Initial response to salvage therapy determines prognosis in relapsed pediatric Hodgkin lymphoma patients

Authors

  • Monika L. Metzger MD, MSc,

    Corresponding author
    1. Department of Oncology, St. Jude Children's Research Hospital, Memphis, Tennessee
    2. University of Tennessee Health Science Center, Memphis, Tennessee
    • Department of Oncology, St. Jude Children's Research Hospital, Memphis, TN 38105-2794
    Search for more papers by this author
    • Fax: (901) 521-9005

  • Melissa M. Hudson MD,

    1. Department of Oncology, St. Jude Children's Research Hospital, Memphis, Tennessee
    2. University of Tennessee Health Science Center, Memphis, Tennessee
    Search for more papers by this author
  • Matthew J. Krasin MD,

    1. Department of Radiological Sciences, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • Jianrong Wu PhD,

    1. Department of Biostatistics, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • Sue C. Kaste DO,

    1. Department of Oncology, St. Jude Children's Research Hospital, Memphis, Tennessee
    2. University of Tennessee Health Science Center, Memphis, Tennessee
    3. Department of Radiological Sciences, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • Larry E. Kun MD,

    1. Department of Radiological Sciences, St. Jude Children's Research Hospital, Memphis, Tennessee
    Search for more papers by this author
  • John T. Sandlund MD,

    1. Department of Oncology, St. Jude Children's Research Hospital, Memphis, Tennessee
    2. University of Tennessee Health Science Center, Memphis, Tennessee
    Search for more papers by this author
  • Scott C. Howard MD, MSc

    1. Department of Oncology, St. Jude Children's Research Hospital, Memphis, Tennessee
    2. University of Tennessee Health Science Center, Memphis, Tennessee
    Search for more papers by this author

Abstract

BACKGROUND:

Pediatric Hodgkin lymphoma (HL) is a highly curable disease; however, prognostic factors for the survival of patients who develop recurrent disease have not been clearly defined.

METHODS:

This was a retrospective analysis of 50 pediatric patients with HL who relapsed or progressed between 1990 and 2006 and who were retrieved with intense cytoreductive treatment regimens followed by autologous stem cell transplantation and radiation therapy. A Cox proportional hazards model was used to determine risk factors for second treatment failure and death.

RESULTS:

The median patient age was 16.1 years (range, 4.9-22.1 years) at the time of HL diagnosis. Fifteen patients developed progressive disease during therapy, 14 patients relapsed early, and 21 patients relapsed late. Patients who remained alive at the time of this study had been followed for a median of 4.4 years (range, 1.2-16.6 years). The 5-year overall survival rate for patients who had an inadequate response (n = 14) to initial salvage therapy was only 17.9% (95% confidence interval [CI], 3.1%-42.5%) compared with 97.2% (95% CI, 81.9%-99.6%) for patients who responded (n = 36; P < .0001). In a multivariate Cox regression analysis of overall survival, an inadequate response to initial salvage therapy was the only significant variable (hazard ratio, 43.6; 95% CI, 5.4-354; P = .0004).

CONCLUSIONS:

The current results indicated that pediatric patients with relapsed HL who have an inadequate response after initial primary salvage chemotherapy have a very poor prognosis and should be considered for novel therapies directed at biologic or immunologic targets. Cancer 2010. © 2010 American Cancer Society.

