Immune checkpoint therapy in pediatric and adolescent lymphomas

Abstract

Immune checkpoint therapy (ICT) is designed to unleash the anti-tumor activity of T-lymphocytes. Cytotoxic T-lymphocyte- associated antigen 4 (CTLA-4) inhibition and programmed death 1 (PD-1) inhibition are the most commonly utilized ICT in clinical cancer therapy, and they enhance anti-tumor immunity by interrupting the inhibitory signals CTLA-4 and programmed death ligand 1 (PD-L1), respectively. In pediatric Hodgkin lymphoma, ICT has demonstrated remarkable efficacy in both high-risk and relapsed disease, with investigation into the efficacy in low-risk disease ongoing. Pediatric mature B-cell lymphomas have variable expression of PD-L1 and there is very limited experience of incorporating ICT in their treatment. Primary mediastinal B-cell lymphoma (PMBCL), anaplastic large cell lymphoma, aggressive natural killer-cell lymphoma, and peripheral T-cell lymphoma, not otherwise specified all consistently express PD-L1, which provides a strong biological rationale for the use of ICT in these diseases. In PMBCL, the Children’s Oncology Group and the National Cancer Institute’s National Clinical Trials Network recently completed a randomized phase III trial of nivolumab in combination with chemo-immunotherapy in children and adults with newly diagnosed PMBCL. Results of this trial are expected in 2027. In anaplastic large cell lymphoma and aggressive natural killer-cell lymphoma, ongoing clinical trials are evaluating the efficacy of ICT. Given the transformational role of ICT in pediatric Hodgkin lymphoma, there is significant promise for the use of ICT in multiple subtypes of pediatric non-Hodgkin lymphoma with increased expression of PD-L1.

Introduction

Immune checkpoint therapy (ICT) is designed to unleash the anti-tumor activity of T-lymphocytes which are “the established lynchpins of anti-tumor immunity”.¹ The United States Food and Drug Administration’s (FDA) approval of the cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) inhibitor ipilimumab in 2011 and the programmed death 1 (PD-1) inhibitors nivolumab and pembrolizumab in 2014 has ushered in a new era of cancer immunotherapy for numerous different forms of cancer.² Mechanistically, CTLA-4 inhibitors increase T-cell expansion by interrupting the inhibitory interaction of CTLA-4 on T cells with B7 on antigen-presenting cells.³ Alternatively, PD-1 inhibitors prevent T-cell exhaustion by interrupting the inhibitory interaction of programmed death ligands 1 and 2 (PD-L1 and PD-L2) on tumor cells with PD-1 receptors on T cells.³ Through both mechanisms, anti-tumor immunity is enhanced (Figure 1). More recently, modulation of additional immune regulatory pathways including those mediated by lymphocyte activation gene-3 (LAG-3), T-cell immunoglobulin and mucin-domain containing-3 (TIM-3), and T-cell immunoglobulin and immunoreceptor-tyrosine-based-inhibitory-motif domain (TIGIT) are being explored in order to build on the success of CTLA-4 and PD-1 inhibitors.¹ Additionally, combination therapy with ICT and various other forms of anti-cancer therapy (cytotoxic chemotherapy, small molecule inhibitors, ionizing radiation, antibody-drug conjugates, cancer vaccines, etc.) are either already approved by the FDA or are being actively investigated (Table 1).⁴ Collectively, these agents represent an established and expanding cancer immunotherapy arsenal with significant therapeutic potential. Unique to ICT agents is the tumor agnostic FDA approval for use when various biomarkers of ICT response are present. PD-L1 tumor positivity by immunohistochemistry (e.g., 22C3 PD-L1 antibody clone for pembrolizumab, 28-8 PD-L1 antibody clone for nivolumab, and SP142 PD-L1 antibody clone for atezolizumab) with either the established tumor proportion score (percentage of tumor cells that are PD-L1-positive) or combined positive score (percentage of tumor and immune cells that are PD-L1-positive) are FDA-approved companion diagnostic assays for PD-1/PD-L1 inhibitors.⁵ Additionally, as an increased quantity of tumor neoantigens has been recognized as a positive predictor of response to ICT, high tumor mutation burden (>10 mutations per Mbp) and high microsatellite instability (often defined as instability of >2 of 5 microsatellite repeats determined by polymerase chain reaction) are two additional tumor agnostic companion biomarkers that support the use of ICT.⁵ This ability to begin to predict which patients will benefit from ICT has led to broad-ranging applications of ICT across a multitude of cancer types.

Figure 1.Mechanisms of T-cell activation via CTLA-4 and PD-1. CTLA-4 is an immune checkpoint receptor expressed on T cells which competes with CD28 to bind B7 which then leads to cell cycle arrest and decreased T-cell proliferation. PD-1 is a negative regulator of T-cell activity that, when bound to its ligands PD-L1 and PD-L2, leads to inhibition of T-cell function. CTLA-4: cytotoxic T-lymphocyte-associated antigen 4; MHC: major histocompatibility complex; PD-1: programmed death 1; PD-L1: programmed death ligand 1; PD-L2: programmed death ligand 2; TCR: T-cell receptor.

