Management of relapsed and refractory classical Hodgkin lymphoma in children and adolescents


Dr Stephen Daw, Children and Young People’s Cancer Service, University College Hospital London, 250 Euston Road, London NW1 2PQ, UK. E-mail:


There are a number of options for salvage treatment in children and adolescents with relapsed and refractory classical Hodgkin Lymphoma. These include salvage with standard dose chemotherapy, high dose chemotherapy with autologous stem cell transplant, allogeneic stem cell transplant or other novel approach. Radiotherapy has an important role in the salvage of some patients as part of a combined modality approach. This review outlines these salvage approaches and discusses whether the evidence from paediatric studies justifies a risk-adapted approach to salvage for individual patients or whether all patients should receive consolidation with high dose chemotherapy and autologous stem cell transplantation, which is often described as standard salvage management in adults. The important prognostic factors and how these may be used to allocate patients to standard versus high dose chemotherapy regimens are discussed. The role of allogeneic transplantation, novel agents and late effects will also be discussed.

The cure rates of first-line treatment for classical Hodgkin lymphoma (HL) in children are very high with the use of modern effective chemotherapy, often used in combination with involved field radiotherapy (IFRT). The most effective regimens achieve cure rates in excess of 90% for early stage and 80% for advanced stage disease, with equivalent outcomes for adolescents and children (Schellong et al, 1986, 1999; Vecchi et al, 1993, 1997; Hunger et al, 1994; Schellong, 1996; Donaldson & Link, 1997; Hutchinson et al, 1998; Dörffel et al, 2003; El- Badawy et al, 2008). Single modality radiotherapy (RT) is rarely used now as primary treatment due to unacceptable failure rates even in low stage patients (Shankar et al, 1997). Paediatric paradigms for primary treatment have evolved to reduce late effects whilst maintaining excellent cure rates (Hodgson et al, 2007), achieved by developing chemotherapeutic regimens that minimize late toxicity and, when given, RT is low dose (20–25 Gy) and involved field. Primary chemotherapy is allocated using a risk-adapted approach utilizing prognostic factors associated with primary treatment failure (Smith et al, 2003), and response to chemotherapy is under study to guide either the final number of chemotherapy cycles (Schwartz et al, 2009), or to limit or remove RT in patients who achieve complete remission with chemotherapy alone (Dörffel et al, 2003).

Salvage therapy – general considerations

Approximately 10% of patients with early stage, and up to 25% with advanced stage disease, relapse after first line therapy (Donaldson et al, 2007; Schwartz et al, 2009). Cure may still be achieved in patients with recurrent disease and all should receive second-line standard dose salvage chemotherapy (SDCT). Salvage single modality RT is rarely used. There are no randomized trials in children defining the ‘best’ salvage chemotherapy regimen or comparing SDCT with high dose chemotherapy (HDCT) and autologous stem cell transplantation (ASCT). Despite these limitations there is evidence that prognostic factors at disease recurrence may be used to allocate a risk stratified salvage approach (Schellong et al, 2005). A combination of pre-treatment prognostic factors at relapse, which defines risk group and assessment of chemosensitivity to re-induction SDCT, which determines curative intent, defines the individual treatment plan.

Salvage in paediatric patients includes a greater use of SDCT and IFRT compared with adult paradigms. Low-risk patients with late relapse and limited stage, may be salvaged with SDCT plus RT and high-risk patients, who include all primary refractory HL, receive consolidation with conventional HDCT/ASCT.

The cut-off point between low- and high-risk strategies is under investigation. The current European ‘EuroNet’ trial (EuroNet-PHL-C1 [HD 2007/10, Eudra CT no. 2006-000995-33]; Fig 1) defines an intermediate risk group, which includes all relapsed HL except late relapse after two cycles of primary chemotherapy (low risk) and primary progression (high risk). For intermediate risk relapsed HL the best salvage approach is unclear between SDCT plus IFRT versus SDCT plus HDCT/ASCT. In the current European strategy, intermediate risk patients who achieve a complete (CR), or partial response (PR) and fluro-deoxy-glucose positron emission tomography (FDG-PET) negativity, with SDCT de-escalate to the low risk strategy (SDCT plus IF-RT) while non-CR, or PR with FDG-PET positivity, receive consolidation with HDCT. Proceeding to HDCT with a positive FDG-PET scan was questioned in a recent adult series, which assessed significance of FDG-PET pre HDCT and found a 4-year event-free survival (EFS) rate of 33% (PET positive) versus 77% (PET negative), suggesting normalization of functional imaging pre-ASCT should be the goal of salvage treatment (Moskowitz et al, 2010). The evidence in paediatric patients is limited and is under investigation in the current EuroNet trial.

Figure 1.

 Risk-adapted treatment plan for relapsed and refractory Hodgkin’s Lymphoma in accordance with the Euronet – PHL – C1 Trial. RG – relapsed group. RG1, low risk group; RG2, intermediate risk group; RG3, high-risk group. TG1, treatment group 1; TG2, treatment group 2; TG3, treatment group 3; IEP, ifosfamide, etoposide, prednisolone; ABVD, adriamycin, bleomycin, vinblastine, dacarbazine; PET, positron emission tomography; RT, radiotherapy; BEAM, BCNU [carmustine], etoposide, cytarabine, melphalan; ASCT, autologous stem cell transplantation; HDCT, high dose chemotherapy.

