Hodgkin lymphoma (HL) is the seventh most common subtype of lymphoma, with approximately 8500 new cases and 1120 estimated US deaths per year in 2016. The majority of US patients achieve cure, with a 5-year overall survival (OS) rate of more than 85%. However, up to 10% of patients are refractory to initial therapy, and up to 30% of patients will eventually relapse after frontline therapy, at which point the expectation of cure can be as low as 30% in patients with high-risk relapse or as high as 70% in patients without poor prognostic markers.
Currently, patients with treatment-naïve HL are classified as having early favorable, early unfavorable, or advanced disease based on several factors. These include: stage; presence or absence of B symptoms (ie, systemic symptoms of fever, night sweats, and weight loss); erythrocyte sedimentation rate elevation; number of nodal sites involved; and presence or absence of bulky disease. In the United States, standard chemotherapy for most patients includes a combination of doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD), with or without consolidative radiation therapy (RT). Several studies have been performed to determine the optimal number of chemotherapy cycles and/or dose of RT for each group of patients.
The HD10, HD11, and EORTC 20012 trials suggest what standard therapy should be: For those with early favorable disease, 2 cycles of ABVD are followed by 20 gray (Gy) of involved site RT (ISRT); the associated OS is 97% and the progression-free survival (PFS) is 93%.1 For those with early unfavorable disease, 4 cycles of ABVD are followed by 30 Gy of ISRT; the associated OS is 94% and the PFS is 86%.2 For those with advanced disease, 6 cycles of ABVD are undertaken; the associated OS is 88% and the PFS is 74%.3
Considering that HL occurs commonly in young patients and that the majority of patients are cured with standard therapy, recent studies have focused on developing algorithms to guide escalation versus de-escalation of treatment based on response, risk of treatment failure, and risk of long-term toxicity.
One evolving concept is the use of PET-adapted therapy. The Deauville score (DS) has become the standard way to evaluate PET response in HL. PET scan results are given a score of 1 to 5 based on a fludeoxyglucose (FDG) uptake compared with background mediastinal and liver uptake. A DS of 1, 2, or 3 is considered negative, and a DS of 4 or 5 is considered positive. In addition to the PET scans performed as a standard part of baseline and end-of-therapy evaluations, early interim PET scan performed after 2 cycles of chemotherapy (PET2) is emerging as a prognostic and predictive marker in HL. In a study of 206 patients with early unfavorable (n = 53) or advanced stage (n = 207) treatment-naïve HL, the 3-year PFS for PET2-negative patients was 95%, compared with 28% for those who were PET2-positive.4 Based on these data, there have been increasing interest in risk-adapted therapy based on PET2 results. In patients who achieve complete metabolic response at the end-of-therapy PET, routine PET surveillance is not recommended. Radiographic surveillance using CT scans no more frequently than every 6 months for the first 2 years following completion of therapy, followed by clinical surveillance only, is a common approach.
PET2 has been used to investigate if RT may be omitted in subgroups of patients with limited-stage HL. The EORTC/ LYSA/FIL H10 noninferiority trial evaluated whether RT could be omitted in patients with early-stage HL, achieving negative PET2 without compromising PFS. Patients with early favorable (HL-F) and early unfavorable (HL-U) HL and a negative PET2 received further ABVD (2 cycles in HL-F; 4 in HL-U) without RT or standard ABVD (1 cycle in HL-F; 2 in HL-U) plus 30 Gy of involved node RT (INRT). The study failed to show noninferiority of ABVD alone: 5-year PFS was 99% versus 87% in HL-F patients and 92% versus 90% in HL-U patients receiving INRT or no RT, respectively.5
In the RAPID trial, patients with early-stage HL received 3 cycles of ABVD, then those with a negative PET3 were randomized to no further therapy versus 30 Gy of involved field RT (IFRT). Five-year PFS was 97% in those who received IFRT versus 91% in the observation arm. However, 7 versus 2 patients died while in complete response (CR) and 5 versus 2 patients died of progressive disease in the IFRT and observation arms, respectively.6 Given the small PFS benefit and suggestion of increased toxicity in these trials, it is possible that IFRT may not be warranted in patients with early-stage HL who achieve a negative interim PET.