Importantly, ICT is but one form of a burgeoning arsenal of anti-lymphoma immunotherapies. While chimeric antigen receptor T-cell (CAR-T) therapy has demonstrated remarkable efficacy in the treatment of pediatric B-cell acute lymphoblastic leukemia, the use of CAR-T therapy for pediatric lymphoma is still under active investigation with early reports of efficacy demonstrated in both mature B-cell lymphomas^6,7 and CD30-positive Hodgkin lymphoma (HL) and non-Hodgkin lymphoma (NHL).⁸ Compared to CAR-T therapy, ICT does not require extensive cellular processing which is advantageous for aggressive lymphomas that require timely therapy, and promotes anti-tumor effects through a variety of neoantigens instead of the one or two tumor targets incorporated in the current iterations of CAR-T therapy. Bispecific T-cell engagers, such as the CD20 x CD3 glofitimab, have also demonstrated early efficacy in pediatric mature B-cell lymphoma.⁹ Finally, antibody-drug conjugates, for example brentuximab vedotin (BV) for anaplastic large cell lymphoma (ALCL)¹⁰ have also demonstrated efficacy in pediatric lymphoma when a suitable drug-target pair is present. Thus, ICT can be viewed as the “tip of the spear” of emerging immunotherapies for pediatric lymphomas.

Table 1.List of immune checkpoint therapies that have received approval from the United States Food and Drug Administration.

Given the extensive immune infiltrates in selected lymphomas, the use of ICT for NHL is a rational extrapolation and extension of the early successes of ICT in melanoma and HL.¹¹ There is differential expression of PD-L1 within the pediatric lymphoma tumor microenvironment (TME) (Figure 2). Interestingly, in adult patients with lymphoma, dramatically different response rates to ICT have been observed based on the histological subtype.¹² Kline et al. have suggested that tumoral expression of PD-L1 combined with the extent of “tumor inflammation,” characterized by immune cell infiltration and activated T-cell phenotypes, are the best predictors of response to ICT in lymphoma.¹³ While much is known about the benefit of ICT in various forms of adult lymphomas, less is known about the benefit of ICT in pediatric lymphomas and the distribution of histological subtypes of NHL and HL is significantly different in pediatric and adult populations. Here, we provide a comprehensive review of the current clinical use of ICT in selected pediatric and adolescent HL and NHL.

Hodgkin lymphoma

As a group, lymphomas are the third most frequent category of malignancy in children aged 0-14 and the most common in patients aged 15-19 years, with HL being the single most common cancer in this age range.¹⁴ The incidence of HL is bimodal with peaks occurring in the 20s and 50s in industrialized countries, with younger peak incidence in lower income countries.¹⁵ Classic HL (cHL) consists of multiple histological subtypes determined by the presence or absence of fibrosis, the number of Hodgkin Reed-Sternberg (HRS) cells, and the characteristics of the inflammatory cellular infiltrate. The malignant cells of cHL are the HRS cells, a clonal population usually derived from germinal center B cells.^16-18 Although HRS cells account for <1% of the reactive inflammatory infiltrate,¹⁹ secretion of pro-inflammatory cytokines, chemokines, and tumor necrosis factor receptor family products determines the histological subtype.^20-22 In cHL the HRS cell immunophenotype is nearly always positive for CD30, and the majority of these cells also express CD15 (~70%), with only 6-10% expressing CD20.^19,22,23 HRS cells do not express CD45, CD19, or CD79A, hematologic markers usually expressed in B-cell NHL.^19,22,23 Recurrent genomic alterations drive proliferation in cHL by upregulating the Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway. These alterations include activating variants in S TAT 6,^24,25 chromosome 9p gains and amplifications,^24-26 and inactivating variants in JAK-STAT negative regulators and inhibitors including SOCS1^25,27 and PTPN1.^24,28 Additional mutations that augment the nuclear factor-kappaB (NF-kB) pathway include inactivating variants of NF-kB inhibitors such as TNFAIP3.^24,29,30 In Epstein-Barr virus (EBV)-positive cases, EBV latent membrane protein (LMP1) functions like a constitutively active tumor necrosis factor receptor family member and signals through NF-kB.³¹ Also seen in HL are mutations that contribute to immune evasion in HL. These include inactivating variants in B2M and gene fusions in CIITA resulting in reduced or absent expression of major histocompatibility complex (MHC) class I^24,32,33 and MHC class II expression,^34,35 respectively. Most cases of HL also include chromosome 9p24 amplification or copy number gains^36,37 which result in increased expression of immune checkpoint genes CD274 and PDC1LG2 which encode PD-L1 and PD-L2, respectively.^36,37 These immunoregulatory checkpoint proteins aid in immune evasion by reducing T-cell proliferation, decreasing cytokine production, impairing T-cell effector function, and inducing T-cell exhaustion and anergy.

Figure 2.Differential PD-L1 expression in the tumor microenvironment of different pediatric non-Hodgkin and Hodgkin lymphomas. EBV: Epstein-Barr virus; GC: germinal center; HL: Hodgkin lymphoma; PD-L1: programmed death ligand 1.

