From molecular pathogenesis to novel Therapeutic approaches - a review of recent advances in the treatment of peripheral T cell Lymphomas

Abstract

Peripheral T-cell lymphomas (PTCLs) comprise approximately 30 distinct subtypes, collectively accounting for around 10–20% of all non-Hodgkin lymphomas, with a geographic distribution that is more predominant in East Asia and South America. They differ significantly in tumor biology, microenvironmental composition, and chemosensitivity. Standard frontline therapy remains anthracycline-based regimens such as CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisolone). With the exception of ALK+ anaplastic large cell lymphoma (ALCL), such treatments typically yield poor clinical outcomes in most histological subtypes. Oncogenesis in PTCL is driven by a convergence of genetic and epigenetic aberrations, T-cell receptor (TCR) signaling dysregulation and microenvironmental remodeling. Together, they disrupt the balance between cell survival and apoptosis, leading to malignant transformation. Advances in molecular characterization have started to shed light on subtype-specific oncogenic targets, thus allowing for rational development of targeted therapies. A major recent breakthrough was the landmark ECHELON-2 trial, which has led to the approval of brentuximab vedotin (BV), an anti-CD30 antibody-drug conjugate, which in combination with CHP (cyclophosphamide, doxorubicin and prednisolone) has become the standard of care for CD30-positive PTCLs. Beyond brentuximab, novel therapeutic agents targeting various aspects of PTCL lymphomagenesis are under active investigation, with some already entering routine clinical practice. Epigenetic modifiers including HDAC inhibitors, hypomethylating agents, and EZH2 inhibitors exploit the epigenetic dysregulation characteristic of T-follicular helper cell-derived lymphomas. In addition, small molecule inhibitors targeting JAK-STAT, PI3K, and ALK pathways, as well as immunotherapy strategies such as monoclonal antibodies, bispecific antibodies, checkpoint inhibitors, and CAR-T cell therapy, are also in varying stages of clinical development. An area of unmet clinical need is the increasing prevalence of PTCL in older patients, where treatment toxicity is a major consideration, and where such tailored treatment might yield the greatest benefit.

Introduction

PTCLs are a heterogeneous group of approximately 30 distinct post-thymic T-cell malignancies [1, 2], with incidences that range from around 5–10% in the Western world to about 15–20% in Asia and South America [3–5]. Today, PTCLs remain a significant challenge for clinicians. Most patients eligible for intensive chemotherapy are still treated with CHOP (cyclophosphamide, doxorubicin, vincristine and prednisolone), a regimen first described over 30 years ago [6, 7]. With the exception of ALK-positive anaplastic large cell lymphoma (ALCL), which is more chemo-sensitive and has a favorable 5-year progression free survival (PFS) and overall survival (OS) of about 60 and 70% respectively on CHOP-based treatment, the remaining PTCL subtypes generally fare poorly, with 5-year progression free survival (PFS) of 20–30% and overall survival (OS) of 30–40% [3, 8–12]. In the relapsed/refractory setting, outcomes are even more dismal, with median PFS of 2–3 months and median OS of 5–7 months [13–15].

Over the past two decades, the management of B-cell lymphomas has seen major advances, from refined risk stratification incorporating tumor biology and cell-of-origin information [16], to improvements in treatment strategies across both frontline [17, 18] and relapsed/refractory settings. The approval of three anti-CD19 CAR T-cell therapies [19–21], three CD3×CD20 bispecific antibodies [22–24], and numerous small molecule inhibitors [25–28] has further transformed clinical outcomes. In contrast, progress in T-cell lymphomas has lagged considerably, hamstrung by their biological complexity and relative rarity. In the past decade, only four new drugs have received FDA approval for peripheral T cell lymphomas (PTCLs) - brentuximab vedotin [29], pralatrexate [30], belinostat [31], and romidepsin [32], the last of which had its PTCL indication subsequently withdrawn after failing to meet its primary endpoint in a confirmatory Phase III study [33]. This modest expansion in our therapeutic armamentarium has yet to translate into meaningful clinical benefit. The recently reported LEO (Lymphoma Epidemiology of Outcomes)-MER(Molecular Epidemiology Resource) prospective cohort, one of the largest of its kind in the contemporary era, demonstrated that despite evolving first-line regimens, including novel agent combinations, outcomes for non-ALCL subtypes remain extremely poor [34], underscoring this persistent unmet clinical need that motivates this review.

This review aims to provide a concise summary of recent advances in PTCL biology and treatment (Fig. 1), focusing on the more common nodal subtypes: PTCL, not otherwise specified (PTCL-NOS), nodal lymphomas of T follicular helper (TFH) cell origin including angioimmunoblastic T-cell lymphoma (AITL), and anaplastic large cell lymphoma (ALCL). Less common subtypes with distinct therapeutic approaches, including adult T-cell leukemia/lymphoma (ATL), T-cell prolymphocytic leukemia (T-PLL), T-cell large granular lymphocytic leukemia (T-LGL), cutaneous T cell lymphomas, and extranodal natural killer/T cell lymphoma, have been previously reviewed elsewhere [35–39] and are not covered here.

Fig. 1.

The future of PTCL treatment - biology-guided algorithm with novel and targeted therapies

References for this review were identified through searches of PubMed with the search terms “Peripheral T cell lymphoma” and its permutations, “therapy”, “genomic”, “transcriptomic” and “epigenetic” from 2018 to present. Articles were also identified through searches of the authors’ own files. Only papers published in English were reviewed. The final reference list was generated on the basis of originality and relevance to the broad scope of this review.

Nodal PTCL subtypes and molecular pathogenesis

Oncogenesis in PTCL is a complex process. A combination of genetic and epigenetic aberrations, microenvironmental factors, dysregulation of T cell receptor (TCR) signaling and in some cases, chronic infection and inflammation result in abnormal T cell activation, leading to an imbalance between survival and death signals, and subsequent oncogenic transformation. This process can be understood in terms of cell of origin, where subtypes of PTCLs are believed to be neoplastic versions of normal T cells, as well as recurrent mutations postulated to drive lymphomagenesis.

Anaplastic large cell lymphomas (ALCL)

ALCL accounts for about 6–25% of nodal PTCLs [40] and is defined histologically by hallmark cells, which are large pleiomorphic cells with strong CD30 expression [41]. The cell-of-origin is believed to be an activated CD4-positive cytotoxic T cell. ALCLs are subdivided into four distinct entities.

ALK-positive ALCL

Rearrangements involving the ALK gene on chromosome 2p23 produce fusions with partner genes such as NPM1, TPM3 and ATIC in ALK-positive ALCL [42]. These ALK fusion proteins homodimerize and constitutively activate ALK. Key downstream effects include upregulation of JAK-STAT3 signaling [43, 44], HIF1-α target genes, and T-cell cytotoxic genes such as TIA1 and GZMB [45]. NOTCH1 upregulation mediated by ALK-NPM fusion is also characteristic, and a transgenic mouse model (Cd4; NPM-ALK) demonstrated that NOTCH1 expression at the early thymic stage allows TCR-β rearrangement and thymic β-selection to be bypassed [46], suggesting a putative mechanism for lymphomagenesis. In addition, recurrent NOTCH1 T349P and T311P mutations reported in ALK-positive ALCL may also be important for pathogenesis [47].

ALK-negative ALCL

ALK-negative ALCL shares morphological features with ALK-positive ALCL but lacks the former’s ALK gene rearrangements. Genetic driver(s) for its tumorigenesis remains unclear. However, two clinically significant recurrent chromosomal rearrangements have been identified: rearrangements of DUSP22 on chromosome 6p25.3 (30%) and TP63 on chromosome 3q28 (8%) [48]. They appear to be mutually exclusive [48–50]. Those with DUSP22 rearrangements display a distinct phenotype lacking in cytotoxic markers and phosphorylated STAT3 [51]. The most frequent TP63 rearrangement produces a TP63-TBL1XR1 fusion that lacks the transactivation domain, structurally homologous to the dominant-negative oncogenic ΔNp63 isoform that inhibits the p53 tumor suppressor pathway [52]. Other tyrosine kinase gene fusions involving TYK2 or ROS1 lead to constitutive JAK-STAT3 activation, and recurrent mutations in TP53, JAK1, STAT3, PRDM1, KMT2D, KRAS, TET2 and NOTCH1 are also reported [53].

Therapeutic implication

Nodal ALCLs are unified by a clear kinase dependency in their pathogenesis. ALK-positive ALCL is driven by chromosomal fusions that result in constitutive ALK activation, while ALK-negative ALCL, though lacking ALK rearrangements, converges on the same downstream effectors through constitutive JAK-STAT pathway upregulation. Ongoing investigations into targeted therapy for this group centers on inhibition of these kinases, with meaningful clinical activity already demonstrated.