Given poor outcomes with continued ABVD therapy in patients with a positive PET2, another question is whether or not therapy should be escalated in patients with limited-stage HL and a positive PET2. In the experimental arms of the EORTC/LYSA/H10 trial, patients with early-stage HL-F or HL-U who were PET2-positive received 2 cycles of escalated bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone (eBEACOPP), followed by INRT. Compared with standard ABVD × 3 or 4 cycles plus INRT in HL-F and HL-U, respectively, pooled analysis revealed that intensification to eBEACOPP was associated with a statistically significant improvement in 5-year PFS from 77% to 91%.5 In the CALGB/ALLIANCE 50604 trial, 164 patients with early-stage HL were given 2 additional cycles of ABVD if PET2-negative (91%), and 2 cycles of eBEACOPP with 306 Gy of IFRT if PET2-positive (9%). With a median of 2 years’ follow-up, 3-year PFS was 92% in PET2-negative patients versus 66% in PET2-positive patients, and it was concluded that the protocol is unlikely to meet its secondary endpoint of improved PFS with eBEACOPP plus IFRT in patients with PET2-positive disease.7
PET2 has also been investigated as a marker to identify patients who may benefit from intensification from ABVD to eBEACOPP in the advanced-stage setting. The SWOG 0816 trial included 358 patients with advanced-stage HL initially treated with 2 cycles of ABVD. PET2-negative patients (DS 1-3, 82%) were treated with an additional 4 cycles of ABVD for a total of 6 cycles of ABVD, and PET2-positive patients (DS 4-5, 18%) were intensified to receive 6 cycles of eBEACOPP after the initial 2 cycles of ABVD. With a median follow-up of 39.7 months, 2-year OS was 98% and 2-year PFS was 79% for the cohort. The 2-year PFS was 82% in PET2-negative patients and 64% in PET2-positive patients, significantly improved compared with historical controls treated who had continued ABVD after a positive PET2. Escalated BEACOPP was associated with increased grade 4/5 toxicity (86% vs 36%) and OS data are immature.8 While improved outcomes for PET2-positive patients who escalated to eBEACOPP cannot be ignored, given the absence of proven OS benefit and increased toxicity, whether or not intensification to eBEACOPP is warranted in patients with PET2-positive advanced-stage HL remains open to discussion. Trials exploring the incorporation of novel agents based on PET response are ongoing.
Another question is whether therapy can be safely de-escalated in patients with advanced-stage HL who achieve a negative PET2. The RATHL study included 1214 patients with treatment-naïve bulky stage II or advanced-stage HL. All patients were treated with 2 cycles of ABVD, and the 952 who achieved negative PET2 were randomized to either receive an additional 4 cycles of ABVD or to omit bleomycin from subsequent doses and de-escalate to 4 cycles of AVD. Four percent of patients received consolidative RT at the treating physician’s discretion in both arms. There was no difference in PFS at 3 years (85.4% vs 84.4%) or OS (97% vs 97.5%) in ABVD versus AVD, and the risk of respiratory adverse events was reduced by omitting bleomycin.9 Based on these data, many have accepted that bleomycin may safely be omitted from ABVD in patients with advanced-stage HL who achieve negative PET2.
Standard treatment of HL in first relapse currently includes salvage chemotherapy, such as ifosfamide, carboplatin, and etoposide (ICE); dexamethasone, cytarabine, and cisplatin (DHAP); gemcitabine, vinorelbine, and liposomal doxorubicin (GVD); and ifosfamide, gemcitabine, vinorelbine (IGEV); all are salvage chemotherapies and can be followed by autologous stem cell transplantation (autoSCT). Response rates to salvage therapy are in the 60% to 80% range; however, at about 20%, CR rates are low. Furthermore, approximately 50% of patients will relapse within 5 years of autoSCT. Patients with a late, localized relapse may instead be treated with a combination of chemotherapy and radiation. For those patients who relapse post autoSCT and are eligible, further salvage therapy followed by allogenic stem cell transplantation (alloSCT) is considered. With the development of novel agents, these historic treatment paradigms are rapidly changing, and currently there is no FDA-approved standard of care (Table 1).
CD30 is universally expressed on Reed-Sternberg (RS) cells. It is a target of interest in HL due to its role in promoting RS cell survival through simulation of nuclear factor kappa-light-chain-enhancer (NF-kB) signaling, as well as its interaction with the tumor microenvironment to promote tumor progression.10 Brentuximab vedotin (BV) is an antibody-drug conjugate in which the microtubule-disrupting agent, monomethyl auristatin (MMAE), is linked via protease-cleavable linker to the anti-CD30 monoclonal antibody, brentuximab. In a pivotal phase II study, BV was administered to 102 patients with relapsed or refractory HL and a median of 3.5 prior therapies, all of whom had prior autoSCT. At a median of 18.5 months of observation, the overall response rate (ORR) was 75% with a median duration of response (DOR) of 6.7 months; the CR rate was 34% with a median DOR of 20.5 months.11 Long-term data published with a median of 69.5 months follow-up revealed a 5-year OS of 41% and 5-year PFS of 22%. For the 34 patients who achieved CR, OS was 64%, and 52% remained in CR at 5 years. Of the 13 patients in CR at study closure, 4 had undergone alloSCT and 9 remained in continued CR without further lymphoma-directed therapy.12 Based on these data, BV was approved by the FDA for use in relapsed or refractory HL in patients who have progressed after autoSCT or who are autoSCT ineligible.
