eur. J. Immunol. 2021; 51; 2555–2567.
52. Mlynarczyk C, Font
an L, Melnick A. Germinal center-derived
lymphomas: the darkest side of humoral immunity. Immunol.
Rev. 2019; 288; 214–239.
53. Caganova M, Carrisi C, Varano G et al. Germinal center dysre-
gulation by histone methyltransferase EZH2 promotes lym-
phomagenesis. J. Clin. Invest. 2013; 123; 5009–5022.
54. Ortega-Molina A, Boss IW, Canela A et al. The histone lysine
methyltransferase KMT2D sustains a gene expression pro-
gram that represses B cell lymphoma development. Nat. Med.
2015; 21; 1199–1208.
55. Zhang J, Dominguez-Sola D, Hussein S et al. Disruption of
KMT2D perturbs germinal center B cell development and pro-
motes lymphomagenesis. Nat. Med. 2015; 21; 1190–1198.
56. B
eguelin W, Popovic R, Teater M et al. EZH2 is required for
germinal center formation and somatic EZH2 mutations pro-
mote lymphoid transformation. Cancer Cell 2013; 23; 677–
692.
57. B
eguelin W, Teater M, Meydan C et al. Mutant EZH2 induces
a pre-malignant lymphoma niche by reprogramming the
immune response. Cancer Cell 2020; 37; 655–673.e11.
58. Attaf N, Dong C, Gil L et al. Functional plasticity and recur-
rent cell states of malignant B cells in follicular lymphoma.
2022.
59. Wang X, Nissen M, Gracias D et al. Single-cell profiling
reveals a memory B cell-like subtype of follicular lymphoma
with increased transformation risk. Nat. Commun. 2022; 13;
6772.
60. Roider T, Seufert J, Uvarovskii A et al. Dissecting intratumour
heterogeneity of nodal B-cell lymphomas at the transcrip-
tional, genetic and drug-response levels. Nat. Cell Biol. 2020;
22; 896–906.
61. Young RM, Staudt LM. Targeting pathological B cell receptor
signalling in lymphoid malignancies. Nat. Rev. Drug Discov.
2013; 12; 229–243.
62. Phelan JD, Young RM, Webster DE et al. A multiprotein
supercomplex controlling oncogenic signalling in lymphoma.
Nature 2018; 560; 387–391.
63. Gopal AK, Kahl BS, de Vos S et al. PI3Kdinhibition by idelali-
sib in patients with relapsed indolent lymphoma. N. Engl. J.
Med. 2014; 370; 1008–1018.
64. Matasar MJ, Capra M,
€
Ozcan M et al. Copanlisib plus rituxi-
mab versus placebo plus rituximab in patients with relapsed
indolent non-Hodgkin lymphoma (CHRONOS-3): a double-
blind, randomised, placebo-controlled, phase 3 trial. Lancet
Oncol. 2021; 22; 678–689.
65. Flinn IW, Miller CB, Ardeshna KM et al. DYNAMO: a phase II
study of duvelisib (IPI-145) in patients with refractory indo-
lent non-Hodgkin lymphoma. J. Clin. Oncol. Off. J. Am. Soc.
Clin. Oncol. 2019; 37; 912–922.
66. Zinzani PL, Mayer J, Flowers CR et al. ROSEWOOD: a phase II
randomized study of Zanubrutinib plus obinutuzumab versus
obinutuzumab monotherapy in patients with relapsed or
refractory follicular lymphoma. J. Clin. Oncol. 2023; 41;
5107–5117.
67. Study details |a study of zanubrutinib plus anti-CD20 versus
lenalidomide plus rituximab in participants with relapsed/
refractory follicular or marginal zone lymphoma.
ClinicalTrials.gov,https://clinicaltrials.gov/study/NCT05100
862?term=NCT05100862&rank=1(accessed 9 September
2024).
68. Amin R, Mourcin F, Uhel F et al. DC-SIGN-expressing macro-
phages trigger activation of mannosylated IgM B-cell receptor
in follicular lymphoma. Blood 2015; 126; 1911–1920.