Beyond the two classical nodal subtypes, two less common, and predominantly extranodal ALCL variants also warrant brief mention. Their biology demonstrates the complex interplay between environmental triggers and acquired genetic alterations in malignant transformation of lymphoid tissue and may yet inspire therapeutic strategies applicable to their nodal counterparts.

Primary cutaneous ALCL

This is a special subset of ALCL that is restricted to the skin [41], It bears the morphologic and immunophenotypic features of systemic ALCL. Most cases are ALK-negative, and carry DUSP22, TP63, and TYK26 rearrangements seen in systemic ALK-negative ALCL [52, 54]. Notably, gene expression profiling showed that cases clustered by genetic subtype rather than anatomic site, highlighting an important principle of lymphomagenesis. [51].

Breast-implant associated ALCL (BIA-ALCL)

BIA-ALCL, another uncommon form of ALCL, arises almost exclusively in women around textured breast implants, with a median onset of 9 years post-implantation [55]. Its development in an environment predisposed to hypoxia, chronic inflammation and infection [56] is associated with upregulation of cell motility genes (e.g. MET) and hypoxia-associated genes (e.g. CA9) [57, 58]. It is one of the best examples of lymphomagenesis where genetic aberrations interact with microenvironmental factors. Morphologically, it is indistinguishable from ALK-negative ALCL [59]. However, unlike the latter, DUSP22 and TP63 rearrangements are absent [60]. Instead, epigenetic modifying genes such as KMT2C and CREBBP are frequently mutated, while JAK-STAT3 pathway activation is similarly constitutive [61].

Follicular helper (TFH) T cell lymphoma

Nodal lymphomas of T follicular helper (TFH) cells derivation (n-TFHLs), comprising angioimmunoblastic-type (AITL); follicular type (n-TFHL-F) and not otherwise specified (n-TFHL-NOS), are believed to originate from memory TH subsets. These CD4+ lymphocytes normally interact with B cells to maintain germinal centers. The TFH immunophenotype is characterized by expression of at least two of: PD1, ICOS, CXCL13, CD10 or BCL6 [1]. The three TFHL subtypes share a similar mutational landscape [62–64] and are recognized as a unified group in both the International Consensus Classification [2] and the World Health Organisation-HAEM5 [1], with the former considering them one single disease entity and the latter grouping them into a family of three related conditions.

Angioimmunoblastic T cell lymphoma (AITL)

AITL is the prototypic and most common TFH lymphoma, its histology characterized by an abundant polymorphous microenvironment with B-cell expansion, vascular proliferation, and Reed-Sternberg-like immunoblasts [65]. The neoplastic cells derive from CD4+ CD8– TCRαβ+ cells, and usually harbor mutations causing epigenetic dysregulation, including TET2, IDH2^R172 and DNMT3A, as well as RHOA small GTPase [66], and TCR signaling pathway genes such as PLCG1, CD28, FYN and VAV1 [62, 63]. TET2 mutation in an early hematopoietic progenitor is thought to drive clonal hematopoiesis [67] that eventually leads to malignant T cell transformation [68].

TFH lymphoma, not otherwise specified (n-TFHL-NOS)

n-TFHL-NOS, previously known as nodal PTCL with TFH phenotype, encompasses nodal CD4-positive PTCLs with the TFH immunophenotypic signature but without the morphologic features of AITL. This histological distinction is not well delineated, reflecting what is likely a biological continuum that is also evident in the overlapping mutational profiles of these two subtypes. Like AITL, it frequently harbors mutations in TET2, DNMT3A and RHOA [64], as well as TCR signaling pathway components, with the exception of CD28, which appears specific to AITL [63].

Follicular T cell lymphoma

The least common TFH subtype, n-TFHL-F, displays a follicular growth pattern mimicking follicular lymphoma [69]. It shares the TFH immunophenotype and mutational profile of AITL and n-TFHL-NOS, frequently harboring the TET2, IDH2^R172, DNMT3A, and RHOA^G17V mutations [70], but is additionally characterized by a t(5;9)(q33;q22) ITK-SYK fusion in ~40% of cases [71], resulting in constitutively active SYK tyrosine kinase signaling independent of antigen stimulation, eventually leading to malignant transformation. Recurrent FER and FES mutations [70] activating STAT3 signaling and leading to lymphomagenesis have also been described.

Therapeutic implication

Unlike the kinase-driven ALCLs, oncogenesis in TFH lymphomas appears to be linked to epigenetic dysregulation. These vulnerabilities are increasingly exploited therapeutically. Epigenetic-targeting agents in PTCL trials consistently demonstrate enriched responses in TFH subtypes, providing early data for a biology-guided approach that warrants further investigation.

PTCL, not otherwise specified

PTCL-NOS is a diagnosis of exclusion, now estimated to account for approximately 20% of all PTCL cases following reclassification of erstwhile cases into newly-defined subtypes [3, 9, 72]. It is a morphologically heterogeneous group, but neoplastic cells are typically CD3+CD2+CD5+CD7+, most commonly CD4+CD8−, with frequent aberrant loss of CD5 or CD7 [73]. The majority arise from αβ T cells, with a minority transforming from γδ T cells [74]. Recent gene expression profiling have identified signatures that are distinct from other PTCL subtypes, such as AITL and ALCL [75],

Despite their heterogeneity, two putative subgroups have recently emerged within PTCL-NOS, defined by their expression of the GATA 3 binding protein (GATA3) and T-box-21 (TBX-21) transcription factors, and their respective target genes, suggesting different cells of origin - PTCL-GATA3 from Th2 cells and PTCL-TBX21 from Th1/cytotoxic T cells. PTCL-GATA3 is associated with high MYC expression, greater genomic complexity and poorer prognosis, while PTCL-TBX21 shows NF-κB pathway upregulation and more favorable outcomes [45, 76, 77]. Other common genetic aberrations include deletions/mutations in key tumor suppressor genes such as TP53, PTEN and CDKN2A (more frequent in PTCL-GATA3), TCR signaling pathway mutations in PLCG1, CD28 and VAV1 [78, 79], and occasional JAK2 rearrangements [80]. Mutations in epigenetic regulators such as DNMT3A, typical of n-TFHL, are rarely seen in PTCL-NOS.

Therapeutic implication

As a histologically defined subgroup encompassing PTCLs that do not fit into more specific entities, PTCL-NOS represents a biologically heterogeneous population in which pathogenesis is pathway-driven rather than defined by a single actionable lesion. Nevertheless, molecular stratification is advancing, with emerging data supporting two broad transcriptional subtypes, MYC-driven PTCL-GATA3 and NF-κB-driven PTCL-TBX21. This raises the prospect of biology-guided therapy even within this most heterogeneous of categories.

Impact of tumor microenvironment

It is increasingly recognized that far from being merely the structural support for neoplastic cells, the microenvironment plays an active and dynamic role in cancer biology. The tumor microenvironment in PTCLs remains poorly understood, largely because of the unique challenge posed by T cells. They are key effectors in the microenvironment and exist as a mixture of reactive and neoplastic populations with overlapping features. Recent single-cell studies have nonetheless begun to illuminate the TME’s contribution to PTCL biology [81, 82].

Using integrated genomic and transcriptomic data, Huang and co-workers [81] defined 4 microenvironment subtypes in PTCLs. The “Tfh-like” subtype, characterized by RHOA, TET2, DNMT3A and IDH2 mutations in the malignant cells and enrichment of B cells and plasma cells in the microenvironment, comprised mainly AITL and PTCL-NOS cases. It also showed increased levels of expression of CRBN, TGIT, and CD52. The “inflammatory” subtype, associated with Th1/Th2-derived lymphomas, features pro-inflammatory M1 macrophages and Treg cells, with upregulation of immune checkpoint genes (PD-L1, LAG3, and IDO1) and immunosuppressive cytokines like IFN-α and IFN-β [81]. Other studies also reported increased VEGF-mediated angiogenesis [83]. The “mesenchymal” microenvironmental subtype is enriched with M2 macrophages, stromal cells like cancer-associated fibroblasts and vascular endothelial cells, as well as extracellular matrix, with neoplastic cells exhibiting a Th17 and stromal signature. This group comprises mainly ALK+ ALCL. Finally, the “depleted” subtype, features overall low expression of most microenvironment signatures compared with the other three subtypes. The “depleted” subtype shows low expression of most microenvironment signatures and lacks association with any specific cell of origin or histological subtype. This field remains in its nascency, with existing studies limited by small sample sizes, limited subtype representation, and absence of spatial and clinical data [81, 82, 84–87]. Further work is needed to clarify the role of the microenvironment in PTCL biology.