Primary progressive HL, which remains refractory to salvage chemotherapy, has a very poor outcome with HDCT/ASCT in children, as in adults, and this group urgently requires a better approach that may include allogeneic transplantation or other novel therapy. At present, however, no national paediatric trial mandates allogeneic SCT (alloSCT) as the first transplant approach. The experience is very limited in paediatric practice and is controversial but there is clearly a need to improve on the outcomes of HDCT/ASCT in chemo-refractory progressive HL.

RT has an important role in salvage but its use should be assessed on an individual basis, taking account of previous radiation exposure, in or out field recurrence, stage at recurrence and acute and long-term toxicities of total treatment burden. Relapse after primary RT alone is rare, as this has virtually disappeared as primary therapy, however in this circumstance patients are retrievable with standard dose chemotherapy ± further RT. (Ruffer et al, 2005). Current approaches to salvage therapy discussed here, address salvage after primary chemotherapy or combined modality regimens.

Standard dose chemotherapy salvage at first relapse

At the point of disease recurrence the first step is re-induction with salvage chemotherapy. There are no randomized studies comparing SDCT regimens. The choice of re-induction regimen must be based on data from single arm studies and retrospective analyzes but should take account of the primary therapy to use at least some non cross-resistant drugs and avoid cumulative drug toxicities. The aims of salvage are to obtain cyto-reduction, demonstrate chemo-sensitivity and facilitate harvesting of peripheral blood stem cells for ASCT. Regimens should have low toxicity and high efficacy and minimize the risk of long-term toxicity, including secondary myelodysplastic syndrome or myeloid leukaemia.

SDCT salvage regimens can be divided into intensive conventional regimens, such as mini-BEAM [BCNU (carmustine), etoposide, cytarabine, melphalan] (Linch et al, 1993); platinum-based regimens [ESHAP (etoposide, methylprednisolone, cytarabine, cisplatin), DHAP (dexamethasone, cytarabine, cisplatin), APE (cytarabine, cisplatin, etoposide)], (Aparicio et al, 1999; Josting et al, 2005a; Wimmer et al, 2006); ifosfamide-/etoposide-based regimens [EPIC (etoposide, prednisolone, ifosfamide and cisplatin), IEP-ABVD (ifosfamide, etoposide, prednisolone, adriamycin, bleomycin, vinblastine and dacarbazine), ICE (ifosfamide, carboplatin, etoposide)] (Hickish et al, 1993; Moskowitz et al, 2001; Schellong et al, 2005) or other novel combinations [GV (gemcitabine, vinorelbine), IGEV (ifosfamide, gemcitabine, vinorelbine, prednisolone)] (Santoro et al, 2007; Cole et al, 2009). Currently, alternating IEP-ABVD is the standard in children in Europe. Overall response rates (ORR) (CR and PR > 50%) of some salvage regimens are reported in Table I. It is difficult to compare outcomes because of the differences in prognostic factors between patients in these series. Chemo-sensitivity is associated with improved EFS and determines whether the salvage strategy should be continued or abandoned. Patients should have response assessment early, after two cycles of SDCT, with conventional imaging and FDG-PET. FDG-PET is more accurate than conventional computerized tomography (CT) in this setting and highly predictive in the pre-transplant setting (Spaepen et al, 2003). The decision to continue salvage therapy with RT for consolidation versus HDCT ASCT is based on assessment of prognostic factors.

Table I.   Salvage chemotherapy regimens for Hodgkin lymphoma.
ReferenceRegimenDrugsPatients (n)ORR (%)Treatment-related deaths (n)
  1. Standard dose salvage chemotherapy regimens can be divided into intensive conventional regimens (mini-BEAM), Cisplatin based regimens (ESHAP, DHAP, APPE, DECAL), Ifosfamide/Etoposide based regimens (EPIC, IEP-ABVD, ICE, IVE, IV) or other novel combinations (GV, IGEV).

  2. BCNU, carmustine; CR, complete response; ORR, overall response rate.

Martin et al (2001)Mini-BEAMBCNU; etoposide; cytarabine; melphalan55841
Aparicio et al (1999)ESHAPEtoposide; methylprednisolone; cytarabine; cisplatin22731
Josting et al (2005a)DHAPDexamethasone; cytarabine; cisplatin102880
Wimmer et al (2006)APECytosine arabinoside, cisplatin, etoposide31680
Hickish et al (1993)EPICEtoposide, prednisolone, ifosfamide, cisplatin40580
Schellong et al (2005)IEP-ABVDIfosfamide, etoposide, prednisolone, adriamycin, bleomycin, vinblastine, dicarbazine176850
Moskowitz et al (2001)ICEIfosfamide; carboplatin; etoposide65880
Proctor et al (2003)IVEIfosfamide; etoposide; epirubicin51840
Bonfante et al (2001)IVIfosfamide; vinorelbine47830
Cole et al (2009)GVGemcitabine, vinorelbine30760
Santoro et al (2007)IGEVIfosfamide; gemcitabine; vinorelbine; prednisolone91810