Children and adolescents with Hodgkin lymphoma (HL) achieve 5-year event-free survival rates of approximately 85% with modern combined-modality therapy.1 Even those who fail frontline therapy have a 10-year overall survival (OS) rate as high as 75%.2 Given the small number of patients that fail primary therapy, no uniform salvage strategy has been developed for this patient population. Most salvage regimens contain new drug combinations that are not included in frontline therapy, additional radiation therapy, and high-dose chemotherapy with autologous stem cell transplantation (SCT).2, 3 Several groups have tried to identify prognostic factors predictive of frontline treatment failure in children and adults.4, 5 In adults, a prognostic score for patients with relapsed HL has been developed in which a shorter time to treatment failure (<12 months after the end of primary therapy), a higher clinical stage at treatment failure (stages III and IV), and anemia at the time of treatment failure predict a poor outcome.6 In children and adolescents, only progressive disease during frontline therapy, as opposed to relapse >1 year after the completion of therapy, has been identified as prognostic for a low rate of second event-free survival.2 The objective of the current study was to identify factors that predicted outcome in pediatric patients with HL who progressed or relapsed after initial treatment with a combined-modality therapy regimen that included both chemotherapy and radiation.

MATERIALS AND METHODS

Between January 1990 and December of 2006, 342 patients with newly diagnosed HL were treated at St. Jude Children's Research Hospital with risk-stratified, response-adapted chemotherapy plus involved-field radiation. Six of the 56 patients who had failed frontline therapy were excluded from the study for the following reasons: One patient who had a late relapsed refused any further therapy and died of disease, and 5 patients received their salvage therapy elsewhere.

Frontline Therapy

The chemotherapy and radiation therapy regimens that were used between 1990 and 2001 have been described elsewhere7-9 and are summarized in Table 1. Patients with high-risk HL who were diagnosed after 2001 and those with intermediate-risk HL who were diagnosed after mid-2004 received 12 weeks of weekly multiagent (vinblastine, nitrogen mustard, doxorubicin, etoposide, vincristine, bleomycin, and prednisone), dose-intensive chemotherapy (the “Stanford V” regimen) followed by response-based, low-dose, involved-field radiation.10 Five patients received a combination of vinblastine, doxorubicin, methotrexate, and prednisone (VEPA) and cyclophosphamide, vincristine, and procarbazine (COP) (see Table 1, footnote), and none of those patients relapsed. All patients were staged according to the Ann Arbor staging classification11 with computed tomography (CT) scans, positron emission tomography (PET) or gallium scans, and bone marrow biopsies when indicated. All patients were restaged to evaluate early response as indicated by their protocol and no later than 2 weeks after they finished chemotherapy in preparation for their radiation therapy. No patient was delayed for >4 weeks between the completion of chemotherapy and the initiation of radiotherapy.

Table 1. Frontline Treatment Regimens
Frontline Treatment RegimensCumulative Doses of Chemotherapy
4 VAMP6 VEPA2 VAMP/2 COP3 VAMP/3 COPStanford V
  • VAMP indicates vinblastine, doxorubicin, methotrexate, and prednisone; VEPA, vinblastine, etoposide, prednisone, and doxorubicin; COP, cyclophosphamide, vincristine, and procarbazine; Stanford V, vinblastine, nitrogen mustard, doxorubicin, etoposide, vincristine, bleomycin, and prednisone; CR, complete response; Gy, grays; PR, partial response; PD, progressive disease.

  • a

    Of the 5 patients who received VEPA/COP, none relapsed.

  • b

    Radiation dose (involved field) was determined by early response to therapy. Numbers in brackets refer to the number of relapsed patients who originally received the given dose.

  • c

    For low-risk patients who received VAMP and were in CR and enrolled after January 2000.

  • d

    To bulky sites, which were defined as mediastinal-to-thoracic ratio ≥33% or peripheral adenopathy ≥6 cm.

  • e

    To bulky sites, which were defined as a mediastinal-to-thoracic ratio ≥33%.