The therapeutic relevance of spatial profiling of PTCL TME is an area of active interest. Each of the four subtypes described above shows differential expression of potentially targetable proteins, including CD52 and cereblon in the Tfh-like group, and PD-L1, IDO1 and TIM3 in the inflammatory group [80], This raises the prospect of microenvironment-informed therapeutic selection. Proof of concept for this approach has emerged. The recently-concluded Phase II GUIDANCE-01 study in B cell lymphoma demonstrated the feasibility of integrating genomic and TME subtyping to guide rational addition of targeted agents in diffuse large B cell lymphoma [88]. Given the importance of the TME to the biology of PTCL subtypes such as TFHL, a similar framework holds considerable promise as the field matures, though prospective clinical validation remains essential.

Current standard of care

First-line treatment

The standard front-line induction regimen for PTCL remains anthracycline-based chemotherapy, with the first-generation combination, CHOP, comprising cyclophosphamide, doxorubicin, vincristine, and prednisolone [6, 7, 89] remaining the most commonly used [90–92] despite decades of research and advancement in our understanding of PTCL biology. Various studies have reported overall response rates of 60–70% and complete response rates of 30–40%, depending on disease subtype [3, 90, 93]. However, sustained remission is harder to achieve, with 5-year progression-free survival (PFS) typically under 30% and 5-year overall survival (OS) under 50% [3, 90, 93].

In general, adding etoposide (CHOEP) offers modest event-free survival (EFS) benefit in younger patients (3-year EFS; CHOEP 70.5% vs CHOP 51.0%) [94], but no significant OS improvement (3-year OS; CHOEP 81.3% vs CHOP 75.2%). There is even less convincing evidence of clinical benefit in older patients. A recent Dutch cohort study largely corroborated this finding, but subgroup analysis did show an OS benefit of using CHOEP in ALK+ ALCL (5-year OS; CHOEP 90% vs CHOP 61%) [95]. Nevertheless, with the widespread adoption of brentuximab-based regimens in upfront treatment of both ALK+ and ALK- ALCL, this finding has not significantly altered clinical practice. At present, CHOEP is not standard-of-care in PTCL, and its selection over CHOP should be individualized, guided by patient characteristics and preferences.

The major breakthrough in PTCL treatment in the past two decades was the development of brentuximab-vedotin (BV), a novel anti-CD30 antibody-monomethyl auristatin E conjugate with potent and selective antitumor activity [96–98]. The addition of BV to cyclophosphamide, doxorubicin and prednisolone (BV-CHP) has now become standard first-line therapy for CD30-positive PTCLs, in particular ALCL, following the ECHELON-2 Phase 3 randomized control trial demonstrated superior 5-year PFS (51.4% vs. 43.0%) and OS (70.1% vs. 61.0%) over CHOP alone [99]. Of note, the study population was enriched for ALCL by design, therefore the study was not powered to independently assess the benefit of BV-CHP on other subtypes such as AITL and PTCL-NOS.

Relapsed/Refractory disease

Unlike in B cell lymphomas, durable remissions are uncommon in PTCLs, with 5-year PFS under 30% in patients treated with standard CHOP-like regimens [3, 90, 93]. In fit patients, salvage therapy with ifosfamide- [100, 101], platinum- [102, 103], or cytarabine-based [102, 104] regimens are used as a bridge to stem cell transplant. Those unable to tolerate intensive therapy should be considered for newer agents or clinical trials. Some of these novel drugs increasingly being used in the R/R setting are described in greater detail in the next section.

Hematopoietic stem cell transplantation

The role of hematopoietic stem cell transplantation in PTCL is not clearly defined. Recommendations of various national and co-operative groups such as the National Comprehensive Cancer Network (NCCN), American Society for Transplantation and Cellular Therapy (ASTCT) and the European Society for Medical Oncology (ESMO) differ. The most important prospective studies to date are the Nordic Lymphoma Group’s NLG-T-01 study [105], and the more recent Comprehensive Oncology Measures for Peripheral T-cell Lymphoma Treatment (COMPLETE) registry in the United States [106]. Many clinicians use stem cell transplant as a means of consolidating responses either in the upfront setting, or after salvage therapy, with the aim of overcoming the generally poor clinical outcomes when treated only with chemotherapy.

Upfront treatment

Clinical practice guidelines do not recommend the indiscriminate use of autologous stem cell transplantation (ASCT) as upfront consolidative therapy across all PTCL subtypes, even though several studies have reported PFS and OS benefit in specific clinical scenarios based on PTCL subtype, prognostic score and disease stage [105–109]. At present, frontline consolidative ASCT is recommended by NCCN/ASTCT/ESMO only for PTCL-NOS, AITL, and ALK-negative ALCL achieving CR or PR following induction chemotherapy [110–112]. Exploratory subgroup analysis of the ECHELON-2 cohort demonstrated a PFS benefit with ASCT, even in patients who have received a more potent induction regimen such as BV-CHP [113]. In general, allogeneic stem cell transplantation (alloSCT) in the upfront setting is reserved only for the most adverse risk cases, such as hepatosplenic T-cell lymphoma or patients failing to achieve at least a partial response (PR) after induction.

Relapsed/Refractory disease

In the relapsed/refractory (R/R) setting, survival outcomes for patients with certain PTCL subtypes such as AITL and PTCL-NOS have not improved in the past quarter of a century, with 5-year OS rates of only about 10% in patients who had not undergone hematopoietic stem cell transplantation [114]. In contrast to upfront management, in R/R PTCLs, alloSCT remains the only reliably curative option, as confirmed by recent EBMT/CIBMTR registry data, which reported 48–52% 3-year PFS and 60–64% 3-year OS 60–64%, supporting alloSCT for fit patients [115, 116], and even then, benefit varies by subtype, with R/R AITL having particularly favorable outcomes (3-year PFS 57.2%, OS ~60.7%) [117–119]. In contrast, other major nodal PTCL subtypes such as PTCL-NOS and ALK- ALCL fare more poorly, with 3-year PFS rates reported at 46.1% and 49.5% respectively [116]. However, this is not universally applicable. A large Japanese cohort demonstrated that some patients derive little or no benefit from alloSCT, particularly those with primary induction failure or those older than 60 years, whose post-transplant survival was inferior [120]. This is supported by multivariate analysis on the EBMT registry cohort, which also found that age (older), disease status at time of transplant (not being in CR), and haploidentical donor, were other independent factors associated with worse outcomes [116]. ASCT generally have no role in R/R PTCLs, but registry data and retrospective studies suggest it may be considered in chemotherapy-sensitive patients who did not receive it frontline.

Towards novel and future therapeutic approaches

Novel drugs

Given the limitations and suboptimal outcomes of standard-of-care chemotherapy described above [121], there is a urgent unmet clinical need for novel therapeutics in the treatment of PTCL. With advances in molecular characterization of PTCLs, emerging treatments that aim to overcome these historically poor outcomes by targeting newly discovered pathogenic mechanisms and targets, have been developed. These new drugs may target lymphomagenesis via canonical oncogenic pathways, epigenetic modifiers, tumor-specific surface antigens, or microenvironmental features (Fig. 2). A handful of these have been approved for routine clinical use (Table 1), whilst the majority remain under active investigation (Table 2).

Fig. 2.

Molecular landscape of nodal PTCL displaying oncogenic pathways, druggable targets and selected novel and targeted therapeutic agents currently under investigation

Table 1.

