Chapter 10: Hodgkin Lymphoma

Hodgkin lymphoma (HL) was recognized in the first half of the nineteenth century by Thomas Hodgkin and Samuel Wilks (¹). HL usually arises in lymph nodes, preferentially in the cervical area, and the majority of HLs manifest clinically in young adults in their third and fourth decades of life. The incidence of HL is 3.0 per 100,000 person-year in the United States; it is higher in the Western countries than Asian countries (^2,3). Biological and clinical studies have shown that HLs are comprised of two disease entities: nodular lymphocyte-predominant HL (NLPHL) and classical HL (cHL) (¹). The two entities differ in their clinical features and behavior. Within cHL, four subtypes have been described: nodular sclerosis, mixed cellularity, lymphocyte-rich, and lymphocyte-depleted. These four subtypes differ in their clinical features, growth pattern, presence of fibrosis, and frequency of Epstein-Barr virus infection but share the immunophenotype of tumor cells.

The management of HL continues to evolve. Before the widespread use of modern polychemotherapy, large-field radiation therapy was able to cure cHL patients. However, reliance on radiation alone required extensive radiation portals to treat nearly the entire lymphatic system with radiation doses up to 44 Gy. With long-term follow-up, many patients developed heart toxicity and second malignancies. Therefore, efforts have been made to reduce the long-term toxicities of treatment for HL while maintaining excellent cure rates. With modern chemotherapy, multiple randomized studies have shown that radiation portals can be safely reduced from extended-field radiation to involved-field radiation and, now, even smaller fields of radiation, including involved-node radiation.

Currently, the treatment of cHL is stratified by risk groups—early-stage favorable, early-stage unfavorable, and advanced-stage disease—according to the clinical stage and the presence or absence of adverse clinical features. In this chapter, we will review advances in the management of HL, including recent publications about strategies of management of the rarer diagnosis of NLPHL.

Over the past decade, investigators have made significant progress in the diagnosis, classification, staging, prognosis, and treatment of HL. In past years, the true lineage of the neoplastic cells in HL was unknown; hence, the term Hodgkin disease was used. It is now recognized that almost all cases of HL are of B-cell lineage; hence, the name changed to Hodgkin lymphoma.

The classification of HL has remained relatively stable over the past 40 years, and the World Health Organization (WHO) classification of lymphoid neoplasms was updated in 2008 (¹) (Table 10-1). The current WHO classification recognizes that NLPHL is distinct from the other types that can be grouped together under the rubric of cHL (¹).

In 2015, it was estimated that 9,050 Americans would be diagnosed with HL, with a median age at diagnosis of 39 years (⁴). Hodgkin lymphoma has been traditionally defined as a hematopoietic neoplasm composed of diagnostic Hodgkin and Reed-Sternberg (HRS) cells within a reactive cell background. An HRS cell is large, 30 to 60 μm, containing a bilobed vesicular nucleus, with each lobe containing a prominent round eosinophilic nucleolus surrounded by a clear zone or halo; it also has abundant cytoplasm. However, HRS cells often comprise less than 1% of the involved tumor tissue and are absent in NLPHL. Hodgkin and Reed-Sternberg cells are believed to be derived from germinal center (GC) B cells that have unfavorable immunoglobulin V gene mutations, whereas lymphocyte-predominant (LP) cells, which were previously termed lymphocytic and histiocytic (LH) cells, are thought to originate from antigen-selected GC B cells (⁵).

Nodular Lymphocyte-Predominant Hodgkin Lymphoma

Clinical Features

Approximately 5% of patients with HL have NLPHL. This disease is usually localized and most often involves cervical or axillary lymph nodes (^1,6). The disease affects patients of all ages, males more often than females, and is clinically indolent (Table 10-2). Systemic symptoms—such as fever, weight loss, and nights sweats (also known as B symptoms)—are infrequent. Patients with NLPHL are characterized as having a more indolent course with delayed relapse compared with cHL, analogous to low-grade non-HL (⁶). Patients with NLPHL are at 5% to 6% risk for developing diffuse large B-cell lymphoma (DLBCL) or T-cell/histiocyte-rich large B-cell lymphoma (¹).

The German Hodgkin Lymphoma Study Group (GHSG) has reported a large retrospective study of 394 patients with NLPHL. Patients were predominantly men (75%), only 9% had B symptoms, and 79% had early-stage disease (⁷). With a median follow-up of 50 months, tumor control (freedom from treatment failure) and overall survival were 88% and 96%, respectively, slightly better than in cHL (82% and 92%, respectively).

Histologic Features

Nodular lymphocyte-predominant Hodgkin lymphoma is characterized by effacement of nodal architecture by variably sized, vague nodules composed of numerous small lymphocytes, histiocytes, and characteristic neoplastic LP cells (Fig. 10-1A) (^1,8). These cells are typically large, with pale cytoplasm and polyploid vesicular nuclei containing inconspicuous nucleoli resembling kernels of popped corn, hence the nickname popcorn cells (Fig. 10-1B). However, LP cells can exhibit a range of cytologic appearances. These cells can be round and/or have relatively prominent nucleoli. Eosinophils, neutrophils, and plasma cells are usually absent in NLPHL, and there is no associated necrosis or fibrosis.

Cases of NLPHL can also have diffuse areas. When diffuse areas are large, their presence often correlates with more aggressive disease. To reflect this change in clinical behavior, many pathologists diagnose such cases as NLPHL with progression to T-cell/histiocyte-rich large B-cell lymphoma, also described as T-cell–rich B-cell lymphoma (TCRBCL). Other pathologists use the term NLPHL with large diffuse areas and suggest that the diffuse areas may represent the beginning stages of progression to DLBCL. The boundary between NLPHL with diffuse areas and TCRBCL remains blurred. Most cases previously designated as diffuse LPHL, as defined previously (⁹), are now classified differently. With appropriate workup, these cases are usually classified as NLPHL with large diffuse areas, lymphocyte-rich cHL, or TCRBCL.

Gene expression profiling of NLPHL to determine the origin and pathogenesis of LP cells found significant similarities between NLPHL, TCRBCL, and cHL (¹⁰). Overall, LP cells are thought to derive from antigen-selected GC B cells (⁵). LP cells also demonstrate deregulation of numerous apoptosis regulators and putative oncogenes and a partial loss of their B-cell phenotype. In addition, there is constitutive activation of nuclear factor-κB (NF-κB), Janus kinases/signal transducers and activator of transcription (JAK/STAT) pathway, and aberrant extracellular-regulated kinase signaling.

Immunophenotypic Findings

Nodular lymphocyte-predominant Hodgkin lymphoma is immunophenotypically distinct from other types of HL. The LP cells usually express leukocyte common antigen (LCA; CD45), immunoglobulin J chain, B-cell antigens (CD19, CD20, CD22, CD79A, and BCL-6), and epithelial membrane antigen (EMA) and are negative for CD15 and CD30 (Figs. 10-1C and D). These results suggest that the LP cells are B cells that arise from the GC. The LP cells are negative for T-cell antigens but are often surrounded by a rosette of small, reactive T cells that may be positive for pan-T-cell antigens and CD57 (Fig. 10-1E). Epstein-Barr virus is almost always absent in the LP cells of NLPHL.

Lymphocyte-Rich Classical Hodgkin Lymphoma

Lymphocyte-rich cHLs (LRHLs) have been recognized that resemble NLPHL histologically but are cHL immunophenotypically (^1,6,8). The frequency of this type of HL is not well known but is most likely low (<5%). Clinically, patients with LRHL are similar to patients with other subtypes of cHL or have clinical findings intermediate between NLPHL and cHL. Unlike patients with NLPHL, late relapse is uncommon in patients with LRHL.

Histologically, these tumors are rich in small lymphocytes and histiocytes. Granulocytes and plasma cells are usually infrequent. Necrosis is usually not present. Lymphocyte-rich cHL may be either nodular or diffuse. The nodular type closely resembles NLPHL. Vague nodules of numerous small lymphocytes are present, and the nodules may have a small compressed GC (Figs. 10-2A and B). The neoplastic cells are present in the mantle zones of the nodules. These neoplastic cells usually resemble Reed-Sternberg and typical mononuclear variant cells (so-called Hodgkin cells) rather than LP cells. The cell composition is similar in diffuse cases of LRHL, but nodularity is minimal or absent. Immunohistochemical studies of LRHL show that the large neoplastic cells have an immunophenotype similar to that of all cHL cases, positive for CD15 and CD30 and negative for LCA (CD45) (Figs. 10-2C,D,E).

