Advances in Biology, Diagnostics, and Treatment of Hodgkin s Disease

Transcription

1 Biology of Blood and Marrow Transplantation 12:66-76 (2006) 2006 American Society for Blood and Marrow Transplantation /06/ $32.00/0 doi: /j.bbmt Advances in Biology, Diagnostics, and Treatment of Hodgkin s Disease Ralf Küppers, 1 Joachim Yahalom, 2 Andreas Josting 3 1 Institute for Cell Biology, Tumor Research, University of Duisburg-Essen, Essen, Germany; 2 Memorial Sloan- Kettering Cancer Center, Department of Radiation Oncology, Cornell University, New York, New York; 3 First Department of Internal Medicine, University Hospital Cologne, Cologne, Germany Correspondence and reprint requests: Andreas Josting, MD, First Department of Internal Medicine, University Hospital Cologne, Joseph-Stelzmann-Str. 9, Cologne, Germany ( andreas.josting@uni-koeln.de). ABSTRACT Hodgkin and Reed-Sternberg cells are derived from germinal center B cells. Amplification of identical rearranged and mutated immunoglobulin genes from different Hodgkin and Reed-Sternberg cells from the same patient also answered the question of the malignant nature because the clonality the key criterion of malignancy of Hodgkin and Reed-Sternberg cells was hereby shown. In addition, it could be demonstrated that Hodgkin and Reed- Sternberg cells do not only expand clonally within 1 affected lymph node, but also clonally disseminate in advanced-stage disease and relapse even after clinical complete remission. Recent publications demonstrate that a probably small subset of Hodgkin disease exists with T-cell derivation of the respective Hodgkin and Reed- Sternberg cells. The management of Hodgkin s disease is undergoing a paradigm shift as a result of very effective drug regimens that are capable of inducing high remission rates, the use of combined chemoradiotherapy with involved-field irradiation in early stages, the introduction of effective salvage chemotherapy of relapsed Hodgkin s disease with peripheral stem cell transplantation, a better understanding of prognostic factors, economic constraints, and a more sensitive realization of the magnitude of late treatment mortality American Society for Blood and Marrow Transplantation KEY WORDS Hodgkin and Reed-Sternberg cells Hodgkin s disease Chemotherapy Radiotherapy INTRODUCTION Hodgkin s disease (HD) is one of the most frequent lymphomas, with an annual incidence rate of 3 to 4 new cases per persons in the Western world. The hallmark of HD is typical large Hodgkin and Reed- Sternberg (HRS) cells, which represent usually 1% of the cellular infiltrate in affected tissues. They are embedded in a microenvironment consisting of T cells, B cells, neutrophils, macrophages, and other cell types. In the World Health Organization lymphoma classification, 5 subtypes of HD are distinguished: namely, nodular sclerosis, mixed cellularity, lymphocyte-rich classical, lymphocyte depletion, and lymphocytepredominant HD. On the basis of differences in lymphoma histology and HRS cell morphology and immunophenotype, lymphocyte-predominant HD is considered as a distinct disease, whereas the other 4 subtypes are grouped together as classic HD. In lymphocyte-predominant HD, the tumor cells, which are called lymphocytic and histiocytic cells, carry rearranged immunoglobulin variable genes a hallmark of B lymphocytes and show an immunophenotype typical for mature B cells, thus demonstrating that they derive from B cells. It is important to note that the rearranged variable genes of the lymphocytic and histiocytic cells carry somatic mutations and often show evidence of ongoing mutational activity during tumor clone expansion. Such variable gene mutations normally happen in antigen-activated B cells participating in T cell dependent immune responses in the microenvironment of the germinal center by the process of somatic hypermutation. Therefore, lymphocytic and histiocytic cells are considered to originate from mutating germinal center B cells [1]. The HRS cells in classical HD show a phenotype that does not resemble any normal hematopoietic cell type. However, HRS cells from nearly all cases of classical HD carry clonal and mutated immunoglobulin gene rearrangements, thus demonstrating that also the HRS cells in classical HD represent transformed mature B cells. On the basis of the detection of destructive somatic mutations in approximately a quarter of cases of classical HD, it is discussed that the HRS cells 66

2 Advances in Hodgkin Disease derive from preapoptotic germinal center B cells [2]. A small fraction of classical HD, likely 2%, originate from T cells. Recently, gene expression profiles of 4 HD-derived cell lines were generated and compared with corresponding profiles from normal mature B cells and various types of B non-hodgkin s lymphoma. Focussing on the expression of B-lineage markers, the analysis revealed decreased messenger RNA levels for nearly all established B lineage specific genes [3]. These findings extended earlier studies, which had already shown downregulation of several B-cell markers, such as CD79, Oct-2, and the B cell receptor (BCR). The downregulation of most B cell specific molecules does not involve factors important for antigen presentation and interaction with T helper cells (eg, CD40, CD80, and major histocompatibility complex class II), thus suggesting that it is important for HRS cells to retain antigen-presenting capacities. Transforming Events Involved in HD Pathogenesis In approximately 40% of cases of classical HD in the Western world, the HRS cell are infected by Epstein-Barr virus (EBV). Because EBV can immortalize B cells in vitro and because the EBV-encoded latent membrane protein 1, which is expressed by EBV HRS cells, is known to have oncogenic potential, it is likely that EBV functions as a tumor virus in HD and is casally involved in the pathogenesis of EBV-positive classic HD. The search for oncogene and tumor-suppressor gene mutations in HRS cells is hampered by the rarity of the HRS cells in the tissue. So far, only a few genes were analyzed for mutations in HRS cells [1]. Translocations of the bcl-2 gene were detected only in very rare cases, and no translocations involving the bcl-6 proto-oncogene were identified. Mutations