Lymphoma classifications, how to develop a future unified taxonomy

Introduction

In 2022 two different lymphoma classifications were published – the 5^th edition of the WHO Classification of Hematolymphoid Tumours (WHO-HAEM5), coordinated by the WHO-IARC Editorial Board¹, and the International Consensus Classification (ICC) developed by the Clinical Advisory Committee (CAC)². The WHO-HAEM5/ICC divide interrupts an almost 30 years era of a single, widely accepted lymphoma classification³. Absence of a consensus for designating lymphoid disorders may hinder communications between physicians, threatening the inclusion of patients in clinical trials and inevitably confounding the trainees.

Despite this apparent fracture in our field, these two lymphoma classifications have more commonalities than differences. The basic concepts and approaches are similar, both aiming toward the goal of defining clinicopathologic entities, with a secondary aim of identifying phenotypic and molecular markers to guide diagnosis and therapy.

Here, a group of experts involved in both WHO-HAEM5 and ICC review and discuss the existing differences in the two systems and the steps required to bridge the gaps leading to a unified future lymphoma classification (Table 1, Suppl. Figures 1,2).

List of main discrepancies between ICC and WHO-HAEM5 classifications with proposed actions needed to resolve outstanding questions

		ICC 2023	WHO-HAEM5	Clinical/Research Implications and Next Steps
Entity	MAIN DISCREPANCIES
Follicular lymphoma (FL)	Grading	Retain morphological grading (with the distinction of 3B FL)	Grading of FL optional in the subtyping of cFL (with the distinction of follicular large B-cell)	1. Assess the utility of morphometry, proliferation index, OMICs data, artificial intelligence or machine learning algorithms in defining robust prognostic subgroups within FL 2. Integrated clinico-molecular studies to investigate outcome of homogenously treated FL based on OMICs features (e.g. mutations, epigenome, transcriptome, single cell analyses) and correlation with histological grading
	Testicular FL	Recognized as distinct subtype of FL in young boys	Not recognized as distinct FL subtype	Determine biologic and clinical features in larger series.
	BCL2r-negative, CD23-positive follicle center lymphoma	Recognized as provisional entity	Partially overlapping with FL with predominantly diffuse growth pattern (dFL)	1. Investigations to determine whether this represents a distinct homogeneous clinicopathology entity within the broader group of BCL2-R-negative FLs 2. Define clinical differences between FL with/without del(1)(p36), STAT6, CREBBP, TNFRSF14 mutation in the context of absence of BCL2r
	FL with unusual cytological features	Not recognized	Recognized	Establish precise diagnostic criteria through integrated clinical, pathological and molecular studies
Diffuse large B-cell lymphoma (DLBCL)	Primary large B-cell lymphoma of immune-privileged sites	No umbrella group included (felt to be premature)	Includes CNS, testis, vitreo-retinal	Comparative genomic studies among different extranodal locations of MCD/C5/MYD88 (°) DLBCL genomic subtypes
	HHV8-negative primary effusion-based lymphoma	EBV-positive cases excluded	EBV-positive cases included	1. Study larger series of cases, with full clinical and pathological annotations, in different countries 2. Asses diagnostic criteria and clinical relevance for EBV positivity 3. Establish precise diagnostic criteria and consensus nomenclature
DLBCL/High grade B-cell lymphoma (HGBL)	HGBL DH BCL6/MYC	Provisional entitiy	No a specific entity; encompassed within DLBCL, NOS or HGBL, NOS	1. Assess the clinical behavior and outcome of BCL6r and MYCr HGBL and underlying biology. 2. Specific research to address the cases with MYC::BCL6 mutual translocation (MYCr with BCL6 locus as partner)
	DHL terminology	High-grade (irrespective of morphology)	High grade or Large cell (based on morphology)	1. Establish more precise diagnostic criteria. Clinico-pathological studies to determine whether cytological subtype influences diagnosis or therapy. 2. Clinical studies/trials to determine the clinical relevance, underlying biology and diagnostic criteria of HGBL, NOS 3. Consensus conference to determine the appropriate terminology for cases with concomitant rearrangements of MYC and BCL2 as well as MYC and BCL6
Immunodeficiency LPDs	Immunodeficiency LPDs	ICC emphasizes the clinical context for diagnosis and treatment	WHO-HAEM5 systematically applies histopathological, viral and immunodeficiency characteristics seen across different clinical settings	1. Consensus conference to determine the appropriate terminology and the impact of the underlying immune scenario in influencing therapy. 2. Determine the prognostic significance of cases EBV+ vs. EBV- 3. Studies to improve diagnostic criteria for polymorphic vs. monomorphic (lymphoma) categories
	Epstein-Barr virus-positive mucocutaneous ulcer
	EBV positive polymorphic B cell LPD, NOS
	Inborn error of immunity-associated lymphoid proliferations and lymphomas
Plasma cell myeloma/Multiple Myeloma	Molecular subclassification	YES Subclassification based on genetic features.	No genetic subclassification	Comprehensive clinical studies or trials with genetic annotations to determine the clinical relevance of genetic classification in patients receiving novel therapies.
B-PLL and HCLv vs. SBLPN		No changes. B-PLL and HCLv remain as distinct entities	SBLPN entity created to encompass B-PLL and HCLv and some other rare splenic B-cell lymphoma cases	1. Study of large cohorts to define clinical, immunophenotypic and genomic features of the different entities 2. Development of blood-based genomic biomarkers to subclassify splenic B cell lymphoid neoplasms