Total no. of patients treateda9042329776
No. of relapsed patients (%)7 (8)8 (19)6 (19)18 (19)11 (14)
Anthracyclines
 Doxorubicin, mg/m2200300100150150
Alkylators
 Cyclophosphamide, mg/m2  24003600 
 Procarbazine, mg/m2  28004200 
 Nitrogen mustard, mg/m2    18
Vinca alkaloids
 Vincristine, mg/m2  5.68.48.4
 Vinblastine, mg/m24848243636
Epipodophyllotoxins
 Etoposide, mg/m2 1600  360
Other
 Prednisone, mg/m222401120112016801680
 Methotrexate, mg/m2160 80120 
 Bleomycin, U/m2    30
Radiation therapyb
 CRNone [1],c 15 Gy [2]15 Gy [1]
 PR25 Gy [2]25 Gy [7]d25 Gy [4]e25 Gy [16]d25 Gy [9]e
 PD before radiation[2][1][2][2][1]

Follow-Up and Treatment Failure

After frontline therapy, patients were evaluated every 3 months for the first year, every 4 months for the next 2 years, every 6 months for the next 2 years, and yearly thereafter. Evaluations consisted of laboratory studies, chest x-rays, and physical examinations at every visit and CT scans with functional imaging (gallium or PET scans, depending on the era of treatment) at the first and second yearly follow-up visits and at the discretion of the treating physician if symptoms developed between screening visits. When treatment failure was suspected, patients were evaluated with CT scans and functional imaging studies as well as bilateral bone marrow biopsies. Treatment failure was confirmed histologically in all patients.

Salvage Therapy

Salvage regimens were chosen by the treating physician according to previous treatment intensity, timing of treatment failure, patient status, and physician preference. In the earlier era, patients were more likely to be retrieved with a nitrogen mustard (cyclophosphamide), vincristine, procarbazine, and prednisone (M[C]OPP)/doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) regimen or with an M(C)OPP/ABVD-derived regimen (n = 11). In recent years, patients more often were retrieved with high-dose methotrexate, ifosfamide, etoposide, and dexamethasone (MIED) or with ifosfamide, carboplatin, and etoposide (ICE) (n = 26); in some cases of localized late treatment failure, patients received Stanford V chemotherapy (n = 11). One patient received VEPA as an initial salvage regimen, and 1 patient received gemcitabine/vinorelbine (G/V). Salvage therapy included high-dose chemotherapy with SCT in all high-risk patients who responded to first or subsequent salvage chemotherapy. Autologous transplantation conditioning regimens varied during the 17-year study period: 5 patients received carboplatin/etoposide, 3 patients received cyclophosphamide/busulfan, 2 patients received cyclophosphamide/topotecan, 2 patients received busulfan/melphalan, 1 patient received MOPP, and 18 patients received combined carmustine, etoposide, cytosine arabinoside, and melphalan. After SCT, radiation therapy was administered to lymph node sites that were affected at the time of treatment failure.

Response Assessment

Adequate responses to initial salvage therapy included complete responses (CR) (ie, a reduction ≥75% reduction in the original tumor on a CT scan [using the product of the 2 greatest tumor dimensions to determine size] and negative gallium or PET scans) or partial responses (PR) (a 50%-74% reduction in tumor size in a CT scans, regardless of PET avidity). An inadequate response was defined as anything that did not meet these criteria. No patient who attained a CR had a positive PET scan.

Statistical Methods

Treatment failure was defined as progressive disease or any form of relapse, and patients were classified with progressive disease if new lesions developed or if an existing lesion increased in size by >25%, or if treatment failure occurred within 3 months of the end of therapy. Early relapse was defined as treatment failure from 3 to 12 months after the completion of therapy, and late relapse was defined as any failure that occurred thereafter. These definitions were chosen in accordance with previous studies of patients with relapsed/progressive Hodgkin lymphoma.2 Cutoff values that were used to categorize white blood cell count (≤13.5 × 109/L vs >13.5 × 109/L), hemoglobin (<10.5 g/dL for girls and <12.0 g/dL for boys vs greater than or equal to those levels), and erythrocyte sedimentation rate (<50 mm per hour vs ≥50 mm per hour) also were based on values that were identified as predictive of outcome in previous studies.5, 6 A Cox proportional hazards model was used to identify risk factors for second treatment failure and death.12 For the multivariate Cox regression analysis, only variables that were identified as significant (P < .05) in the univariate Cox regression analysis were considered. Freedom from second treatment failure was calculated from the date of first treatment failure to the date of an event (second treatment failure, disease progression, or death, whichever occurred earliest), and patients who did not have an event were censored at the last follow-up date. OS was calculated from the date of first treatment failure to the date of death from any cause. Living patients were censored at the last follow-up date. Freedom from second treatment failure and OS were estimated using the Kaplan-Meier method.13 Log-rank tests were used to compare freedom from second treatment failure and OS distributions according to demographic and treatment characteristics. All analyses were performed in SAS software (version 9.1l SAS Institute, Inc., Cary, NC). After receiving institutional review board approval, we reviewed medical records for demographic information, details of therapy, and disease and survival outcomes.