69. Radtke AJ, Postovalova E, Varlamova A et al. Multi-omic pro-
filing of follicular lymphoma reveals changes in tissue archi-
tecture and enhanced stromal remodeling in high-risk
patients. Cancer Cell 2024; 42; 444–463.e10.
70. Linley A, Krysov S, Ponzoni M, Johnson PW, Packham G,
Stevenson FK. Lectin binding to surface Ig variable regions
provides a universal persistent activating signal for follicular
lymphoma cells. Blood 2015; 126; 1902–1910.
71. Rojekar S, Gholap AD, Togre N et al. Current status of man-
nose receptor-targeted drug delivery for improved anti-HIV
therapy. J. Control. Release 2024; 372; 494–521.
72. Morin RD, Mendez-Lago M, Mungall AJ et al. Frequent muta-
tion of histone-modifying genes in non-Hodgkin lymphoma.
Nature 2011; 476; 298–303.
73. Green MR, Kihira S, Liu CL et al. Mutations in early follicular
lymphoma progenitors are associated with suppressed antigen
presentation. Proc. Natl. Acad. Sci. USA 2015; 112; E1116–
E1125.
74. Li J, Chin CR, Ying H-Y et al. Loss of CREBBP and KMT2D
cooperate to accelerate lymphomagenesis and shape the lym-
phoma immune microenvironment. Nat. Commun. 2024; 15;
2879.
75. B€
od€
or C, Grossmann V, Popov N et al. EZH2 mutations are
frequent and represent an early event in follicular lymphoma.
Blood 2013; 122; 3165–3168.
76. Isshiki Y, Porazzi P, Chen X et al. EZH2 inhibitors enhance
CART cell quality, efficacy, in vivo homing, tumor cell binding
and killing of fully syngeneic primary B cell lymphomas, As
well As reprogramming lymphoma cells to a highly immuno-
genic and T cell adherent phenotype. Blood 2023; 142; 432.
77. Heward J, Konali L, D’Avola A et al. KDM5 inhibition offers a
novel therapeutic strategy for the treatment of KMT2D
mutant lymphomas. Blood 2021; 138; 370–381.
78. Dave SS, Wright G, Tan B et al. Prediction of survival in fol-
licular lymphoma based on molecular features of tumor-
infiltrating immune cells. N. Engl. J. Med. 2004; 351; 2159–
2169.
79. Tobin JWD, Keane C, Gunawardana J et al. Progression of dis-
ease within 24 months in follicular lymphoma is associated
with reduced intratumoral immune infiltration. J. Clin. Oncol.
2019; 37; 3300–3309.
80. Scott DW, Gascoyne RD. The tumour microenvironment in B
cell lymphomas. Nat. Rev. Cancer 2014; 14; 517–534.
81. Yang Z-Z, Kim HJ, Villasboas JC et al. Mass cytometry analysis
reveals that specific intratumoral CD4+T cell subsets corre-
late with patient survival in follicular lymphoma. Cell Rep.
2019; 26; 2178–2193.e3.
82. Pandey S, Mourcin F, Marchand T et al. IL-4/CXCL12 loop is
a key regulator of lymphoid stroma function in follicular lym-
phoma. Blood 2017; 129; 2507–2518.
83. B
elanger S, Crotty S. Dances with cytokines, featuring TFH
cells, IL-21, IL-4 and B cells. Nat. Immunol. 2016; 17; 1135–
1136.
84. Eto D, Lao C, DiToro D et al. IL-21 and IL-6 are critical for
different aspects of B cell immunity and redundantly induce
optimal follicular helper CD4 T cell (Tfh) differentiation. PLoS
One 2011; 6; e17739.
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
A follicular lymphoma roadmap 91
85. Yang Z-Z, Novak AJ, Ziesmer SC, Witzig TE, Ansell SM. Atten-
uation of CD8(+) T-cell function by CD4(+)CD25(+) regulatory
T cells in B-cell non-Hodgkin’s lymphoma. Cancer Res. 2006;
66; 10145–10152.
86. Rodriguez S, Alizadeh M, Lamaison C et al. Follicular lym-
phoma regulatory T-cell origin and function. Front. Immunol.
2024; 15; 1391404.
