Introduction to the peripheral T-cell lymphoma review series: advances in molecular characterization, classification refinement and treatment optimization

Peripheral T-cell lymphomas (PTCL), also known as mature T- and NK-cell lymphomas, are a diverse group of diseases with variable clinical presentations, including predominantly nodal or extranodal involvement, as well as rare leukemic forms. In Western populations PTCL represent 10% of all non-Hodgkin lymphomas; however, 20% of all non-Hodgkin lymphomas in Asia have a T/NK-cell phenotype with a higher proportion of Epstein-Barr virus (EBV)- or human T-lymphoma virus 1 (HTV1)-associated entities also seen. In the present review series, four papers cover our current understanding of PTCL disease biology, as well as optimal therapeutics in 2023 based on best available evidence, with a focus on nodal and selected extranodal entities.

Eleven distinct or provisional PTCL entities were described in the Revised European American Lymphoma classification in 1994, and the subsequent World Health Organization (WHO) updates evolved to describe over 30 different PTCL subtypes today. In 2022, two separate classifications were developed. The 5^th edition of the WHO classification (WHO-HAEM5), led by experts designated by the International Agency for Research on Cancer (IARC), was published as a review¹ and a beta version of the blue book is available online pending publication in a definitive form. A separate International Consensus Classification (ICC) was proposed by the Clinical Advisory Committee, a group which had directed the WHO classifications in the past, and is led by international lymphoma experts encompassing clinical, pathology and basic science expertise.² Both classifications are referred to here for completeness, and differences from, as well as comparisons to, the 2017 4^th revised WHO classification (WHO-HAEM4R) are highlighted in Table 1. While the ICC maintains the distinction between provisional and definitive disease entities, all diseases are established along the same rank in the WHO-HAEM5. Several entities introduced in the WHO-HAEM4R as provisional have been confirmed in both classifications.

From the first identification of a T-cell receptor rearrangement to establish monoclonality in T-cell lymphomas,³ genetics has always been integral to the diagnosis of PTCL. However, apart from the fusion of NPM::ALK as a result of t(2;5)(p23;q35) in anaplastic lymphoma kinase (ALK)-positive (ALK⁺) anaplastic large cell lymphoma (ALCL),⁴ and the more recently discovered DUSP22 and TP63 rearrangements in ALK-negative (ALK^–) ALCL,^5,6 very few structural disease-defining aberrations have been described. Comprehensive studies evaluating the molecular profiles and mutational landscapes of PTCL have markedly enriched the characterization and defining criteria of several disease entities and enhanced our understanding of their pathobiology and pathogenesis, translating into significant classification changes. In some cases, genomic analysis has also provided prognostic information and helped to identify key oncogenic drivers which have informed therapeutics. Importantly, the ICC published a companion paper describing the current role as well as the promise of genomics and molecular testing across all lymphomas, including PTCL.⁷

Although there are a daunting number of PTCL entities and the projection is further expansion, being disease ‘splitters’ (vs. ‘lumpers’) has the advantage of highlighting the peculiar biology and natural history of each entity with parallel optimization of treatment approaches. The so-called ‘nodal’ PTCL include PTCL not otherwise specified (NOS), ALCL and follicular helper T-cell lymphoma (TFHL), this last being referred to as nodal T-follicular helper (TFH) cell lymphomas in the WHO-HAEM5. A new entity has also been added to nodal PTCL: ‘(primary) nodal EBV-positive T/NK-cell lymphoma’ and is deemed a provisional entity in the ICC (Table 1). In addition to EBV positivity, tumor cells are positive for cytotoxic markers and limited studies have suggested a very poor outcome, not unlike that of advanced stage extranodal NK/T-cell lymphoma.⁸ Although the term ‘nodal’ is something of a misnomer as extranodal sites of involvement are not uncommon across all subtypes, this terminology is useful to distinguish these subtypes of lymphoma from those arising primarily in extranodal sites or skin as well as leukemic PTCL subtypes.

Table 1.Mature T- and NK-cell neoplasms in the International Consensus Classification (ICC) and World Health Organization (WHO)-HAEM5 classification (2022) in reference to the WHO-HAEM4R classification (2017) (adapted from Alaggio et al.¹ and Campo et al.²).

PTCL-NOS has long been recognized as the ‘wastebasket’ diagnosis since it captures entities that do not meet the criteria for a defined PTCL subtype. As such, PTCL-NOS is a heterogenous group that still has not been fully delineated. Bisig and de Leval⁹ describe the proposed subtypes, GATA3 and TBX21, which were first identified through Affymetrix molecular profiling.¹⁰ This distinction is reproduced using an immunohistochemistry algorithm¹¹ and it is hoped that in the future, a molecular Nanostring assay¹² utilizing RNA derived from formalin-fixed, paraffin-embedded tissue can be applied. The GATA3 subtype with a signature of TH2 cells, including GATA3 and targets as well as components of the PI3K pathway, has an inferior prognosis compared to the TB2X1 subtype which has higher expression of TBX21 (T-bet) and EOMES and their targets (e.g., CCL3, IFNγ).¹⁰ However, the story has emerged as being more complicated, because a subset of TBX21 associated with the presence of DNMT3A mutations and cytotoxic gene expression signature has also an inferior prognosis.¹³ Furthermore, several cases remain ‘unclassified’. Although not yet incorporated in current classifications, correlative studies integrating molecular classification of PTCL-NOS in clinical trials may elucidate whether the GATA3 and TBX21 subtypes have a differential response to novel therapies.

