Over 90% of children with Hodgkin lymphoma (HL) now reach long term remission with conventional treatment (Behrendt et al, 1987; Oberlin et al, 1992; Hudson et al, 1993; Schellong et al, 1999; Landman-Parker et al, 2000). Nevertheless, 2–15% of children with HL relapse or have refractory disease (progression during initial therapy or within 3 months following treatment) after initial response. Patients with relapsed or refractory HL have several different treatment options available (Byrne & Gockerman, 2007; Cashen & Bartlett, 2007). During the 1990s, autologous stem cell transplantation (ASCT) was shown to improve progression-free survival (PFS) and overall survival (OS) in adult patients with relapse/refractory disease (Chopra et al, 1993; Horning et al, 1997; Yuen et al, 1997; Schmitz et al, 2002). Consequently, the most frequently used therapies for relapsed/refractory HL include salvage chemotherapy with second-line drugs and high-dose therapy with ASCT. Indications and dose of radiotherapy depends on primary therapy and relapse locations. Relapse outcome depends on first line treatment, stage at relapse, and time to relapse, but only a few reports have investigated long term results and determined factors associated with survival in children in the context of initial response, risk adapted strategy treatment, and ASCT in relapse (Williams et al, 1993; Baker et al, 1999; Verdeguer et al, 2000; Lieskovsky et al, 2004; Harris et al, 2011; Shafer et al, 2011). Given that the cumulative incidence of HL relapse is around 10%, collecting data over a long period is a challenge. This study retrospectively analysed the databases of 31 French centres of the Société Française de lutte contre les Cancers de l'Enfant et de l'Adolescent (SFCE). Data from 70 children with relapsed or refractory HL was recorded over a period of 16 years. All these children were initially treated by combined modality treatment with standard dose radiotherapy of 20 Gy. The purpose of the study was to evaluate the outcome of these patients and determine prognostic factors in order to propose alternative therapy and define, in this particular context, a group at high risk for adverse outcomes who would be eligible for experimental strategies.
There is little data available regarding children and adolescents with Hodgkin lymphoma (HL) who relapse after combined-modality treatment, even though they have a substantial chance of cure. The purpose of this national retrospective study was to evaluate the outcome of patients with recurrent/refractory HL and determine adverse prognostic factors. From 1990 to 2006, 70 patients (median age 13·9 years) with refractory (n = 31) or first relapse (n = 39) HL were identified. Median time from end of treatment to relapse was 6 months (3–56). Relapses occurred in irradiated areas in 43/70 patients. Salvage therapy consisted of chemotherapy and 50 patients received high-dose chemotherapy with autologous stem cell transplantation. Radiotherapy was performed in 29 cases, tandem autologous transplantation in five and allograft in three. With a median follow-up of 40 months (2–140), significant prognostic factors were time to progression/relapse and response to therapy before autograft. Event-free survival and overall survival in patients with refractory disease, early relapse and late relapse were 35 ± 9%, 67 ± 11%, 76 ± 10% and 48 ± 11%, 89 ± 7% and 80 ± 10%, respectively. As progression <3 months was a major adverse prognostic factor, novel therapeutic approaches are needed for this group of patients. By contrast, patients have substantial chance of long term second remission in case of relapse >3 months.
French Paediatric Haematology Oncology centres were asked to report all HL relapses in children that occurred between January 1990 and December 2006. Patients with lymphocyte predominant HL were excluded from the study. Reports at relapse included initial staging and therapy, relapse stage, treatment, and outcome. Progressive/refractory disease was defined as the following: an increase from the initial tumour diameter or the appearance of any new lesion during or <90 days after the end of therapy. Relapsed HL was defined as the appearance of any new lesion more than 3 months after the end of treatment following complete disappearance of HL with first-line therapy. Relapse between 3 and 12 months after completion of therapy was defined as early relapse and relapse after 12 months as late relapse. Informed consent to register data was obtained according to French national health ethics rules.