RESULTS

Patient Characteristics at Diagnosis of Treatment Failure

Patient characteristics at diagnosis and second treatment failure are described in Table 2. The median age for the 50 patients (30 males and 20 females) was 16.1 years (range, 4.9-22.1 years) at the diagnosis of HL and 17.5 years (range, 7.1-24.6 years) at the time of first treatment failure. Seventy-two percent of patients (n = 36) were white, and 28% (n = 14) were black. Seventeen patients (34%) had localized disease (stage IA or IIA), and 33 patients (66%) had advanced disease (stage IIB, IIIA, IIIB, or IV) at initial diagnosis. B symptoms (drenching night sweats, unexplained fevers, or weight loss >10%) were present in 44% of patients at initial diagnosis. Forty patients (80%) had nodular sclerosing histology, and 19 patients (38%) had extranodal disease (lung, liver, bone, or bone marrow) at the time of treatment failure. Fifteen patients (30%) progressed during frontline therapy, 14 patients (28%) relapsed early (3-12 months after the completion of initial therapy), and 21 patients (42%) relapsed late (>12 months after the completion of therapy). First treatment failures occurred after a median of 14.6 months (range, 1.1-121 months). Forty-nine of 50 patients (98%) had follow-up within the last year.

Table 2. Patient Characteristics
CharacteristicNo. of Patients, n = 50%
  • a

    Progressive disease was defined as the development of new lesions, an increase >25% in the size of an existing lesion, or treatment failure within 3 months of the end of therapy.

  • b

    Early relapse was defined as treatment failure between 3 months to 12 months after the completion of therapy.

  • c

    Late relapse was defined as treatment failure >12 months after the of completion of therapy.

Age at diagnosis, y
 Median16.1 
 Range4.9-22.1 
Age at first treatment failure, y
 Median17.5 
 Range7.1-24.6 
Race
 Black1428
 White3672
Sex
 Male3060
 Female2040
Ann Arbor stage at initial diagnosis
 I12
 II2448
 III714
 IV1836
B symptoms at initial diagnosis
 No2856
 Yes2244
Histology
 Nodular sclerosing4080
 Other1020
Sites of first treatment failure
 Lymph node only3162
 Extranodal1938
Type of treatment failure
 Progressive diseasea1530
 Early relapseb1428
 Late relapsec2142

Response to Frontline and Salvage Regimens

Table 3 summarizes disease response and ultimate outcomes for specific patients according to frontline therapy, failure type (primary progressive, early relapse, or late relapse), salvage therapy, and use of SCT. Response rates to salvage regimens (MIED, 76%; Stanford V, 91%; MOPP and derivatives, 55%; and ICE, G/V, and VEPA, 33%) varied and potentially were biased by patient characteristics (timing and bulk of recurrence as well as intensity of frontline therapy), so response rates could not be compared. Of the 15 patients who had primary progressive disease, 7 patients (54%) did not have an adequate response to primary salvage therapy; all received additional cycles of intense chemotherapy, but 5 patients never demonstrated sensitivity to any salvage chemotherapy and died of their disease. Two patients eventually responded to salvage therapy and underwent consolidation therapy with autologous SCT and radiation therapy. One of those patients died of disease, and the other patient has remained disease-free 4 years after his initial treatment failure.