In the WHO-HAEM4R 2017 classification, a new category for TFHL was created to expand beyond the prototype, angioimmunoblastic T-cell lymphoma (AITL), to encompass cases previously classified as PTCL-NOS that have a TFH phenotype, currently defined as the presence of two or more of BCL6, CD10, CXCL13, PD1 and ICOS by immunohistochemistry, but lacking the morphological criteria for AITL. In addition, the TFHL category includes the less common morphological variant, follicular T-cell lymphoma. Minor nomenclature differences can be noted when comparing the ICC and WHO-HAEM5 (Table 1). The genomic landscape of TFHL has been well characterized with recurrent mutations in TET2, DNMT3A, IDH2 and RHOA. Although not mandatory, mutational profiling is now routinely performed at many centers and is helpful for the diagnosis of TFHL,^{7,1 4} especially for borderline cases, and can guide therapeutics.¹⁵

ALCL is broadly divided into ALK⁺ ALCL, identified by the presence of the ALK protein detected by immunohistochemistry, and the morphologically similar ALK^– ALCL. ALK^– ALCL was once proposed to be combined with PTCL-NOS because of its poor outcome,¹⁶ but later studies confirmed a more favorable prognosis.¹⁷ Almost 10 years ago, further genetic heterogeneity was discovered with the identification of a DUSP22 rearrangement in 20-30% cases.¹⁸ As described by Bisig and de Leval, DUSP22-rearranged cases have distinct morphological, immunohistochemical and molecular features, establishing them as a genetic subtype in the ICC, but not in the WHO-HAEM5 classification, due to disparate prognoses in studies which collectively suggest that they are likely more intermediate between ALK⁺ ALCL and ‘triple negative’ ALCL which lack any rearrangement of ALK, DUSP22 or TP63.^19-21 Approximately 2-8% of cases of ALK^– ALCL harbor a TP63 rearrangement which is associated with a dismal outcome.¹⁸ These cases are not yet a distinct genetic subtype as further studies are needed to fully characterize them.

Extranodal entities, as detailed by Lewis, Zhou and Dogan,²² with a clinical lens provided by Stuver, Epstein-Peterson and Horwitz,²³ are identified by their primary location. They constitute a diverse and rare group of diseases with a limited number of dedicated clinical trials, making therapeutic advancement very challenging. Subcutaneous panniculitis-like T-cell lymphoma was first recognized in the 2001 WHO classification which was later refined to exclude cases with a γδ phenotype that are lumped into primary cutaneous γδ T-cell lymphoma because of a similar poor prognosis.²⁴ Approximately 20% are associated with hemophagocytic lymphohistiocytosis, which can be particularly severe in those harboring a germline mutation in HAVCR2.²⁵ Hepatosplenic T-cell lymphoma often occurs in the setting of chronic immunosuppression, especially in patients receiving combination therapy with tumor necrosis-α inhibitor therapy and thiopurines (e.g., azathioprine). Intestinal T/NK cell lymphoma includes the familiar enteropathy-associated T-cell lymphoma (previously referred to as type 1 EATL) linked to celiac disease, which was importantly distinguished from monomorphic epitheliotropic intestinal T-cell lymphoma (MEITL) (previously referred to as ‘type 2’ EATL) in the WHO-HAEM4R classification, each having a distinct genetic landscape, with highly recurrent SETD2 mutations observed in the latter. Intestinal T-cell lymphoma NOS was introduced in the WHO-HAEM4R classification to capture cases that would have previously been included with PTCL-NOS and is retained in the 2022 proposals. Indolent T- and NK-cell lymphoproliferative disorders of the gastrointestinal tract (with the term lymphoma used in the WHO-HAEM5 for T-cell cases) are critical to recognize in order to avoid use of chemotherapy which does not appear to impact the natural history. Historically, treatment paradigms in mature T- and NK-cell neoplasms have been largely modelled on those for aggressive B-cell lymphomas, leading to default use of CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) or CHOPlike chemotherapy,²⁶ which remains the standard comparison arm in clinical trials today. With a still high risk of relapse, etoposide is often added to CHOP chemotherapy (CHOEP) coupled with consolidative autologous stem cell transplant; however, in the absence of a randomized trial, practices vary. With improved understanding of the molecular basis of disease and exciting expansion of novel therapies under investigation in relapsed/refractory PTCL (Figure 1, Table 2), as outlined by Ngu and Savage,¹⁵ more recent studies have integrated these agents in the front-line setting, including chemotherapy-free approaches, with a particular interest in TFHL.