Overall survival (OS) was measured from the date of diagnosis of first recurrence to the last visit or death. Disease progression, second malignancies, death from any causes, or last follow-up defined events for the calculation of event-free survival (EFS). Progression of HL was the only event defined in disease-free survival (DFS). OS, EFS, and DFS were calculated according to the Kaplan-Meier method using XLStat life software. Differences between groups were analysed by using the log-rang test and a significant P value < 0·05. Variables analysed as prognostics factors included the following: gender, initial stage, interval between the end of treatment and relapse, stage at relapse, site of relapse (nodal vs. extranodal), systemic symptoms at relapse, and response to chemotherapy prior to high-dose therapy. Response evaluation after salvage chemotherapy was available for all patients and based on computed tomography evaluation. Definition of response was: complete remission (CR) or complete remission unconfirmed (CRu) for response >70%, partial remission (PR) with a response < 70% or progressive disease.
Characteristics of the patients with progression/first relapse
We collected the data of 70 patients (39 males) with refractory HL (n = 31) or first relapse of HL (n = 39). Median age at diagnosis was 13·6 ± 2·9 years (range 2·9–17·7) and 15 ± 3 years at relapse (range 4–20) Characteristics of the cohort at diagnosis are detailed in Table 1. Diagnosis of relapse was confirmed by histological analysis in all cases. Characteristics of the cohort at first relapse are detailed in Table 2. Thirty-one patients (44%) had refractory/progressive disease, 20 patients (29%) had early relapse and 19 patients (27%) had late relapse. Excluding patients with refractory disease, the median time of relapse was 6 months (3–56 months). Only five relapses were observed after 24 months.
|Stage I||3 (4)|
|Stage II||30 (43)|
|Stage IIE||4 (6)|
|Stage III||12 (17)|
|Stage IIIE||3 (4)|
|Stage IV||18 (26)|
|B symptoms||45 (64)|
|Nodular sclerosis||64 (91)|
|Mixed cellularity||6 (9)|
|MDH 90||49 (70)|
|MDH 03||20 (29)|
|Bone marrow involvement||2||7|
|Relapse < 3 months||6||8|
|Relapse 3–12 months (early)||20||28|
|Relapse > 12 months (late)||19||28|
|Initial response of salvage therapy|
|High-dose chemotherapy with ASCT||50||71|
After clinical and radiological work up, eight patients (11%) were classified with stage I disease, 27 patients (39%) had stage II disease, eight patients (11%) had stage III disease, and 27 patients (39%) had stage IV disease. Pulmonary involvement was the predominant extra-nodal localization (n = 16) followed by bone involvement (n = 11), bone marrow involvement (n = 2), and liver involvement (n = 1). Sixteen out of 70 patients (23%) had clinical systemic symptoms. Forty-three patients had relapsed in-field radiotherapy sites. Considering clinical evolution between initial diagnosis and relapse, we observed 17 cases (24%) of upstaging and 56% of patients with stage II disease relapsed with stage II disease with mediastinal involvement. Fourteen out of 18 patients with stage IV disease at diagnosis relapsed with stage IV disease (P < 0·001).
We did not observe any histopathological HL subtype modification between diagnoses and relapse. Almost all of the patients had the nodular sclerosis form of HL (n = 64; 91%), whereas mixed cellularity (n = 6; 9%) was uncommon. Initial treatment varied according to the initial HL stage and the protocol in use at the time of diagnosis (French Society of Paediatric Oncology trials MDH 90[Landman-Parker et al, 2000] and MDH 03, and ABVD [doxorubicin, bleomycin, vinblastine, dicarbazine]). At initial staging, patients were classified as follows: stage I (n = 3), stage II (n = 30), stage IIE (n = 4), stage III (n = 12), stage IIIE (n = 3), and stage IV (n = 18). Forty-five children (64%) had clinical systemic symptoms. Initial treatment included chemotherapy in all patients and 50 patients (71%) additionally received involved field radiotherapy at 20 Gy. Twenty children did not receive radiotherapy at first treatment because of progression during or immediately after chemotherapy.