Selected novel therapeutic agents that are at various stages of approval for use in nodal PTCL

Drug	Class/Target	Key Trial	ORR/CR	Median DOR	Approval Jurisdiction & Year	Reference
Frontline treatment
Standard-of-care
Antibody-Drug Conjugate
Brentuximab vedotin (BV)	ADC; anti-CD30	ECHELON-2 Phase 3 RCT (N = 452; BV-CHP n = 226 vs CHOP n = 226; CD30+ PTCL)	83%/68% (BV-CHP) vs 72%/68% (CHOP); ITT population	Median PFS 48.2 mo (BV-CHP) vs 20.8 mo (CHOP); 5-yr OS 70.1% vs 61.0%	FDA 2018 (CD30+ PTCL); EMA	[29, 99]
Relapsed/refractory disease
Standard-of-care
Antibody-Drug Conjugate
Brentuximab vedotin (BV)	ADC; anti-CD30	Phase 2 (ALCL, n = 58; PTCL n = 34)	86%/57% (ALCL); 41% (PTCL)	12.6 mo (ALCL)	FDA 2011 (sALCL); expanded CD30+ PTCL FDA 2018	[122]
Anti-metabolite
Pralatrexate	Antifolate; DHFR inhibitor	PROPEL. Phase 2 (n = 111)	29%/11%	10.1 mo	FDA 2009 (R/R PTCL)	[30]
HDAC Inhibitors
Belinostat	Pan-HDACi	BELIEF (Phase 2, n = 129)	26%/11%	13.6 mo	FDA accelerated approval 2014 (R/R PTCL)	[31, 123]
Approved in Specific Jurisdictions
Monoclonal Antibody
Mogamulizumab	Anti-CCR4 mAb	Phase 2 (n = 38 CCR4+ PTCL)	35%/14%	NR	PMDA (Japan) 2014 (CCR4+ R/R PTCL and CTCL)	[124]
HDAC Inhibitor
Tucidinostat (Chidamide)	HDACi (class I selective; oral)	Phase 2 (n = 79)	28%/14%	NR	NMPA (China) 2014 (R/R PTCL)	[125]
JAK Inhibitor
Golidocitinib	Selective JAK1 inhibitor (oral)	JACKPOT8 Part B (Phase 2, n = 88)	44%/24%	20.7 mo	NMPA (China) 2024 (R/R PTCL ≥ 1 prior line); FDA fast-track designation	[126, 127]
EZH1/2 Inhibitor
Valemetostat	Dual EZH1/2 inhibitor (oral)	VALENTINE-PTCL01 (Phase 2, n = 119)	44%/14%	11.9 mo	PMDA (Japan) 2021 (R/R ATL),2024 (R/R PTCL); Phase 2 data in PTCL; regulatory submission pending	[128, 129]
Clinically Active but Not Yet Approved/Compendium-Listed
Frontline treatment
Anti-metabolite
Praletraxate	Antifolate; DHFR inhibitor	CRESCENDO (Phase 3, target n = 504, Praletrexate-CHOP vs Belinostat-CHOP)	Not reported yet	Not reported yet	Ongoing trial for frontline; FDA approved for R/R	[130]
HDAC Inhibitors
Belinostat	Pan-HDACi	CRESCENDO (Phase 3, target n = 504, Praletrexate-CHOP vs Belinostat-CHOP)	Not reported yet	Not reported yet	Ongoing trial for frontline; FDA approved for R/R	[130]
Relapsed/refractory disease
JAK inhibitor
Ruxolitinib	JAK1/2 inhibitor	Phase 2 (N = 52, PTCL n = 45, mycosis fungoides n = 7; n = 21 activating JAK and/or STAT mutations, n = 15 pSTAT expression by iHC, n = 17 neither)	25%/6%(overall) 33%/5% (activating JAK and/or STAT mutations) 29%/14% (pSTAT expression by iHC) 12%/0% (neither)	Not reported	Not yet FDA approved. NCCN recommended second-line therapy in R/R	[131]
Hypomethylating agent
Azacitidine	Nucleoside analog	ORACLE, Phase 3 RCT (N = 86; R/R TFHL)	33.3%/11.9%	Not reported	Not yet FDA approved. NCCN recommended second-line therapy in R/R	[132]
HDAC Inhibitors
Romidepsin	HDACi (class I selective)	Phase 2 (n = 130)	25%/15%	28 mo	FDA 2011 (R/R PTCL); voluntarily withdrawn 2021 after confirmatory Phase 3 failure	[32, 133]
PI3K Inhibitor
Duvelisib	PI3Kδ/γ dual inhibitor (oral)	PRIMO (Phase 2, n = 123)	48%/33%	7.9 mo	EMA orphan designation; Phase 3 TERZO ongoing; not yet approved	[134, 135]

Abbreviations: ADC, antibody-drug conjugate; ALCL, anaplastic large cell lymphoma; ATL, adult T-cell leukemia/lymphoma; BV, brentuximab vedotin; CR, complete response; DHFR, dihydrofolate reductase; DOR, duration of response; EZH, enhancer of zeste homolog; FDA, US Food and Drug Administration; HDACi, histone deacetylase inhibitor; IC, investigator’s choice; JAK, Janus kinase; NMPA, National Medical Products Administration (China); NR, not reached; ORR, overall response rate; PI3K, phosphoinositide 3-kinase; PMDA,Pharmaceuticals and Medical Devices Agency (Japan); PTCL, peripheral T-cell lymphoma; R/R, relapsed/refractory; sALCL, systemic anaplastic large cell lymphoma

Table 2.

Selected clinical trials in nodal PTCL

Trial Name/NCT No.	Phase	Setting	Population/Subtype	Experimental Regimen	Primary Endpoint	Status/Notes	Reference
Frontline Treatment
CRESCENDONCT06072131	3 (randomized)	1 L	All PTCL subtypes	5-arm: Belinostat 600 mg/m2-CHOP vs Belinostat 1000 mg/m2-CHOP vs Praletrexate 20 mg/m2-COP vs Pralextrexate 30 mg/m2-COP vs CHOP	PFS	Actively recruiting	[130]
Alliance A051902 NCT04803201	2 (randomized)	1 L	CD30-negative (<10%) nodal PTCL	3-arm: CHOP(E) vs azacitidine + CHOP(E) vs duvelisib + CHOP(E)	CRR	Suspended as of Feb 2026	[136]
ECHELON-2 Extension NCT01777152	3 (randomized)	1 L	CD30+ PTCL (≥10%)	BV-CHP vs CHOP	PFS (mature OS follow-up)	Ongoing long-term follow-up; BV-CHP now standard for CD30+ disease	[29, 99]
SGN35-032 NCT04569032	2	1 L	CD30-low (<10%) PTCL	BV-CHP	ORR	Completed January 2026, pending final analysis	[137, 138]
Relapsed/Refractory disease
TERZO NCT06522737	3 (randomized)	R/R	Nodal TFH lymphoma (AITL/nodal TFHL)	Duvelisib vs gemcitabine or bendamustine (IC)	PFS	Actively recruiting	[139]
VALENTINE-PTCL01NCT04703192	2	R/R (≥1 prior line)	All PTCL subtypes	Valemetostat	ORR	Completed enrolment; ORR 44%, CR 14% in early analysis	[128]
ORACLE	3 (randomized)	R/R	TFH lymphoma/AITL	Azacitidine vs IC (bendamustine/gemcitabine/romidepsin)	PFS	Completed; Exploratory analyses ongoing	[140]
SoquelitinibNCT06561048	3 (randomized)	R/R	All PTCL subtypes	Soquelitinib (ITK inhibitor) vs IC	ORR/PFS	Registrational Phase 3; FDA fast-track and orphan designation
NCT05225584 (KT-333)	1a/1b	R/R	T-cell lymphomas (incl. nodal PTCL)	KT-333 (STAT3 degrader)	Safety/MTD	Phase 1 dose escalation; early-phase; novel mechanism	[141]
Cellular Therapy
TRANSCRIPTNCT05444712	3 (randomized)	1 L consolidation	PTCL in CR/PR after induction	ASCT vs observation	PFS	Actively recruiting
TRANSFER (JCOG2210)jRCTs031240169	3 (randomized)	1 L consolidation	PTCL in CR/PR after induction	ASCT vs observation	PFS	Actively recruiting	[142]
Multiple CAR-T trials (CD7, CD5, TRBC1-targeting)	1/2	R/R	R/R nodal PTCL (CD7+, CD5+, or TRBC1+)	Autologous/allogeneic CAR-T; fratricide-protected constructs	Safety, ORR	Multiple trials at major centres; early-phase; proof-of-concept responses reported

Abbreviations: AITL, angioimmunoblastic T-cell lymphoma; ALCL, anaplastic large cell lymphoma; ALK, anaplastic lymphoma kinase; ASCT, autologous stem cell transplantation; BV, brentuximab vedotin; CAR-T, chimeric antigen receptor T-cell; CHOP, cyclophosphamide, doxorubicin, vincristine, prednisolone; CHOEP, CHOP + etoposide; CHP, cyclophosphamide, doxorubicin, prednisolone; CR, complete response; EFS, event-free survival; EBMT, European Blood and Marrow Transplantation Society; EZH, enhancer of zeste homolog; IC, investigator’s choice; ORR, overall response rate; PFS, progression-free survival; PTCL, peripheral T-cell lymphoma; R/R, relapsed/refractory; SOC, standard of care; STAT3, signal transducer and activator of transcription 3; TRBC1, T-cell receptor beta chain constant 1

Targeting epigenetics modifiers

Epigenetic dysregulation through histone modifications, DNA methylation, and chromatin remodeling, result in downstream silencing of tumor suppressor genes and aberrant activation of oncogenic pathways. It is increasingly recognized as a major driver of PTCL pathogenesis.