in the tumor-suppressor genes p53 and CD95 occur in only a few cases. Because HRS cells mostly lack CD95 mutations but seem to be resistant to CD95-induced apoptosis, other members of the CD95 signaling pathway were also analyzed for inactivating mutations, but no mutations were identified in the caspase-8, caspase- 10, or Fas-associated death domain (FADD) genes. The identification of constitutive activity of the nuclear factor- B(NF- B) transcription factor, which plays an important role in the survival of B cells, in the HRS cells of all cases of classic HD prompted studies of mutations in members of the NF- B signaling pathway [4]. Inactivating mutations in the NF- B inhibitors I B and I B were identified in approximately 20% and 10% of cases, respectively. Moreover, genomic amplifications of the c-rel proto-oncogene, a member of the NF- B family, are present in approximately 30% of cases of classic HD. Thus, it seems BB&MT that several distinct aberrations can contribute to constitutive NF- B activity in HRS cells. Aberrant Activation of Multiple Signaling Pathways in HRS Cells Numerous studies in the last years revealed deregulated activity of multiple signaling pathways and transcription factors in HRS cells of classic HD. As was already mentioned previously, constitutive NF- B activity is observed in nearly all cases of classical HD [4]. Besides the involvement of transforming events as potential causes for this activation, NF- B function could also be mediated by stimulating CD40 signaling through CD40 ligand expressing T cells in the microenvironment or by activating CD30, which is expressed in all cases of classic HD. The importance of NF- B activity for HRS cell survival has been demonstrated by showing that its inhibition in HD cell lines induces their apoptosis [4]. The Jak-STAT pathway is the main mediator of cytokine signaling. In HRS cells, the STAT3 and STAT6 transcription factors often show constitutive activation. For STAT6, this activation seems to result from an autocrine stimulation pathway, because HRS cells express interleukin (IL) 13 and the IL-13 receptor, and inhibition of IL-13 signaling reduced STAT6 activity [5]. Inhibiting IL-13 also leads to a reduced proliferation of HD cell lines, thus indicating a role of IL-13 in HRS cell proliferation [6]. Notch1 is a transmembrane receptor that, upon binding of ligand, is cleaved and translocates to the nucleus, where it activates expression of multiple genes. Notch1 is not expressed by normal B cells, but it was found at high levels in HRS cells of classic HD. Triggering of HD cell lines by the Notch1 ligand Jagged1 induces strong proliferation [7]. Because Jagged1 is detectable in vivo in HRS and surrounding cells, activated Notch1 signaling may contribute to HRS cell proliferation. A further transcription factor constitutively activated in HRS cells is AP-1. In HRS cells, AP-1 seems to be mainly composed of the c-jun and JunB components [8]. Inactivation of AP-1 in HD cell lines indicates an important role of this factor in the growth of HRS cells. Whereas c-jun seems to be upregulated by an autoregulatory mechanism, JunB is a target of NF- B in HRS cells. Deregulated activation of receptor tyrosine kinases (RTKs) is frequently involved in cellular transformation of various malignancies. A recent analysis revealed aberrant expression and activation of several RTKs platelet derived growth factor receptor alpha (PDGFRA), (DDR2, EPHB1, RON), tropomyosin related kinase B (TRKB), and tropomyosin receptor kinase A (TRKA) in HRS cells of classic HD; these were most pronounced in the nodular sclerosis subtype [9]. Antibodies detecting phosphotyrosine in sev- 67

3 R. Küppers et al. eral proteins indeed revealed that the HRS cells in many cases had increased phosphotyrosine contents, thus indicating the RTK signal in the HRS cells. Activation seems to occur mainly by their ligands in a paracrine or autocrine fashion. Such coexpression and activation of multiple RTK seems to be unique to HD among lymphomas and indicates that this may contribute to HD pathogenesis. It will be important to identify the targets of the RTK activity in HRS cells. The Search for Novel Therapeutic Targets Although approximately 90% of HD patients can be cured with current chemotherapy and radiation therapy, the remaining patients respond poorly to standard treatment. Moreover, the current therapies are associated with toxic side effects and a risk for secondary neoplasms. Therefore, there is a need to search for novel therapeutic approaches that are less toxic and that are helpful for treating patients who are not currently curable. Because the constitutively activated signaling pathways discussed previously most likely have important functions in the survival and proliferation of HRS cells, some may become valuable targets for therapeutic intervention. Given the strong indication for a key role of NF- B in the survival of HRS cells, inhibiting NF- B seems to be a promising approach. Several compounds are currently being developed and tested that target NF- B, such as the proteasome inhibitor bortezomib. Impairing RTK activity may also be an attractive target for therapy, because RTK inhibitors are already in clinical studies for various cancers. However, because HRS cells often express multiple RTKs, targeting a single RTK by a specific inhibitor may not be sufficient to induce apoptosis of HRS cells. There is still little known about the genetic transforming events that cause the generation of HRS cells and their rescue from apoptosis. If oncogenes involved in the pathogenesis of HD are identified, one may try to develop therapeutic strategies to inhibit their function and thereby induce growth arrest and apoptosis of HRS cells. Finally, much more is to be learned also regarding the role of the typical microenvironment in HD. If we better understand what role the T cells, macrophages, eosinophils, and other nontumor cells play in supporting the growth of the HRS cells or why they fail to eliminate the tumor cells, this may also lead to the development of novel strategies to cure HD. CURRENT TREATMENT STRATEGIES INCORPORATING POSITRON EMISSION TOMOGRAPHY IMAGING AND INTEITY-MODULATED RADIOTHERAPY