Historically, treatment for pediatric HL has been multimodal and consisted of multiagent non-cross-resistant chemotherapy and radiation therapy. This multimodal approach has been associated with late effects including secondary malignancies, cardiovascular disease, pulmonary dysfunction, endocrinopathies (thyroid dysfunction, infertility) and persistent fatigue.³⁸ Contemporary trials in HL are risk-stratified and response-adapted.^39-42 Cure rates for even high-risk HL exceed 90%, and therefore contemporary clinical trials must balance maintaining high cure rates with limiting acute toxicities and late effects of treatment. Strategies include use of less toxic agents, lower cumulative doses of cytotoxic chemotherapy, eliminating or reducing toxic treatment modalities (such as radiation), and the addition of targeted chemoimmunotherapy and immunotherapy.⁴³ Although there are age-related differences in the frequency of histological subtypes, the majority of pediatric and adult cases of HL share overlapping morphology, immunophenotype and genetic alterations, and recent clinical trials have expanded eligibility to include children, adolescents, and young adults (CAYA).

In the HL TME, PD-L1 is expressed both on HRS cells and on tissue-associated macrophages. Nivolumab and pembrolizumab are anti-PD-1 monoclonal antibodies that interrupt the interaction of PD-L1 and PD-L2 with PD-1 on cytotoxic T cells in the HL TME. These ICT restore T-cell effector function and permit immune-mediated elimination of HRS cells.^36,44,45 Among lymphomas in children, the use of ICT has been best studied and most established in HL. Early phase I/II trials with nivolumab and pembrolizumab generated pediatric pharmacokinetic and safety data and demonstrated activity in pediatric patients with relapsed or refractory (R/R) solid tumors and HL (Table 2).^46,47 Subsequently, the Children’s Oncology Group (COG) and EuroNet collaborative trial AHOD1721 (CheckMate 744) was the first risk-stratified, response-adapted phase II trial using ICT in CAYA aged 5-30 years with R/R HL.^40,41 In this trial, in patients with a low-risk relapse (R1 cohort), treatment with four cycles of BV combined with nivolumab (BvN) resulted in complete metabolic remission (CMR) in 23 of 28 patients (82%).⁴⁰ Intensification with BV and bendamustine (BvB) was reserved only for patients with an incomplete response to BvN. In the total R1 cohort 93% achieved a CMR prior to consolidation with radiation therapy and without autologous hematopoietic stem cell transplantation (AHSCT). With a 3-year event-free survival (EFS) of 87% and 3-year progression-free survival (PFS) of 95%, this trial reduced the use of conventional chemotherapy and demonstrated that a subset of patients with low-risk relapse can be cured without AHSCT. Future randomized trials comparing AHSCT with a transplant-free approach are needed. In the second cohort (R2) of CheckMate 744, patients with standard-risk R/R cHL had a 59% CMR rate (23 of 43 patients) after four induction cycles of BvN.³⁷ As in the R1 cohort, intensification therapy with BvB was reserved for patients without CMR and resulted in 94% of all R2 patients achieving a CMR prior to consolidation with AHSCT. The 1-year PFS was 91% for this cohort.

The safety and efficacy of ICT in R/R HL have resulted in the incorporation of ICT into clinical trials of front-line therapy for patients with high-risk HL. Most notable is the S1826 phase III clinical trial which included adolescents and adults, with pediatric patients (12 to 17 years) accounting for ~24% of the study population.⁴⁰ Patients with newly diagnosed advanced stage (stage III and IV) cHL were randomized to receive six cycles of doxorubicin, vinblastine, and dacarbazine (AVD) with either nivolumab or brentuximab vedotin (N-AVD, Bv-AVD). The N-AVD arm was better tolerated than the Bv-AVD arm with discontinuation in 9.4% and 22.2% of patients, respectively and also demonstrated superior outcomes with 2-year PFS of 92% versus 83%. Three-year PFS and OS rates of 93% and 100% were recently reported for adolescent patients treated on this study.⁴⁸ These excellent outcomes combined with <1% of patients in either arm receiving consolidative radiation therapy has established the N-AVD arm as the current standard of care for advanced-stage HL in both adolescent and adult populations. An additional ongoing study in patients with high-risk HL is the KEYNOTE-667 trial (NCT03407144) which is a phase II trial enrolling patients aged 3-25 years of age with newly diagnosed high-risk cHL. In this trial, patients receive two cycles of standard induction with vincristine, etoposide, prednisone, and doxorubicin (OEPA), and those with a slow early response, defined as a Deauville score 4-5 by positron emission tomography (PET), are treated with pembrolizumab in addition to standard cyclophosphamide, vincristine, prednisone, and dacarbazine (COPDAC-28) consolidation. In the interim analysis, PET-negative responses at the end of therapy were observed in 27 of 42 patients (64%) suggesting that the addition of pembrolizumab to consolidation may augment responses in this high-risk population.⁴⁹ Given the dramatic success of ICT in high-risk HL, ICT has been incorporated into additional clinical trials for patients with low-risk and intermediate-risk HL.