Prognostic factors at relapse

Time to disease recurrence and response to salvage chemotherapy are the pre-eminent prognostic factors (Claviez et al, 2004; Lieskovsky et al, 2004). An important distinction is between primary progressive disease and disease that relapses after a remission. Time to recurrence is highly significant for overall survival (OS) and EFS with primary progression consistently adverse for OS (James et al, 1992; Baker et al, 1999; Balwierz et al, 2000). Time to treatment failure was the strongest prognostic factor for survival in the ST-HD-86 trial (P = 0·0001), the largest published salvage series in paediatric HL (Schellong et al, 2005). Three risk groups emerged, defined as progressive disease on or within 3 months of primary treatment which had the worst prognosis (disease-free survival [DFS] 41%, OS 51%), early relapse 3–12 months from primary treatment had a better OS (DFS 55%, OS 78%), and late relapse over 12 months from primary treatment had a significantly better DFS (DFS 86%, OS 90%).

Response to re-induction chemotherapy and disease status at transplant are also highly predictive of outcome. Patients who are refractory to salvage chemotherapy have a very poor prognosis even with HDCT consolidation: 5-year failure-free survival (FFS) of 0% vs. 59% in chemo-sensitive patients (P < 0.001) (Claviez et al, 2004); 5-year FFS of 9% vs. 44% in chemo-sensitive patients (P = 0·06) (Baker et al, 1999); 5-year OS 18% and FFS 0% vs. 68% OS and 59% FFS in chemo-sensitive patients (Claviez et al, 2008). Several series have assessed disease status at transplant. Pre-transplant elevated lactate dehydrogenase (LDH), a marker of active disease, was associated with a 5-year FFS of 0% vs. 42% in those with normal a pre-transplant LDH (P < 0·001) (Baker et al, 1999). A large residual mediastinal mass at time of transplant predicted poor OS, EFS and progression-free survival (PFS) (P < 0·05) (Lieskovsky et al, 2004). All studies used conventional imaging and there is no published data regarding FDG-PET in relapsed paediatric HL.

Additional significant parameters for DFS and OS are original stage, stage at relapse, local versus distant site relapse and amount and type of primary therapy. Stage IV and extra-nodal disease at relapse are significant predictors of poor OS, EFS and PFS (Claviez et al, 2004; Lieskovsky et al, 2004). Adolescents with B symptoms at recurrence had an 11-year OS of 27% vs. 60% without after HDCT/ASCT, (Akhtar et al, 2010). Relapse in original and distant sites had a 3-year OS of 0%, while local relapse only had a 5-year DFS of 90%. (P < 0·001) (James et al, 1992). Salvage success rate is influenced by primary chemotherapy with improved DFS after primary treatment with two cycles of chemotherapy, rather than after four or more cycles and after primary regimens which did not include Procarbazine/Etoposide (Schellong et al, 2005).

Prognostic factors in adult HL are similar. Time to disease recurrence, stage, anaemia at relapse, B symptoms, extra-nodal disease and chemo-sensitivity are significant poor prognostic factors identified in relapsed HL in adults. (Brice et al, 1997; Moskowitz et al, 2001, 2004; Josting et al, 2002, 2005a). Treatment decisions are less driven by prognostic factors in adult trials as most patients with relapsed HL are considered candidates for autologous transplantation. Some adult guidelines do advocate a risk-stratified approach with favourable risk relapse receiving SDCT and IFRT, intermediate risk receiving HDCT/ASCT and poor risk receiving consolidation with tandem HDCT or reduced intensity conditioning (RIC)-alloSCT (Brice, 2008). SDCT ± RT achieved a 5-year FFS of 58% and 10-year OS of 68% in adults without adverse features who had relapsed from a previous CR, demonstrating that selection of patients at relapse for a risk-adapted approach is appropriate (Vassilakopoulos et al, 2002).

Paediatric trials utilize prognostic factors to allocate salvage strategies but there is no internationally agreed, reproducible prognostic model, which would allow comparison across clinical trials. The small numbers in some studies, the variability in primary treatment, variation in RT use, field and dose and inclusion of patients with variable clinical characteristics in salvage studies are some of the reasons why there remains debate about the optimal approach to salvage therapy.