Abbreviations: ICC, International Consensus Classification; WHO-HAEM5, 5th edition of the WHO Classification of Hematolymphoid Tumours; CNS, central nervous system; NSCHL, nodular sclerosis classic Hodgkin lymphoma; LPD, lymphoproliferative disorder; MGZL, mediastinal grey zone lymphoma; PMBL, primary mediastinal B-cell lymphoma; PLL, prolymphocytic leukemia; HCLv, Hairy cell leukemia, variant; SBLPN, splenic B-cell lymphoma/leukemia with prominent nucleoli; NOS, not otherwise specified; LBCL, large B-cell lymphoma. (°) refers to genetic subgroups of DLBCL as summarized in ¹¹

Follicular Lymphoma

Essential diagnostic criteria are similar for the recognition of follicular lymphoma (FL) in both the WHO-HAEM5 and the ICC. Conceptual differences are the identification of discrete subclasses of follicular lymphomas and the requirement for histological grading.

Both classifications recognize a core category of FL, which in the WHO-HAEM5 is termed “classic FL” encompassing the former FL grades 1, 2 and 3A. The discussion as to whether FL should be graded or not dates back decades. Regarding clinical applicability of the updated classification systems, grading (1–3A) is not required for the diagnosis of FL. It remains debatable whether reporting grade 3A might influence, with other parameters, therapeutic decision-making. Finally, a better delineation of grade 3B or large-cell lymphoma FL is also warranted.

Objective methodologies for reproducible FL grading are needed, integrating data from molecular studies (e.g. mutations, epigenetics, single cell RNA/DNA sequencing, cell-free DNA). Extensive molecular testing, immunologic characterization and correlation with clinical and imaging information would also clarify the biologic features of the different FL subtypes and grades, including their prognostic impact (Table 1).

Rare FL subtypes with distinct biological features are being increasingly recognized and may require unique management paradigms. Duodenal-type FL and pediatric-type FL are the same in both WHO-HAEM5 and ICC. In contrast, primary testicular FL is delineated as a specific entity in the ICC, based on distinctive clinical and biological features, whereas it is included with other FL variants in the WHO-HAEM5. The WHO-HAEM5 includes subtypes of predominantly diffuse follicular lymphoma (dFL) and follicular lymphomas with unusual cytologic features. In contrast, the ICC identifies – as a provisional entity – BCL2 rearrangement-negative, CD23-positive follicle center lymphoma, partially overlapping with dFL according to WHO-HAEM5. These differences also highlight the different approaches chosen by the two classifications: a more morphologically based approach of the WHO-HAEM5 and a more molecularly based approach by the ICC.

The specific position for large B-cell lymphoma (LBCL) with IRF4 rearrangement is a controversial issue. Because of its frequent large cell morphology, WHO-HAEM5 has moved it to the LBCL family, while ICC placed it together with FL due to the almost constant presence of follicular areas and the better outcome of the patients compared to DLBCL, NOS in the same age group ^4–8.

Diffuse Large B cell lymphoma (DLBCL) and High-grade B cell lymphoma (HGBL)

While most aggressive B-cell lymphoma categories remain nearly identical in the two classifications, several differences exist between the WHO-HAEM5 and ICC about DLBCL and HGBL (Table 1, Suppl. Figure 1). While some are terminological issues, conceptual issues can be divided into 3 broad categories: 1) classification of LBCL in immune-privileged sites; 2) inclusion criteria for so-called “double hit” lymphoma and its designation as high grade versus LBCL; 3) HHV8-negative effusion-based/fluid overload associated LBCL.