Table 3. Response to Frontline and Salvage Regimensa
Frontline Treatment RegimenDisease Stage at Initial DiagnosisType of First Treatment FailureFirst Salvage TherapyAdequate Response to First Salvage Therapy?Autologous Stem-Cell TransplantationOutcome
  • VAMP indicates vinblastine, doxorubicin, methotrexate, and prednisone; MOPP, nitrogen mustard, vincristine, prednisone, and procarbazine; ABVD, doxorubicin, bleomycin, vinblastine, and dacarbazine; DD, dead of disease; MIED, high-dose methotrexate, ifosfamide, etoposide, and dexamethasone; CR2, second complete remission; COPP, cyclophosphamide, vincristine, prednisone, and procarbazine; Stanford V, vinblastine, nitrogen mustard, doxorubicin, etoposide, vincristine, bleomycin, and prednisone; VEPA, vinblastine, etoposide, prednisone, doxorubicin; COP, cyclophosphamide, vincristine, and procarbazine; ICE, ifosfamide, carboplatin, and etoposide; AWD, alive with disease; CR3, third complete remission; Gem/Vin, gemcitabine/vinorelbine.

  • a

    Numbers in parentheses indicate the numbers of patients in each category.

VAMP (7)IA (1), IIA (6)Progression (2)MOPP/ABVD (1)No (1)Yes (1)DD (1)
MIED (1)Yes (1)Yes (1)CR2 (1)
Early (2)COPP/ABVD (2)Yes (1)No (1)CR2 (1)
No (1)No (1)DD (1)
Late (3)Stanford V (3)Yes (3)No (3)CR2 (3)
VEPA (8)IIA (1), IIB (2), IIIA (1), IVA (1), IVB (3)Progression (2)MOPP (2)Yes (2)No (1)CR2 (1)
Yes (1)CR3 (1)
Early (3)MOPP (3)Yes (2)Yes (2)CR2 (2)
No (1)Yes (1)DD (1)
Late (3)MOPP (3)Yes (1)Yes (1)CR2 (1)
No (2)Yes (2)DD (2)
VAMP/COP ×4 (6)IIA (4), IIIA (2)Progression (2)ICE (1)No (1)Yes (1)CR3 (1)
MIED (1)Yes (1)Yes (1)CR2 (1)
Early (2)Stanford V (2)Yes (2)No (2)CR2 (2)
Late (2)MIED (2)Yes (2)Yes (2)CR2 (2)
VAMP/COP ×6 (18)IIA (4), IIB (4), IIIB (3), IVA (4), IVB (3)Progression (6)MIED (4)No (3)No (3)DD (3)
Yes (1)Yes (1)DD (1)
VEPA (1)No (1)No (1)DD (1)
Stanford V (1)No (1)No (1)DD (1)
Early (2)MIED (2)Yes (2)Yes (1)CR2 (1)
Yes (1)DD (1)
Late (10)Stanford V (5)Yes (5)No (3)CR2 (3)
Yes (2)CR3 (2)
MIED (5)Yes (3)No (1)CR2 (1)
Yes (2)CR2 (1)
AWD (1)
No (2)No (1)DD (1)
Yes (1)CR3 (1)
Stanford V (11)IIA (1), IIB (2), IIIB (1), IVA (3), IVB (4)Progression (3)MIED (3)Yes (3)Yes (3)CR2 (3)
Early (5)MIED (5)Yes (4)Yes (4)CR2 (4)
No (1)Yes (1)CR2 (1)
Late (3)MIED (2)Yes (2)Yes (2)CR2 (2)
Gem/Vin (1)Yes (1)Yes (1)CR2 (1)