Relapse treatment and response evaluation
Relapse treatment was according to physicians' decisions and adapted to the initial therapy received by the patients (Table 3). Sixty-nine patients received salvage chemotherapy with either mitoguazone, ifosfamide, vinorelbine, and etoposide (MINE) (Ferme et al, 1995); ifosfamide, etoposide, and doxorubicin (IVA) (Morschhauser et al, 2008); vincristine, prednisone, procarbazine, and doxorubicin (OPPA) (Schellong et al, 1999); or other classical polychemotherapies, such ABVD or cyclophosphamide, vincristine, procarbazine, and prednisone (COPP). Thirty-five patients received MINE alone for a median of two courses (1–4) or MINE in combination with IVA (n = 2), OPPA (n = 8), or ABVD (n = 3). One patient received MINE combined with rituximab. Twelve patients received IVA alone for a median of two courses (1–4) or IVA combined with MINE (n = 2). Twelve patients received OPPA in combination with MINE (n = 6) or COPP (n = 5). Overall, the MINE strategy was the main choice for 35 first treatments at relapse. Aspects of the disease influenced the physician choice, mainly initial response and delay with previous anthracyclin-based treatment. A trend to use MINE in refractory cases was observed. By contrast, 10/12 patients receiving OPPA were initially treated without any anthracyclin treatment as in the MDH 90 protocol for localized disease. Nine patients received other chemotherapies including ABVD (n = 2), mustargen, vincristine, procarbazine, and prednisone (MOPP) with ABVD (n = 2), high-dose cyclophosphamide (n = 1), doxorubicin, methylprednisolone, high-dose cytarabine, and cisplatin (ASHAP; n = 1), and others (n = 3). One patient received radiotherapy only as initial treatment of the relapse. The median number of salvage chemotherapy cycles given was 3 (range 1–8 cycles).
|Therapy (number of courses)||Total||Progressive disease||Early relapse||Late relapse|
|MINE (2–4) + ASCT||28||40||12||17||11||16||5||7|
|IVA (1–4) + ASCT||11||16||6||9||2||3||3||4|
|OPPA (2) + ASCT||6||9||2||4||1||1||3||4|
|Other + ASCT||5||6||1||1||3||4||1||1|
|MINE (2) + Other||7||10||3||4||2||3||2||3|
|IVA (2) + Other||1||1||1||1||0||0||0||0|
|OPPA + Other||6||9||3||4||1||1||2||3|
Fifty-one patients (73%) achieved CR or CRu. Eight patients (11%) achieved PR. Eleven patients (16%) failed to respond and exhibited evidence of progression in the peripheral node (n = 3), mediastinum (n = 5), or pulmonary localization (n = 3). We observed CR+CRu in 20 (57%) cases treated with MINE and 7 (58%) cases treated with IVA. Early evaluation of salvage therapy was performed after two cycles in 54 (77%) cases. Twenty-nine out of 54 patients (54%) achieved CR or CRu after two cycles of salvage therapy, and 24 of these patients (83%) are in second CR. Nineteen out of 54 patients (35%) had achieved PR after two cycles; however, only five achieved CR after salvage therapy and only four of these patients are still in second CR. Seven out of 54 (13%) had progressive disease after two cycles of chemotherapy and only two patients achieved CR with subsequent treatment (Fig 1).
Eighteen patients (26%) did not receive ASCT. Seven of these patients (39%) had progressive disease after salvage therapy and six of these patients received palliative care after three lines of treatment. Four patients died and two patients had progressed or had stable disease after palliative care. One child with PR after salvage therapy received complementary radiotherapy and was treated for a second relapse at 8 months after the first relapse. Ten patients were in second CR after salvage therapy: seven patients in the late relapse group, two in early relapse group and one in the progressive disease group. All 10 patients received complementary radiotherapy and eight are in second CR after a median follow-up of 90 months.