In a study of NLPHL by the European Task Force on Lymphoma (⁸), a large number of tumors that had been classified as NLPHL were reviewed; the diagnosis was confirmed in only half of these cases. Most of those excluded were reclassified as LRHL.

Nodular Sclerosis Hodgkin Lymphoma

Clinical Features

Nodular sclerosis (NS) is the most common form of cHL, representing 60% to 70% of all cases in Western countries; it is also the most common type of cHL in patients younger than age 50 years. Whites are affected more often than others, and nodular sclerosis HL (NSHL) is much less frequent in Asian countries (^2,3). The age-adjusted incidence rate of NSHL has been stable since 1993 to 2008 in the United States according to the National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) data (²). The male-to-female ratio is approximately equal. Nodular sclerosis HL has a marked predilection for involving mediastinal, supraclavicular, and cervical lymph nodes. A mediastinal mass is very common, and the thymus may be involved.

Histologic Features

Nodular sclerosis HL is characterized by a triad of findings: (1) a nodular pattern, (2) broad bands of fibrosis that outline the nodules, and (3) characteristic mononuclear cell variants known as lacunar cells (Fig. 10-3). A lacunar cell has abundant clear cytoplasm with a sharply demarcated cell membrane. In formalin-fixed tissue, a characteristic artifact occurs. The cell cytoplasm retracts, leaving a clear space or lacuna surrounding the cell, hence the origin of the name. The typical lacunar cell has a polylobulated nucleus with one or multiple small nucleoli. However, lacunar cells can show great morphologic variability and can be round with prominent nucleoli, or they may resemble large noncleaved cells. A heterogeneous mixture of reactive cells may be seen in HL, including small lymphocytes, histiocytes, eosinophils, neutrophils, and plasma cells in variable numbers.

Nodular sclerosis HL has been graded (as 1 or 2) by the British National Lymphoma Investigation (BNLI) group (¹¹) according to the numbers of neoplastic cells and reactive cells present, and this grading system has been adopted by the WHO classification (¹). Grade 2 cases of NSHL show numerous neoplastic (lacunar) cells and depletion of reactive lymphocytes. Lymphocyte-depleted and syncytial variants (Fig. 10-4) of NSHL have been described; these cases are the outermost examples of grade 2 NSHL. The syncytial variant of NSHL is composed of sheets of neoplastic cells and necrosis.

Mixed Cellularity Hodgkin Lymphoma

Clinical Features

The mixed cellularity variant of HL (MCHL), the second most common type, affects 15% to 25% of all patients with the disease and is the most common form in patients older than age 50 years (^1,12). Males are affected more often than females. According to NCI SEER data, MCHL is relatively more common in African Americans and Hispanic Americans than in whites in the United States. A substantial percentage of patients with MCHL have clinical stage III or stage IV disease and B symptoms.

Histologic Features

Mixed cellularity HL is characterized by a large number of Reed-Sternberg cells and Hodgkin cells in a background of numerous eosinophils, plasma cells, histiocytes, and granulocytes in varying proportions (¹) (Fig. 10-5). The lymph node architecture is usually diffusely replaced. Partially involved lymph nodes show selective paracortical infiltration. Disorderly fibrosis may be seen, but the broad fibrous bands and capsular fibrosis characteristic of NSHL are absent.

Two variants of MCHL can be relatively more difficult to diagnose. In the interfollicular variant, which most likely represents partial involvement of lymph nodes by HL, the tumor is located in the interfollicular region and is often associated with reactive follicular hyperplasia and marked plasmacytosis. In the epithelioid histiocyte-rich variant, numerous epithelioid histiocytes and granulomas are present; these can be so numerous as to obscure the neoplastic cells. The importance of these variants of MCHL lies in their unusual histologic findings.

Lymphocyte-Depleted Hodgkin Lymphoma

Clinical Features

Lymphocyte-depleted HL (LDHL) is the least common type, representing 1% of all cases (¹²). In the NCI SEER study, it was shown that the age-adjusted incidence rate for LDHL has decreased. This decrease is most likely explained by the recognition by pathologists that many tumors previously classified as LDHL are, in fact, non-HLs (such as anaplastic large-cell lymphoma). Improved classification is the result of application of immunohistochemical and molecular methods to the study of these tumors.

Patients with LDHL are usually elderly, and LDHL is rare in individuals younger than 40 years old. There is a slight male predominance. Whites and African Americans are equally affected. Most patients have advanced clinical stage and B symptoms. Patients commonly have a large contiguous mass of matted lymph nodes or diffuse visceral involvement. The diffuse fibrosis type of LDHL commonly has a subdiaphragmatic distribution. Lymphocyte-depleted HL has the poorest prognosis of all types of HL (¹²).

Histologic Features

The LDHL category includes two variants originally recognized by Lukes and Butler: diffuse fibrosis and reticular (Fig. 10-6). The diffuse fibrosis variant of LDHL is characterized by an extensive proliferation of disordered, hypocellular fibrosis. Diagnostic HRS cells can be difficult to find and may be spindled within dense collagen. Reactive inflammatory cells are relatively few. The reticular variant of LDHL has numerous HRS cells and bizarre variants that have been termed pleomorphic variants. These cells may exhibit marked variations in nuclear number and shape, often with giant nucleoli. Normal small lymphocytes are infrequent compared with the other subtypes of HL. Necrosis is common and may be extensive. The HRS cells and pleomorphic variants may be found in sheets. Mitotic figures are usually numerous.

Immunophenotypic Findings in Classical Hodgkin Lymphoma

Overall the mature B-cell origin of the HRS cells is not readily apparent because HRS cells have a very unusual phenotype and have expression of genes that are seen on many hematopoietic cell types. The neoplastic cells are positive for CD15 and CD30 and negative for LCA (CD45) (Fig. 10-7) and EMA (¹). B-cell antigens, such as CD20, CD79A, PAX-5/BSAP, and MUM1/IRF4, are expressed in a subset of cases. CD20 expression is often weak. T-cell antigens are usually not expressed by the neoplastic cells. BCL-2 is positive in up to half the cases and has been correlated with poorer prognosis (¹³). Epstein-Barr virus is relatively common in cHL, but its frequency varies greatly among different types.

The University of Texas MD Anderson Cancer Center (MDACC) diagnostic and treatment algorithms for HL are shown in Fig. 10-8.

Staging of the Patients

The Ann Arbor system for staging patients with HL at the time of initial presentation forms the basis for the treatment of disease and has allowed comparison of results achieved by different investigators for more than two decades. Important modifications of the Ann Arbor system were developed at the Cotswold Conference in 1989 (Table 10-3) (¹⁴). Since then, the staging evaluation has been changing. A recent recommendation for the staging procedure in lymphoma described the importance of positron emission tomography (PET)–computed tomography (CT) scan for fluorodeoxyglucose (FDG)-avid lymphomas, of which HL is one example (¹⁵). For HL and other FDG-avid lymphomas, PET-CT improves the accuracy of staging compared to CT scans for nodal and extranodal sites (¹⁶). Positron emission tomography–CT leads to change in stage in 10% to 30% of patients, more often upstaging, although alteration in treatment occurs in fewer patients with no demonstrated impact on overall outcome. However, PET-CT is critical as a baseline measurement before therapy to increase the accuracy of subsequent response assessment (¹⁷). In addition, contrast-enhanced CT scan should be included for a more accurate measurement of nodal size if required for clinical trials.

Patient Evaluation

The initial evaluation of patients with HL has both prognostic and therapeutic significance (see Fig. 10-8). Routine studies that should be performed include a complete blood cell count with differential, electrolytes, blood urea nitrogen (BUN), creatinine, liver function tests, lactate dehydrogenase (LDH), albumin, pregnancy test in women of childbearing age, erythrocyte sedimentation rate (ESR), pulmonary function test (PFT) with carbon monoxide diffusing capacity (DLCO), evaluation of cardiac ejection fraction, chest x-ray, CT of neck, chest, abdomen, and pelvis, and PET-CT (Table 10-4).

Bone marrow biopsy has been standard in lymphoma staging. However, the high sensitivity of PET-CT for bone marrow involvement has recently led to questioning the use of bone marrow biopsy as a staging procedure in several common histologies, including HL (¹⁵). In one study in HL, 18% of patients had focal skeletal lesion on PET-CT, but only 6% had positive bone marrow biopsy, all with advanced-stage disease on PET-CT (¹⁸). Patients with early-stage disease rarely have marrow involvement in the absence of a suggestive PET finding, and those with advanced-stage disease rarely have marrow involvement in the absence of disease-related symptoms. Although the issue is controversial and some institutions still perform bone marrow biopsy for the initial staging evaluation, almost all patients would not have been allocated to another treatment based on bone marrow biopsy results. Thus, the recommendation states that bone marrow biopsy is no longer indicated for the routine staging of HL.