IN HD Over the last 2 decades, and more intensively during the past 5 years, the practice of radiotherapy of HD has drastically changed. Small involved fields replaced the large extended fields such as mantle and inverted Y and allowed multidirectional configuration of the radiation beam rather than the 2 opposed field approach. The smaller volume of disease remaining after chemotherapy decreased the target size for radiotherapy and provided a better opportunity for decreasing the exposure of healthy organs. The small radiation volumes mandated better identification of targets and improved precision of delivery. Fortunately, reductions in treatment volumes coincided with new developments in imaging: the availability of high-resolution computed tomographic (CT) and positron emission tomography (PET) scanning and progress in the planning and delivery of radiotherapy. Of most importance are the wider availability of 3-dimensional conformal radiotherapy and the more recent introduction of intensity-modulated radiotherapy (IMRT) that allow precise targeting of the involved site with major reduction of radiation to the surrounding tissues. Integrating the modern imaging tools into the planning systems of the new radiation technology (PET-CT simulator) has enhanced accuracy and safer treatment of challenging volume sites and reirradiation of relapsed sites. More recent developments in radiotherapy, such as respiratory gating and imageguided adaptive radiotherapy, may also be relevant to HD and will be discussed. Although the potential long-term benefits of this paradigm shift will require longer observation, it is very likely that the reduction of healthy tissue exposure, such as breast, heart, and lungs, will improve the therapeutic ratio. The Involved Field: Reasons and Implications of a Major Change in Radiotherapy of HD For more than 3 decades, the use of extended-field radiotherapy, often encompassing all nodal sites from the upper neck to the groin, including the spleen, was the central principle guiding the cure of HD with radiation alone. This concept is now obsolete. The radiation field of today is limited and restricted to the site of clinical disease, and when the disease volume had been reduced after chemotherapy, it is often limited to the new postchemotherapy residual volume [10]. The definition of involved field for different sites is still evolving, and some radiation oncologists, particularly in Europe, even advocate limiting the field to the initially involved individual lymph node(s) and not to the full lymph node group site. The reason for the drastic change is quite simple and is well supported by solid data. Because radiation is no longer used as a single agent, there is no need to treat clinically uninvolved lymph node sites the sites that, if left unirradiated when effective chemotherapy was not available, were the most prone to relapse. Several randomized studies have confirmed the adequacy of using involved-field radiotherapy (IFRT) in comparison with extended-field radiotherapy and are summarized in Table 1. The studies clearly indicate that 68

4 Advances in Hodgkin Disease Table 1. Studies Comparing Involved Field Radiation with Extended Radiation in Combined-Modality Programs for Favorable and Unfavorable Early-Stage HL Study Treatment Regimens FFTF or RFS OS (y) Milan (133 patients) [2] ABVD (4) STLI 97% 93% (5) ABVD (4) IFRT 94% 94% GHSG HD8 (1064 patients) [3] COPP/ABVD (4) EFRT 86% 91% (5) COPP/ABVD (4) IFRT 84% 92% EORTC/GELA H8U (995 patients) [4] MOPP/ABV (6) IFRT 94% 90% (4) MOPP/ABV (4) IFRT 95% 95% MOPP/ABV (4) STLI 96% 93% RFS indicates relapse-free survival; HL, Hodgkin s lymphoma; EFRT, extended-field radiotherapy; IFRT, involved-field radiotherapy; STLI, subtotal lymphoid irradiation;, not significant. BB&MT reducing the radiation field has not detracted from the excellent disease Freedom from treatment failure (FTTF) or relapse-free survival rates (84%-95%) or overall survival (OS; 92%-94%) in patients with favorable [11] or unfavorable (German Hodgkin s Disease Study Group [GHSG] [12], European Organization for Research and Treatment of Cancer [EORTC]/Groupe d Etudes Des Lymphomes De l Adulte (GELA) [13,14], and Milan [11]) early-stage HD. The detailed analysis of the GHSG HD8 study demonstrated that a smaller radiation field was associated with a reduction in acute side effects and a trend for a lower risk of second malignancies (2.8% versus 4.5%) that may strengthen with a longer follow-up [12]. Combined-modality therapy not only allows for a drastic restriction of the irradiated volume, but it also permits a meaningful reduction (by up to 50%) in the effective prescribed radiation dose. The GHSG studies of combined-modality therapy in patients with unfavorable early-stage disease indicated that disease control with 20 Gy was as effective as that with 40 Gy, provided that bulky disease sites were irradiated to 40 Gy [12]. The recently completed GHSG study HD10 for patients with favorable disease randomized patients to either 20 Gy of IFRT or 30 Gy of IFRT after a short course (2 or 4 cycles) of doxorubicin (Adriamycin), bleomycin, vinblastine, and dacarbazine (ABVD) [15]. At a median follow-up of 24 months, the overall results are excellent (FFTF of 97%), with no difference among the 4 arms. The GHSG study of unfavorable patients (HD11), with a similar radiotherapy dose design and a 24-month FFTF of 90%, also suggested that radiation dose reduction is safe [16]. More mature and detailed results are expected soon. The EORTC/GELA current H9F trial for favorable patients is evaluating epirubicin, bleomycine, vinblastine, prednisone (EBVP) 6 to complete remission (CR), followed by either IFRT 36 Gy or involved field radiotherapy (IFRT) 20 Gy. The third arm of patients with CR after 6 cycles of EBVP and no radiation was closed early because of an excessive number of relapses. The combined-modality arms remained open until the study ended. The conversion from large multisite radiation fields to a smaller and better defined radiation field allowed also for accurate conformal radiation therapy. The large fields of the past limited the radiation technique to 2 simple opposed anterior and posterior fields. The