High doses chemotherapy and stem cell transplantation
Fifty-two children received stem cell transplantation after salvage therapy: 45 received single ASCT, five received double ASCT, and two were allografted after salvage therapy. ASCT was performed in 21/31 patients in the progressive disease group, 17/20 patients in the early relapse group and 12/19 in the late relapse group. Of the 45 patients who received single ASCT, 15 were in second CR, 22 were in CRu, five were in PR, and three had disease progression before ASCT. Forty patients (89%) underwent a conditioning regimen with BEAM (carmustine 300 mg/m2 [day-5], cytarabine 800 mg/m2 [days-5 to -2], etoposide 800 mg/m2 [days-5 to -2], and melphalan 140 mg/m2 [day-1]) and 5 patients underwent a conditioning regimen with CBV (cyclophosphamide 1·2–1·8 g/m2 for 4 days, etoposide 125–400 mg/m2 twice daily for 3 days, and carmustine 300–600 mg/m2 for 1 day). After ASCT, 29 patients received radiotherapy. ASCT-related mortality was reported in one case with veno-occlusive disease after BEAM therapy.
Tandem ASCT was conducted in five patients in patients with refractory disease (Morschhauser et al, 2008), two of whom were in CR or CRu, 3/5 were in PR. Patients received tandem ASCT with CBV followed by BEAM (n = 1), CBV followed by total body irradiation, cytarabine, and melphalan (TAM; n = 1), and BEAM followed by TAM therapy (n = 3). Four out of five patients achieved second CR and one patient relapsed. Two patients died: one from treatment-related toxicity for myelodysplasia and one due to a second malignancy.
Two patients were allografted, one in second CR before transplant (alive in second CR with a follow-up of 24 months), and one in PR after salvage chemotherapy; however, this patient, died from treatment-related toxicity.
With a median follow-up time from relapse of 40 months (range 2–140), 27 patients (39%) suffered progression/relapse: there were 10 progressions during salvage therapy and 17 relapses with a median time of second relapse of 9 months (range 2–26). Nineteen patients (27%) died, with 14 patients (20%) dying from HL and five patients dying from another event (treatment toxicity [n = 2], second malignancy [n = 1], accidental death [n = 1], or sudden death [n = 1]). One patient is receiving treatment for secondary acute myeloid leukaemia.
Fifty-one patients are still alive: 38 in second CR, seven in third CR, three are currently receiving treatment for subsequent relapse, and three are receiving palliative care. In these patients, the OS was 69 ± 6%, 5-year EFS was 55 ± 6% and DFS was 60 ± 6%.
We examined the impact of the presence or absence of various risk factors on outcome in a univariate analysis. The following factors had no significant impact on DFS: age, gender, stage, systemic symptoms at relapse, and ASCT. Relapse in a previous radiation field was documented in 43 cases but not associated with inferior outcome (EFS 80% P = 0·13). Stage IV disease at relapse was associated with inferior outcome but this was not significant (EFS 36%; P = 0·08) as well as stage IV at initial diagnosis (EFS 61% P = 0·68). Mediastinal involvement at relapse was present in 25 cases and was not associated with inferior outcome (EFS 76% P = 0·28). A risk factor analysis revealed time to progression/relapse as a strong and unique prognostic factor (Table 4). Thus, patients with progressive disease experienced a significantly (P ≤ 0·05) shorter OS and EFS compared with patients with early or late relapse. In contrast, OS and EFS were similar between patients with early or late relapse (Table 3 and Fig 2). When we consider only late relapse, median delay to relapse was 19 and 20 months before and after 1996, respectively (median inclusion time in the study), and therefore we can expect relatively stable results considering that no second relapse was observed after 26 months follow up.