Magnetic resonance imaging (MRI) has not superseded CT scanning of the chest and abdomen in the evaluation of HL. It is largely restricted to the assessment of specific situations such as bony involvement and spinal cord compression as well as in lieu of CT scans in pregnant patients.

Prognostic Factors

In HL patients at a low clinical stage (CS)—that is, CS I or CS II—several prognostic factors, based largely on patients treated only with radiotherapy, have been identified through retrospective studies. Adverse factors are: (1) advanced age, which correlates with the presence of occult abdominal disease and with poor results of salvage therapy; (2) male sex; (3) MC histologic type, which is associated with the presence of occult abdominal disease; (4) B symptoms, also associated with the presence of occult abdominal disease; (5) large mediastinal mass, defined as a mass measuring greater than one-third the chest diameter on a standard chest radiograph; (6) a larger number of involved nodal regions; (7) an elevated ESR; (8) anemia; and (9) a low serum albumin level (^19,20).

International organizations have defined various systems that calculate the risk of recurrence of disease or, in some cases, death, after treatment for HL. The European Organization for the Research and Treatment of Cancer (EORTC) has defined CS I and CS II patients as having an unfavorable risk of development of recurrence if any of the following factors apply: (1) age >50 years, (2) no symptoms present with ESR >50 mm/h or B symptoms with ESR >30 mm/h, (3) large mediastinal mass, (4) stage II, or (5) at least four nodal regions involved (²¹). The GHSG has assigned CS I and CS II patients to the category of unfavorable disease with any of the following adverse factors: (1) large mediastinal mass, (2) at least three nodal regions involved, (3) no symptoms present with ESR >50 mm/h or B symptoms with ESR >30 mm/h, or (4) localized extranodal infiltration (so-called E lesions) (Table 10-5) (²²). In advanced disease, the International Prognostic Score (IPS) was developed on the basis of an analysis of 5,141 patients most of whom were initially treated with an anthracycline-containing chemotherapy regimen. Seven factors were identified, as shown in Table 10-6 (²³).

Response Assessment

Prior to 1999, the criteria used to assess response to therapy were not routinely standardized. The International Working Group (IWG) formulated guidelines for the assessment of response to therapy in 1999 (²⁴). The criteria were based on CT scan and have been used internationally. With the introduction of the PET scan, the guideline has been updated twice, in 2007 and in 2014 (^15,25). Based on the high negative predictive value (95%-100%) and positive predictive value (>90%) of PET scan in the response evaluation for HL (²⁶), current recommendations for response evaluation clearly state that PET-CT is more accurate than CT for end-of-treatment assessment. Previous guidelines for reviewing PET scan were based on imprecise visual interpretation, whether the scan is positive or negative, and whether it is intended for end-of-treatment evaluation using mediastinal blood pool as the comparator (²⁷). More recent guidelines recommend using a 5-point scale assessment (Deauville criteria, Table 10-7) for clinical trials including interim analysis and end-of-treatment assessment. A score of 1 or 2 is considered to represent complete metabolic response. A score of 4 or 5 is considered to be treatment failure at the end-of-treatment evaluation. There are difficulties in the interpretation of a score of 3, in which the uptake is higher than mediastinum but less than or equivalent to liver. Recent data suggest that most patients with a score of 3 have good prognosis at the end of treatment (²⁸). However, in response-adapted trials exploring treatment de-escalation, a more cautious approach may be preferred.

Treatment of Hodgkin Lymphoma

Nodular Lymphocyte-Predominant Hodgkin Lymphoma

Because of the rarity of this disease, it is difficult to derive the information by randomized prospective clinical trials. Recently, several well-designed single-arm phase II trials and large retrospective analyses have been reported.

Early-Stage Disease

Although radiation as a single modality for treatment would be considered inferior treatment for patients with early-stage cHL, multiple studies have observed excellent outcomes using radiation therapy (RT) alone for early-stage NLPHL.

In the retrospective review by the GHSG on the HD-4 and HD-7 trials, the 2-year freedom from treatment failure (FFTF) and overall survival (OS) rates were 92% and 100% respectively, with involved-field RT (IFRT), compared with 100% and 94% respectively, for extended-field RT (EFRT) (²⁹). Also, our center (MDACC) reported excellent outcomes with RT alone for stage IA and IIA patients (³⁰). With a median follow-up of 8.8 years, only 1 of 20 patients who received limited-field RT experienced relapse. The best outcome was noted in stage IA patients, who had a 5-year relapse-free survival rate of 95%. The Harvard study group reported a retrospective analysis of long-term outcomes of 113 patients with early-stage NLPHL (³¹). Ten-year progression-free survival (PFS) and OS rates were 64% and 100%, respectively, with limited-field RT, and 81% and 95%, respectively, with EFRT. Of note, 86% of patients who received chemotherapy alone had relapse of disease.

These observations indicate that: (1) chemotherapy alone is not indicated for NLPHL, (2) RT alone would be accepted as the standard of management for early-stage NLPHL without bulky disease or B symptoms, and (3) limited-field RT is appropriate to reduce the toxicity and mortality. With these retrospective studies, no improvement was seen with combined-modality treatment (chemotherapy and RT) compared with RT alone. However, the data from the British Columbia Cancer Agency (BCCA) have suggested a potential improvement in the outcomes for adding a brief course of ABVD (doxorubicin, bleomycin, vinblastine, and dacarbazine) before RT in patients with early-stage NLPHL (³²). Ten-year PFS and OS rates were 65% and 84%, respectively, for RT alone, and 91% and 93%, respectively, for combined-modality treatment. As with retrospective studies, cautious interpretation is needed because of possible selection bias, variable staging procedures, availability of supportive care, and differences in duration of follow-up for the different treatments.

Because of high CD20 expression in NLPHL, rituximab monotherapy was evaluated for the treatment of early-stage NLPHL. Two prospective studies have been reported by the GHSG and the Stanford group (^33,34). Overall response rates (ORRs) were high (100% in both studies). However, the responses were not durable. Currently, the National Comprehensive Cancer Network (NCCN) guidelines recommend IFRT alone for early-stage NLPHL without B symptom. B symptom and bulky disease are uncommon presentations for NLPHL and would be treated with combined-modality treatment as for cHL.

Advanced-Stage Disease

Because at least 70% to 80% of patients with NLPHL are diagnosed with early-stage disease, defining the optimal treatment regimen for advanced-stage disease is challenging. Chemotherapy is the mainstay of treatment for advanced-stage disease.

The GHSG compared the outcome of patients with NLPHL and cHL enrolled in prospective trials (⁷). There were no significant differences in FFTF between NLPHL and cHL, with 50-month FFTF rates of 77% and 75%, respectively. Of note, the chemotherapy regimens used in the GHSG trials were COPP (cyclophosphamide, vincristine, procarbazine, and prednisone), COPP/ABVD, and BEACOPP (bleomycin, etoposide, doxorubicin, cyclophosphamide, vincristine, procarbazine, and prednisone), which contain higher dose of alkylating agents than ABVD. The BCCA reported a matched-control analysis of patients with NLPHL and cHL treated with ABVD or ABVD-like chemotherapy (³⁵). Although not statistically significant, there was a trend toward an inferior PFS for patients with NLPHL versus cHL (44% vs 77% at 15 years; P = 0.096). These studies have suggested that alkylating agents may provide some therapeutic advantage. We have reported the results of R-CHOP (rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone) in patients with advanced-stage NLPHL (³⁶). The ORR with R-CHOP was 100%, with a complete response (CR) rate of 90%, and no relapses or transformations have been observed at a median follow-up of 42 months.

Currently, NCCN guidelines list therapeutic options such as CVP (cyclophosphamide, vincristine, and prednisone), CHOP, and ABVD with or without rituximab. The standard approach to patients with advanced-stage NLPHL at MDACC is R-CHOP based on our data.

Relapsed and Transformed Disease

Patients with NLPHL may have late relapse or transform to B-cell lymphoma, for which standard treatment is not well defined. Rituximab has been evaluated for treatment of relapsed NLPHL. In a study by the GHSG that enrolled 14 patients, rituximab therapy resulted in an ORR of 100%, a CR rate of 57%, and a median time to progression of 33 months (³⁷). The Stanford group examined the benefit of limited versus extended rituximab therapy in the frontline and relapsed settings (³³). Eighteen patients with relapsed NLPHL were enrolled in the study. The ORR with rituximab monotherapy was 100%, and the 5-year PFS was 71.4% with rituximab maintenance therapy of 4 weekly infusions every 6 months for 2 years. These results indicate that rituximab monotherapy is effective in relapsed NLPHL.