Histone deacetylase inhibitors (HDACi)

HDACi are the most extensively studied class of drugs targeting epigenetic modifiers. HDACs remove acetyl groups from histones, remodeling chromatin structure and limiting transcription factor access to target promoters [143, 144]. HDACi restores acetylation balance, inducing cell cycle arrest, apoptosis, and autophagy.

Amongst these, romidepsin, became the first HDACi approved for use in R/R PTCL in 2011 by the United States Food and Drug Administration (FDA). The accelerated approval was based on results from a phase II study showing 25% overall response rate (ORR), 15% complete response (CR) and a median response duration of 28 months [32, 133]. Notably, subgroup analysis found that 15% of AITL patients remained in CR for more than 3 years [145]. Beyond single-agent activity, romidepsin was also evaluated in combination with other agents such as pralatrexate [146], pembrolizumab [147], azacitidine [148], lenalidomide [149], and various small kinase inhibitors like alisertib [150], duvelisib [151] and tenalisib [152], with varying results. Promising results in R/R PTCL led to the pivotal Phase III study by the Lymphoma Study Association (LYSA) to evaluate the feasibility of adding romidepsin to standard CHOP therapy (Ro-CHOP) in the frontline setting (NCT01796002) [153]. Unfortunately, Ro-CHOP was not found to be more efficacious (median PFS 12.0 vs. 10.2 months), and was also associated with increased Grade 3–4 toxicity. Bristol Myers Squibb subsequently withdrew the PTCL indication in the US and Canada [154].

Newer HDACi includes vorinostat, which was originally approved for R/R CTCLs [155, 156]. A small Phase I single arm study combined it with CHOP in treatment of previously untreated PTCL, but the small cohort size and lack of control arm limit interpretation [157]. In R/R PTCLs, vorinostat combinations with lenalidomide-dexamethasone [158] and alisertib [159] demonstrated only modest efficacy, and no further clinical development appears to be ongoing despite potent in vitro tumoricidal activity [160].

On the other hand, belinostat, a vorinostat analogue, received FDA approval for in 2014 for treatment of R/R PTCL, supported by results of the Phase II BELIEF trial (25.8% ORR, 10.8% CR in 120 evaluable patients) [31]. A later Phase I study of belinostat plus CHOP reported encouraging frontline results [161], and the currently ongoing Phase III CRESCENDO trial is evaluating optimized dosing of belinostat in combination with CHOP for first-line treatment, with results awaited [130].

Meanwhile in China, tucidinostat (chidamide), a selective inhibitor of HDAC1, 2, 3, and 10, was approved by the National Medical Products Administration (NMPA) for R/R PTCL in 2014 following a Phase II trial demonstrating 28% ORR and 14% CR. Consistent with the central role of epigenetic dysregulation in TFH lymphomas, responses were notably higher in AITL patients (50% ORR, 40% CR) [125]. Several tucidinostat-based frontline combinations, such as tucidinostat-CHOP [162], tucidinostat-CHOEP [163], and tucidinostat-prednisone-etoposide-thalidomide [164], have reported encouraging response rates. However, small cohort sizes and high rates of hematological toxicity have limited broader adoption.

Several other novel HDACi, including panobinostat [165], nanatinostat [166], ricolinostat [167] and citarinostat [168], are still being studied and have not yet obtained regulatory approval. Challenges common to all HDACi include bioavailability and resistance [169], the latter likely arising from heterogeneous antigen expression and differential acetylation of gene regulatory elements [170]. Dual kinase-HDAC inhibitors are being explored as a strategy to simultaneously target multiple pathways and overcome resistance [171], though these too remain in early stages of development.

Hypomethylating agents

Aberrant DNA methylation is another important epigenetic driver of tumor progression [172]. Hypomethylating agents reverse these modifications and restore expression of silenced tumor suppressor genes. The first-generation DNA hypomethylating agents, azacitidine and decitabine, were approved by FDA in 2004 and 2006 respectively for the treatment of MDS [173–176], with azacitidine subsequently approved for use for acute myeloid leukemia [177, 178]. Discovery of epigenetic modulator mutations in PTCL pathogenesis has prompted interest in applying these agents to PTCL treatment.

A French retrospective study of single agent 5-azacitidine in R/R PTCL initially reported 53% ORR over all subtypes, rising to 75% in AITL [68]. However, the subsequent ORACLE Phase II trial evaluating the use of 5-azacitidine monotherapy in R/R TFH lymphomas (NCT03593018) [132, 140] failed to meet its primary endpoint, with a median PFS of only 5.6 months versus investigator’s choice therapy. In combination studies, azacitidine plus CHOP in the frontline setting fared better, particularly in TFH subtypes [179], and is now under further evaluation in the randomized Phase II ALLIANCE study A051902 [136]. Another combination, azacitidine with romidepsin, demonstrated 61% ORR, 48% CR, and median PFS of 8 months in R/R disease [148]. More recently, azacitidine combined with anti-PD-1 checkpoint inhibitors, with [180] and without lenalidomide [181], have shown preliminary ORR exceeding 60% without significant added toxicity, suggesting a promising and potentially well-tolerated salvage approach, especially in AITL, which is now under investigation.

EZH2 inhibitors

Enhancer of Zeste Homolog 2 (EZH2) is the catalytic subunit of polycomb repressive complex 2 (PRC2), contributing to cancer pathogenesis through epigenetic silencing of tumor suppressor genes [182, 183]. The discovery of EZH1/2 mutations in T cell lymphomas [184, 185] has ignited interest in EZH2 inhibitors for the treatment of PTCLs. In this category, only valemetostat, a selective dual EZH1/2 inhibitor, is well-studied in PTCL. A Phase I study reported a promising 70.6% ORR in AITL patients, leading to the multicenter, open-label, single-arm, Phase II VALENTINE-PTCL01 trial to evaluate single agent valemetostat in R/R PTCLs [128], which demonstrated 44% ORR (14% CR, 29% PR), median PFS of 5.5 months, and median OS of 17 months. As of 2026, it is approved only in Japan for treatment of R/R Adult T-cell Leukemia/Lymphoma (ATL) [186] and R/R PTCL [129].

Targeting the JAK/STAT pathway

Outside of epigenetic modifiers, a number of targetable oncogenic pathways have also been identified in PTCL. The JAK/STAT signaling pathway is critical to normal T cell function, acting downstream of common gamma receptor–dependent cytokines IL-2, −4, −7, −9, −15, −21 and their receptors [187]. In T cell neoplasms, a range of mechanisms, from activating mutations in the JAK and STAT proteins [53], to loss of negative regulation [188], and constitutive cytokine secretion [189], have been implicated in the pathogenesis of ALCL [53, 190], large granular lymphocytic leukemia (LGL) [191], NK/T lymphoma [192, 193], and ATL [194]. This provides the rationale for using JAK inhibitors for the treatment of patients with T cell lymphomas.

Ruxolitinib was the first JAK inhibitor approved by FDA, initially for the treatment of myelofibrosis, and was later evaluated in a Phase II trial for R/R PTCL, with ORR ranging from 13% in patients without identifiable JAK/STAT pathway activation, to 45–53% in those with pSTAT3 overexpression or activating JAK/STAT mutations [131]. Nevertheless, the small cohort size and reliance on non-routine biomarker assessments precluded FDA approval for treatment of PTCL. A selective JAK1 inhibitor, golidoctinib, was evaluated in the Phase II JACKPOT8 study for treatment of R/R PTCL, which reported 44.3% ORR and 29.5% CRR, with manageable toxicity, leading to approval in China in 2024 [126, 127]. Another novel drug that targets JAK/STAT activation is cerdulatinib, a pan-JAK/SYK inhibitor. It was studied in the Phase II CELTIC-1 trial. Among 58 evaluable R/R PTCL patients, ORR was 36.2% overall but notably higher in AITL/TFH subtypes (51.9%) versus PTCL-NOS (0%), with median response durations of 12.9 and 5.3 months respectively. These results suggest JAK inhibitor activity is subtype-specific and highlights the need for better understanding of resistance mechanisms and response predictors, as well as individualized treatment approaches in PTCLs [195].