|Male||39||64 ± 10||0·46||58 ± 8||0·64|
|Female||31||73 ± 8||63 ± 8|
|Stage at diagnosis|
|Lymphatic involvement||45||76 ± 7||0·22||65 ± 7||0·25|
|Extranodal (IV+E)||25||54 ± 13||52 ± 10|
|Systemic symptoms at diagnosis|
|A||45||76 ± 11||0·48||58 ± 7||0·9|
|B||23||64 ± 8||62 ± 10|
|Stage at relapse|
|Lymphatic involvement||43||74 ± 8||0·25||58 ± 8||0·81|
|Extranodal (IV+E)||27||60 ± 12||64 ± 10|
|Time to disease recurrence|
|Progressive disease||31||48 ± 11||0·05||40 ± 9||0·002|
|Relapse 3–12 months (early)||20||89 ± 7||74 ± 10|
|Relapse > 12 months (late)||19||80 ± 10||78 ± 10|
|Yes||50||66 ± 8||0·6||64 ± 7||0·21|
|No||18||61 ± 13||52 ± 11|
Overall outcome was better in patients with CR or CRu after salvage chemotherapy (n = 51): DFS was observed in 75% of patients with CR or CRu compared with 10% of patients with PR or progressive disease (P < 0·0001). Initial response to therapy after two cycles of salvage chemotherapy also seemed to be a prognostic factor: DFS was seen in 24/29 patients (81%) with CR or CRu compared with six patients (17%) with PR or progressive disease (P < 0·001) (Fig 3).
Four patients with progressive disease underwent transplantation (ASCT n = 3, allograft n = 1). Only one is alive in CR2.
Global prognosis of HL in children with combined-modality therapy is successful more than 90% but about 10% fail to attain a complete remission or subsequently relapse (Behrendt et al, 1987; Schellong et al, 1999; Hudson et al, 2004). Remission rates for patients who relapse after initial treatment have improved since the introduction of high-dose chemotherapy with ASCT, with OS and DFS rates of approximately 65% and 60%, respectively (Chopra et al, 1993; Horning et al, 1997; Yuen et al, 1997; Byrne & Gockerman, 2007). Published reports on salvage therapy in paediatric patients have been limited, but the results of OS and DFS were similar (Baker et al, 1999; Verdeguer et al, 2000; Lieskovsky et al, 2004). A retrospective study by the German Paediatric Hodgkin Study Group involved 176 children with progressive disease or first relapse, with 10-year OS and DFS of 75 ± 4% and 62 ± 4%, respectively (Schellong et al, 2005). Median time from second relapse was 9 months (range 2–26 months). Univariate analyses identified time to relapse as a risk factor with a strong predictive impact on outcome. EFS at 5 years was 40% in the group with refractory disease (progression or relapse <3 months) and 74% and 76% in the groups with relapse between 3 and 12 months (early) or >12 months (late), respectively (P = 0·01). However, this risk factor is probably related to the introduction of a high-dose chemotherapy and ASCT treatment strategy for early relapse, which was introduced in the 1990's (Lohri et al, 1991; Stoneham et al, 2004; Brice, 2008). More recently, Shafer et al (2011) reported the outcome of teenagers and young adults treated by stem cell transplantation; OS was 73% and PFS was 60%. In our study, the only relevant prognostic factor appears to be related to early relapse and chemoresistance. Several additional prognostic factors for poor outcome were identified in adults. Moskowitz et al (2004) developed a prognostic model to predict the outcomes of patients with relapsed or refractory HL and found that three factors were also associated with outcome: extra nodal sites of disease, early relapse and B symptoms. Lieskovsky et al (2004) determined three factors to be significant predictors of poor outcome: extra nodal disease at first diagnosis, primary induction failure, and mediastinal mass at time of ASCT. We found no significant contribution of these reported prognostic factors, but OS and DFS for patients with stage IV at relapse seemed to be inferior.