Three of 14 patients who relapsed early (21%) did not have an adequate response to primary therapy; 1 of those patients refused further therapy and died of disease, whereas the other 2 patients received intense second-line salvage therapy with autologous SCT, after which 1 patient died of disease and the other remained alive and in remission for 2 years. Four of 21 patients (19%) who relapsed late did not have an adequate response to initial salvage therapy: Only 1 of those patients remained alive and disease free >5 years after he had further intensive chemotherapy, autologous SCT, and radiation therapy. Of the other 3 patients who died—all from disease or treatment-related complications—2 patients died after undergoing autologous SCT despite an inadequate response to initial therapy.

Of the 36 patients who had an adequate initial response to their primary salvage regimen, only 2 patients died of disease after subsequent treatment failure, and 1 patient has remained alive with disease for the past 10 years with short-lasting responses to various therapies. Four other patients who had an initial adequate response to salvage therapy had second treatment failures but were retrieved with a third-line regimen and remain disease-free.

Freedom From Second Treatment Failure and Survival

For all patients, the 5-year rate of freedom from second treatment failure was 57.8% (95% confidence interval [CI], 42.1%-70.6%), and the 5-year OS rate was 74.2% (95% CI, 58.7%-84.6%) (Fig. 1). The 5-year OS rate was significantly lower for patients who had primary progressive disease (53.3%; 95% CI, 26.3%-74.3%) than for patients who relapsed early or late (83.2%; 95% CI, 63.6%-92.8%; P = .01) (Fig. 2, top). When analyzing OS according to adequacy of response to primary salvage therapy, patients who did not respond to their primary salvage therapy had a 5-year OS rate of only 17.9% (95% CI, 3.1%-42.5%) compared with 97.2% (95% CI, 81.9%-99.6%) for patients who did respond (P < .0001) (Fig. 2, bottom). Patients who remained alive have been followed for a median 4.9 years (range, 1.5-18.1 years), and no patients have been lost to follow-up. All living patients have been seen or contacted within the last 15 months.

Figure 1.

Freedom from second treatment failure and overall survival are illustrated for all patients. The 5-year overall survival rate was 74.2% (95% confidence interval [CI], 58.7%-84.6%), and the rate of 5-year freedom from second treatment failure was 57.8% (95% CI, 42.1%-70.6%).

Figure 2.

(Top) Overall survival is illustrated for all patients according to type of treatment failure after frontline therapy. The 5-year overall survival rate for patients who had primary progressive disease was 53.3% (95% confidence interval [CI], 26.3%-74.3%) versus patients who relapsed early or late (83.2%; 95% CI, 63.6%-92.8%). (Bottom) Overall survival is for all patients according to the adequacy of response to primary salvage therapy. The 5-year overall survival rate according to adequacy of response to primary salvage therapy was 97.2% for patients who had a complete or partial response (95% CI, 81.9%-99.6%) and 17.9% for patients who had stable or progressive disease (95% CI, 3.1%-42.5%).

Prognostic Factors for Second Treatment Failure and Death

Several prognostic factors were analyzed for their potential impact on second treatment failure. Neither white blood cell count (≤13.5 × 109/L vs >13.5 × 109/L) at treatment failure nor lymph node treatment failure versus extranodal treatment failure was associated significantly with second treatment failure. The only significant prognostic factors for second treatment failure in univariate analysis were initial progressive disease (hazard ratio [HR], 2.8; 95% CI, 1.1-7.1; P = .02), anemia (hemoglobin <10.5 g/dL for girls and <12 g/dL for boys vs greater than or equal to those levels) at treatment failure (HR, 2.6; 95% CI, 1.0-6.5; P = .04), erythrocyte sedimentation rate (<50 mm per hour vs ≥50 mm per hour) at treatment failure (HR, 2.6; 95% CI, 1.0-6.8; P = .04), and inadequate response to initial salvage therapy (HR, 65.4; 95% CI, 13.7-311; P < .0001). In a multivariate Cox regression analysis that was controlled for the type of first treatment failure (progressive disease, early, late), the erythrocyte sedimentation rate, and anemia, only an inadequate response to initial salvage therapy was significant (HR, 73; 95% CI, 14-381; P < .0001).