87. Mondello P, Fama A, Larson MC et al. Lack of intrafollicular
memory CD4
+
T cells is predictive of early clinical failure in
newly diagnosed follicular lymphoma. Blood Cancer J. 2021;
11; 130.
88. Wahlin BE, Sander B, Christensson B et al. CD8
+
T-cell con-
tent in diagnostic lymph nodes measured by flow cytometry
is a predictor of survival in follicular lymphoma. Clin. Cancer
Res. 2007; 13; 388–397.
89. Nastoupil LJ, Chin CK, Westin JR et al. Safety and activity of
pembrolizumab in combination with rituximab in relapsed or
refractory follicular lymphoma. Blood Adv. 2022; 6; 1143–
1151.
90. Ho CI, Gopal AK, Ujjani CS et al. Pembrolizumab with rituxi-
mab in relapsed/refractory follicular lymphoma and diffuse
large B cell lymphoma. Blood 2023; 142; 6144.
91. Armand P, Janssens A, Gritti G et al. Efficacy and safety
results from CheckMate 140, a phase 2 study of nivolumab
for relapsed/refractory follicular lymphoma. Blood 2021; 137;
637–645.
92. Linton KM, Vitolo U, Jurczak W et al. Epcoritamab mono-
therapy in patients with relapsed or refractory follicular
lymphoma (EPCORE NHL-1): a phase 2 cohort of a single-
arm, multicentre study. Lancet Haematol. 2024; 11; e593–
e605.
93. Kim TM, Taszner M, Novelli S et al. Safety and efficacy of
odronextamab in patients with relapsed or refractory follicular
lymphoma. Ann. Oncol. 2024; Aug 13:S0923-7534(24)
03759-1. https://doi.org/10.1016/j.annonc.2024.08.2239.
94. Morschhauser F, Bishton M, Eyre TA et al. Mosunetuzumab
in combination with lenalidomide has a manageable safety
profile and encouraging activity in patients with relapsed/
refractory follicular lymphoma: initial results from a phase Ib
study. Blood 2021; 138; 129.
95. Belada D, Falchi L, Lepp€
aSet al. Epcoritamab with rituximab
+lenalidomide (R2) provides durable responses in high-risk
follicular lymphoma, regardless of POD24 status. Hematol.
Oncol. 2023; 41; 125–127.
96. Morschhauser F, Dahiya S, Palomba ML et al. Lisocabtagene
maraleucel in follicular lymphoma: the phase 2 TRANSCEND
FL study. Nat. Med. 2024; 30; 2199–2207.
97. Neelapu SS, Chavez JC, Sehgal AR et al. Three-year follow-up
analysis of axicabtagene ciloleucel in relapsed/refractory indo-
lent non-Hodgkin lymphoma (ZUMA-5). Blood 2024; 143;
496–506.
98. Fowler NH, Dickinson M, Dreyling M et al. Tisagenlecleucel in
adult relapsed or refractory follicular lymphoma: the phase 2
ELARA trial. Nat. Med. 2022; 28; 325–332.
99. Plaks V, Chou J, Goyal L et al. Abstract CT036: axicabtagene
ciloleucel (axi-cel) product attributes and immune biomarkers
associated with clinical outcomes in patients (pts) with
relapsed/refractory (R/R) indolent non-Hodgkin lymphoma
(iNHL) in ZUMA-5. Cancer Res. 2021; 81; CT036.
100. Scholler N, Perbost R, Locke FL et al. Tumor immune con-
texture is a determinant of anti-CD19 CAR T cell efficacy
in large B cell lymphoma. Nat. Med. 2022; 28; 1872–
1882.
101. Toninelli M, Rossetti G, Pagani M. Charting the tumor micro-
environment with spatial profiling technologies. Trends Cancer
2023; 9; 1085–1096.
102. Tian L, Chen F, Macosko EZ. The expanding vistas of spatial
transcriptomics. Nat. Biotechnol. 2023; 41; 773–782.
103. Russell AJC, Weir JA, Nadaf NM et al. Slide-tags enables
single-nucleus barcoding for multimodal spatial genomics.
Nature 2024; 625; 101–109.