BV is also approved as early consolidation after autoSCT in patients with high-risk relapsed or primary refractory HL, based on the AETHERA trial demonstrating a PFS benefit. In this phase III, multicenter, international double-blinded study, 329 patients with high-risk HL were randomized to receive BV 1.8 mg/kg intravenously (IV) every 3 weeks versus placebo for 16 doses starting 30 to 45 days after autoSCT. High risk was defined as primary refractory disease, relapse within 12 months of frontline therapy, or relapse with extranodal disease more than 12 months after frontline therapy. Median PFS was 42.9 months in the BV arm versus 24.1 months in the placebo arm. There were no statistically significant differences in 3-year OS (81% vs 79%), and treatment was well tolerated.13
Immune Checkpoint Inhibitors
PD-1 is a transmembrane protein that functions as a negative immunoregulator, promotes self-tolerance, and has been shown to be crucial in the pathobiology of classical HL. Amplifications of chromosome 9p24.1 in RS cells lead to overexpression of PD-1 ligands as well as JAK2, JAK/STAT pathway activation, further PD-1 ligand transcription, and subsequent immune evasion.14
Nivolumab is a fully human anti–PD-1 immunoglobulin G4 (IgG4) monoclonal antibody. In a phase I study, 23 patients with relapsed or refractory HL received nivolumab 3 mg/kg IV every 2 weeks. Seventy-eight percent were post autoSCT and 78% were previously treated with BV. Treatment was well tolerated, the ORR was 87% (CR, 17%) and the remaining 13% of patients had stable disease. Responses were durable, with 24-week PFS of 86%.15 In a single-arm phase II study, 80 patients with HL who were post auto-SCT and relapsed or refractory to BV were treated with nivolumab, resulting in an ORR of 66% (CR, 9%) with a median DOR of 7.8 months.16 Based on these results, nivolumab was granted FDA approval for use in patients with HL who have relapsed post autoSCT and post-transplant BV.
Pembrolizumab is a humanized mouse IgG4 anti–PD-1 mono-clonal antibody. KEYNOTE-087 was a single-arm phase II trial in which 210 patients with relapsed or refractory HL received pembrolizumab at 200 mg every 3 weeks. Patients were heavily pretreated, 83% had received BV, and 61% were post autoSCT. The ORR was 69% (CR, 22%), median DOR was 11 months, and treatment was well tolerated.17 Based on these data, the FDA granted pembrolizumab accelerated approval for use in patients with HL that has relapsed after 3 or more lines of therapy.
The incidence curve of HL is bimodal, with 1 peak around age 20 years and a second around age 65 years. Those diagnosed with HL over age 60 years are considered older and more likely to experience bleomycin lung toxicity, to have comorbidities limiting anthracycline use, and to have inferior outcomes compared with younger patients. Their reported 5-year PFS is 30% to 45%, with 5-year OS of 40% to 60%.18 As a number of novel agents are entering the HL armamentarium, several upfront alternatives to ABVD and less-toxic salvage regimens of interest have emerged, including BV monotherapy, BV-dacarbazine, BV-nivolumab, AVD-lenalidomide, and sequential therapy with BV-AVD. BV-bendamustine showed promising efficacy in this population, but further study has been halted due to unacceptable toxicity in older adults (Table 2).
The field of HL is rapidly changing. In the treatment-naïve setting, radiation therapy will be used more selectively, as novel agents are incorporated into upfront treatment regimens with the goal of decreasing long-term toxicity. The understanding of how to best use a PET-adapted approach to escalate versus de-escalate treatment will continue to evolve. Another area of investigation is the use of novel agents as consolidative therapy post autoSCT, including BV-checkpoint inhibitor combination therapy. AlloSCT will remain an option for eligible patients. With the availability of novel agents, outcomes in older and unfit patients with HL will continue to improve.
Author affiliations: All authors are with the John Theurer Cancer Center, Hackensack, New Jersey.
Address correspondence to: Lori A. Leslie, MD, John Theurer Cancer Center, 92 2nd St, Hackensack, NJ 07601. E-mail: Lori.Ann. Leslie@hackensackmeridian.org.
Financial disclosures: Tatyana Feldman has received consultancy fees and has participated on paid advisory boards for Bristol-Myers Squibb. She has received honoraria and lecture fees on speaker’s bureaus for Celgene, Seattle Genetics, AbbVie, Pharmacyclics, Janssen, and Kite Pharma. Lori Leslie has received lecture fees for participating in speaker’s bureaus for Seattle Genetics, Celgene, and Kite Pharma.