Targeting the PI3K/AKT pathway

The phosphatidylinositol 3-kinase (PIK3)/AKT pathway is also central to cell survival and proliferation, and its dysregulation contributes to tumorigenesis and chemoresistance in lymphomas including PTCL [196–198], where increased activated AKT expression is associated with poor prognosis [199]. Idelalisib, a PIK3δ-selective inhibitor, and copanlisib, a PIK3α/δ inhibitor, have already demonstrated the utility of PI3K inhibition in indolent B cell lymphomas [200, 201], providing rationale for their evaluation in PTCL. Tenalisib inhibits both PI3Kδ/γ and salt-inducible kinase 3 (SIK3). Single-agent studies in R/R T-cell lymphoma reported 45.7% ORR [202], and a Phase I/II study by the same group combining tenalisib with romidepsin demonstrated 63% ORR (PTCL 75%, CTCL 53.3%) with a favorable safety profile [203]. Larger prospective studies are needed to corroborate these findings. Of the PI3K inhibitors, duvelisib, an oral dual PI3Kδ/γ inhibitor, has had the greatest success. A Phase I study in PTCL patients reported 50% ORR [204]. This was followed by a Phase II trial in heavily pretreated patients showing 49% ORR (34% CR) and median response duration of 7.7 months, with AITL patients performing particularly well (67% ORR) [205]. Currently, only duvelisib monotherapy is listed in the NCCN Drugs & Biologics Compendium as a second-line agent, and duvelisib-based combinations remain investigational [206].

Targeting anaplastic lymphoma kinase (ALK)

ALK is a receptor tyrosine kinase first identified as the oncogenic NPM-ALK fusion protein in ALCL [207]. It is expressed in 60–80% of ALCLs [208], making it an attractive therapeutic target, with limited off-target toxicity expected, given its restricted expression in adult tissue [207, 209, 210].

A number of ALK tyrosine kinase inhibitors (TKIs), most of which small molecular inhibitors, have been evaluated for the treatment of various types of cancers. Only five of these, crizotinib [211–213], ceritinib [214], alectinib [215], brigatinib, and lorlatinib, are FDA-approved for use in ALK-positive ALCL in pediatric and young adult populations. At present, ALK inhibitors are not routinely used in adult lymphomas, as most of the data supporting their use have been reported in children and young adults [216], and studies involving adult subjects have largely been limited to small Phase I trials [217]. It is important to note, however, that resistance mutations have been described with ALK inhibitors and may limit long-term efficacy [218].

Targeting farnesyltransferases

Farnesyltransferase is an enzyme that catalyzes the addition of a C₁₅-farnesyl lipid group to the carboxyl terminal of several intracellular proteins. This enables Ras membrane association and functional maturation [219]. Its inhibition has been explored as a means of suppressing Ras-dependent oncogenesis [220]. A number of farnesyltransferases have been developed, including antroquinonol, tipifarnib, and lonafarnib. Among them, tipifarnib is the most widely investigated and has entered phase III clinical trials on cancers [221], and is the only one evaluated in T-cell lymphomas, with a Phase II study in heavily pretreated AITL/PTCL-NOS reporting 45% ORR in AITL patients [222].

Immunotherapy

Other than directly targeting oncogenic pathways, immunotherapeutics represent another novel frontier in PTCL treatment. Immunotherapy broadly refers to treatments that harness the patient’s own immune system to fight cancer. Named Science magazine’s Breakthrough of the Year in 2013 [223], its origins trace back to late nineteenth century, when American surgeon William Coley used heat-killed bacterial mixtures as immune stimulants for inoperable sarcomas [224], an approach that fell out of favor with the advent of radio- and chemotherapy. Interest was revived when James Allison at University of Texas MD Anderson Cancer Centre proposed immune evasion as a key cancer mechanism in 1996 [225]. Since then, immunotherapy has grown to encompass a broad range of therapeutics ranging from monoclonal antibodies to cellular therapy, most of which remain in their nascency in T cell lymphomas.

Monoclonal antibodies (mAbs) and antibody-drug conjugated (ADCs)

Rituximab, an anti-CD20 monoclonal antibody, became the first monoclonal antibody approved by the FDA for treatment of cancer in 1997 [226], and is now standard-of-care in most B-cell lymphomas. Similar efforts to develop antibody-based therapeutics against T cell targets have only seen moderate success. The only antibody-based therapy approved for frontline nodal PTCL is brentuximab vedotin, an anti-CD30 ADC, used in combination with chemotherapy for ALCL which make up only about 15% of PTCLs [2, 48]. Other agents remain in various stages of development and evaluation.

Mogamulizumab-kpkc is an anti-CCR4 IgG1 antibody approved in the US for R/R mycosis fungoides (MF) and Sezary Syndrome (SS) [227], and in Japan additionally for adult T cell leukemia/lymphoma (ATL) [228–230]. However, as CCR4 expression is heterogeneous in nodal PTCLs [231, 232], mogamulizumab currently has a very limited role in treatment of nodal PTCLs.

Similarly, another mAb used in non-nodal PTCL, alemtuzumab (Campath), targets CD52 and is the treatment of choice for T-prolymphocytic leukemia (T-PLL) in both the upfront and relapsed setting [206, 233, 234], but is not currently used routinely for nodal PTCL. Its addition to CHOP yielded 75% ORR, but had significant hematological toxicity [235]. Another phase II trial combining alemtuzumab with etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (EPOCH) in T and NK cell lymphomas also found the alemtuzumab-EPOCH combination to have an unacceptable toxicity profile [236].

Daratumumab, a first-in-class anti-CD38 mAb, is at present only FDA-approved for treatment of plasma cell dyscrasias, but the increased expression of CD38 on neoplastic NK and T cells [237, 238] has raised interest in extending its use to NK and T cell lymphomas. A Phase II trial in R/R NK/T lymphoma showed early promise [239], but a subsequent study combining it with gemcitabine, cisplatin, and dexamethasone in R/R nodal PTCLs reported unacceptable toxicity and had to be terminated early (NCT04251065).

Basiliximab is a ⁹⁰Yttrium–labelled chimeric human-mouse mAb that targets CD25, the alpha subunit of interleukin-2 receptor, which is widely expressed on mature T cells [240] and is reported to be upregulated in certain nodal PTCL subtypes [241]. It was evaluated as an addition to BEAM conditioning for ASCT in a Phase I study with no dose-limiting toxicities observed but has not been developed further [242]. At present, there has been no further development in this therapeutic approach.

Additional targets under early investigation include CD70, CD37, CD94, and KIR3DL2, underscoring the expanding role of mAbs and ADCs in the treatment of T-cell lymphoid malignancies.

Bispecific antibodies

Bispecific antibodies (BsAbs) are engineered, monoclonal, dual-specificity antibodies that simultaneously engage tumor antigens and cytotoxic immune cells [243]. About a dozen BsAbs have now entered regular clinical practice in the treatment of a range of conditions [244, 245], but currently, there is none approved by the FDA for treating T-cell lymphomas. A key challenge is the antigenic overlap between normal and neoplastic T cells, leading to the risk of fratricide [246].

One of the best studied BsAb in PTCL is acimtamig (AFM13), a tetravalent CD30×CD16A antibody that recruits NK cells to CD30-expressing T cell lymphoma cells. Single-agent studies showed modest activity in R/R PTCL [247, 248], failing to meet the primary endpoint in the Phase II REDIRECT study [249]. A novel approach pre-complexing acimtamig with cord blood-derived NK cells showed promising Phase II results (88% ORR in 24 evaluable patients) [250], but was terminated by the sponsor in June 2025 [249, 251, 252]. Several other BsAbs are under early investigation, including GNC-038 (tetra specific; CD19/CD3/PD-L1/4-1BB) (NCT05627856) and ONO-4685 (PD-1×CD3 (NCT05079282) [253], with a number of Phase I studies recently completed or currently recruiting.

Chimeric antigen receptor T (CAR-T) cells

The latest addition to the immunotherapy armamentarium are CAR-T cells, which are generated by incorporating antibody-derived variable regions (Fv) directed against tumor antigens into the T-cell receptor, forming a chimeric antigen receptor (CAR). This construct is then introduced into the patient’s own T cells, which are expanded ex vivo and re-infused to mount a targeted anti-tumor response. There are now several FDA-approved CAR-T products for treatment of B-cell and plasma-cell cancers [254, 255], but none currently licensed for T-cell lymphomas. Key obstacles include fratricide from shared antigen expression on normal and malignant T cells, T-cell aplasia, and product contamination during apheresis. Various strategies including CRISPR-based gene editing, protein expression blocking, and use of NK cells as CAR vehicles are being explored to overcome these challenges across multiple antigenic targets.