First, a North American Hodgkin Consortium phase II trial compared standard doxorubicin, bleomycin, vinblastine, dacarbazine (ABVD) therapy with chemoimmunotherapy (BvN). This trial included adolescents and adults (>16 years) with newly diagnosed cHL.⁵⁰ Patients with non-bulky stage I and II cHL with CMR after two cycles of standard ABVD were randomized to two additional ABVD cycles followed by consolidation with six doses of nivolumab over 3 months (N=36) or three cycles of BvN (N=45). PFS at 24 months was 100% in both groups. Patients with bulky disease and CMR after two cycles (N=35) were non-randomly assigned two additional cycles of standard ABVD therapy followed by 3 months of nivolumab consolidation and had a 2-year PFS of 86% (95% confidence interval [95% CI]: 67-95%). All patients without CMR after two cycles (N=37) were non-randomly assigned to receive intensified therapy with four cycles of Bv-ABVD followed by consolidation nivolumab therapy and had a 2-year PFS of 77% (95% CI: 57-89%). The second notable trial in the low-stage HL population is the ongoing PET-adapted COG-led Intergroup study AHOD2131 (NCT05675410). This is a phase III trial in children and adults (age 5-60 years) with newly diagnosed stage I or II cHL, and compares standard therapy to the incorporation of BvN in early-stage cHL.⁴⁹ This trial stratifies patients based on favorable or unfavorable features (large mediastinal mass, >3 nodal sites, B symptoms with erythrocyte sedimentation rate [ESR] >30, ESR >50 without B symptoms, and age >50 years) and metabolic response after two cycles of standard ABVD therapy. Rapid early responders are randomized to receive four cycles of BvN versus two cycles of ABVD (favorable risk) or four cycles of AVD (unfavorable risk). All patients with a slow early response are randomized to receive four cycles of BvN versus two cycles of eBEACOPP (escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, prednisone) or two cycles of eBPDac (escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, prednisone, dacarbazine). All patients with a slow early response receive involved-site radiation therapy. The aim of the trial is to use immunotherapy to improve PFS, maintain OS, and minimize long-term morbidity and treatment-related mortality by reducing exposure to radiation therapy and cumulative doses of cytotoxic chemotherapy.

Table 2.Comparative analysis of immune checkpoint therapy in pediatric lymphomas.

In an ongoing investigator-initiated study (RADICAL, NCT05253495), nivolumab replaces doxorubicin after induction therapy in CAYA with de novo intermediate- or high-risk cHL. The chemoimmunotherapy approach includes the use of BV to target HRS cells along with the anti-CD20 antibody rituximab, nivolumab, doxorubicin, dacarbazine, and vinblastine (Figure 3A, B). Fifty patients have completed therapy (N=20 as part of the RADICAL study). All patients have achieved complete remission and there has been no evidence of nivolumab toxicity. Only four patients have required radiation therapy, with an EFS and OS of 100% (median follow-up >90 months; range, 2-159 months).⁵¹

Figure 3.Study schema for intermediate- and high-risk participants in the RADICAL trial. (A) Participants with intermediate-risk Hodgkin lymphoma will receive two cycles of BV-AVD-R 1 and 2. Response assessment will be performed with FDG-PET scan after two cycles of BV-AVD-R. Rapid early responders will continue therapy with two cycles of BV-NVD-R. Patients deemed to be slow early responders after two cycles of BV-AVD-R will continue therapy with four cycles of BV-NVD-R. Radiation therapy will be given at completion of therapy only for slow early responders not achieving complete remission by the end of chemoimmunotherapy. (B) Participants with high-risk Hodgkin lymphoma will receive two cycles of BV-AVD-R. Response will be assessed with FDG-PET scan after two cycles of BV-AVD-R. Rapid early responders will continue therapy with four cycles of BV-NVD-R. Patients deemed to be slow early responders after two cycles of BV-AVD-R will receive two cycles of BV-NAVD-R, followed by four cycles of BV-NVD-R. Radiation therapy will be given at completion of therapy only for slow early responders not achieving complete remission by the end of chemoimmunotherapy. BV-AVD-R: brentuximab vedotin, doxorubicin, vinblastine, dactinomycin, and rituximab; BV-NVD-R: brentuximab vedotin, vinblastine, dactinomycin, nivolumab, and rituximab; CR: complete remission; EOT: end of treatment; HL: Hodgkin lymphoma; FDG-PET: fluorodeoxyglucose positron emission tomography; RER: rapid early responder; RT: radiation therapy; SER: slow early responder.

The last 20 years have witnessed a dramatic shift in the treatment of HL from multimodal chemotherapy with radiation therapy to chemoimmunotherapy with ICT and minimal (or involved-field) radiation as the standard of care (Figure 4). Future trials involving CAYA will undoubtably continue to incorporate ICT for newly diagnosed patients with the goals of improving or maintaining outcomes and reducing long-term toxicities for increasing numbers of patients. In addition, in CAYA with relapsed cHL, trials using ICT will seek to increase remission rates, better define patients who do not require consolidative AHSCT for cure, and improve outcomes for patients who proceed with AHSCT without a CMR. In addition, ICT may be explored as consolidation for newly diagnosed patients or combined with BV as maintenance following AHSCT.⁵² As ICT becomes the standard of care in front-line therapy, new clinical trials will be needed to determine optimal retrieval approaches in R/R patients with prior exposure to ICT.