Patients with late relapse, defined as over 12 months from primary treatment, who are responsive to salvage chemotherapy achieve excellent survival with SDCT and RT. In the Deutsche Arbeitsgemeinschaft für Leukämieforschung und –behandlungim Kindesalter (DAL)-ST-HD-86 study, SDCT plus RT achieved a 10-year DFS of 86% and OS of 90% in all late relapse patients, and in low stage disease originally treated with two cycles of chemotherapy, 10-year DFS was 96% (Schellong et al, 2005). Salvage chemotherapy consisted of alternating ifosfamide, etoposide and prednisolone (IEP) and doxorubicin, bleomycin, vinblastine and dacarbazine (ABVD) followed by RT to regions involved at relapse and original sites if primary therapy had not included RT. The primary therapy for patients in this trial had been the very effective chemo-RT regimens of the DAL/Gesellschaft für Pädiatrische Onkologie und Hämatologie (GPOH) HD trials. One hundred and seventy-six patients were enrolled with progressive disease (n = 51) and first relapse (n = 125). The probabilities at 10 years for the whole group were 75% (OS), 62% (DFS) and 57% (EFS). This study showed that salvage may be risk-adapted and that SDCT has a fundamental role in the salvage of late relapse patients. In a British study, patients salvaged with SDCT with late and local relapse, had a 5-year DFS of 90% (James et al, 1992). In another British study, salvage with SDCT and RT achieved excellent 10-year OS except in stage IV patients (Shankar et al, 1997). Several other series have demonstrated no benefit in outcome in patients salvaged with HDCT versus SDCT regardless of stage or duration of first remission (Stoneham et al, 2004; Belgaumi et al, 2009).

The DAL-ST-HD-86 trial forms the basis of the current risk-adapted relapse strategies for paediatric classical HL within the European inter-group ‘EuroNet’ PHL-C1 trial (Fig 1). Low risk patients receive only SDCT plus IFRT. The North American Children’s Oncology Group (COG) protocols also studied SDCT salvage in late asymptomatic and non-bulky relapse in those initially treated for IA/IIA disease (Chen et al, 2007).


Patients who may be salvaged with SDCT plus RT are those with low risk relapse defined as late relapse, limited stage and chemotherapy responsive disease.


HDCT is confined to salvage of higher risk relapse and primary progressive patients in most international paediatric studies. The lack of any randomized trial comparing SDCT and HDCT in paediatric patients combined with very limited evidence for a survival benefit of HDCT in first relapse has limited the universal adoption of HDCT in paediatric studies. Most studies have failed to show a survival benefit with HDCT except in patients with primary progressive disease and multiply-relapsed disease. (Claviez et al, 2004; Stoneham et al, 2004; Schellong et al, 2005; Belgaumi et al, 2009). Evidence for the use of HDCT has been extrapolated from adult trials, which showed poor OS with conventional salvage (Longo et al, 1992; Bonfante et al, 1997), and the promising early studies that compared HDCT/ASCT with SDCT in two Phase III randomized trials, which showed an improved disease-free survival with HDCT (Linch et al, 1993; Schmitz et al, 2002). These provide the most compelling evidence for a higher FFS after HDCT in chemo-sensitive relapse and refractory disease.

The British National Lymphoma Investigation (BNLI) group reported 3-year EFS of 53% (HDCT arm) versus 10% (SDCT arm) but no significant OS advantage was demonstrated (Linch et al, 1993). The German Hodgkin Study Group (GHSG) HD-R1 trial (Schmitz et al, 2002) reported 3-year FFS of 55% (HDCT arm) versus 34% (SDCT arm). At a median follow up of 83 months there remained no difference in OS. The lack of a survival benefit in these trials is attributed to patients in the non-transplant arm undergoing transplant at second relapse, which suggests that HDCT at first or second relapse provides comparable outcomes. However, avoiding toxicities of multiple salvage regimens and the potential anxiety associated with multiple relapses, in combination with acceptable morbidity and mortality of current HDCT are compelling reasons why many physicians recommend auto-SCT at first relapse.

Results of paediatric salvage studies with HDCT

Series with HDCT/ASCT in paediatric and adolescent patients are small and report EFS rates of 31–67% (Table II) (Williams et al, 1993; Baker et al, 1999; Verdeguer et al, 2000; Frankovich et al, 2001; Claviez et al, 2004, 2008; Lieskovsky et al, 2004; Stoneham et al, 2004; Schellong et al, 2005).

Table II.   Outcome after autologous SCT for paediatric relapsed/refractory HL.
ReferencenOS (%)PFS/DFS/FFS/EFSFollow-up (years)
  1. SCT, stem cell transplantation; HL, Hodgkin lymphoma; OS, overall survival; PFS, progression-free survival; DFS, disease-free survival; FFS, failure-free survival; EFS, event-free survival.

Williams et al (1993)816439% PFS3
Baker et al (1999)534331% FFS5
Verdeguer et al (2000)259562% EFS5
Frankovich et al (2001)347667% DFS5
Stoneham et al (2004)517266% DFS4·5
Lieskovsky et al (2004)416863% PFS4·2
Claviez et al (2004)745950% FFS2·8
Prete et al (2005)916763% EFS5
Harris et al (2006)386843% PFS2·2
Akhtar et al (2010)586253%3·6

In the DAL-ST-HD-86 trial (Schellong et al, 2005) in early and progressive HL at first relapse, the 6-year DFS was 51% and OS 66% (HDCT group), compared with 47% and 65% respectively, for the SDCT group, which was not significant. In contrast, a survival advantage was observed for HDCT in multiply-relapsed patients whose 6-year OS was 62% compared with 29% in those salvaged with SDCT (P = 0·04).