The same prognostic factors were analyzed for their potential association with death (Table 4). Factors that were associated significantly with death in univariate analysis were an erythrocyte sedimentation rate ≥50 mm per hour (HR, 3.7; 95% CI, 1.2-11; P = .02), initial progressive disease (HR, 4.3; 95% CI, 1.1-17; P = .04), and an inadequate response to initial salvage therapy (HR, 50; 95% CI, 6.3-388; P = .0002). In a multivariate Cox regression analysis that was controlled for erythrocyte sedimentation rate and type of first treatment failure, only an inadequate response to initial salvage therapy was significant (HR, 44; 95% CI, 5.4-354; P = .0004). Patients who progressed during chemotherapy before receiving their prescribed radiotherapy were not at greater risk of death compared with patients who progressed after receiving their prescribed frontline radiotherapy (3 of 8 patients whose disease progressed before frontline radiation therapy died vs 4 of 7 patients whose disease progressed after radiation therapy; P = .38).

Table 4. Prognostic Factors for Death After First Treatment Failure of Pediatric Hodgkin Lymphoma
FactorNo. of Patients (%)Univariate AnalysisMultivariate Analysis
All, n = 50Deaths, n = 13HR [95% CI]PHR [95% CI]P
  • HR indicates hazard ratio; CI, confidence interval.

  • a

    Anemia was defined as a hemoglobin level <10.5 g/dL for females and <12.0 g/dL for males.

Anemia at first treatment failurea
 Yes136 (46)3.0 [1.0-8.9].05
 No377 (19)1   
Erythrocyte sedimentation rate at first treatment failure, mm/h
 <50387 (18)1 1 
 ≥50126 (50)3.7 [1.2-11.1].021.2 [0.3-4.6].81
Sites of first treatment failure
 Lymph node only317 (23)1 
 Extranodal196 (32)1.4 [0.5-4.3].52  
Type of first treatment failure
 Progressive disease157 (47)4.3 [1.2-16.7].042.2 [0.5-9.4].27
 Early relapse143 (21)1.2 [0.2-6.0].831.1 [0.2-6.0].92
 Late relapse213 (14)1 1 
Adequate response to primary retrieval therapy
 Yes362 (6)1 1 
 No1411 (79)49.6 [6.3-388].000243.6 [5.4-354].0004

DISCUSSION

The strongest predictor of survival in our study was whether patients had an adequate response to their first 2 cycles of intense salvage chemotherapy (termed “chemosensitivity”). However, the role of chemosensitivity as a predictor of outcome has been controversial. Earlier studies in adults suggested that chemosensitivity before autologous SCT does not predict progression-free survival,14, 15 whereas more recent studies evaluating prognostic factors and treatment outcome in adult patients with primary progressive HL clearly demonstrate that sensitivity to conventional salvage chemotherapy regimens predicts survival.16-20 Comparison of these studies, however, is not straightforward, because the study by Chopra et al.14 evaluated a cohort of adult patients who underwent autologous SCT despite the absence of chemosensitive disease; whereas most patients who had chemoresistant disease in the study by Josting et al.21 either died rapidly from their disease or were deemed ineligible for autologous SCT. In a cohort of pediatric patients with HL who had recurrent or progressive disease and underwent autologous SCT, the factors associated with a poor outcome were extranodal disease at the time of treatment failure, mediastinal mass at the time of transplantation, and progression during frontline therapy (risk ratio, 6.4; 95% CI 1.6-25.3; P = .008).22, 23 Extranodal disease in our study was not associated with a poorer prognosis. In another retrospective study of 51 pediatric patients with relapsed or refractory HL, autologous SCT was no better than conventional salvage therapy in patients with relapsed disease but was beneficial for patients with primary refractory HL.24