Figure 1.Systemic agents approved, previously studied, or under investigation in peripheral T-cell lymphomas/mature T/NK-lymphomas. *Approved in the USA; brentuximab vedotin (BV) is globally approved in relapsed/refractory (R/R) anaplastic large cell lymphoma (ALCL) and for front-line treatment in CD30⁺ PTCL; BV-CHP (BV, cyclophosphamide, doxorubicin, prednisolone) in Europe approved in ALCL only. ^#Romidepsin was previously approved in the USA/Canada but withdrawn following report of the negative findings of the phase III, RO-CHOP (romidepsin, cyclophosphamide, doxorubicin, vincristine, predisone) versus CHOP study; pralatrexate is not approved in Europe. ^Approved outside the USA, mogamulizumab is approved in Japan only for CCR4⁺ R/R PTCL (and adult T-cell leukemia/lymphoma [positive for human T-lymphotropic virus 1]); chidamide is approved in China only for R/R PTCL. Please also see Table 2 and the reviews by Ngu and Savage¹⁵ and Stuver, Epstein and Horwitz²⁵ for drug approvals and disease rationale. ALK: anaplastic lymphoma kinase; CAR: chimeric antigen receptor; cIAP1: cytoplasmic inhibitor of apoptosis 1; CRBN: cereblon; DHFR: dihydrofolate reductase; EBV: Epstein-Barr virus; EZH: enhancer of zest

Table 2.Selected drugs, their targets and disease rationale under study in peripheral T-cell lymphoma/mature NK/T-cell neoplasms.

CHOP (and CHOEP) has the best track record in ALCL, especially ALK⁺ ALCL. The practice-changing ECHELON-2 study in CD30⁺ PTCL evaluated the addition of the anti-CD30 antibody drug conjugate brentuximab vedotin (BV) to CHP (CHOP with the omission of vincristine because of the overlapping peripheral neuropathy), and demonstrated improved progression-free and overall survival compared to CHOP chemotherapy, leading to a new treatment paradigm with particular relevance in ALCL.²⁷ Uncertainties remaining around the efficacy of CHP-BV in non-ALCL CD30⁺ PTCL subtypes have led to differences in regulatory approval, with use of this regimen in Europe restricted to ALCL alone¹⁵ (Table 2).

With a growing realization that a ‘one size fits all’ approach is suboptimal in PTCL, entity-specific studies have evolved, leading to the complete abandonment of CHOP in some entities, even in the absence of a randomized controlled trial. This is best exemplified in the extranodal subtypes.²³ Extra-nodal NK/T-cell lymphoma is typified by overexpression of P-glycoprotein which imparts a multi-drug resistance phenotype, including resistance to anthracyclines. As outlined by Stuver and colleagues,²³ this led to the exploration of non-CHOP chemotherapy, with clear improvement in outcomes with asparaginase-based regimens as well as cisplatin-based combinations in early-stage disease along with early integration of radiotherapy for disease that could be encompassed in a radiation field. Since extranodal NK/T-cell lymphoma is an EBV-driven disease, programmed death ligand (PDL) and PD1 inhibitors show clear therapeutic promise and are under investigation in a variety of settings (Figure 1, Table 2).²³ Non-CHOP regimens (e.g., ICE - ifosfamide, carboplatin, etoposide) are also preferred in the primary therapy of hepatosplenic T-cell lymphoma followed typically by an allogeneic stem cell transplant. At the other end of the spectrum, immunomodulator therapy, such as cyclosporine, has evolved to be the treatment of choice in subcutaneous panniculitis like T-cell lymphoma with a recent small series also indicating that the JAK1/2 inhibitor ruxolitinib is highly effective in those with refractory hemophagocytic lymphohistiocytosis.^23,28

With an increased understanding of the genetic and molecular basis of PTCL, more rational, biologically driven and subtype-specific research is underway. As highlighted by Ngu and Savage,¹⁵ TFHL represent the ‘poster child’ for more personalized therapies as studies emerge showing sensitivity to epigenetic modifiers, with some long-term remissions being achieved²⁹ (Figure 1, Table 2). This has led to studies of combinations of novel agents, most often integrating epigenetic therapies and even including cohorts of treatment-naïve patients; however, the curative potential of these combinations is still unknown.

Overall, although progress in mature NK/T-cell neoplasms has lagged behind that in B-cell lymphomas, a renewed focus has led to important advances in understanding the molecular basis of these diseases as well as success in shifting the therapeutic bar in many subtypes; despite this, there is still work to be done. Modern clinical trials have evolved in some cases to focus on specific entities or pathways, often integrating rich correlative studies to elucidate markers of response and resistance. Ultimately, for our community of disease splitters, this ‘divide and conquer’ approach will lead to more customized treatment approaches and improved outcomes across the whole spectrum of PTCL.

Footnotes

• Received June 16, 2023
• Accepted August 29, 2023

Correspondence

K.J. Savage ksavage@bccancer.bc.ca

Disclosures

KJS has received honoraria from and provided consulting for BMS, Merck, Seagen, and Janssen; has sat on a steering committee for Beigene; has received research funding from BMS; has received institutional research funding from Roche; and sat on a Data Safety and Monitoring Committee for DSMC. LdeL has no conflicts of interest to disclose.

Contributions

KJS and LdL wrote the manuscript.