conversion to smaller and better defined radiation volumes allowed the use of more conformal radiation therapy, based on better imaging, computerized planning programs, and, when indicated, advanced tools such as IMRT. Modern breakthroughs in radiotherapy technology can now be implemented in HD to increase accuracy even further, avoid healthy organ irradiation, and thus improve the therapeutic ratio [17]. The only clear indication to use radiation alone in HD is in early-stage lymphocyte predominance HD. Yet this rare subtype of HD does not require extensive-field radiotherapy. The disease is commonly limited to 1 peripheral site (neck, axilla, or groin), and involvement of the mediastinum is extremely rare. The treatment recommendations for lymphocyte predominance HD differ markedly from those for classic HD [18]. Chemotherapy is not indicated for earlystage disease, and it should be emphasized that even if regional radiation fields are selected, the uninvolved mediastinum should not be irradiated, thus avoiding the site most prone to radiation-related short- and longterm side effects. Although there has not been a study that compared extended-field radiotherapy (commonly used in the past) with IFRT, retrospective data suggest that IFRT is adequate [19]. The radiation dose recommended is between 30 to 36 Gy, with an optional additional boost of 4 Gy to a (rare) bulky site. Although the decreases in radiation fields and doses described previously maintained disease control rates that are difficult to top, the concern regarding the contribution of radiation to long-term complications and late mortality from second tumors and coronary heart disease still prevails [20]. However, it should be recognized that data regarding these late 69

5 R. Küppers et al. complications are generated from series of patients cured of HD during the 1960s and 1970s by full-dose extended-field radiation alone [21]. It will take at least another decade to assess the risk from low-dose miniradiotherapy. Early data suggest that it is likely to be markedly reduced [10,11]. It was also clearly established by international case-control studies that the risk of both breast cancer and lung cancer is radiation dose dependent, and it is predictable that if most of the organ is not exposed, the total risk is reduced [22-24]. Here, the change to IFRT is critical [25]. This could be illustrated in reference to the radiationrelated risk of breast cancer in young women. By using the extended mantle field of the past, a substantial amount of breast tissue has been exposed to either the full prescribed dose or to an attenuated dose (at the field margins or under the lung shields) in almost all women irradiated for HD. Most breast exposure in the mantle era resulted from the radiation of the axillae (65% of tumors in this study developed in the outer part of the breast) [22] and, to a lesser extent, from wide mediastinal and hilar irradiation. Approximately two thirds of women with early-stage HD do not require radiation of the axillae, and additional protection to the upper and medial aspects of the breast could be provided by further reducing the field size by using careful CT-based planning, which usually allows for smaller mediastinal volumes, especially after chemotherapy. At the same time, omitting radiotherapy altogether not only requires longer and potentially toxic chemotherapy, but it also results in inferior and worrisome outcomes, as several recent randomized studies have documented [26-28]. Modern Involved-Field Treatment Planning: New Options, New Tools The conversion from large multisite radiation fields to a smaller and better defined small radiation field with reduced uncertainty margins mandated accurate identification of the involved volume and optimal and consistent radiation targeting. At the same time, it allowed the option of 3-dimensional conformal radiation therapy. The old extensive radiation fields of the past, such as mantle or inverted Y, included multiple sites at various depths (from the body surface), and each site had different limitations of access and tolerance of healthy tissue. The only way to include these sites in 1 radiation field (and, thus, avoid undesired overlaps and gaps when radiation fields were matched) was to treat the entire field from only 2 opposed directions: anterior and posterior. This technique ensured the inclusion of most lymph nodes in 1 field, yet it also resulted in the exposure of large volumes of healthy organs (eg, heart, lungs, breasts, and spinal cord) to the full prescribed radiation dose. Because the notion of treating large areas of involved and uninvolved areas has changed in favor of treating only the involved lymph node group, new options of more conformal radiotherapy have opened up. When the target is limited, a multiple noncoplanar beam approach is possible. Over the last decade, the imaging of HD has markedly improved. Fast high-resolution CT scanners and fluorodeoxyglucose PET scanning are invaluable for identification of involved sites. New technology allowed integration of the 2 scanners (PET and CT) and has provided an incremental advantage to radiotherapy planning that has already integrated CT simulation as a standard planning tool. Software that allows simultaneous treatment planning on CT images and PET images or, even better, the new generation of PET-CT simulator equipment has upgraded our ability to precisely outline tumor volumes while the patient is in the radiation treatment position. Dependable immobilization devices are used to ensure consistency throughout the treatment course of fractionated radiotherapy. New experimental tools such as respiratory gating and online image-guided adaptive radiotherapy may also have a role in the treatment of selected cases of HD in the future. The most important upgrade in radiation oncology of recent years has been the development and implementation of IMRT. IMRT incorporated new planning concepts with powerful computing and required a new generation of linear accelerators equipped with computer-controlled dynamic multileaf collimators to allow accurate planning of challenging targets that lie very close to critical healthy organs. We have found that IMRT is