Transformation at time of relapse can also occur. In a retrospective study by the BCCA, 95 patients were identified as diagnosed with NLPHL over a 40-year time period (³⁸). Median time of follow-up was 6.5 years, and 14% of patients experienced transformation. Median time to transformation was 8.1 years, with 4:1 ratio of DLBCL to TCRBCL. In the 10 patients with transformed lymphoma, the 10-year PFS and OS rates were 52% and 62%, respectively.

The rarity of the disease makes it difficult to prospectively evaluate the role of autologous stem cell transplantation (ASCT) for patients with relapsed or refractory NLPHL. However, patients who relapse with transformation should be managed according to algorithms for DLBCL. An MDACC retrospective study reviewed the outcomes for 26 patients who underwent ASCT. At time of transplantation, many had transformation to TCRBCL. At time of ASCT, 85% were in remission, with 35% in CR. At a median follow-up of 50 months, the event-free survival (EFS) rate was 69% (³⁹).

MD Anderson Approach

We treat stage IA and IIA LPHL patients with IFRT. It is rare for a stage I or II patient to present with B symptoms, but if a patient does, we treat the patient with (particularly for stage IIB) combined-modality therapy with an anthracycline-containing chemotherapy regimen followed by IFRT. Our preferred regimen is R-CHOP. For advanced-stage patients, we treat with R-CHOP for six cycles. Patients who relapse can be considered for extended rituximab therapy. For patients with evidence of transformation to DLBCL or TCRBCL, if anthracycline-containing chemotherapy had already been given, we use salvage chemotherapy with regimens such as rituximab plus ICE (ifosfamide, carboplatin, and etoposide) followed by ASCT.

Classical Hodgkin Lymphoma

The common practice for treatment and participation in clinical trials is to divide cHL patients into three treatment groups: early stage favorable, early stage unfavorable, and advanced stage.

Early-Stage Favorable Hodgkin Lymphoma

Treatment of early-stage favorable HL is evolving. Historically, wide-field RT or EFRT without chemotherapy was the standard of care (⁴⁰). Extended-field RT produced superior disease-free survival (DFS) compared with IFRT (⁴¹). More than 90% of patients achieved CR with this approach; however, the relapse rate was unacceptably high (≥30%). In addition, EFRT had considerable long-term side effects. In a large prospective analysis of over 15,000 HL patients, the actuarial risk of developing a solid tumor was 21.9% at 25 years after HL diagnosis, with the absolute risk being nearly 50%. Common secondary solid tumors were female breast and lung cancers (⁴²). The key studies comparing RT alone with combined-modality strategies were conducted by the GHSG and EORTC. In the GHSG HD-7 trial, patients were randomly assigned to receive either 30 Gy of EFRT alone or two cycles of ABVD followed by the same RT (⁴³). Although, response rates did not differ between the two treatment arms, the 7-year FFTF rate was significantly better in the combined-modality arm (88% vs 67%). The results of the randomized EORTC H8F trial were similar. Treatment arms consisted of three cycles of MOPP (mechlorethamine, vincristine, procarbazine, prednisone)/ABV (doxorubicin, bleomycin, vinblastine) followed by IFRT or subtotal nodal irradiation (STNI) alone (⁴⁴). Patients receiving combined-modality treatment had a significantly superior 5-year EFS (98% vs 74%) and better 10-year OS estimates (97% vs 92%). As a result of these two large randomized controlled trials and the recognition of notable long-term side effects and high relapse rates, EFRT monotherapy has now been abandoned in favor of combined-modality therapy, which is now the standard treatment for early-stage HL.

Combined-modality therapy has evolved based on the premise that this approach results in high freedom from recurrence in early-stage HL and that efficacy can be maintained using less toxic chemotherapy and RT regimens. At MDACC, investigators performed a retrospective analysis of 286 patients with early-stage HL treated with chemotherapy followed by IFRT or EFRT with a median dose of 40 Gy (⁴⁵). Five-year relapse-free survival (RFS) and OS rates were 88% and 93%, respectively. The type and number of chemotherapy cycles used did not significantly affect RFS and OS. However, the 5-, 10-, and 15-year cumulative risks of developing solid tumors in patients treated with chemotherapy and IFRT were 0%, 6.9%, and 11.4%, respectively. These results were strikingly more favorable than those of chemotherapy plus EFRT (2.7%, 11.1%, and 28.7%, respectively).

There are many completed and ongoing trials addressing issues of the best modality, best RT field, optimal dose of RT, optimal combination of drugs, number of cycles, and optimal timing of chemotherapy, with the goals being to maintain efficacy and minimize toxicities (^{22,46,47,48,49,50}) (Table 10-8).

The key study in combined-modality therapy for the current standard treatment is the HD-10 trial by the GHSG (²²). The GHSG HD-10 trial had four arms testing two versus four cycles of ABVD followed by 20 versus 30 Gy of IFRT in patients with favorable early-stage HL. This trial addressed both the optimal dose of RT and the optimal number of cycles of chemotherapy. The ABVD two- and four-cycle arms both had CR rates of 97%. The 20- and 30-Gy IFRT groups had CR rates of 97% and 98%, respectively. With a median follow-up of 7.5 years, there were no differences among the four groups in PFS, FFTF, and OS. The four-cycle ABVD and 30-Gy IFRT treatment groups had more toxicity than the less intensive treatment groups. Based on these data, the least toxic regimen, two cycles of ABVD and 20 Gy of IFRT, is the current standard approach for favorable early-stage HL.

In the next trial by the GHSG, the HD-13 trial, the aim was to determine whether bleomycin or dacarbazine can be omitted from chemotherapy (⁴⁷). This four-arm trial investigated ABVD, AVD (doxorubicin, vinblastine, dacarbazine), ABV, and AV (doxorubicin, vinblastine) plus 30 Gy of IFRT. In this trial, the ABV and AV plus IFRT arms were closed because of concern for higher relapses. Five-year FFTF rates were 93%, 81%, 89%, and 77% with ABVD, ABV, AVD, and AV, respectively. Based on this trial, both dacarbazine and bleomycin cannot be omitted from ABVD without a substantial loss of efficacy. The standard treatment for patients with early-stage favorable HL should remain ABVD followed by IFRT.

Recently, several studies have evaluated the use of interim PET scan for treatment stratification. The EORTC/Lymphoma Study Association (LYSA)/Fondazione Italiana Linfomi (FIL) H10 trial was conducted to assess whether involved-node radiotherapy (INRT) could be omitted without compromising PFS in patients attaining a negative early PET scan after two cycles of ABVD as compared with standard combined-modality treatment (⁴⁶). The patients were randomized to a standard treatment giving RT irrespective of PET status after two cycles of ABVD or to an experimental arm that omitted RT if the PET was negative after two cycles of ABVD. Patients with a positive interim PET continued treatment with two cycles of escalated BEACOPP. The chemotherapy-only arm (four cycles of ABVD) was closed due to an increased number of events, and all patients with a negative PET received additional RT. Although the median follow-up time was very short (1.1 years), 1-year PFS was significantly lower in the experimental arm than the standard arm (94.9% vs 100.0%). In contrast, the UK RAPID trial showed noninferior outcome for patients who omitted RT after negative PET scan (⁵¹). In this trial, patients were randomized to IFRT or to the no further treatment arm if they had a negative PET scan after three cycles of ABVD. The 3-year PFS rates were 93.8% versus 90.7%, and the 3-year OS rates in an intent-to-treat analysis were 97.0% and 99.5% in patients who received IFRT and no further treatment, respectively. Thus, there was a trend toward improved PFS for patients who received IFRT. Another trial is ongoing by the GHSG. The randomized GHSG HD-16 trial resembles the H10 trial. Patients are randomized to a standard treatment arm or an experimental arm that omits RT if the PET scan is negative after two cycles of ABVD. The question of whether treatment can be further reduced based on the results of the PET scan is the subject of ongoing clinical trials.

The MD Anderson Approach

The treatment for favorable early-stage HL is still evolving. Patients are typically screened for clinical protocol options if available. As standard therapy, we use two cycles of ABVD plus 20 Gy of IFRT for this group of patients.