Mature B-cell lymphomas

Mature B-cell lymphomas, including Burkitt lymphoma (BL) and diffuse large B-cell lymphoma (DLBCL), are the most common NHL in children and adolescents. OS in pediatric patients treated with intensive multiagent chemotherapy and, for those with high-risk disease, intensive chemotherapy with the addition of rituximab is excellent.⁵³ However, current therapy is associated with substantive acute or long-term toxicities and for the small percentage of patients with R/R disease, chance of cure remains limited.^54-56 In pediatric and adult BL, including post-transplant lymphoproliferative disease with Burkitt morphology, PD-L1 expression is generally low or absent (Figure 2).^57-59 Conversely, PD-L1 expression in DLBCL is variable and in adult patients is enriched in specific disease subtypes, including primary mediastinal B-cell lymphoma (PMBCL, see below), T-cell/ histiocyte-rich large-B-cell lymphoma, and tumors that are EBV-positive.⁶⁰ In a retrospective case series of pediatric patients, 19/33 (57%) of those with DLBCL had PD-L1-positive disease, with a higher proportion in those whose pathology demonstrated EBV positivity and 9p24.1 copy gains.⁵⁷

Immunotherapy as a strategy for improvement of outcomes in children with R/R mature B-cell lymphoma is an active area of investigation, currently focused primarily on bispecific T-cell engager therapy, antibody-drug conjugates, and CAR-T therapy.⁶¹ Experience with ICT in pediatric mature B-cell lymphomas is extremely limited. A study of nivolumab in pediatric patients with R/R solid tumors included three patients with BL and three with DLBCL, none of whom had a complete response (CR).⁶² Ongoing studies in children with DLBCL include studies investigating combination therapies including PD-1 agents with tumor-associated antigen-specific T cells (NCT03843294), the LAG-3 inhibitor relatlimab (NCT05255601) and decitabine (NCT05816746).

Primary mediastinal B-cell lymphoma

PMBCL is a rare and aggressive subtype of NHL which originates from thymic B cells. Patients typically present with a large mediastinal mass that can involve the chest wall or lung. Pleural or pericardial effusions are also common.

Figure 4.The evolution of Hodgkin lymphoma clinical trials from chemo-radiotherapy to chemo-immunotherapy. Artificial intelligence was used to create this figure.

PMBCL accounts for 2-4% of NHL, is more common in females, and affects predominantly adolescents and young adults.⁶³ PMBCL was previously classified as a subtype of DLBCL, however, recent molecular characterization has demonstrated similarities to HL including dysregulation of JAK-STAT and NF-kB signaling, and amplification of 9p24 which leads to overexpression of PD-L1.⁶⁴

Currently, there is no single standard of care for the treatment of PMBCL in pediatric or adult patients. In pediatrics, patients with PMBCL have historically been included on prospective trials for mature B-NHL, which were designed to treat BL, the most common subtype of B-NHL in children.^63,65 In contrast, PMBCL in adults is typically managed similarly to DLBCL, the most common B-NHL in adults. The most common chemotherapy regimens used to treat adults with PMBCL in North America are rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) with or without radiation therapy and dose-adjusted rituximab, etoposide, prednisone, vincristine, cyclophosphamide, and doxorubicin (DA-EPOCH-R).⁶⁶ Recently, DA-EPOCH-R was also studied in pediatric patients with PMBCL. In a prospective international study, among 46 pediatric patients treated with DA-EPOCH-R, EFS was 69.6% (95% CI: 55.2%-80.9%), similar to that observed in prior trials with Lymphomes Malins B (LMB)-based pediatric regimens.⁶⁷ Retrospective series have also reported outcomes for children treated with DA-EPOCH-R with EFS ranging from 81% to 84%.^68,69 Among the 15-30% of patients who experience relapse or primary refractory disease, outcomes are poor and novel therapies are needed.⁷⁰

Given the overexpression of PD-L1/PD-L2 in PMBCL, ICT represents a promising therapeutic approach. Anti-PD-1 agents have been studied in adults with R/R PMBCL with promising results. KEYNOTE-013 (NCT01953692) and KEY-NOTE-170 (NCT02576990) were international, open-label, phase Ib and II trials, respectively, in which patients ≥18 years with R/R PMBCL received pembrolizumab for up to 2 years. The objective response rate was 48% (33% CR) in the phase Ib trial and 45% (13% CR) in the phase II trial, both in heavily pre-treated populations.⁷¹ Long-term follow-up (median 48.7 months) of the phase II trial KEYNOTE-170 demonstrated that pembrolizumab monotherapy had 4-year durable responses. Among the 11 patients with a CR, all remained in CR at this late follow-up without additional therapy.⁷² Results from these trials, as well as safety data for pembrolizumab in children,⁴⁷ supported FDA approval for pembrolizumab in adult and pediatric patients with refractory PMBCL or those who have relapsed after two or more prior lines of therapy. Combination strategies with anti-PD-1 therapy have also been explored. Nivolumab in combination with the anti-CD30 antibody-drug conjugate BV was studied in a phase II trial that included 29 adults with R/R PMBCL. With a median follow-up of 39.6 months, the objective response rate was 73.3% and the CR rate was 37%. PFS and OS rates at 24 months were 55.5% and 75.5%, respectively.⁷³