a valuable tool in planning certain large thoracic volumes and is critical for reirradiating relapsed disease as part of a salvage program [17]. The change in the lymphoma radiotherapy paradigm thus coincided with substantial improvements in imaging and treatment-planning technology that have revolutionized the field of radiotherapy over the last 15 years. IMRT allows for accurately enveloping the tumor with either a homogenous radiation dose ( sculpting ) or delivering higher doses to predetermined areas in the tumor volume ( painting ). The end result of this new modality is highly accurate treatment with maximal sparing of healthy tissues. TREATMENT OF ADVANCED AND RELAPSED HD Advanced HD Up to the middle of the last century, patients with advanced stages of HD were incurable. With the advent of more effective drugs used early in childhood leukemia, Moxley et al. [29] at the National Cancer Institute were the pioneers who paved the road for the incredible success of modern chemotherapy in oncology, with a 50% cure rate for advanced-stage HD patients, purely with the drug combination MOPP 70

6 Advances in Hodgkin Disease (mechlorethamine, vincristine [Oncovin], procarbazine, and prednisone). Despite the great accomplishments with MOPP and MOPP-like regimens, there were major obstacles: (1) 15% to 30% of the patients did not reach a CR, and only approximately 50% of patients could be cured, and (2) MOPP had significant acute toxicity and an increased risk of sterility and acute leukemia because of the alkylating agents. In 1975, Bonadonna et al. [30] introduced the ABVD regimen in an attempt to develop a regimen for patients who had experienced treatment failure after MOPP. Vinblastine had demonstrated high activity as a single agent and lacked cross-resistance with vincristine in human tumors. Both doxorubicin and bleomycin were very active drugs and showed objective responses in approximately 50% of patients. Dacarbazine was added because it was active as a single agent and also showed synergism with doxorubicin. The results of prospective multicenter trials using the traditional standard regimens MOPP or ABVD or alternating, sequential, or hybrid combinations of these 2 very effective, non cross-resistant drug regimens, mostly assisted by additive radiotherapy have shown satisfactory CR rates (up to 80%-90%). However, the failure-free survival and OS rates at 5 years were only 65% to 70% and 75% to 85%, respectively. The pivotal Cancer and Leukemia Group B trial in advanced HD, which compared MOPP, ABVD, and alternating MOPP/ABVD without additive radiotherapy, showed equal therapeutic results for ABVD and MOPP/ABVD as far as progression-free survival and OS were concerned. Both regimens were superior to MOPP. ABVD had less germ cell and hematopoietic stem cell toxicity. A long-term follow-up of this study over 15 years has recently been published demonstrating a 45% to 50% progression-free survival and a 65% overall survival for ABVD and MOPP/ABVD [31]. There are single-center reports with more favorable results using ABVD, MOPP/ABV (adriamycin, bleomycin, vincristine), or similar regimens, but the numbers of patients are small and seem to carry a bias of selection. These rather disappointing results raised a number of questions: 1. Is ABVD good enough, with 30% to 35% failures within 5 to 10 years? 2. How can we improve initial tumor control without compromising long-term cure? 3. Do we need consolidating radiotherapy after effective chemotherapy? 4. Are the fourth-generation regimens Stanford V, bleomycin, etoposide, adriamycin, cyclophosphamide, oncovin [vincristine], procarbazine, prednisone (BEACOPP), chlorambucil, vinblastine, procarbazine, prednisone, etoposide, vincristine (ChlVPP/EVA), and mechlorethamine (MEC) better than ABVD, both short and long term? BB&MT Table 2. BEACOPP Regimens Regimen BEACOPP (baseline) 21 Bleomycin 10 IV 8 Etoposide 100 IV 1-3 Adriamycin (doxorubicin) 25 IV 1 Cyclophosphamide 650 IV 1 Oncovin (vincristine) 1.4* IV 8 Procarbazine 100 PO 1-7 Prednisone 40 PO 1-14 BEACOPP (escalated) 21 Bleomycin 10 IV 8 Etoposide 200 IV 1-3 Adriamycin (doxorubicin) 35 IV Cyclophosphamide 1250 IV 1 Oncovin (vincristine) 1.4* IV 8 Procarbazine 100 PO 1-7 Prednisone 40 PO 1-14 G-CSF SQ 8 BEACOPP Bleomycin 10 IV 8 Etoposide 100 IV 1-3 Adriamycin (doxorubicin) 25 IV Cyclophosphamide 650 IV 1 Oncovin (vincristine) 1.4* IV 8 Procarbazine 100 PO 1-7 Prednisone 40 PO 1-7 G-CSF SQ 8-13 IV indicates intravenous; PO, by mouth; SQ, subcutaneously; G-CSF, granulocyte colony-stimulating factor. *Vincristine dose capped at 2 mg. 5. How can we improve the quality and quantity of life by reducing acute and long-term toxicity? These questions led the GHSG in 1992 to investigate a new principle on the basis of the cyclophosphamide, vincristine (Oncovin), procarbazine, prednisone (COPP)/ABVD regimen by using the same drugs, except for dacarbazine and vinblastine, adding etoposide, and introducing a time- and dose-escalated schedule in lymphoma treatment for the first time, thus formulating the BEACOPP principle (Table 2). This regimen allows administration of the most tumoricidal drugs on day 1 to 3 and recycling on day 21 instead on day 28 or, in the BEACOPP-14 regimen, on day 15 and, by the help of granulocyte colonystimulating factor (G-CSF), to escalate the dosage of etoposide, cyclophosphamide, and doxorubicin by 100%, approximately 90%, and approximately 40%, respectively. In a randomized 3-arm phase III study, COPP/ ABVD was tested against baseline BEACOPP and escalated BEACOPP. Radiotherapy in this study was given to initial bulk and residual tumors with 30 Gy in approximately 65% to 70% of patients [32]. At the last analysis in June 2003, after a median follow-up of 7 years, the escalated BEACOPP arm showed highly significant superiority over the COPP/ ABVD arm for FFTF (85% versus 67%; P.0001) and OS (90% versus 79%; P.0001), despite the higher number of 11 acute myeloid leukemia (AML)/ 71