Two studies have shown no difference in outcome between paediatric and adult patients treated for relapsed and refractory HL with HDCT. A case control study in 81 children and 81 adults with relapsed HL, matched for prognostic factors, who received HDCT/ASCT found no difference in outcome between the two groups with similar PFS rates of 39% (paediatric) versus 48% (adult) and notably no significant difference in incidence or causes of procedure related morbidity and mortality (Williams et al, 1993). Another study found no significant difference in 5-year FFS or OS in 53 children (31% and 43% respectively) who received HDCT/ASCT when compared to an adult cohort (29% and 42% respectively) (Baker et al, 1999). The outcome for children with chemotherapy-resistant relapse consolidated with HDCT was dismal in this study (<10%) and it is difficult to determine if autologous transplantation is justified in this chemo-refractory group.

A small study of 20 children with relapsed, refractory and very poor prognosis HL evaluated the outcome of myeloablative chemo-radiotherapy and ASCT (Verdeguer et al, 2000). The projected 5-year OS and EFS rates were 95% and 62%. Disease status at ASCT was not a significant factor in this small series.

A 5-year DFS of 67% was reported in 34 refractory and relapsed HL patients who received HDCT (Frankovich et al, 2001). Extra-nodal disease was adverse for DFS and OS and bulky disease at time of HDCT was adverse for DFS. Idiopathic diffuse lung injury developed in 44% of patients in this study. The timing of thoracic RT (pre-, post- or peri-ASCT) was not a significant risk factor for this however a pre-existing history of atopy was highly predictive of pulmonary complications (P = 0·0001).

A British retrospective comparison of relapsed and refractory patients salvaged with HDCT (n = 51) versus SDCT (n = 78) found no difference in OS after first relapse (P = 0·4) regardless of duration of first remission (Stoneham et al, 2004). The 5-year OS with duration of first remission < or >1 year in the HDCT cohort was 74% and 84% respectively (P = 0·6), and 67% and 86% respectively (P = 0·07), in the SDCT cohort. This study concluded that HDCT did not offer any significant survival advantage over SDCT in children with relapsed HL. However, first-line treatment was relatively non-intensive, single modality RT or non-alternating ChlVPP (chlorambucil, vinblastine, procarbazine, prednisolone), and children who received the HDCT were a worse prognosis group (P < 0·001). A survival benefit for HDCT was found however, in primary progressive HL, with OS of 70% with HDCT versus 0% with SDCT.

A 5-year OS, EFS and PFS of 68%, 53% and 63% respectively, was reported in 41 children who underwent HDCT/ASCT for relapsed and refractory HL (Lieskovsky et al, 2004). Multivariate analysis showed three significant predictors for poor OS and EFS: extranodal disease at first relapse, presence of a mediastinal mass at time of HDCT and primary induction failure.

One study reported the influence of salvage chemotherapy response and time to relapse on outcome in paediatric HDCT patients (Claviez et al, 2004). For the whole group the 5-year OS was 59% and FFS was 50%. The 5-year OS and FFS was 68% and 59% in chemo-sensitive patients versus 18% and 0% in chemo-resistant patients respectively. Therefore, HDCT/ASCT was inadequate treatment for patients with chemo-refractory disease and alternative approaches should be explored in these patients. Time to relapse also attributed to survival significantly with OS of 43% in progressive disease, 70% in early relapse and 75% in late relapse (P = 0·02). A larger series reported 5-year OS of 67% and EFS of 63% in 91 patients with HDCT ASCT in relapsed and refractory HL (Prete et al, 2005). Chemo-sensitive patients had a better outcome, with patients transplanted in second CR having5-year EFS of 83% but this was not significant. A recent study of HDCT in 58 patients with relapsed and refractory HL reported EFS of 45% and OS of 55% at 11 years with the only negative prognostic factor for OS being the presence of B symptoms at relapse or progression, OS 27% with versus 60% without B symptoms (P = 0·003) (Akhtar et al, 2010).

The conclusions from these paediatric series are that HDCT and ASCT is recommended in all primary progressive HL who are chemo-sensitive. HDCT also provides a survival benefit in multiply relapsed patients after SDCT. Patients with primary progressive and chemo-refractory disease at salvage have poor outcomes with HDCT and alternative approaches are urgently required in this group. For patients with early relapse, the outcome is intermediate between late relapse and progressive HL and the best salvage strategy is unclear as no survival benefit has been demonstrated with HDCT in this group. The outcome with HDCT is similar to adult HL and adult experience may inform paediatric paradigms.

HDCT conditioning regimes

No randomized trial has established the superiority of one conditioning regimen over another in HL patients consolidated with HDCT ASCT. Currently chemotherapy only regimens are favoured over total body irradiation (TBI)-containing conditioning. There is an increased risk of myelodysplastic syndrome and second cancers associated with TBI and no benefit has been demonstrated in TBI-containing regimens. The CBV (cyclophosphamide, carmustine and etoposide) regimen has been most widely used in North America, whereas in Europe the BEAM [BCNU (carmustine), etoposide, cytarabine, melphalan] regimen is used.