Early-Stage Unfavorable Hodgkin Lymphoma

Combined-modality approaches consisting of four cycles of chemotherapy followed by IFRT represent the standard of care for patients with early-stage unfavorable HL. Multiple trials have shown that reduction of radiation field does not lead to inferior clinical outcomes. In a retrospective analysis conducted at MDACC, 286 patients (1980-1995) with early-stage HL were treated with chemotherapy followed by IFRT or EFRT. The type and number of chemotherapy regimens used did not significantly affect RFS and OS. There was a trend toward higher risks of secondary tumors in the EFRT group (⁴⁵). In the EORTC H8U trial, three different regimens were randomly compared (⁴⁴). Patients were assigned to receive either six cycles of MOPP/ABV plus IFRT, four cycles of MOPP/ABV plus IFRT, or four cycles of MOPP/ABV plus subtotal nodal plus spleen irradiation. The MOPP/ABV regimen followed by IFRT resulted in 88% EFS at 5 years and 85% OS at 10 years with no difference noted compared to the other treatment arms. Thus, four cycles of chemotherapy is the standard for patients with early-stage unfavorable HL. Similar to early-stage favorable HL, ABVD alone was more effective than MOPP and equally as effective as, but less toxic than, the alternating regimen MOPP/ABVD (⁵²). Given the relapse rates with ABVD, there is interest in evaluating alternative more intensive regimens (^{46,53,54,55,56}) (Table 10-9).

To address whether ABVD or the Stanford V regimen (mechlorethamine, doxorubicin, vinblastine, vincristine, bleomycin, etoposide, prednisone) would be the best approach for patients with early-stage bulky unfavorable HL, the intergroup Southwest Oncology Group (SWOG)/Eastern Cooperative Oncology Group (ECOG) 2496 trial was conducted (⁵³). In this trial, patients with early-stage bulky unfavorable disease (they also included advanced-stage disease in the trial) were randomized to six cycles of ABVD followed by IFRT at 36 Gy to bulk greater than 5 cm versus 12 weeks of Stanford V followed by the same IFRT plan of care. In this patients group, there were no difference in ORR and FFS between ABVD and Stanford V.

To improve the tumor control in this patient group, the BEACOPP regimen was compared to ABVD in trials by both the GHSG and the EORTC. In the EORTC H9U trial, patients were randomized to receive either six cycles of ABVD followed by IFRT, four cycles of ABVD followed by IFRT, or four cycles of BEACOPP baseline followed by IFRT. All patients received 30 Gy of IFRT. At a median follow-up of 4 years, EFS and OS remain statistically equivalent in all arms, with EFS ranging from 87% to 91% and OS ranging from 93% to 95% (⁵⁶). Although the final results of this trial are pending, BEACOPP could not show a benefit over ABVD at the time of analysis. The GHSG HD-11 trial randomized patients to four arms of therapy and evaluated four cycles of ABVD followed by 30 versus 20 Gy of IFRT and compared outcomes to four cycles of BEACOPP baseline followed by 30 versus 20 Gy of IFRT. The FFTF with BEACOPP was superior in patients who received 20 Gy of IFRT, whereas there were no differences between BEACOPP and ABVD in patients who received 30 Gy of IFRT. Overall survival did not differ significantly between the four treatment arms. Thus, BEACOPP was not adopted as a standard chemotherapy regimen for patients with early-stage unfavorable HL due to increased toxicity observed in comparison with ABVD (⁵⁵).

The GHSG HD-14 trial introduced escalated BEACOPP to evaluate a more intensive regimen (⁵⁴). Patients received four cycles of ABVD followed by 30 Gy of IFRT or two cycles of BEACOPP escalated followed by two cycles of ABVD (2 + 2) and then 30 Gy of IFRT. At a median follow-up of 43 months, there was better tumor control (5-year FFTF estimate of 94.8%) with the 2 + 2 protocol, compared with the ABVD arm (5-year FFTF of 87.7%). There was no significant difference in OS between the two arms. Based on these trials, ABVD remains the standard chemotherapy for patients with early-stage unfavorable HL.

Similar to early-stage favorable HL, current trials for patients with early-stage unfavorable HL, such as the EORTC/Groupe d’Etude des Lymphomes de l’Adulte H10U and GHSG HD-17, are evaluating the treatment stratification according to the result of an interim PET scan. The standard arm in the EORTC/LYSA/FIL H10U trial consisted of four cycles of ABVD followed by 30 Gy of INRT irrespective of the result of an interim PET scan performed after the second cycle of ABVD. In the experimental arm, patients with a negative PET received a total of six cycles of ABVD without consolidation RT, whereas patients with a positive PET continued treatment with two cycles of escalated BEACOPP before receiving INRT. However, as for patients with early-stage favorable HL, the chemotherapy-only arm (six cycles of ABVD) was closed due to an increased number of events, so that all patients with a negative PET received additional RT. There was no difference in the 1-year PFS between the standard and experimental arms (97.3% vs 94.7%). In the GHSG HD-17 trial, all patients received chemotherapy according to the 2 + 2 regimen before a PET scan was performed. In the standard arm, patients received an additional 30 Gy of IFRT irrespective of the results of the PET scan. In the experimental arm, patients with a negative PET scan stopped treatment, whereas patients with a positive PET scan received 30 Gy of INRT. This ongoing trial plans to evaluate whether it is possible to spare RT in patients with a negative PET scan after intensive escalated BEACOPP.

A trial conducted by the National Cancer Institute of Canada (NCIC) and the ECOG indicated that chemotherapy-only approaches appear possible in patients with early unfavorable HL, at least in patients with nonbulky disease (⁵⁷). The trial randomized patients with early-stage unfavorable clinical features to receive either four to six cycles of ABVD or two cycles of ABVD followed by STNI. At a median follow-up of 11.3 years, freedom from disease progression was better in patients who receiving combined-modality treatment; however, OS was better for patients treated with chemotherapy alone. This was mainly caused by the increased number of deaths from secondary neoplasia among patients who had received combined-modality treatment. Nevertheless, we should keep in mind that STNI is outdated and no longer used. Chemotherapy alone might be a treatment option in patients with nonbulky early-stage unfavorable HL, but combined-modality therapy should remain standard until further data support that the chemotherapy-only approach is feasible.

The MD Anderson Approach

In summary, treatment with four cycles of ABVD plus 30 Gy of IFRT is presently a standard-of-care option for early-stage unfavorable HL. We screen patients for any available clinical protocols. If a patient has a bulky mediastinal mass of 10 cm or greater, we typically treat with six cycles of ABVD followed by IFRT.

Advanced-Stage Hodgkin Lymphoma

Treatment of advanced-stage HL usually consists of six to eight cycles of chemotherapy. The ABVD regimen was shown to be effective and less toxic than MOPP and MOPP/ABVD in a randomized clinical trial by the Cancer and Leukemia Group B (CALGB) (⁵²). With ABVD, MOPP, and MOPP/ABVD, 5-year failure-free survival rates were 61%, 50%, and 65%, and 5-year OS rates were 73%, 66%, and 75%, respectively. Based on the trial, the chemotherapy regimen most often used in the United States is ABVD. However, the GHSG has established escalated BEACOPP as a standard treatment for advanced-stage HL. Many trials are addressing whether one regimen may be more suitable for the treatment of advanced HL than another, and the issue has been the subject of ongoing debate for more than decade. Various chemotherapy regimens have been developed in an attempt to improve outcomes (^{26,52,58,59,60,61}) (Table 10-10).

The GHSG HD-9 trial, a three-arm randomized trial, evaluated four cycles of COPP/ABVD versus eight cycles of BEACOPP baseline versus eight cycles of BEACOPP escalated (^61,62). The BEACOPP escalated regimen showed significantly better survival than the other two arms. With BEACOPP escalated, COPP/ABVD, and BEACOPP baseline, 5-year FFTF rates were 87%, 69%, and 76%, respectively. The 5-year OS rates were 91%, 83%, and 88%, respectively.

The GHSG HD-12 trial investigated whether the number of cycles of BEACOPP escalated could be de-escalated by evaluating eight cycles of BEACOPP escalated versus four cycles of BEACOPP escalated plus four cycles of BEACOPP baseline (4 + 4) and what the potential added benefit of consolidation RT would be in treating sites of initial bulk or residual disease (⁶⁰). Severe toxicity and therapy-related death rates were similar in both arms, and the survival outcome was slightly inferior in the 4 + 4 regimen. Thus, the trial could not address how to decrease the toxicity while maintaining the efficacy of eight cycles of BEACOPP escalated.