Due to its increased expression in T-cell acute lymphoblastic leukemia and a number of T cell lymphomas [256, 257], CD7, a member of the immunoglobulin superfamily, is the most extensively targeted antigen in T-cell lymphoma CAR T studies. Strategies to prevent fratricide include selecting naturally CD7-negative T cells [258–260], CRISPR-mediated CD7 deletion [261], and intracellular retention signals [262–264]. Multiple Phase I trials (NCT05377827, NCT05043571, NCT05620680, NCT03690011, NCT05979792) across Asia, Europe, and the United States have reported encouraging early results in R/R T-ALL/LBL, with several ongoing studies expected to generate further data.

Another target being evaluated is CD5, which is ubiquitously expressed on T lymphocytes [265], and is upregulated in a number of T cell lymphomas, including PTCL-NOS, TFH lymphomas, ATL, and more than half of ALCLs [266]. The ongoing MAGENTA Phase I trial uses an unmodified anti-CD5 CAR-T product with a CD28 costimulatory endodomain, with preliminary data reported short-lived responses [267]. Another study evaluating CRISPR-edited CD5-deleted product with a cetuximab-based safety switch has been tested in 3 patients with R/R PTCL [268]. Additional trials involving CD5-deleted anti-CD5 CAR T products, including the Phase I VIPER101 trial (NCT06420089) are ongoing.

T-cell receptor beta-chain constant domain 1 (TRBC1) represents another tumor-specific target, as malignant T cells clonally express either TRBC1 or TRBC2 [269]. A Phase I/II study evaluating anti-TRBC1 CAR T cell in R/R T-cell lymphomas reported 67% ORR and 40% CR. Although the study investigators found CAR persistence in peripheral blood to be limited, they were able to demonstrate persistent localization of the CAR-T cells in tumor sites. This is still the subject of ongoing investigation [270].

CD30, a transmembrane member of the tumor necrosis factor (TNF) superfamily, is characteristically expressed in Hodgkin lymphoma, but also in ALCLs, several forms of cutaneous T cell lymphomas, and less consistently in other PTCL subtypes [271]. CD30 CAR-T products have been developed primarily for R/R Hodgkin lymphoma, with several trials also including CD30-positive PTCLs [272–276]. Results across these studies varied considerably in efficacy and toxicity, and were generally limited by modest response durations, small cohort sizes, and heterogeneous patient populations. At present, a number of Phase I anti-CD30 CAR-T trials are open for recruitment for patients with R/R cHL and CD30+ PTCLs (NCT02917083, NCT04288726, NCT04083495).

Another TNF family transmembrane protein, CD70, is absent on resting lymphocytes but upregulated upon activation [277], and highly expressed in both B- [278] and T-cell [279] lymphomas. A Phase I dose-escalation study of an allogeneic anti-CD70 CAR T in 39 evaluable R/R T-cell lymphoma patients reported 46.2% ORR overall, rising to 51.6% at higher dose levels (19.4% CR, 32.3% PR) [280]. Preclinical work using CD70-directed cord blood-derived CAR-NK cells has also demonstrated potent cytotoxic activity against T-cell lymphoma cell lines in vitro [281].

Other than these, investigational CAR-T products against CD4 [282, 283], CD3 [284, 285], CCR4 [286], and CD37 [287, 288] have also shown preclinical promise, but clinical development has been limited by safety concerns including CD4 T-cell aplasia and lack of tumor specificity. Indeed, clinical testing of LB1901, an investigational anti-CD4 CAR T, in a phase I study resulted in CD4 T-cell aplasia in a patient, resulting in clinical hold on the trial [289, 290].

Beyond direct antigen targeting, certain receptors can also be co-expressed on CAR T cells to exploit their physiological properties. In a Phase I trial of R/R CD30-positive cHL and CTCL, co-expression of CCR4 on anti-CD30 CAR T cells harnessed CCR4‘s native lymphocyte-homing function to enhance CAR T persistence and anti-tumour activity at sites of disease [291].

Immune checkpoint inhibitors (ICI)

One of the fundamental hallmarks of cancer is immune evasion [292]. Tumor cells evade immunity by upregulating checkpoint ligands, leading to T-cell exhaustion [293]. Another aspect of immunotherapy is the restoration of anti-tumor immunity by reversing T-cell anergy using immune checkpoint inhibitors.

One of the best studied immune checkpoints is the programmed cell death protein-1 (PD-1), an inhibitory receptor expressed on activated B, T, and NK cells. Binding of PD-1 with its ligands PD-L1/2 suppresses T-cell-mediated cytotoxicity against PD-L1/2-expressing cells [294]. Targeting PD-1 or PD-L1 disrupts this interaction and restores T-cell-mediated destruction of neoplastic cells. Anti-PD-1 monotherapy has shown modest activity in most PTCL subtypes. The largest prospective study to date is a Chinese Phase II trial of geptanolimab in 89 R/R PTCL patients, which reported 40.4% ORR (14.6% CR). Subgroup analysis found superior responses in tumors with high PD-L1 expression levels such as ENKTL, while others such as PTCL-NOS only had modest responses [295]. Phase II studies of pembrolizumab [296] and nivolumab [14] similarly demonstrated limited single-agent activity against R/R T-cell lymphomas, with the pembrolizumab study halted early at interim analysis [296]. Combination of romidepsin with pembrolizumab improved outcomes, with 39.5% ORR, 34.2% CRR, and median PFS of 29.5 months [297]. A notable safety concern across PD-1 inhibitor studies is hyperprogression, as malignant T cells may themselves express PD-L1 and expand in response to treatment, with 4 cases reported in the nivolumab study alone [14]. PD-L1 inhibitors are even less established. The AVAIL-T trial found that avelumab monotherapy failed to demonstrate meaningful activity in R/R PTCL [298]. A more recent Phase I/II trial of durvalumab with or without lenalidomide reported promising interim CTCL data (best overall and skin response rates of 58.3 and 75%), with final results that include the PTCL group pending [299].

Another frequently targeted immune checkpoint is the CD47/SIRPα pair. CD47 is expressed on the surface of most cells and functions as a “don’t-eat-me” signal [300]. Its overexpression on tumor cells enables immune evasion [301–304]. Clinical development of the anti-CD47 antibody magrolimab suffered a major setback in 2024 when Gilead halted further development after the Phase III ENHANCE study failed to show survival benefit and demonstrated higher mortality rates [305–307]. Phase II trials in aggressive NHL (PRISM Study; NCT03527147) and T-cell lymphomas are presumed terminated (NCT04541017). Nevertheless, several other agents targeting this pathway remain under investigation, including alternative anti-CD47 antibodies (NCT04541017, NCT04728334), bispecific antibodies [308], SIRPα/Fc fusion protein antibodies [309] (CTR20212227, NCT03530683, NCT02890368) and small-molecule inhibitors (NCT01359982).

Additionally, cytotoxic T lymphocyte-associated protein 4 (CTLA-4), constitutively expressed in Tregs and pathologically upregulated on some malignant cells, competes with CD28 to bind CD80/CD86, thus preventing the co-stimulatory signal for T-cell activation. Unlike in solid tumors, data on CTLA-4 inhibition in PTCL is limited, and is an area of ongoing investigation. In the CheckMate 039 study, addition of anti-CTLA-4 antibody, ipilimumab, to nivolumab did not significantly improve outcomes over nivolumab monotherapy in the PTCL subgroup [310, 311].

Personalized therapy approaches

Biomarkers

In addition to novel drugs and new targets, the way with which we approach cancer treatment is also rapidly evolving, with an increasing emphasis on tailored therapy for the individual patient. PTCL’s genetic and phenotypic heterogeneity makes an individualized treatment approach more attractive than the traditional “one-size-fits-all” strategy. In a number of other hematological malignancies, imaging biomarkers such as tumor volume on FDG-PET [312, 313], genomic [314, 315] and circulating tumor DNA [316], immunohistochemical [317–319] and gene expression-defined subtypes [320] are increasingly used to inform clinical decision-making. However, similar approaches in PTCL remain in their nascency. Therefore, the identification of relevant biomarkers in PTCLs, and their incorporation into personalized therapeutic strategies to guide appropriate escalation or de-escalation of treatment is a major unmet need.

Prognostic

Prognostic markers are measurable clinical or molecular characteristics that reflect a disease’s natural history, independent of the treatment received [321]. Although the International Prognostic Index (IPI), first proposed in 1993, remains widely used for risk stratification in NHL [322], it lacks sufficient granularity, particularly in T-cell lymphomas. There is, therefore, a need for more disease-specific prognostic markers.