In paediatric patients RT has been omitted by some groups as part of first-line therapy in selected, low stage disease. Low dose RT and IFRT are becoming standard management when combined modality treatment is indicated. A number of ongoing trials are exploring omitting RT in patients with a good response to chemotherapy. At relapse therefore some patients will be RT naive and others will have received only low dose RT, which enables further RT to be incorporated into relapse strategies. A minority of patients with relapse HL achieve long-term disease control with salvage RT alone and this approach is rarely used. It is not currently a standard recommendation in paediatric practice at relapse. It may be appropriate in selected patients with limited stage disease, relapse at sites of original disease only and who are considered poor candidates for intensive chemotherapy. The 5-year DFS in several series with salvage RT is around 40% (Leigh et al, 1993; Campbell et al, 2005). In another study the GHLSG (Josting et al, 2005b) found a 5-year FFS of 28% and OS 51%. Adverse factors for survival were short duration of initial remission, B symptoms, advanced disease and extra nodal disease. Patients enrolled in these studies are a highly selected patient population, only 2% of patients in the HD4-HD9 trials underwent this approach. In some studies IFRT has been recommended prior to HDCT in order to reduce disease bulk and improve local disease control although this approach has been questioned due to concerns regarding radiation-induced pneumonitis and is not currently standard management in children. RT is not currently incorporated into the conditioning regimes due to concerns regarding toxicity although this remains an approach in some adult series (Moskowitz et al, 2001).

Most relapses after HDCT occur at sites of previous disease, particularly bulk disease (Poen et al, 1996), suggesting that the addition of RT post-HDCT may be beneficial. Use of IFRT pre- and post-HDCT was shown to improve freedom from relapse in patients who had never received prior RT and patients with stage I-III disease in one study (Poen et al, 1996). For patients with persistent disease post-HDCT, a significantly better PFS (40% vs. 12.1%) was seen in those who received IFRT in another study (Mundt et al, 1995). The current European paediatric salvage approach incorporates post-transplant FDG-PET to guide decisions regarding RT in the HDCT group, IFRT is limited to those patients who are FDG-PET positive after HDCT. In the SDCT group, the addition of RT is standard in all patients, applied to regions involved at relapse or regions involved at relapse plus original sites if primary therapy had not included RT.

Allogeneic stem cell transplantation in relapsed and refractory Hodgkin lymphoma

There is increasing interest in the use of alloSCT in highly selected paediatric patients with HL. Traditionally this has been in the context of relapse post-autologous transplant but there is increasing recognition that patients with primary progressive HL, which does not achieve a CR with SDCT, have a poor outcome with conventional HDCT/ASCT consolidation. Interest in alloSCT derives from the potential of an immune-mediated graft-versus-lymphoma (GVL) effect and the potential role of post-transplant immunotherapy using donor lymphocyte infusions to activate the GVL effects. Use of alloSCT has been extremely limited in paediatric practice as so few patients fail to respond to primary and conventional salvage strategies and the prohibitively high rates of non-relapse mortality (NRM). RIC was introduced to ameliorate NRM while maintaining the GVL effect (Peggs et al, 2005).

Paediatric data regarding the role of alloSCT for HL is limited. The largest paediatric and adolescent series published are by the European Group for Blood and Marrow Transplantation (EBMT). Claviez et al (2007) reported the experience of 151 patients with relapsed and refractory HL aged <21 years at transplantation. Forty percent had myeloablative conditioning (MAC) and 60% RIC. The OS at 25 months was 50%. The probability of PFS at 2 and 5 years was 39% and 29% respectively, with a Cumulative relapse rate of 37% and 44% at 2 and 5 years, respectively. The cumulative transplant-related mortality (TRM) at 2 and 5 years was 24% and 27%. An increased relapse rate was associated with RIC, which became significant at 12 months follow up. Inferior PFS was associated with refractory disease and poor performance status.

A follow up EBMT series analyzed 91 children and adolescents <18 years with refractory and relapsed HL (Claviez et al, 2009). Fifty-one received RIC and 40 MAC regimens. The 5-year OS and PFS was 45% and 30%, respectively. Relapse remained the major reason for treatment failure with cumulative relapse rates at 2 and 5 years of 36% and 44%, respectively. The NRM at 1 year was 21% and 26% at 5 years, with no difference between MAC and RIC regimens. Highly encouraging results were described in a subgroup, defined by chemo-sensitive disease and good performance status, who achieved 3-year OS of 83% and PFS of 60%. RIC was associated with an increased risk of relapse, apparent from 9 months following transplant. There was no demonstrable effect of chronic graft-versus-host disease (GvHD) on treatment outcomes, in contrast to some adult series.

Other paediatric series include a tandem myeloablative ASCT followed by RIC alloSCT in 10 patients, of which 8/10 proceeded to RIC alloSCT (Bradley et al, 2006). Six were alive with no evidence of disease at 1 year and 1-year OS was 67%. MAC autoSCT followed by RIC alloSCT was feasible and well tolerated. A recent single centre retrospective study in children and young adults demonstrated 2-year PFS in chemo-refractory relapse of 40% with RIC alloSCT and 19% with autologous SCT although this was not significant (Shafer et al, 2010).