The next GHSG trial also aimed to reduce treatment toxicity without compromising efficacy (²⁶). The trial evaluated the role of response evaluation based on PET scan in assessing the need for IFRT. Chemotherapy consisted of eight cycles of BEACOPP escalated, six cycles of BEACOPP escalated, or eight cycles of BEACOPP-14, a time-dense variant of the BEACOPP baseline protocol. Additional localized RT was only applied to patients who had PET-positive residual lymphoma larger than 2.5 cm at the end of chemotherapy. With eight cycles of BEACOPP escalated, six cycles of BEACOPP escalated, and eight cycles of BEACOPP-14, the 5-year FFTF rates were 85%, 89%, and 85%, respectively, and the 5-year OS rates were 92%, 95%, and 95%, respectively. The negative predictive value for the postchemotherapy PET scan so that IFRT to PET-negative lesions would be omitted was very high (94.1% at 12 months). This superiority for six cycles of BEACOPP escalated was mainly attributed to the lower rate of treatment-related adverse events and fewer deaths due to secondary neoplasia. Based on this trial, treatment with six cycles of BEACOPP escalated was adopted as a standard chemotherapy for advanced-stage HL by the GHSG.

The interim analysis of the most recent trial by the GHSG, HD-18, was presented (⁶³). In this trial, patients initially receive two cycles of BEACOPP escalated. Then, interim PET scan is performed, and patients are randomized. The standard arm consists of a total of six cycles of BEACOPP escalated irrespective of the result of the interim PET. In the experimental treatment arm, patients with a CR by interim PET scan are randomized to receive either four or two cycles of BEACOPP escalated. Patients with PET-positive residual disease after two cycles of chemotherapy are randomized to receive either a total of four additional cycles of BEACOPP escalated or rituximab plus BEACOPP escalated. There were no significant differences in the survival between the rituximab plus BEACOPP escalated and the BEACOPP escalated arms. The PFS in PET-positive patients receiving standard treatment with BEACOPP escalated was higher than expected, with a 3-year PFS of over 90%. In this trial, PET result after cycle 2 of therapy was not able to determine a high-risk patient group.

In the SWOG S0816 trial, patients with stage III or IV disease underwent a baseline PET scan (⁶⁴). They then received two cycles of ABVD, and the PET scan (PET-2) was repeated. If the PET-2 scan was negative, four further cycles of ABVD were given. If the PET-2 scan was positive, treatment was changed and intensified to BEACOPP escalated for six cycles for patients who were human immunodeficiency virus (HIV) negative and to BEACOPP standard for six cycles for patients who were HIV positive. This trial was also the first American study to use centralized real-time intergroup review (SWOG, ECOG, CALGB) of the PET scan results for treatment decisions. Response-adapted therapy with centralized interim PET review was highly feasible, even in an intergroup setting. However, the 2-year PFS of PET-2–positive patients was still lower than that of PET-2–negative patients of (61% vs 79%), even though they received six more cycles of BEACOPP escalated after PET-2.

The Gruppo Italiano Terapie Innovative nei Linfomie (GITIL) conducted a similar trial (⁶⁵). In GITIL HD0607, patients started with two cycles of ABVD chemotherapy, and then the PET-2 scan was performed. If the patients were PET negative, they received another four cycles of ABVD with or without RT. If the patients were PET positive, they received four cycles of BEACOPP escalated and two cycles of BEACOPP baseline. The 1-year PFS rates of patients with PET-2–positive and PET-2–negative scans were 81% and 95%, respectively.

At this point, response-adapted therapy based on interim PET scan should be performed in well-designed clinical trials, and longer follow-up data of the completed trials are essential.

Even with the results of the GHSG HD-9 trial, first-line chemotherapy for advanced-stage HL is still a matter of debate. The standard arm used in the HD-9 trial was COPP/ABVD, not ABVD alone. Three randomized trials have been conducted to address this issue. One Italian group conducted the HD2000 trial and LYSA conducted the H34 trial to compare the outcome between BEACOPP and ABVD (^58,66). In the HD2000 trial, patients were randomly assigned to receive six cycles of ABVD, four cycles of BEACOPP escalated plus two cycles of BEACOPP baseline, or six courses of CEC (cyclophosphamide, lomustine, vindesine, melphalan, prednisone, epidoxorubicin, vincristine, procarbazine, vinblastine, and bleomycin). Patients who received BEACOPP had higher PFS and OS rates than patients who received ABVD (5-year PFS, 81% vs 68%; 5-year OS, 92% vs 84%). In the H34 trial, patients were randomly assigned to receive eight cycles of ABVD or four cycles of BEACOPP escalated plus four cycles of BEACOPP baseline. Both PFS and OS were higher in the BEACOPP arm than the ABVD arm (5-year PFS, 93% vs 75%; 5-year OS, 99% vs 92%). These two trials showed a higher efficacy with BEACOPP than ABVD.

However, GITIL reported that ABVD has a similar efficacy to BEACOPP if high-dose chemotherapy (HDCT) is planned at the time of relapse of refractory disease (⁵⁹). In the GITIL trial, patients were randomly assigned to either four cycles of BEACOPP escalated plus four cycles of BEACOPP baseline or to six to eight cycles of ABVD, each followed by local RT when indicated. Patients with residual or progressive disease after the initial therapy were to be treated with high-dose salvage therapy with ASCT. The 7-year FFTF was significantly better with BEACOPP than ABVD (85% vs 73%); however, there was no significant difference in the 7-year OS between arms after completion of the overall planned treatment (89% vs 84%). Although two out of three randomized trials showed some benefit in survival with BEACOPP, longer follow-up is essential to confirm the conclusion because the toxicities such as secondary malignancy would be an issue for the long-term survival in young patients. Most of the institutes in the United States are still using ABVD as first-line chemotherapy, mostly because of its high efficacy, high tolerability, and lower toxicities compared with BEACOPP escalated (^66a).

The MD Anderson Approach

We screen patients for available protocols for initial treatment of advanced-stage disease. As a standard approach off clinical protocol, we treat these patients with six to eight cycles of ABVD. Although the data supporting IFRT for advanced-stage disease is controversial, we sometimes consider IFRT for patients who have presented with an initial bulky mass and who continue to have a residual mass at the end of therapy with PET-negative status.

The Value of Positron Emission Tomography Scan in Hodgkin Lymphoma

A PET scan is useful not only for staging but also for response evaluation and evaluation of expected outcome in patients with HL. An interim PET scan obtained after two cycles of therapy (PET-2) was a stronger prediction of outcome than the IPS, with 2-year PFS for patients with a positive PET-2 of 13% compared with 95% for those with a negative PET-2 (⁶⁷). The PET-2 also strongly predicts treatment failure (⁶⁸). In a meta-analysis, a positive PET-2 in low-intermediate–risk advanced HL patients was a reliable predictor of poor response (⁶⁹). The value of the interim PET scan is now being evaluated in prospective clinical trials. The SWOG S0816 phase II intergroup trial is evaluating interim PET in stage III or IV cHL patients treated with two cycles of ABVD. Patients who have a negative PET receive four additional cycles of ABVD; PET-positive patients receive BEACOPP baseline if HIV positive and BEACOPP escalated if HIV negative.

The HD-15 trial showed a negative predictive value of 94% for PET after BEACOPP-based therapy in advanced-stage HL (⁷⁰). The PFS at 12 months was 96% for PET-negative patients and 86% for PET-positive patients. At the time of posttreatment examination, PET has higher diagnostic and prognostic value than conventional CT (^71,72,73). The role of PET prior to transplantation was also evaluated: a negative PET prior to ASCT is significantly associated with higher EFS. In a study by the Memorial Sloan-Kettering Cancer Center, the 5-year EFS was 80% in patients with PET-negative status and 40% in patients with PET-positive status before transplantation (⁷⁴).

Refractory or Relapsed Hodgkin Lymphoma

Although many patients with HL are cured with frontline therapy, 10% to 15% of patients with early-stage disease with unfavorable risk factors and 40% of patients with advanced-stage disease with high-risk factors can develop relapse or refractory disease.

Relapsed HL can be divided into three subgroups: early relapse within 12 months of CR after first-line chemotherapy; late relapse after CR >12 months after first-line chemotherapy; and primary refractory HL (ie, patients who never achieve a CR). Moskowitz et al identified the following three prognostic factors associated with EFS in patients receiving ICE, followed by HDCT and ASCT: CR less than 1 year, extranodal disease, and presence of B symptoms. The 5-year EFS was 83% in patients with zero to one factor compared with 10% if all three factors were present (⁷⁵).

For patients with relapsed or refractory disease after standard frontline management, additional salvage chemotherapy followed by HDCT plus ASCT is the standard approach. One of the key goals of salvage chemotherapy is to achieve CR prior to ASCT. The response rates of multiple salvage regimens are listed in Table 10-11. It is difficult to directly compare these regimens because they have not been evaluated in randomized clinical trials.