Primary CNS DLBCL (PCNSLBL) is well-established as a distinct clinical and biologic entity and remains virtually unchanged in both new classifications compared to the WHO-HAEM4R3. These tumors, along with their biologically similar primary testicular counterparts, have distinct biology with recurrent NFkB pathway mutations (MYD88, CD79B, CARD11, PIM1/2), as well as CDKN2A/B and HLA deletions⁹. While the ICC now considers primary testicular LBCL (PTLBL) a distinct entity and acknowledges the biologic similarities with PCNSLBL, the WHO-HAEM5 has created the umbrella entity “Primary LBCL of immune privileged sites” including both subtypes and, in addition, vitreoretinal LBCLs. While there is no disagreement between the classifications regarding their close biologic relationship, the ICC opted to defer an umbrella category considering the distinct clinical management of PTLBL and PCNSLBL and the fact that some other extranodal DLBCL (Primary cutaneous DLBCL, leg-type; intravascular LBCL) share a similar genetic underpinning, an issue that still requires additional molecular and clinical studies.

HGBL with double hit (DHL) were first acknowledged in the WHO-HAEM4R as lymphomas with either large cell or high-grade morphology harboring both MYC and either BCL2 or/and BCL6 rearrangements. Since that time, the distinct biology of those with MYC and BCL2 rearrangements has been well established, although their aggressiveness may also depend in part on additional factors (e.g. dark zone/high molecular grade signature; MYC partner). In contrast, the MYC and BCL6 DHL cases are biologically different from the MYC and BCL2-rearranged cases and less well-studied ^10,11. Both classifications have thus opted to specifically recognize tumors with dual MYC and BCL2 rearrangements. The ICC chose to retain the cases with DH MYC and BCL6 translocations as a provisional entity while WHO-HAEM5 incorporates such lymphomas – depending on the cytomorphology – as a genetic subtype of DLBCL, NOS or HGBL, NOS. Some controversies also remain about the terminology (i.e. large cell versus high grade) in DHL. While both classifications consider all aggressive lymphomas with MYC and BCL2 rearrangements as a single entity, the ICC classifies them as HGBL – accepting some variation in the cytologic appearance, while the WHO-HAEM5 acknowledges this variation in the name of the entity (DLBCL/HGBL). These differences in terminology may have the potential to cause confusion which could impact therapy assignment and design of clinical trials (Table 1).

In addition to the well-established HHV8-positive primary effusion lymphoma (PEL), both classifications have introduced a new category to be used for HHV8-negative effusion-based lymphomas without a solid component. The ICC category excludes cases that are positive for EBV, while WHO-HAEM5 accepts variants that may be positive for EBV. More study is clearly needed on these rare lymphomas, with multi-institutional, and ideally international, cohorts and genetic evaluation to determine which cases may represent a distinct disease versus a site-specific variant of DLBCL (Table 1).

Lymphoid proliferations and lymphomas associated with immune deficiency and dysregulation

The WHO-HAEM5 and ICC use different approaches to classify “Lymphoid proliferations and lymphomas associated with immune deficiency and dysregulation” (Table 1, Suppl. Figure 2). The WHO-HAEM5 approaches these processes by algorithmically annotating the diverse morphology (e.g. polymorphic lymphoproliferation), viral association (e.g., KSHV/HHV8, EBV), and basis of the immunodeficiency (e.g., acquired immune deficiency, post-transplant, inborn errors of immunity), respectively. In the ICC, there is an emphasis on the specific clinical condition (post-transplant, HIV, iatrogenic, congenital immune deficiency) since types of LPD/lymphoma and management approaches may differ accordingly. The use of the term “post-transplant lymphoproliferative disorder (PTLD)” as a classificatory term is also a controversial issue between the WHO-HAEM5 and ICC^1,2. More precise clinical and histological criteria in the recognition of polymorphic lymphoproliferative disorders vs frank lymphomas (e.g. monomorphic PTLD in the ICC) is also an area that requires further research (Table 1).

Plasma cell myeloma/multiple myeloma (PCM/MM) molecular classification

The ICC subdivides PCM/MM into distinct mutually exclusive genetic subtypes, based on cytogenetic/FISH studies; (1) multiple myeloma (MM) with CCND family translocations, (2) MM with NSD2 translocations; (3) MM with MAF family translocations and (4) MM with hyperdiploidy. Those without cytogenetic abnormalities are defined as MM, NOS. This genetic subclassification aims to provide opportunities for future studies to assess their clinical relevance. The WHO-HAEM5, in contrast, acknowledges the different genetic patterns but does not recognize separate genetic subtypes as it remains to be determined whether these groups represent distinct clinico-biologic variants. Responses to new drugs such as immunomodulatory agents, proteasome inhibitors, monoclonal antibodies and CAR-T cells based on cytogenetic features or mutational subtypes may help deciphering the importance of genetic subclassification of PCM/MM. Moreover, new classifications based on emerging biological features continue to be developed¹².