The larger alloSCT series in children compare favourably to survival rates in adult series after alloSCT. In 167 adults who received an alloSCT as their first transplant procedure, 4-year OS was only 24%. The procedure-related mortality at 4-years was 51% and the relapse rate was higher in the alloSCT group compared with autoSCT (Peniket et al, 2003). Outcomes after MAC and RIC alloSCT regimens were compared in a later EBMT study (Sureda et al, 2008). The 1-year NRM after RIC alloSCT was 23% vs. 46% after MAC alloSCT. The 5-year OS and PFS rates for RIC alloSCT were 28% and 18% respectively, and 22% and 20% respectively, for MAC alloSCT. The largest adult RIC alloSCT series in HL was reported recently (Robinson et al, 2008). In 285 adults the 3-year NRM was 21%, OS was 43% and the PFS was 25%. Disease progression was the major cause of treatment failure following RIC alloSCT with 59% relapse rate at 5-years. NRM was associated with chemo-refractory disease, poor performance status and age > 45 years. Patients with no risk factors had a 3-year NRM of 12% compared to 46% with two or more risk factors. PFS and OS were associated with performance status and disease status at transplant. Those with neither risk factor had a 3-year PFS and OS of 42% and 56% respectively, compared to a dismal 8% and 25% for patients with one or more risk factor.


AlloSCT represents a curative option in a subset of highest risk patients in whom there are probably no other realistic options for cure at present. However much remains to be achieved to reduce NRM rates of over 20% and even higher relapse rates in paediatric and adult series. In the EBMT series (Claviez et al, 2009) the NRM of MAC and RIC procedures was similar in children. It may be reasonable to offer MAC or ‘intermediate’ intensity regimens to children that can tolerate them in order to reduce the relapse rate post-alloSCT.

Prospective multicentre studies are required to better define the indications, optimal timing conditioning and GvHD prophylactic regimens in children and adolescents.

Late effects, fertility preservation and second malignancies

Whilst effective disease control is the priority aim for any salvage treatment regimen, an understanding and assessment of the late effects of treatment remains important because of the good salvage rate. The development of late effects of treatment is directly related to treatment exposure although some studies have shown HL itself to be an independent risk factor in developing late effects, including second cancers. Late effects of treatment include psychosocial morbidities, endocrine deficiency including infertility, cardiovascular toxicity and second solid and haematological malignancies.

Second cancer is the leading cause of death in long-term survivors of HL, with exceptionally high rates of breast cancer in survivors treated with radiation therapy to the chest at a young age, and a 7–18 times higher risk of malignancies compared to the general population (van Leeuwen et al, 2003). The risk of secondary leukaemia is early and plateaus at 10–15 years after therapy, whereas the risk of second solid tumours, including sarcoma, melanoma, lung, breast and thyroid, continues to increase with follow up. Young age at treatment has a major effect on risk of second malignancy, with children developing more second cancers than adults, and the pattern of solid tumour type is also influenced by treatment age, with breast cancer more common after childhood and young adult HL and lung cancer after adult HL (Travis et al, 2005). Breast cancer risk increases with increasing radiation exposure up to at least 40 Gy (Travis et al, 2003) however a dose of 4 Gy or more delivered to the breast was associated in one study with a 3·2-fold excess risk (95% confidence interval [CI] 1·4–8·2). The risk of second cancer is elevated post-HDCT/SCT and may be influenced by conditioning regimen and type of graft (Metayer et al, 2003).

Anthracycline-associated cardiomyopathy, with or without mediastinal radiation, remains an important cause of mortality and morbidity in survivors and has been reviewed by the Scottish Intercollegiate Guidelines Network (SIGN, 2004). Data from paediatric HL survivors show overall standardized mortality ratios of 5·8–8·2 for cardiac mortality (Mertens et al, 2001; Moller et al, 2001).

Thyroid dysfunction may be seen after neck RT in HL and hypothyroidism, which peaks 3–5 years after RT, shows a dose response risk.

Male infertility is associated with alkylating agent exposure and the risk is dose-dependent. Azoospermia rates up to 86% have been reported (Ben Arush et al, 2000). Use of alternating and multi-agent protocols with reduced cumulative doses of individual drugs has reduced infertility risks. Menstrual irregularity, ovarian failure and infertility occur in females and risk increases with age at treatment (Bath et al, 2002). Fertility preservation must be considered in young patients who are likely to be sterilized by their planned second-line treatment. For boys and young men able to produce semen, sperm should be cryopreserved. The options for young girls and women without a partner are experimental and include ovarian strip cryopreservation (Wallace et al, 2005). To date, there have been at least 10 pregnancies worldwide after othotopic re-implantation of frozen-thawed ovarian cortex. The success rate is unclear as the denominator (the number of women in whom frozen-thawed ovarian tissue has been re-implanted) is unknown. Ovarian cryoprepservation remains a challenging but important area that requires to be considered in young women who relapse and require further treatment that may induce a premature menopause (Wallace & Barr, 2010). Long-term follow up of survivors of childhood cancer, particularly those who have survived second-line treatment, remains essential (Skinner et al, 2006).