Although we screen all patients for available protocols at relapse, the most common salvage chemotherapies outside clinical trials are the platinum-containing regimens such as ICE or DHAP (cisplatin, cytarabine, dexamethasone). With ICE, the ORR was 84% and the CR rate was 26%. The DHAP regimen showed similar results, with ORR of 89% and CR rate of 21%. Gemcitabine-containing regimens are also effective. With GND (gemcitabine, vinorelbine, pegylated liposomal doxorubicin), the ORR was 70%, with a CR rate of 19%. With GDP (gemcitabine, dexamethasone, platinum), the ORR was 62%, with a CR rate of 10%.

High-Dose Chemotherapy With Autologous Stem Cell Transplantation for Relapsed Hodgkin Lymphoma

For patients with chemotherapy-sensitive disease, the treatment of choice after relapse is HDCT followed by ASCT. This recommendation is based on reports from two randomized clinical trials (^76,77). In the BNLI study, patients with relapsed or refractory HL received BEAM (carmustine, etoposide, cytarabine, and melphalan) at high doses followed by an ASCT or at lower doses (mini-BEAM) without an ASCT. The 3-year freedom from second treatment failure was significantly better for patients who received HDCT (53% vs 10%). The GHSG/European Group for Blood and Marrow Transplantation (EBMT) randomized trial compared four cycles of Dexa-BEAM (dexamethasone plus BEAM) versus two cycles of Dexa-BEAM followed by ASCT. At 3 years, the FFTF in the high-dose therapy group was 55% versus 34% with four cycles of chemotherapy.

Multiple investigators have shown that response to salvage chemotherapy is a strong predictor of long-term outcome after ASCT. The 5-year OS for patients who were in CR at the time of ASCT was 79% compared with 59% for those in PR and 17% for those with resistant disease at the time of ASCT (⁷⁸). Studies have shown the impact of pre-ASCT PET scan results on EFS. Patients with negative pre-ASCT PET scans have significantly higher EFS and failure-free survival rates compared to patients with positive pre-ASCT PET scans (^79,80). A European intergroup evaluated a dose-intensified regimen before ASCT (⁸¹). Patients were randomly assigned after two cycles of DHAP to ASCT or sequential cyclophosphamide, methotrexate, and etoposide before ASCT. There were no significant differences between the two treatment arms in terms of mortality, FFTF, and OS. Thus, the less toxic approach consisting of two cycles of DHAP (or other salvage regimen such as ICE) followed by HDCT and ASCT remains the standard of care for patients with relapsed HL.

Treatment of Relapse After Autologous Stem Cell Transplantation

Patients with disease progression after ASCT uniformly have a poor outcome. In a study of HL patients who failed ASCT, the median time to progression after the next therapy was only 3.8 months, and the median survival after ASCT failure was 26 months (⁸²). An international multicenter retrospective study showed that the survival of patients who relapsed after an ASCT did not improve from 1981 to 2007 (⁸³). However, there has been a major advance in the treatment of relapsed or refractory HL in the last 5 years.

Brentuximab Vedotin

CD30 was considered an ideal target for monoclonal antibody therapy for HL, because its expression is highly restricted to the HRS cells. Brentuximab vedotin (BV), or SGN-35, is an intravenously administered antibody-drug conjugate that consists of the CD30-specific monoclonal antibody conjugated with monomethyl auristatin E (MMAE) by linker peptide. Binding of the antibody-drug conjugate to CD30 on the cell surface causes internalization of the drug by endocytosis, and the drug subsequently travels to the lysosome, where proteases cleave the linker and release MMAE to the cytosol (⁸⁴). Released MMAE binds to tubulin and disrupts the microtubule polymerization, resulting in cell cycle arrest and apoptotic death of CD30-expressing cells. After efficacy was shown in a phase I trial including 45 patients with relapsed or refractory CD30-positive hematologic malignancies, a pivotal phase II study with 102 patients with HL who had relapsed after HDCT and ASCT was conducted (^85,86). Patients received BV 1.8 mg/kg every 3 weeks up to a maximum of 16 cycles. The ORR was 75%, with a CR rate of 34%. These data led to the first drug approval by the US Food and Drug Administration (FDA) for the treatment of HL in more than 30 years. Durable remission was reported with longer follow-up (⁸⁷), and the median OS and PFS were 40.5 months and 9.3 months, respectively. The 3-year PFS rate of patients who achieved CR with BV was 58%. This survival outcome is notable considering that the patients enrolled in this trial had disease progression after ASCT.

Achieving CR before ASCT is the key to better outcomes in patients with relapsed or refractory HL. Therefore, BV is often used as a third-line therapy in patients who have not achieved CR after second-line salvage chemotherapy such as ICE. The Seattle group retrospectively evaluated the efficacy of BV in patients who were refractory to platinum-based salvage chemotherapy (⁸⁸). Fifteen patients who had PET-positive disease after platinum-based salvage therapy were treated with a median of four cycles of BV. Normalization of PET scan occurred in 8 (53%) of 15 patients but was only observed in patients who had achieved partial remission or stable disease after salvage therapy. This suggests that BV can achieve PET-CR in a considerable subset of patients with platinum-refractory HL prior to ASCT.

BV is also effective in patients who relapse after allogeneic stem cell transplantation (allo-SCT). Twenty-five patients who relapsed after allo-SCT received BV. The ORR and CR rates were 50% and 38%, respectively. The median PFS was 7.8 months, and the median OS was not reached.

Many clinical trials are ongoing to evaluate BV in various settings; these include as salvage combinations with chemotherapy prior to ASCT, as initial therapy in combinations with chemotherapy and as maintenance therapy after ASCT for high-risk patients.

A phase II study evaluating single-agent BV and augmented ICE salvage therapy prior to ASCT was conducted. Patients received BV for two cycles, followed by PET. Patients who achieved PET-CR proceeded to ASCT. Patients who failed to achieve PET-CR received two cycles of augmented ICE prior to consideration for ASCT. Preliminary results showed that among 28 patients who underwent ASCT, 9 patients (32%) achieved PET-CR with two cycles of BV (⁸⁹). Maintenance therapy with BV after ASCT was evaluated in a placebo-controlled randomized phase III study (AETHERA) (⁹⁰). Patients were enrolled in this study if they were (1) refractory to frontline therapy, (2) had relapse <12 months after frontline therapy, or (3) had relapse ≥12 months after frontline therapy with extranodal disease. The median PFS was 43 months with BV and 24 months with placebo. This represents a significant 43% reduction in the risk of disease progression with BV. Once finalized, these results will potentially change the standard treatment of high-risk patients who relapse after first-line chemotherapy.

A phase III trial comparing BV plus AVD (ABVD without bleomycin) versus standard ABVD in patients with newly diagnosed advanced-stage HL is ongoing (NCT01712490). A phase I study comparing BV in combination with ABVD or AVD treatment showed a high CR rate of 96%. However, BV combined with bleomycin resulted in a high rate of pulmonary toxicity (44%) (⁹¹). Based on this phase I trial, BV is combined with AVD. This BV+AVD combination, if superior to ABVD, may change the standard of care in patients with newly diagnosed advanced-stage HL.

Gemcitabine

Gemcitabine-containing regimens are effective. A phase II study of single-agent gemcitabine, 200 mg/m² given on days 1, 8, and 15 of a 28-day schedule, showed an ORR of 43% with a CR rate of 14% (⁹²). The GVD regimen (gemcitabine, vinorelbine, and pegylated liposomal doxorubicin) was evaluated by the CALGB in 91 patients with relapsed or refractory HL. The ORR was 70%, with a CR rate of 19% (⁹³). The 4-year PFS and OS rates in transplant-naive patients treated with GVD followed by ASCT were 52% and 70%, respectively. In patients in whom prior transplant failed, the 4-year DFS and OS rates were 10% and 34%, respectively. The GDP regimen produced similar results, with an ORR of 62% and a CR rate of 10%. A combination regimen named IGEV (ifosfamide, gemcitabine, vinorelbine) was evaluated in 91 patients and produced an ORR of 81% with a high CR rate of 54%; 60% of primary refractory patients responded to IGEV (⁹⁴).

Bendamustine

Bendamustine is a bifunctional alkylating agent derived from 2-chloroethylamine that had been a standard chemotherapy for indolent lymphoma (follicular lymphoma and mantle cell lymphoma) (^95,96). The Memorial Sloan-Kettering Cancer Center conducted a phase II trial of bendamustine in patients with HL who relapsed after ASCT or who were not eligible for ASCT (⁹⁷). Patients received bendamustine of 120 mg/m² on days 1 and 2 of a 28-day cycle, for six planned cycles. The ORR was 53%, including a 33% CR rate; the median PFS was 5.2 months. Preliminary data of a phase I/II study for the combination of bendamustine and BV for relapsed/refractory transplant-naïve patients showed an ORR of 94% with CR rate of 82% (⁹⁸). At the time of report, 20 of 34 patients who had a response to this combination had undergone ASCT.