104. Sarkozy C, Wu S, Takata K et al. Integrated single cell analy-
sis reveals co-evolution of malignant B cells and tumor
micro-environment in transformed follicular lymphoma. Can-
cer Cell 2024; 42; 1003–1017.e6.
105. Dimitrov D, Sch€
afer PSL, Farr E et al. LIANA+provides an
all-in-one framework for cell-cell communication inference.
Nat. Cell Biol. 2024; 26; 1613–1622.
106. Cang Z, Zhao Y, Almet AA et al. Screening cell-cell communi-
cation in spatial transcriptomics via collective optimal trans-
port. Nat. Methods 2023; 20; 218–228.
107. Liu Z, Sun D, Wang C. Evaluation of cell-cell interaction
methods by integrating single-cell RNA sequencing data with
spatial information. Genome Biol. 2022; 23; 218.
108. Gribben JG, Freedman AS, Woo SD et al. All advanced stage
non-Hodgkin’s lymphomas with a polymerase chain reaction
amplifiable breakpoint of bcl-2 have residual cells containing
the bcl-2 rearrangement at evaluation and after treatment.
Blood 1991; 78; 3275–3280.
109. Galimberti S, Luminari S, Ciabatti E et al. Minimal residual
disease after conventional treatment significantly impacts on
progression-free survival of patients with follicular lym-
phoma: the FIL FOLL05 trial. Clin. Cancer Res. 2014; 20;
6398–6405.
110. Luminari S, Manni M, Galimberti S et al. Response-adapted
postinduction strategy in patients with advanced-stage follic-
ular lymphoma: the FOLL12 study. J. Clin. Oncol. 2022; 40;
729–739.
111. Pott C, Sehn LH, Belada D et al. MRD response in relapsed/
refractory FL after obinutuzumab plus bendamustine or bend-
amustine alone in the GADOLIN trial. Leukemia 2020; 34;
522–532.
112. Pott C, Jurinovic V, Trotman J et al. Minimal residual disease
status predicts outcome in patients with previously untreated
follicular lymphoma: a prospective analysis of the phase III
GALLIUM study. J. Clin. Oncol. 2024; 42; 550–561.
113. Kurtz DM, Scherer F, Jin MC et al. Circulating tumor DNA
measurements as early outcome predictors in diffuse large B-
cell lymphoma. J. Clin. Oncol. 2018; 36; 2845–2853.
114. Roschewski M, Dunleavy K, Pittaluga S et al. Circulating
tumour DNA and CT monitoring in patients with untreated
diffuse large B-cell lymphoma: a correlative biomarker study.
Lancet Oncol. 2015; 16; 541–549.
115. Delfau-Larue M-H, van der Gucht A, Dupuis J et al. Total
metabolic tumor volume, circulating tumor cells, cell-free
DNA: distinct prognostic value in follicular lymphoma. Blood
Adv. 2018; 2; 807–816.
116. Sarkozy C, Huet S, Carlton VEH et al. The prognostic value
of clonal heterogeneity and quantitative assessment of
plasma circulating clonal IG-VDJ sequences at diagnosis in
patients with follicular lymphoma. Oncotarget 2017; 8;
8765–8774.
117. Fern
andez-Miranda I, Pedrosa L, Llanos M et al. Monitoring
of circulating tumor DNA predicts response to treatment and
early progression in follicular lymphoma: results of a prospec-
tive pilot study. Clin. Cancer Res. 2023; 29; 209–220.
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
92 M Collin et al.
118. Jim
enez-Ubieto A, Poza M, Martin-Mu~
noz A et al. Real-life
disease monitoring in follicular lymphoma patients using liq-
uid biopsy ultra-deep sequencing and PET/CT. Leukemia
2023; 37; 659–669.
119. Scherer F, Kurtz DM, Newman AM et al. Distinct biological
subtypes and patterns of genome evolution in lymphoma
revealed by circulating tumor DNA. Sci. Transl. Med. 2016;
8; 364ra155.
Ó2024 The Author(s). Histopathology published by John Wiley & Sons Ltd., Histopathology,86, 79–93.
A follicular lymphoma roadmap 93