With evidence for the safety and efficacy of anti-PD-1 in the relapsed setting, the next advance in PMBCL will be to understand the role of anti-PD-1 in upfront management. The COG, in collaboration with the National Cancer Institute’s National Clinical Trials Network (NCI NCTN), completed a randomized phase III trial of nivolumab in combination with chemo-immunotherapy in children and adults age ≥2 years with newly diagnosed PMBCL (NCT04759586). Patients on this trial received treatment using a chemo-immunotherapy backbone of either DA-EPOCH-R or R-CHOP based on physicians’ preference. Patients were randomized 1:1 to receive chemo-immunotherapy alone or chemo-immunotherapy + nivolumab (Figure 5A). This trial completed accrual in May 2025 and the read-out is expected in early 2027. Results from this study are expected to provide critical insight into the integration of immune checkpoint inhibitors to the upfront management of PMBCL and may ultimately allow future trials to reduce the use of standard chemotherapy.

Anaplastic large cell lymphoma

ALCL is a T-cell NHL that accounts for 10-15% of de novo pediatric NHL.⁷⁴ ALCL is subdivided based on the expression of anaplastic lymphoma kinase (ALK) with over 95% of pediatric ALCL cases expressing this protein.⁷⁵ ALK expression is a result of a fusion oncoprotein, with NPM-ALK t(2;5)(p23;q35) being the most common fusion, present in approximately 75% of ALK⁺ ALCL cases.⁷⁶ ALK is a known oncogenic driver leading to cell proliferation and survival through RAS/RAF/MEK/ERK, PI3K/AKT, and JAK/STAT signaling pathways.⁷⁷ The presence of the NPM-ALK fusion oncoprotein stimulates an anti-ALK autoantibody in patients with ALK⁺ ALCL and the presence of high-titer (often defined as >1:750) anti-ALK autoantibodies at the time of ALK⁺ ALCL diagnosis is associated with a decreased risk of relapse in patients treated with traditional chemotherapeutic regimens (ALCL99, NHL-BFM95, AIEOP-LNH97).^78,81 Additionally, PD-L1 expression, assessed by immunohisto-chemistry, is seen in 76%-100% of ALK⁺ ALCL specimens and Iwafuchi et al. demonstrated that elevated PD-L1 expression is associated with a poor prognosis in pediatric ALK⁺ ALCL.^57,82,83 These observations highlight the potential of ICT to augment anti-tumor immunity and mitigate the immunosuppressive TME in pediatric ALK⁺ ALCL.

Although there is a strong biological rationale for the use of ICT in ALK⁺ ALCL, a robust evaluation of the clinical utility of ICT in children with ALK⁺ ALCL has not been reported. However, three different case reports have described dramatic responses (CR, CR, CR) to PD-1 inhibitors in heavily pre-treated patients with R/R ALK⁺ ALCL.^84-86 This promising early clinical experience led to the development of the European NIVO-ALCL trial (NCT03703050) in which nivolumab monotherapy is given to patients with active R/R ALK⁺ ALCL (cohort 1) and to patients in a second or greater remission after treatment with BV or an ALK inhibitor (cohort 2). Early results of that trial indicate that for patients in cohort 1, single-agent nivolumab led to an overall response rate of 50% and a 2-year PFS of only 17%.⁸⁷ Encouragingly, no severe or unexpected side effects were reported. Interestingly, the cohort 2 part of the study was closed early because relapses among these patients triggered stopping rules. These results demonstrate that nivolumab is an active agent in ALK⁺ ALCL but is not a curative monotherapy, suggesting that it cannot replace allogeneic stem cell transplantation (allo-SCT) for definitive curative consolidative therapy. One additional trial, RELATIVITY-069 (NCT05255601), recently evaluated the efficacy of nivolumab in combination with the LAG-3 inhibitor relatlimab in children and young adults with R/R lymphoma. This trial included patients with ALCL but was closed due to slow accrual. The ACCELERATE ALCL trial (NCT07013562) began in 2025 and is evaluating the use of nivolumab in a combinatorial, risk-adapted, re-induction therapy followed by reduced toxicity conditioning and allo-SCT in CAYA with R/R high-risk ALK⁺ ALCL (Figure 5B). To the authors’ knowledge, there are no additional active trials evaluating ICT in pediatric ALCL. Thus, there remains a strong biological rationale for further clinical evaluation of the benefit of ICT in pediatric ALCL.