High-dose chemotherapy is the current treatment option for patients with early relapsed or refractory HL. In our study, 52 of 70 patients had received high-dose chemotherapy with ASCT. The most used conditioning regimen used with ASCT in our study was BEAM, similar to previously published reports (Chopra et al, 1993; Horning et al, 1997; Yuen et al, 1997; Schmitz et al, 2002). Alternative experience of high-dose chemotherapy by CBV (cyclophosphamide, carmustine and etoposide) with ASCT in recurrent HL was also reported in children, with 5-year OS and DFS of 63% and 45%, respectively (Harris et al, 2011).
Response to initial salvage treatment was also a very strong prognostic factor in relapse HL patients and the importance of cytoreduction before ASCT in prognostic outcome was demonstrated in adults (Moskowitz et al, 2004; Sirohi et al, 2008) The role of chemoresistance in paediatric patients was also reported (Metzger et al, 2010): the EFS with inadequate response at salvage therapy was only 18% compared with 97% for those who responded. In our study, DFS for patients who obtained a good response with salvage chemotherapy (complete remission or partial remission >70%) was 77% compared with 10% for patients with poor response or progression (P < 0·0001). Advances in supportive care during ASCT have subsequently reduced the treatment-related mortality to <3%. In our study, one patient developed myelodysplasia and died from allograft-related toxicity, and one died from second malignancy.
Patients with relapsed HL > 12 months generally have a good prognosis and can be cured in approximately 80% of cases. In our study, this was clearly the case of the patients initially treated with a VBVP regimen and 20 Gy radiation therapy. For patients with late relapse the most appropriate treatment in order to limit adverse effects of treatment will depend on initial treatment (with or without radiation therapy) and response to first line salvage treatment, nowadays evaluated by early F18-fluorodeoxyglucose positron emission computed tomography. In these patients with early good response, salvage therapy without ASCT could be recommended (Josting et al, 2002; Schellong et al, 2005; Brice, 2008). The place of radiation therapy in relapse strategy, as well as in first line treatment for advanced stage disease, is still a matter of debate. Usually, physicians are prefer radiation therapy in early stage relapse with residual disease and initial radiation therapy is limited or absent (Stoneham et al, 2004; Josting et al, 2005). Unfortunately low numbers of patients and treatment heterogeneity did not allow any conclusion in our study.
Patients with progressive disease had poor outcomes (EFS < 40%) despite ASCT. Curative treatment options are limited for patients with progressive disease after first salvage treatment. New approaches are needed and prospective Phase II studies should be proposed, such as tandem ASCT, reported in a study with 46% DFS and 57% OS in patients with very poor prognostic factors (Morschhauser et al, 2008). They authors suggested that the benefit of tandem ASCT was associated with good responders after one ASCT but not to chemoresistant patients. In our study, only five patients were treated with tandem ASCT. The response to treatment was good and four out of five patients achieved second CR but with significant toxicity.
The use of allogenic stem cell transplantation for treatment of relapsed or refractory HL with poor risk factors prognostic remains controversial although an allogenic effect is demonstrated in clinical observations (Milpied et al, 1996; Sureda & Schmitz, 2002; Peggs et al, 2005; Sureda et al, 2008; Robinson et al, 2009). There are a few published studies in children or adolescents (Claviez et al, 2009; Shafer et al, 2011) that demonstrate high treatment-related mortality (16–61%) and low PFS rate (21–45%). This approach has to be investigated by clinical research. Recently, innovative drugs have been reported in Phase II studies with promising results (Younes et al, 2010, 2011) Their place in the relapse strategy treatment has now to be defined. In conclusion, high intensity treatment with ASCT and risk-adapted strategies has improved the global success rate of relapsed HL in children. However, early relapse with poor response to treatment and/or progression of disease before 3 months is still associated with poor outcomes. This fact justifies further research into innovative treatment strategies for children and adolescents with Hodgkin lymphoma.
The authors would like to thank Geneviève Vaudre for support in data collection and Simone Boniface for English language editing.
Conflict of interest statement
The authors report no potential conflicts of interest.