ALCL has the most validated prognostic biomarkers among PTCL subtypes. Positivity for ALK gene rearrangements (ALK-positive ALCL) confers a favorable prognosis (5-year OS 70%) [323, 324], such that consolidative ASCT is no longer recommended in first CR, whereas ALK-negative ALCL carries a worse prognosis (5-year OS 50%) [48] and ASCT remains standard-of-care in eligible patients [110]. Within ALK-negative ALCL, TP63 rearrangements are associated with particularly poor survival (5-year OS 17%) [48], while the prognostic significance of DUSP22 rearrangements is controversial, with an early study initially reporting a more favorable prognosis [48] but more recent studies disputing this [325, 326]. Current guidelines do not suggest using TP63 nor DUSP22 status to guide treatment decisions, however, some clinicians may choose to omit ASCT in patients with DUSP22 rearrangements [327].

Gene expression profiling has identified two molecular subgroups in PTCL-NOS; GATA3- and TBX21-overexpressing. The former carrying an inferior prognosis (5-year OS 19% vs. 38%) [45, 76, 77]. In AITL, a 34-gene signature has defined four prognostic groups, with the B-cell signature group associated with better outcomes [45].

In diagnostic imaging, baseline total metabolic tumor volume and total lesion glycolysis on PET imaging have also been identified as independent predictors of PFS and OS [328].

Predictive

Predictive markers, on the other hand, are those that help to forecast the likelihood of benefit from a specific treatment [321]. One of the most successful implementation of a predictive biomarker in PTCL to date is CD30 expression, which informed the ECHELON-2 trial and led to brentuximab vedotin plus CHP replacing CHOP as standard-of-care in treatment-naive CD30-positive PTCL (5-year PFS 51.4% vs. 43%; 5-year OS 70.1% vs. 61%) [99, 206]. Beyond CD30, the clinical use of predictive biomarkers in PTCL is still in its infancy. Advances in next-generation sequencing have improved circulating tumor DNA detection, and MRD-guided therapy is gaining traction in other hematological malignancies [316, 329–331], where negative MRD status may allow for rational de-escalation of therapy. Its application in PTCL is currently under investigation in both the frontline [332, 333] and relapsed settings [334]. Early studies have found circulating tumor DNA (ct DNA) to be a sensitive and minimally-invasive biomarker for early response assessment that correlated well to traditional response indices such as PFS. A number of trials are now formally evaluating its use in clinical practice (NCT06362148, NCT06089941, NCT06550427).

Ex vivo drug sensitivity testing

The idea of molecularly guided therapy has its roots in the rapidly developing DNA sequencing technology of the early 2000s, which allowed identification of an increasing number of potentially targetable mutations in cancers. However, early studies using genomics-based precision medicine in solid tumors were not successful, highlighting the complex gap between genomic profile and clinical phenotype. [335].

As such, increasingly innovative methods were developed to predict individual response to treatment. One such approach is ex vivo drug sensitivity testing, in which patient-derived neoplastic cells are cultured and treated with various therapeutic compounds in vitro either as monotherapy or in combination to assess response. Initially applied in myeloid malignancies a [88, 336–340], its use has since extended to multiple myeloma [341] and non-Hodgkin lymphomas [342, 343].

A key limitation is turnaround time. For the technique to be implementable in routine practice, the results must be available in a clinically relevant time frame. In addition, in vitro single-agent results may not accurately reflect responses to combination regimens. To address this, computational algorithms have been developed to improve the speed and accuracy of combinatorial drug sensitivity predictions [344–347]. Using one such platform, the Quadratic Phenotypic Optimization Platform (QPOP), a Singaporean group successfully guided treatment selection in a patient with refractory PTCL, achieving complete remission and enabling consolidative ASCT [348]. A multicenter Phase II trial of ex vivo-guided combinatorial drug selection in R/R T-cell lymphomas is currently recruiting (ACTRN12626000088347).

Discussion and future perspectives

The case for Personalized medicine

Nodal PTCLs are a heterogeneous group of diseases that, despite sharing a T cell origin, differ markedly in tumor biology, microenvironmental signatures, chemosensitivity and clinical outcomes. This biological diversity highlight the inadequacy of the traditional “one-size-fits-all” treatment approach, with standard CHOP-based upfront treatment yielding modest responses rates across most subtypes [3, 8–12]. In addition, frontline consolidative stem cell transplantation, currently recommended in many cases [110, 349], carry with it substantial morbidity and mortality risks. An ageing patient population further demands greater personalization of treatment strategies, calibrated to individual fitness and tolerability. In this population, novel targeted agents with potentially more favorable off-target toxicity profiles, may extend the reach of effective treatment to patients previously considered unfit for intensive or curative-intent therapy.

The future of PTCL treatment

Growing insights into PTCL biology have driven advancement in its therapeutic pipeline, with promising early-phase results in novel agents ranging from kinase inhibitors, epigenetic modifiers to cellular therapies. Ex vivo sensitivity testing and computational approaches could further refine rational drug selection. Conventionally, novel therapeutic approaches are first evaluated in the R/R setting, where the risk-benefit ratio more readily justifies experimental treatment. In PTCL, however, the efficacy of frontline CHOP-based chemotherapy fundamentally challenges this this paradigm and makes a compelling case for earlier integration of targeted and biology-driven approaches into first-line treatment.

Translating these advances into clinical practice, however, requires robust trial evidence, and trial design in PTCL is inherently challenging. The combination of biological diversity across subtypes and the low incidence of each individual entity makes adequate patient accrual one of the greatest hurdles in the field. Two complementary solutions present themselves. Greater collaboration across regional and international consortia, and a fundamental re-evaluation of trial methodology. In many respects, these diseases are uniquely suited to adaptive and precision medicine-informed designs that offer more flexible frameworks than conventional randomized approaches [350–352]. Basket trials enroll patients across subtypes sharing a common molecular target irrespective of histological diagnosis; umbrella trials evaluate multiple targeted agents within a single disease entity stratified by biomarker profile; and n-of-1 designs accommodate the smallest of patient cohorts. Embracing such methodological innovation alongside renewed international collaborative commitment may prove decisive in advancing the field.

The prospective validation of biomarkers capable of guiding clinical decision making at each disease stage represents an equally important frontier. For instance, CD30 expression already guides brentuximab vedotin use in both frontline and relapsed/refractory settings [99, 122]. ALK gene translocations in ALCL, associated with a more favorable prognosis [353, 354], informs upfront autologous transplant decisions in ALCL.

In addition, the maturation of liquid biopsy technology [355, 356] has led to mounting evidence for the use of circulating tumor DNA or RNA as clinically actionable measurable residual disease (MRD) markers. For example, in favorable-risk acute myeloid leukemia [329], end-of-induction MRD negativity permits omission of consolidative transplantation; in chronic myeloid leukemia [357–360], sustained deep molecular remission underpins treatment-free remission strategies; and in Philadelphia-positive acute lymphoblastic leukemia (B-ALL), MRD positivity after induction prompts therapeutic escalation with blinatumomab [361]. Validated MRD markers in PTCL [332, 334] could enable a similarly adaptive approach, permitting evidence-based de-escalation in responding patients and early salvage intervention in those at high risk of relapse, with particular relevance in older patients where treatment intensity carries the greatest consequence.

Conclusion

Looking ahead, given their inherent biological heterogeneity and limited efficacy of standard treatment, the convergence of molecular and microenvironmental profiling, ex vivo drug sensitivity testing and dynamic MRD monitoring holds the potential to guide genuinely individualised treatment. Meanwhile, our therapeutic armamentarium continues to expand, with novel agents targeting specific pathways advancing through clinical evaluation. Some, such as valemetostat, have already secured regulatory approval in select jurisdictions. T lymphoid malignancy-directed cellular therapeutics, once constrained by problems of fratricide and T cell aplasia, are also yielding encouraging early safety and efficacy data.

In the longer term, the rational integration of these expanding therapeutic options with individualised biological profiling and evidence-based MRD-guided escalation/de-escalation of treatment holds the potential to fundamentally transform outcomes. Whether this promise is fully realised, the scientific foundations for a more biomarker-driven therapeutic paradigm in PTCL are firmly in place.

Ethics

No ethical approval was required for this study as it did not involve the collection of new primary data from human or animal subjects.

Author contributions

WLC wrote the main manuscript text and prepared the figures. WLC, WJC and QL conceived the review. WLC, WJC and QL reviewed the manuscript.

Funding

The authors have not received any external funding in relation to this work.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.