New chemotherapy combinations and new agents

The success rates of conventional therapies in paediatric HL are excellent and this has limited the search for novel therapies. Most experience with new chemotherapy combinations has been gained with Gemcitabine regimens. The efficacy and toxicity of weekly Gemcitabine and Vinorelbine in relapsed and refractory paediatric and young adult HL demonstrated an ORR of 76% with good tolerability (Cole et al, 2009). The response rate in this study was 50% after two cycles, rising to 76% with application of a further two cycles, demonstrating the importance of not stopping too early, especially if proceeding to HDCT. The IGEV regimen has also shown encouraging results with an ORR of 80% (Santoro et al, 2007).

The search for other novel treatment strategies have focussed on the Hodgkin Reed-Sternberg (HRS) cells antigens and receptors, which can be targeted using naked or conjugated monoclonal antibodies. CD30 is highly expressed on HRS cells but has very limited expression in normal tissues and provides an obvious target. The activity of anti-CD30 antibodies has been disappointing when used as humanized or chimeric antibody (Leonard et al, 2004). Others have attempted to increase activity using radiolabelled I-131-antiCD30 and one study achieved responses in 9/22 patients, however toxicity was significant (Schnell et al, 2005). A novel immunotoxin SGN-35, the antiCD30 antibody conjugated to a synthetic microtubule agent, is under study. One phase I study achieved PR or CR in 37% and some documented tumour reduction in 88% of relapsed HL patients (Younes et al, 2008).

Intracellular survival pathways that favour survival of HRS cells may be targeted by small molecules. Histone deacetylase (HDAC) inhibitors are under investigation in a number of studies. The balance between histone acetyltransferases and histone deacetylation is critical for regulating expression and functional status of proteins involved in cell survival and proliferation. Panobinostat is a pan-HDAC inhibitor that is currently under study in a large international phase II study in relapsed HL (Younes et al, 2009). This follows a phase I trial in 13 patients with haematological malignancies including relapsed HL in which 38% of patients achieved partial remissions (Prince et al, 2007). The oral mammalian target of rapamycin (mTOR) inhibitor everolimus achieved PR in 47% of patients with relapsed HL in a small study, which will be very promising if confirmed in larger studies (Johnston et al, 2007). The first attempt to inhibit nuclear factor (NF)-κB pathways in HL used the proteasome inhibitor bortezomib. Bortezomib has shown disappointing response and has no place as a single agent (Younes et al, 2006) but there is interest in exploring the synergy between bortezomib and chemotherapy. A COG phase II study, currently closed, is evaluating bortezomib in combination with ifosfamide and vinorelbine in children and young adults.

Targeting the microenvironment around the HRS cells is also of interest to disrupt its supportive role because, when removed from this environment, HRS cells are extremely difficult to grow in culture. There is interest in the anti-CD20 antibody rituximab to deplete B cells in the microenvironment, to target those few HRS cells that are CD20 positive and, theoretically, to target the putative HRS stem cell. A pilot study of 22 classical HL patients treated with single agent rituximab demonstrated 22% partial or complete remission and eight additional patients had stable disease (Younes et al, 2003). There is interest in combining rituximab with chemotherapy in managing classical HL. ABVD plus rituximab (R-ABVD) demonstrated improved EFS in all risk groups in newly diagnosed HL, indicating its activity in classical HL (Wedgwood et al, 2007). Lenalidomide has shown promising activity in relapsed HL (Fehniger et al, 2008; Kuruvilla et al, 2008).


The approach to salvage therapy for classical HL in children and adolescents may be risk adapted. Low risk patients with late nodal relapse without adverse features achieve excellent cure rates with SDCT plus IFRT and there is little room for improvement in survival in this group. In intermediate risk patients, with early relapse or late relapse with adverse features, the best salvage approach between SDCT and HDCT is unknown. The challenge in this group is to identify those who will be cured with SDCT and those who require HDCT.

The greatest challenge is patients with primary progressive HL who have the worst survival. For primary progressive HL conventional HDCT is the treatment of choice for chemotherapy responsive HL but chemo-refractory patients have a sub-optimal survival with conventional HDCT and an alternative approach is urgently required to improve the outcome in this group. Emerging evidence from studies incorporating FDG-PET pre-HDCT confirms significantly sub-optimal outcomes in patients who remain PET positive, while those who achieve CR/FDG-PET negativity show substantially improved outcomes. Achieving remission on functional imaging should be a goal of re-induction therapy and conventional HDCT will remain standard practice in this setting. FDG-PET positivity should not, however, preclude patients from potentially curative therapies. AlloSCT may have a role in this setting, although truly refractory patients have a suboptimal outcome in many series regardless of which transplant approach is applied. A major goal of future studies must therefore be in defining new chemotherapy combinations and incorporating novel agents to achieve maximal response pre-SCT and to clarify the indications for alloSCT as first transplant approach.