7 R. Küppers et al. Figure 1. HD12 study design of the GHSG for patients with advanced Hodgkin s disease. RT indicates radiotherapy. myelodysplastic syndrome patients in the BEACOPP escalated arm versus 1 in the COPP/ABVD arm. However, the death rate due to progression of HD was 9.6% in the COPP/BVD arm and only 2.4% in the BEACOPP escalated group of patients. Furthermore, there was no survival difference after salvage treatment among the 3 treatment arms. However, the superiority of 8 cycles of escalated BEACOPP over the former standard-arm COPP/ ABVD had the prize of a considerably higher acute and late toxicity. Therefore, the GHSG started to undertake 2 successive studies, HD12 and HD15 (Figures 1 and 2), to de-escalate BEACOPP and reduce radiotherapy, hence reducing the toxic burden of this very effective principle. In the HD12 study, 8 cycles of escalated BEA- COPP were compared with 4 cycles of escalated and 4 cycles of baseline BEACOPP, and by following a factorial design, there was a further randomization for radiotherapy (yes or no). A total of 65% of patients in the radiotherapy arms received 30 Gy of IFRT, whereas in the nonradiotherapy arms, 10% were radiated because of the decision of a panel that, independently of the randomization, judged every CT after the end of chemotherapy. In the second interim analysis in June 2004, 1396 patients were evaluable. After a median follow-up of 2 years in this sequential analysis, there was neither a difference among the 8 BEACOPP escalated and 4 BEACOPP escalated 4 BEACOPP baseline arms nor a difference for the radiotherapy (yes and no) arms, either for FFTF or OS. At a median follow-up of 30 months, the FFTF for the total cohort was 86.9%, and the OS was 93.8%. At this time of observation, the rate for AML/myelodysplastic syndrome was only half of that observed in the HD9 trial at the same time point. These results favor the continuation of the global study, chaired by Dr. Patrice Cadre from the EORTC, comparing ABVD with BEACOPP (4 4) with or without radiotherapy. In conclusion, ABVD is not generally accepted worldwide as the gold standard chemotherapy regimen for advanced stages of Hodgkin lymphoma, whereas it is the most favored regimen for early and intermediate stages of Hodgkin lymphoma. The new generation of dose-intensified or dose-dense chemotherapy regimens such as BEACOPP have to be tested in different cultural and health structure settings to prove their expected superiority over conventional schemes, such as ABVD or MOPP/ABV, to improve the long-term quantity and quality of life of Hodgkin lymphoma patients. The role of consolidating radiotherapy after reaching clinically confirmed, CT/magnetic resonance confirmed, or PET-confirmed CR has to be assessed in ongoing studies, eg, the HD15 study of the GHSG. Prognostic Factors in Relapsed and Refractory HD It was first noted in 1979 that the length of remission to first-line chemotherapy had a marked effect on the ability of patients to respond to subsequent salvage treatment [33]. In 1992, the National Cancer Institute updated their experience with the long-term follow-up of patients who relapsed after polychemotherapy [34]. Derived primarily from investigations involving failures after MOPP and MOPP variants, the conclusions are thought to be relevant to other chemotherapy programs as well. On this basis, chemotherapy failures can be divided into 3 subgroups: 1. Primary progressive HD, ie, patients who never achieve a CR. 2. Early relapses within 12 months of CR. 3. Late relapses after CR lasting 12 months. With use of conventional chemotherapy for pa- Figure 2. HD15 study of the GHSG for patients with advanced Hodgkin s disease. RT indicates radiotherapy. EPO Erythropoietin. 72

8 Advances in Hodgkin Disease tients with primary progressive disease, virtually no patient survives 8 years. In contrast, the projected 20-year survival for patients with early relapse or late relapse was 11% and 22%, respectively [34]. Primary Progressive HD The GHSG retrospectively analyzed 206 patients with progressive disease, defined as progression during induction treatment or within 90 days after the end of treatment, to determine outcome after salvage therapy and to identify prognostic factors [35]. The 5-year freedom from second failure (FF2F) and OS for all patients were 17% and 26%, respectively. As reported from transplant centers, the 5-year FF2F and OS for patients treated with high-dose chemotherapy (HDCT) are 42% and 48%, respectively, but only 33% of all patients received HDCT. The low percentage of patients who received HDCT was due to rapidly fatal disease or life-threatening severe toxicity after salvage therapy. Other reasons not to proceed to HDCT were an insufficient stem cell harvest, poor performance status, and older age. In a multivariate analysis, Karnofsky performance score at progression (P.0001), age (P.019), and attainment of a temporary remission to first-line chemotherapy (P.0003) were significant prognostic factors for survival. Patients with none of these risk factors had a 5-year OS of 55%, compared with 0% for patients with all 3 of these unfavorable prognostic factors. Early and Late Relapsed HD Although the results reported with HDCT in patients with late relapse have been superior to those reported in most series of conventional chemotherapy, the use of HDCT in late relapses had been an area of controversy because patients with late relapse have satisfactory second CR rates when treated with conventional chemotherapy, with OS ranging from 40% to 55%. However, the HDR-1 trial of the GHSG also showed improved FFTF after HDCT compared with conventional chemotherapy in patients with late relapse. The GHSG has recently performed a retrospective analysis including a much larger number of relapsed patients (n 422) than previously reported. The analysis of prognostic factors suggests that the prognosis of a patient with relapsed HD can be estimated according to several factors. The most relevant factors were combined into a prognostic score. This score was calculated on the basis of the duration of first remission, the stage at relapse, and the presence or absence of anemia at relapse. Early recurrence within 3 to 12 months after the end of primary treatment, relapse stage III or IV, and hemoglobin 10.5 g/dl in female or 12 g/dl in male patients contribute to a score with possible values of 0, 1, 2, and 3, in BB&MT order of worsening prognosis [36]. This prognostic