B-prolymphocytic leukemia (B-PLL), hairy cell leukemia variant (HCLv) and splenic B-cell lymphoma

This group of rare B-cell neoplasms is approached differently by the WHO-HAEM5 and ICC. In the WHO-HAEM5, the term B-PLL was felt to include a heterogeneous group of entities, including leukemic non-nodal mantle cell lymphoma (already recognized in WHO-HAEM4R), accelerated phase of CLL/SLL, and progressed marginal zone lymphomas, among others. Cases formerly classified as CD5-negative B-PLL are now included in a new broad category named “splenic B-cell lymphoma/leukemia with prominent nucleoli”, which mainly consists of the former hairy cell leukemia variant of the WHO-HAEM4R. In contrast, the ICC continues to recognize B-PLL as a rare but distinct entity, with a different mutational profile from HCLv in the available studies. Both groups acknowledge that classification of these diseases remains challenging and requires further clinical, phenotypical and molecular studies (Table 1).

Other differences

Several entities have different names in the WHO-HAEM5 and ICC. These include LBCL with 11q aberration (ICC) vs. HGBL with 11q aberration (WHO-HAEM5); primary cutaneous marginal zone lymphoproliferative disorder (ICC) vs primary cutaneous marginal zone lymphoma (WHO-HAEM5), nodular lymphocyte predominant B-cell lymphoma (ICC) vs nodular lymphocyte predominant Hodgkin lymphoma (WHO-HAEM5), and multiple myeloma (ICC) vs plasma cell myeloma (WHO-HAEM5). Some of these differences reflect subtle differences in the interpretation of already existing data, and in some situations, such as “LBCL vs. HGBL” with 11q aberration, may have consequences in therapy decisions and the design of clinical trials.

There are minor differences in the diagnostic criteria for lymphoplasmacytic lymphoma (LPL). WHO-HAEM5 requires >10% clonal lymphoplasmacytic cells and subclassifies cases into IGM-LPL/Waldenström macroglobulinemia vs non-IGM-LPL. In contrast, the presence of aggregates of clonal LPL cells, even when representing <10% of the BM cellularity, is sufficient for diagnosing LPL according to ICC, based on the proposal of the Second International Workshop on Waldenström Macroglobulinemia (IWWM-2)¹³. Solving these differences in LPL diagnostic criteria may require large clinicopathological studies with appropriate follow-up, where the clinical relevance of different thresholds may be assessed. Cooperative clinical groups should also clarify, through studies incorporating clinical and biological data, whether monoclonal gammopathy of renal or clinical significance represent separate disease entities (WHO-HAEM5), or just are descriptive terms (ICC).

The classification of peripheral T-cell lymphomas (PTCL) in both ICC and WHO-HAEM5 is very similar. Integrated molecular studies and international clinical trials will be required to determine whether the molecular subclassification of PTCL, NOS into Th1 cell (PTCL-TBX21 subgroup) vs Th2 cells (PTCL-GATA3 subgroup with worse outcome and greater genomic complexity) is clinically relevant and whether the subset of PTCL, NOS with expression of cytotoxic molecules commonly observed in patients with impaired immunity should be recognized. Precise diagnostic criteria for the recognition of T-cell subtypes could help to refine diagnostic criteria for these disorders.

WHO-HAEM5 has also included several preneoplastic or tumor-like disorders. The fact that these have not been included in the ICC does not represent a conceptual discrepancy, but merely a methodological difference in the scope of the two classifications.

Conclusion & perspectives

In this article, lymphoma experts participating in both ICC and WHO-HAEM5 have joined their efforts and highlighted the main issues raised by the two lymphoma classifications published in 2022 ^1,2. While several discrepancies do not reflect major scientific issues, others warrant more attention and additional investigations.

Three of the more controversial issues relate to the identification of different subgroups of follicular lymphoma, high-grade/diffuse large B-cell lymphoma subclasses and terminology and precise diagnostic criteria for lymphoproliferative processes arising in immunodeficiency states (Figure 1). Resolving these differences will require pursuing the in-depth morphologic, immunologic, and molecular characterization of detailed and diverse clinical cohorts coming from retrospective, or ideally prospective, well-annotated patients. This process will also necessitate standardization of techniques and histological and molecular data interpretation⁴ (Table 1). Broader input from the clinical community will be important in resolving these differences since changes will impact clinical trial development and standardized therapeutic practices.

Finally, lymphoma classifications should facilitate the administration of optimal treatment for lymphoma patients. Different approaches could be envisioned, such as the development of appropriate clinical trials for rare entities or, in contrast, broad inclusion criteria but with precise and comprehensive characterization of those cases. This may clarify whether follicular lymphoma grading, recognition of specific follicular lymphoma subtypes, the concept of extranodal LBCL in immune-privileged sites umbrella or the existence of specific molecular alterations in large cell lymphoma and PTCL-NOS may optimally offer the best approach to stratify patients.

Current differences in lymphoma classification reflect controversial areas where gaining further knowledge is mandatory. Addressing these gaps will require the active collaborative participation of researchers and cooperative clinical groups with expertise on lymphoma biology, diagnosis, and treatment.