Figure 5.Study schema for the ANHL1931 and ACCELERATE ALCL trials. (A) In the ANHL1931 trial, participants (N=244) with de novo primary mediastinal B-cell lymphoma were randomized 1:1 to receive chemoimmunotherapy alone or chemoimmunotherapy with nivolumab for a total of six cycles. Disease evaluation was conducted at baseline, after cycle 2 and after cycle 6. Radiation therapy (RT) at the completion of chemotherapy was restricted to patients who chose R-CHOP+RT upfront or those with biopsy-proven disease at the end of therapy. (B) In the ACCELERATE ALCL trial, participants with high-risk anaplastic lymphoma kinase-positive anaplastic large T-cell lymphoma (ALK+ ALCL) not previously exposed to brentuximab vedotin (BV) (cohort 1: BV-naïve) will receive two cycles of induction therapy with BV and nivolumab, then undergo disease assessment. Patients in complete remission (CR) will proceed with consolidation with reduced-toxicity conditioning allogeneic stem cell transplantation (RTC + AlloSCT). Patients with a partial response (PR) or stable disease (SD) will receive two cycles of BV, vinblastine, and nivolumab, then undergo disease evaluation. Patients in CR will proceed with consolidation with RTC + AlloSCT. If their response is PR/SD, they will be eligible to receive salvage chemotherapy with ifosfamide, carboplatin and etoposide (ICE) followed by disease evaluation and consolidation with RTC + AlloSCT if in CR/PR. Patients with SD/progressive disease (PD) will be off study. High-risk ALK+ ALCL participants with previous exposure to BV (cohort 2: Prior BV) will receive two cycles of induction therapy with vinblastine and nivolumab, then undergo disease assessment. Patients in CR will proceed with consolidation with RTC + AlloSCT . If the patient has PR/SD, he or she will receive two cycles of BV, vinblastine, and nivolumab, then undergo disease evaluation. Patients in CR will proceed with consolidation with RTC AlloSCT. If the response is PR/SD, they will be eligible to receive salvage with ICE followed by disease evaluation and consolidation with RTC + AlloSCT if in CR/PR. Patients with SD/PD after ICE will be off study. ALCL: anaplastic large T-cell lymphoma; ALK: anaplastic lymphoma kinase; AlloSCT: allogeneic stem cell transplantation; BV; brentuximab vedotin; BV-NAVD-R: brentuximab vedotin, nivolumab, doxorubicin, vinblastine, dactinomycin, and rituximab; DA-EPOCH-R: dose-adjusted rituximab, etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin; EFS: event-free survival; HR: high risk; ICE: ifosfamide, carboplatin and etoposide; NIVO: nivolumab; OS: overall survival; PMBCL: primary mediastinal B-cell lymphoma; PFS: progression-free survival; PR: partial response; R-CHOP: rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone; R/R: relapsed or refractory; RT: radiation therapy; RTC: reduced-toxicity conditioning; SD: stable disease; VBL: vinblastine.

Rare mature lymphomas of T/NK-cell origin

Excluding ALK⁺ ALCL, mature lymphomas of T/natural killer (NK)-cell origin account for less than 10% of all lymphomas occurring in children. Given their rarity and diversity of subtypes, pediatric treatment standards have not been established yet, and most of the available information regarding their pathophysiology and treatment is derived from adult literature. In this review, we comment on two of those entities, aggressive NK-cell leukemia/lymphoma (ANKL) and peripheral T-cell lymphoma, not otherwise specified (PTCL, NOS) highlighting the known aspects of the TME relationship and the potential for ICT use. Both ANKL and PTCL, NOS are the mature T/NK-cell neoplasms currently recognized by the World Health Organization Classification of Paediatric Tumours.⁸⁸

ANKL is an EBV-driven malignancy that is exceedingly rare in Western countries. The adult counterpart of ANKL is extranodal NK-cell lymphoma. In children, ANKL can arise from chronic active EBV infection.⁸⁸ EBV latency genes II (LMP1/2A/2B, EBNA1, EBER1/2) pattern is found in a clonal episomal form, and EBV-encoded small RNA (EBER) is almost universally detected in lymphoma cells by in situ hybridization. The importance of EBV in this cancer goes beyond lymphomagenesis. Its relationship with the host is represented by the correlation between detectable plasma EBV viremia (by polymerase chain reaction) and disease activity, dissemination, response to therapy and outcomes (Figure 6).^89,90 The EBV integration leads to cell division, migration, apoptosis, and activation of multiple diverse signaling pathways.⁹¹ In fact, three ANKL molecular subtypes have been described in association with specific genomic alterations: TSIM (mutations in the JAK/STAT pathway, TP53, and amp9p24.1/JAK2 locus, amp17q21.2/STAT3/5B/5A locus, amp9p24.1/PD-L1/2 locus, and del6q21), MB (MGA mutation and 1p22.1/BRDT loss of heterozygosity) and HEA (HDAC9, EP300, and ARID1A mutations).⁹² These molecular subtypes are not only associated with the cell-of-origin – NK cells (TSIM, MB) or T cells (HEA, MB) – but also with survival. Adult patients with ANKL MB subtype have a poor outcome in comparison to those with TSIM or HAE subtypes independently of stage or performance status at presentation.⁹² To date, molecular ANKL subtypes in pediatric cases have not been described.