Allogeneic Stem Cell Transplantation

The main advantage of an allo-SCT is its graft-versus-HL effect. Retrospective studies have shown this benefit by documenting lower relapse rates in allo-SCT patients who have chronic graft-versus-host disease (GVHD) and by showing that donor lymphocyte infusion (DLI) can induce relatively long-lasting remissions (⁹⁹). Initial studies of allo-SCT in HL patients described high rates of nonrelapse mortality (NRM), up to 61%. More recent studies evaluated reduced-intensity conditioning (RIC) and have shown decreases in treatment-related mortality (TRM). Overall, RIC allo-SCT induces modest long-term remissions with a PFS rate of 30% (^100,101).

The EBMT reviewed 168 patients who had undergone allo-SCT (¹⁰¹). Seventy-nine patients received myeloablative conditioning, and 89 patients received RIC. Fifty-two percent of patients had undergone a prior ASCT and 45% had chemosensitive disease. The NRM was significantly lower and OS was significantly better with RIC versus myeloablative conditioning. One-year NRM was 23%, and 5-year PFS and OS were 18% and 28%, respectively, in patients who received RIC.

At MDACC, we reviewed the outcomes of 58 patients who received RIC with fludarabine-melphalan in preparation for allo-SCT (¹⁰²). Overall, 83% of patients had undergone a prior ASCT and 52% had chemotherapy-sensitive disease at the time of allo-SCT. The TRM at 2 years was 15%, with nearly half of the TRM occurring within the first 100 days after allo-SCT. The incidence of chronic GVHD was 73%. The 2-year PFS and OS rates were 32% and 64%, respectively. There was a trend toward improvement in PFS for those with chemotherapy-sensitive disease but not for OS. Allogeneic stem cell transplantation is still an important option for eligible patients who relapse after ASCT.

Novel Agents

Advances in our understanding of HL pathology and biology have led to the development of promising targeted agents.

Programmed Death-1 Inhibitors

Nivolumab is a programmed death (PD)-1 immune checkpoint inhibitor antibody that selectively blocks the interaction between PD-1 and its ligands PD-L1 and PD-L2. This PD-1 pathway has a mechanism that normally leads to downregulation of cellular immune response. By inhibiting this interaction, nivolumab can enhance T-cell function, which may result in antitumor activity. Nivolumab has been evaluated in a phase II trial for patients with relapsed or refractory HL (¹⁰³). Twenty-three patients were treated; 78% had received ASCT and 78% had received BV before nivolumab. Nivolumab was given at a dose of 3 mg/kg every 2 weeks. The ORR was 87%, with a CR rate of 17%. The 6-month PFS rate was 86%. Nivolumab thus showed substantial therapeutic activity and an acceptable safety profile in patients with previously heavily treated relapsed or refractory HL. Nivolumab has been granted “Breakthrough Therapy Designation” by the FDA, and a pivotal trial is ongoing (NCT02181738).

Another PD-1 inhibitor, pembrolizumab, was evaluated in a phase IB trial in patients who had disease progression or relapse with BV (¹⁰⁴). Similar to nivolumab, a promising level of efficacy was seen, with an ORR of 53% and a CR rate of 20% at 12 weeks in 15 patients who were evaluable for response at the time of the preliminary report.

Histone Deacetylase Inhibitor

Histone deacetylases (HDACs) act on lysine amino acid groups on multiple proteins including many transcription factors. The HDACs are grouped into four classes, with classes I, II, and IV being zinc dependent. Several HDAC inhibitors are being investigated as therapies for relapsed or refractory HL. Panobinostat is an HDAC class I and II or pan-HDAC inhibitor and has been evaluated in a phase II trial in patients with relapsed or refractory HL after ASCT (¹⁰⁵). Patients received panobinostat 40 mg orally three times a week for 21-day cycles. For the 129 patients enrolled in the trial, the ORR was 21%, with a CR rate of 4%. Treatment was well tolerated; the most common grade 3 to 4 adverse event was thrombocytopenia. The median PFS was 6.1 months, and 1-year OS was 78%.

We have conducted a phase I/II randomized trial of ICE with or without panobinostat in patients with relapsed or refractory HL (NCT01169636). Preliminary results of the trial showed an ORR of 86% with a CR rate of 71%. All patients who achieved response were able to proceed to ASCT.

Mocetinostat, another HDAC inhibitor, was evaluated in a phase II trial in patients with relapsed or refractory HL. The initial dose of 110 mg orally three times a week for 28-day cycles was reduced to 85 mg because 70% of patients required dose reduction for toxicity. Among 51 patients treated, 60% had a reduction in tumor measurements, with 24% achieving partial response. Toxicities included thrombocytopenia, fatigue, pneumonia, anemia, and pericardial effusion.

Mammalian Target of Rapamycin Inhibitors

Everolimus is an oral agent that targets the mammalian target of rapamycin (mTOR) complex 1 (mTORC1). Hodgkin lymphoma cells have an activated PI3K pathway (upstream of mTOR) and may be susceptible to inhibition of this pathway. Everolimus was evaluated in a phase II trial in patients with relapsed HL (¹⁰⁶). Nineteen patients were enrolled, and 87% had received ASCT before everolimus. The ORR was 47%, with eight partial responses and one CR. The median time to progression was 7.2 months, with four responders remaining progression free at 12 months. Synergistic activity between everolimus and panobinostat was suggested by in vitro studies. We conducted a phase I trial of everolimus plus panobinostat in patients with relapsed or refractory lymphoma (¹⁰⁷). Among the 30 patients treated, 14 patients had HL. The ORR was 43%, with a CR rate of 15%. However, grade 3 to 4 thrombocytopenia was reported in 64% of patients, limiting the future development of this combination.

Several clinical trials evaluating the combination of mTOR inhibitors and other drugs are ongoing. Sirolimus was evaluated in combination with an HDAC inhibitor, vorinostat, in a phase I trial at MDACC (¹⁰⁸). The ORR was 57%, with a CR rate of 32% in heavily pretreated patients. Brentuximab vedotin will be evaluated in combination with mTOR inhibitors such as sirolimus, temsirolimus, and everolimus (NCT01902160, NCT02254239).

The MD Anderson Approach

Patients with relapsed or refractory HL are planned for second-line or salvage chemotherapy followed by an ASCT. We screen patients who have relapsed or refractory HL for current clinical trial options including our current randomized phase II clinical trial of panobinostat plus ICE (P-ICE) versus ICE. Patients who respond to salvage chemotherapy are planned to undergo ASCT. The role of BV for maintenance in patients with high risk of relapse after ASCT is an evolving topic. Based on the positive data recently presented for the AETHERA trial, this could become a standard of care. We screen patients with relapsed HL after an ASCT for novel agent clinical trial options. The preference is to treat with either BV if not previously given before ASCT or with a BV combination treatment on protocol including our planned BV plus dual mTORC1 inhibitor MLN0128 phase I trial. For patients who have early disease relapse or refractory disease after BV treatment, we consider novel agent protocols such as the PD-1 inhibitor trials. For patients not eligible or who do not wish to enroll in clinical trials, we consider chemotherapy regimens such as GND, bendamustine, or others. Given the benefits versus risks of allo-SCT, a subset of patients can be potentially considered for this approach, particularly otherwise healthy patients who achieve complete remission with addition therapies.

Although standard frontline chemotherapy with or without radiation therapy offers a high cure rate for cHL, approximately 20% of patients will develop refractory or relapsed disease. Thus, the challenge that remains is how to best develop strategies of therapy that increase the cure rates for the refractory/relapsed group of patients, while decreasing both short- and long-term toxicities, including secondary malignancies, for patients who are cured of their disease with current standard approaches. Huge recent achievements have occurred on both fronts, including a decrease of both the number of cycles of chemotherapy and the doses of RT for patients with early-stage disease, and introduction of novel therapeutics such as BV into frontline regimens paired with chemotherapy. Despite the fact that BV is a clear therapeutic advance, there remains a continued need to develop new therapies. Recently completed trials have shown significant promise for several new molecularly targeted therapeutic agents, with the PD-1 inhibitors demonstrating the ability to have significant efficacy even in patients with BV-resistant disease. Future directions are anticipated to increase the focus on evaluating the efficacy of combinations of targeted agents, with likely future trials exploring the potential activity of chemotherapy-free targeted treatment combinations in the frontline setting. Better understanding of the molecular biology of both cHL and NLPHL will also lead to more rationally designed novel agent trials and allow us to best select treatment strategies. The future of HL treatment has evolved significantly over the past decade, and these successes will only be significantly multiplied over the next decade to follow.