score allows distinguishing patients with different FF2F and OS. The actuarial 4-year FF2F and OS for patients relapsing after chemotherapy with 3 unfavorable factors were 17% and 27%, respectively. In contrast, patients with none of the unfavorable factors had FF2F and OS at 4 years of 48% and 83%, respectively. In addition, the prognostic score was also predictive for patients who relapsed after radiotherapy, for patients who relapsed after chemotherapy who were treated with conventional therapies or with HDCT followed by autologous stem cell transplantation (ASCT), and for patients 60 years and a Karnofsky performance status 90%, which were the major candidate groups for dose intensification. The prognostic factor score uses clinical characteristics that can be easily collected at the time of relapse. It separates groups of patients with substantially different outcomes. Treatment Strategies Patients who relapse after radiation therapy alone for localized HD have satisfactory results with combination chemotherapy and are not considered candidates for HDCT and ASCT [12,13]. HDCT followed by ASCT has been shown to produce 30% to 65% long-term disease-free survival in selected patients with refractory and relapsed HD. In addition, the reduction of early transplant-related mortality from the 10% to 25% reported in earlier studies to 5% in more recent studies has led to the widespread acceptance of HDCT and ASCT. Although results of HDCT have generally been better than those observed after conventional-dose salvage therapy, the validity of these results has been questioned because of the lack of randomized trials. The most compelling evidence for the superiority of HDCT and ASCT in relapsed HD comes from 2 reports from the British National Lymphoma Investigation and the GHSG, together with the European Group for Blood and Marrow Transplantation (EBMT). In the British National Lymphoma Investigation trial, patients with relapsed or refractory HD were treated with a combination of carmustine (BCNU), etoposide, cytarabine, and melphalan at a conventional dose level (mini-beam) or a high-dose level (BEAM) with autologous bone marrow transplantation [37]. The actuarial 3-year event-free survival was significantly better in patients who received HDCT (53% versus 10%). The largest randomized, multicenter trial was performed by the GHSG/EBMT to determine the benefit of HDCT in relapsed HD. Patients with relapse after polychemotherapy were randomly assigned between 4 cycles of dexa-beam (dexamethasone BEAM) and 2 cycles of dexa-beam followed by HDCT (BEAM) and autologous bone marrow transplantation/periph- 73

9 R. Küppers et al. eral blood stem cell transplantation (PBSCT). The final analysis of 144 evaluable patients revealed that of 117 patients with partial remission (PR) or CR after 2 cycles of chemotherapy, FFTF in the HDCT group was 55%, versus 34% for the patients who received an additional 2 cycles of chemotherapy. OS was not significantly different [38]. Sequential HDCT Figure 3. HDR-2: study design (GHSG, EORTC, EBMT, and GEL/TAMO). CTX indicates cyclophosphamide, MTX, methotrexate; VP-16, etoposide. In recent years, sequential HDCT has increasingly been used in the treatment of solid tumors and hematologic and lymphoproliferative disorders. Initial results from phase I/II studies indicate that this kind of therapy offers safe and effective treatment. In accordance with the Norton-Simon hypothesis, after initial cytoreduction, few non cross-resistant agents are given at short intervals. In general, PBSCT and the use of growth factors allow the application of the most effective drugs at the highest possible doses at intervals of 1 to 3 weeks. Sequential HDCT thereby enables the highest possible dosing over a minimum period of time (dose intensification). In 1997, a multicenter phase II trial with a highdose sequential chemotherapy program and a final myeloablative course was started to evaluate the feasibility and efficacy of this novel regimen in patients with relapsed HD [26]. Eligibility criteria included age 18 to 60 years, histologically proven relapsed or primary progressive HD, second relapse with no prior HDCT, and an Eastern Cooperative Oncology Group performance status of 0 or 1. The treatment program consists of 2 cycles of DHAP (dexamethasone, cytarabine, and cisplatin) in the first phase to reduce the tumor burden before HDCT. Patients with PR or CR after 2 cycles of DHAP receive sequential HDCT consisting of cyclophosphamide 4 g/m 2 intravenously (IV), methotrexate 8 g/m 2 IV plus vincristine 1.4 mg/m 2 IV, and etoposide 2 g/m 2 IV. The final myeloablative course was BEAM followed by PBSCT with at least CD34 cells per kilogram [39]. At the last interim analysis, 102 patients were available for the final evaluation. The state of remission was multiple relapses in 10 patients, progressive disease in 16 patients, early relapse in 29 patients, and late relapse in 44 patients. At 30 months of median follow-up (range, 3-61 months), results are as follows: response rate after DHAP, 87% (23% CR and 64% PR); response rate at final evaluation, 77% (68% CR and 9% PR). Toxicity was tolerable, with no treatment-related deaths. FFTF and OS for patients with early relapse were 64% and 87%, respectively, for early relapse; 68% and 81% for late relapse; 30% and 58% for patients with progressive disease; and 55% and 88% for patients with multiple relapses [40]. In conclusion, sequential administration of high doses of cyclophosphamide, methotrexate, and etoposide is feasible and did not affect the tolerability of final myeloablative BEAM. This new 3-phase treatment regimen is well tolerated and feasible in patients with relapsed and primary progressive HD. The preliminary data suggest a high efficacy in relapsed HD patients; thus, further randomized studies are warranted. HDR-2 Protocol In January 2001, the GHSG, together with the EORTC, the GEL/TAMO, and the EBMT, started a prospective randomized study to compare the effectiveness of standard HDCT (BEAM) with sequential HDCT after initial cytoreduction with 2 cycles of DHAP (HD-R2 protocol; Figure 3). Patients with histologically confirmed early or late relapsed HD and patients in second relapse with no prior HDCT fulfilling the entry criteria receive 2 cycles of DHAP followed by G-CSF. Patients who achieve no change in their disease status, a PR, or a CR after DHAP are centrally randomized to receive either BEAM followed by PBSCT (arm A of the study) or high-dose cyclophosphamide G-CSF, followed by high-dose methotrexate vincristine, followed by high-dose etoposide G-CSF and a final myeloablative course with BEAM (arm B of the study). Allogeneic Transplantation after Reduced Conditioning in HD Allogeneic transplantation (allobmt) has clear advantages compared with autologous transplantation: donor marrow cells uninvolved by malignancy are used, thus avoiding the risk of infusing occult lymphoma cells, which may contribute to relapse in patients who undergo autologous transplantation. In addition, donor lymphoid cells can potentially mediate a graft-versus-lymphoma effect. 74