Upregulation of PD-L1/2 is observed in the most common subtype, TSIM, resulting in potential sensitivity to ICT. In fact, PD-L1 expression is often found in R/R ANKL samples, also suggesting a potential mechanism of tumor evasion. The use of ICT has been tested in patients with refractory extranodal NK-cell lymphoma, including post use of L-asparaginase-containing regimens.^93-95 A systematic review of 14 articles on pembrolizumab use in adult NK/T-cell lymphoma found an objective response rate of 84.5% with CR in 61% and disease-free survival ranging from 2 to 48 months.⁹⁶ For CAYA with ANKL, the NCT03719105 study is using pembrolizumab to rescue patients not in CR after induction with daratumumab + modified SMILE (dexamethasone, methotrexate, ifosfamide, L-asparaginase, etoposide) chemotherapy prior to consolidation with an allo-SCT (NCT03719105).⁹⁷

In PTCL, NOS the expression of PD-1/PD-L1 is variable, and dependent on the immunohistochemistry cutoff of detection, with reported ranges from 40% to near 100%.^98-100 PD-L1 expression seems to correlate with outcomes, with highly PD-L1-positive tumor expressors resulting in better responses to ICT¹⁰⁰ and affecting staging and outcomes as well.^101,102 Interestingly, gene-expression signatures linked to tumor-infiltrating immune cells, but not to the cancerous T cells, correlate with outcome.¹⁰³ Favorable outcomes are associated with both B- and dendritic-cell signatures, with a macrophage signature portending a worse prognosis. Interestingly, PTCL, NOS cases with a macrophage infiltration had elevated expression of PD-L1/2, suggesting that ICT could potentially result in therapeutic responses in this subset of patients.¹⁰³ In children, there is a lack of information regarding the genomic profile, TME, treatment approaches, and outcomes in PTCL, NOS.¹⁰⁴ In the NCT03719105 study, testing BV (anti-CD30ADC) and pralatrexate as novel therapeutic agents, CAYA with PTCL, NOS are also eligible for enrollment.⁹⁷ Of note, two pediatric patients with refractory PTCL, NOS were treated with nivolumab leading to CR (patient 1, off therapy: immunohistochemistry determined PD-L1 tumor expression, 5%) and very good partial response (patient 2, therapy ongoing: immunohistochemistry determined PD-L1 tumor expression, 80%) (ACX, personal communication).

Figure 6.Epstein-Barr virus (EBV) viremia is associated with malignant transformation of NK/T lymphoma and persistence of EBV viremia following therapy is significantly associated with inferior overall response rate, progression-free survival, and overall survival. EBV: Epstein-Barr virus; ESA: etoposide, dexamethasone, pegaspargase; IL-17: interleukin 17; MESA: methotrexate, etoposide, dexamethasone, pegaspargase; NK: natural killer; ORR: overall response rate; OS: overall survival; PFS: progression-free survival; PPAR: peroxisome proliferator–activated receptor; RNAseq: RNA sequencing; TME: tumor microenvironment; WGS: whole-genome sequencing. Reprinted, with permission, from Xavier et al.⁸⁹ Copyright 2025 by the Ferrata Storti Foundation.

Conclusion and summary

We are at the “end of the beginning” of the investigations of ICT in pediatric and adolescent lymphomas. Based on PD-L1 and PD-L2 expression and phase I/II safety and efficacy data, the histological subtypes of lymphoma most likely to respond positively to ICT (PD-L1/L2 inhibitors) include cHL, PMBCL, ALCL and mature EBV-associated NK/T subtypes in pediatric and adolescent patients. Conversely, BL, germinal center B-cell-like DLBCL, and B- and T-lymphoblastic lymphomas are unlikely to respond to ICT in pediatric and adolescent patients. ICT investigations in cHL, PMBCL, ALCL and mature EBV-associated T/NK-cell lymphomas should be initially focused in the R/R setting but also subsequently in the upfront setting to reduce the burden of acute and long-term effects of chemoradio-therapy, while maintaining ≥90% EFS and OS. Finally, the role of CTLA-4 inhibitors is relatively unknown in pediatric and adult lymphomas. However, based on the lymphoma TME in cHL, PMBCL and ALCL and the presence of antigen-presenting cells in the TME, future investigations may be warranted to determine the safety and efficacy of CTLA-4 inhibitors in the different histological subtypes of pediatric and adolescent lymphomas.

Footnotes

• Received January 19, 2026
• Accepted May 27, 2026

Correspondence

M.S. Cairo mitchell_cairo@nymc.edu

Disclosures

DJH is a consultant for both the Merck Pediatric Hematology and Oncology Scientific Advisory Committee and Ymabs (SERB) Radiopharmaceutical Protocol Steering Committee. LG-R is a consultant for Merck (fees paid to Weill Cornell), BMS, Roche, and Pierre Fabre. MSC serves as a consultant for Jazz Pharmaceuticals and AbbVie; is a member of speakers’ bureau for Jazz Pharmaceuticals and Amgen, Inc.; and receives research funding from Jazz Pharmaceuticals, AbbVie, Merck, Miltenyi Biotec, Servier, Omeros, and Janssen.

Artificial intelligence

Figure 4 was created with the aid of artificial intelligence.

Contributions

MSC conceptualized the study and supervised the review. DJH conceptualized the study and wrote the first version of the manuscript. All authors contributed to writing the manuscript.

Funding

This work was supported in part by 1U54CA232561-01A1 (to MSC) and the Haworth Family Endowment (to DJH).

Acknowledgments

The authors thank Ginny Davenport, RN, Erin Morris, RN, BSN, and Lauren Harrison, RN, MSN for their assistance in the preparation of this manuscript.