In our cohort, of the 14 patients who had an inadequate response to primary salvage therapy, 10 patients died of progressive disease (despite additional aggressive salvage chemotherapy and SCT in 4 patients), and 3 patients remained alive and disease free (2.5 years, 3.5 years, and 6 years after initial treatment failure) after additional cycles of a different salvage therapy regimen followed by autologous SCT. Among the 36 patients who had an adequate response to first salvage therapy, 11 patients (31%) did not undergo autologous SCT, and all of them remain alive and disease free, casting doubt on the supposition that all patients who fail frontline therapy require SCT, particularly if their response to initial salvage therapy is adequate. A randomized trial of SCT versus consolidation radiation therapy or additional chemotherapy will be required to determine whether SCT improves outcomes for specific patient groups.

Our current results also confirm that primary progressive disease indicates a poor prognosis for pediatric patients with HL. This finding is in agreement with previous studies in pediatric and adult patients with relapsed HL indicating the prognostic significance of the length of time between primary diagnosis and treatment failure.2, 6, 21, 25 In the current study, we also attempted to validate the prognostic score developed by the German Hodgkin Lymphoma Study Group for adults6 in our pediatric HL population (data not shown). However, in our study, stage at the time of treatment failure did not predict survival, and hemoglobin concentration was significant for second treatment failure but not for OS and, thus, was not used in our multivariate Cox regression analysis. A salvage therapy study by the German Pediatric Hodgkin Lymphoma Study Group demonstrated that the length of time from diagnosis to first treatment failure was the strongest prognostic factor.2 Multivariate analysis in that study indicated that primary chemotherapy without procarbazine and etoposide predicted a higher disease-free survival, and male sex was associated with higher OS. Neither study analyzed the association between response after 2 cycles of intense salvage chemotherapy and patient outcome.

The role of reducing disease bulk in HL before transplantation also remains controversial.14, 18, 26-29 The most commonly used method of assessment for residual disease has been CT, which does not provide information regarding the presence of persistent active disease within residual lymph node masses and is especially problematic in the interpretation of residual fibrosis in HL, which is slow to resolve and confounds response assessment. More recent studies have evaluated the role of PET for staging, restaging, and assessing tumor chemosensitivity.30-37 Future prospective studies are needed to fully assess the role of PET in predicting outcome and guiding salvage therapy in patients with pediatric HL who fail frontline therapy.

Our current study was limited because patients had failed after several different frontline therapy regimens and were treated by different physicians over a 17-year period. The choice of salvage regimen changed over time and depended on the intensity of the frontline therapy as well as timing of the relapse, and the use of SCT was left to the discretion of the primary physician without pre-existing guidelines. However, the choice of salvage regimen did not influence the rate of adequate response (70% with MIED and 70% with all other regimens), and the practice of all clinicians during this time was similar: Patients who ultimately achieved a response were referred for autologous SCT unless they had late, localized relapses with an adequate response to initial salvage therapy. In this study, we have documented the important new finding that an inadequate response to initial salvage therapy in pediatric patients with HL is correlated with a very unfavorable ultimate outcome.

In conclusion, our results suggest that pediatric patients with HL who do not respond to initial salvage should not receive a series of alternative cytoreduction regimens in the hope of finding a regimen to which the disease is sensitive; rather, such patients should be referred early for experimental therapies directed at biologic or immunologic targets. Thus, pediatric patients with relapsed HL who have an inadequate response after initial primary salvage chemotherapy have a very poor prognosis and should be considered for such novel therapies.

CONFLICT OF INTEREST DISCLOSURES

This work was supported by grants CA-21765, CA-51001, CA-36401, CA-78224, CA-60419, and GM-61393 from the National Institutes of Health and by the American Lebanese Syrian Associated Charities.

Ancillary