10 Advances in Hodgkin Disease Generally, donor availability and age constraints have limited a broader application of allobmt in HD. Moreover, allobmt is associated with a high treatment-related mortality rate of up to 75% in patients with induction failure, which casts doubt upon the feasibility of this approach in HD patients. In most cases, allobmt from HLA-identical siblings is not recommended for patients with HD. The reduced relapse rate associated with a potential graft-versustumor effect is offset by lethal graft-versus-host toxicity. Nevertheless, patients with induction failure and relapsed patients with additional risk factors also have a poor prognosis after HDCT and ASCT. Therefore, the role of allobmt should be further evaluated in these patients by taking advantage of new developments such as nonmyeloablative conditioning regimens and allogeneic PBSCT. To circumvent the problems inherent to the toxicity and treatment-related mortality associated with allografting, the possibility of achieving engraftment of allogeneic stem cells after immunosuppressive therapy combined with myelosuppressive but nonmyeloablative therapy has been assessed. Several groups have recently updated their experience with nonmyeloablative conditioning regimens [40]. The EBMT, together with the GEL/TAMO, the EORTC, and the GHSG, activated a multicenter phase II study to evaluate the treatment-related mortality of patients with primary progressive or relapsed HD (early relapse, multiple relapse, and relapse after ASCT). Patients with an HLA-compatible sibling donor or an HLA-matched unrelated donor will be initially treated with 1 or 2 cycles of DHAP or other salvage protocols to reduce their tumor burden before allogeneic PBSCT. PBSCs will be collected after G-CSF priming of the donor and reinfused after conditioning with fludarabine and melphalan. FUTURE DIRECTIO Alternative strategies have been developed to improve the outcome of relapsed and resistant HD. These approaches include the development of new cytostatic drugs and biological agents with proven efficacy in preclinical models. One of the most promising new cytostatic drugs is the new vinca alkaloid vinorelbine, which has demonstrated activity in HD even in patients pretreated with vincristine or vinblastine. The use of vinorelbine in first- and second-line therapy of HD to improve the frequency and duration of responses is still under investigation. The pyrimidine analogue gemcitabine is the only drug currently under investigation that represents a new cytostatic mechanism of action. The self-potentiating mechanism of action leads to an BB&MT enhanced accumulation and prolonged retention of gemcitabine in the malignant cell. The results of gemcitabine in advanced relapsed HD are promising, with an overall response rate of 53% in heavily pretreated patients. Although some clinical efficacy has been demonstrated in clinical trials with immunotoxins, none of the current available immunotoxins seems to be suited for a clinical phase III study. Bispecific monoclonal antibodies, such as the recently reported CD30 CD64 bispecific monoclonal antibodies, look more promising, with clinical development programs scheduled that include phase III. The use of recombinant DNA technology for site-directed modifications of the immunotoxins and the development of humanized immunotoxins and bispecific monoclonal antibodies might optimize their efficacy. In the future, combining standard chemotherapy/radiotherapy with biological agents might result in the elimination of residual tumor cells and, subsequently, more relapse-free longterm survivors. REFERENCES 1. Küppers R. Molecular biology of Hodgkin s lymphoma. Adv Cancer Res. 2002;84: Kanzler H, Küppers R, Hansmann ML, Rajewsky K. Hodgkin and Reed-Sternberg cells in Hodgkin s disease represent the outgrowth of a dominant tumor clone derived from (crippled) germinal center B cells. J Exp Med. 1996;184: Schwering I, Bräuninger A, Klein U, et al. Loss of the B- lineage-specific gene expression program in Hodgkin and Reed-Sternberg cells of Hodgkin lymphoma. Blood. 2003;101: Bargou RC, Emmerich F, Krappmann D, et al. Constitutive nuclear factor-kappab-rela activation is required for proliferation and survival of Hodgkin s disease tumor cells. J Clin Invest. 1997;100: Skinnider BF, Elia AJ, Gascoyne RD, et al. Interleukin 13 and interleukin 13 receptor are frequently expressed by Hodgkin and Reed-Sternberg cells of Hodgkin lymphoma. Blood. 2001; 97: Kapp U, Yeh WC, Patterson B, et al. Interleukin 13 is secreted by and stimulates the growth of Hodgkin and Reed-Sternberg cells. J Exp Med. 1999;189: Jundt F, Anagnostopoulos I, Forster R, Mathas S, Stein H, Dörken B. Activated Notch1 signaling promotes tumor cell proliferation and survival in Hodgkin and anaplastic large cell lymphoma. Blood. 2002;99: Mathas S, Hinz M, Anagnostopoulos I, et al. Aberrantly expressed c-jun and JunB are a hallmark of Hodgkin lymphoma cells, stimulate proliferation and synergize with NF-kappa B. EMBO J. 2002;21: Renné C, Willenbrock K, Küppers R, Hansmann ML, Bräuninger A. Autocrine- and paracrine-activated receptor tyrosine kinases in classic Hodgkin lymphoma. Blood. 2005;105: Yahalom J, Mauch P. The involved field is back: issues in delineating the radiation field in Hodgkin s disease. Ann Oncol 2002;13(suppl 1):