Association Between Radiotherapy vs No Radiotherapy Based on Early Response to VAMP Chemotherapy and Survival Among Children With Favorable-Risk Hodgkin Lymphoma FREE

Currently more than 90% of children with favorable-risk Hodgkin lymphoma will achieve long-term survival. However, studies demonstrate excess mortality among patients followed up beyond 10 years from their Hodgkin lymphoma diagnosis as a result of late toxic effects of therapy,^1- 2 including the development of second malignant neoplasms and nonneoplastic treatment complications.³ Risk-adapted combined-modality therapy (combined chemotherapy and radiotherapy according to predetermined risk stratification) has therefore been tailored to minimize therapy while maintaining excellent outcome.^4- 8 Response-adapted therapies (tailored according to early initial response) aim to identify patients for whom it would be safe to reduce radiation therapy dose, volume, or both.^4,6,9- 10

We previously reported the results of treatment with 4 cycles of vinblastine, Adriamycin (doxorubicin), methotrexate, and prednisone (VAMP) chemotherapy and low-dose, involved field radiotherapy,^6,11 and we identified a group of patients with a favorable profile including early complete response to VAMP chemotherapy who might be curable without irradiation. We now report the results of a multi-institutional, nonrandomized study testing the efficacy of VAMP with or without radiotherapy based on early response to chemotherapy for children with favorable−risk Hodgkin lymphoma. Evaluation of quality of life during and after treatment was a secondary objective and will be reported in a forthcoming manuscript.

This trial was open to accrual from March 2, 2000, through February 10, 2009. There were no competing trials open at participating institutions during the study period and no eligible patient chose not to participate. The study was designed to assess whether a 2-year event-free survival of more than 90% could be maintained in children with Hodgkin lymphoma and favorable-risk features treated with 4 cycles of VAMP with or without radiotherapy according to their early response after 2 cycles. Eligibility criteria included age younger than 21 years, previously untreated Hodgkin lymphoma, Ann Arbor stage¹² IA or IIA, nonbulky mediastinal mass (ratio of the size of the mass compared with the widest intrathoracic diameter measured on an upright posteroanterior chest radiograph, <33%), no extranodal extension of disease, and fewer than 3 involved nodal regions.¹³ Race and ethnicity are routinely captured on every patient and defined according to open-ended self-reported categories. The protocol was approved by the institutional review board at each participating institution and monitored by the St Jude Children's Research Hospital Data and Safety Monitoring Board. Written informed consent was obtained from parents, legal guardians, or patients, as appropriate.

Initial evaluation included history and physical examination; complete blood count, C-reactive protein, or erythrocyte sedimentation rate; renal and hepatic function; lactate dehydrogenase; alkaline phosphatase; albumin; chest radiograph; contrast-enhanced computed tomography (CT) of neck, chest, abdomen, and pelvis; and functional imaging (gallium, later replaced by [¹⁸F]fluoro-2-deoxyglucose positron emission tomography [PET]). Bone marrow biopsies were not performed.

Chemotherapy consisted of four 28-day cycles of VAMP: vinblastine, 6 mg/m²; doxorubicin, 25 mg/m²; and methotrexate 20 mg/m²; administered intravenously on days 1 and 15 and oral prednisone, 40 mg/m² on days 1 through 14. During therapy patients were evaluated weekly.

All patients underwent early response evaluation by CT scan and functional imaging of originally involved sites of disease just before the third cycle of VAMP. Complete response was defined as a negative gallium scan or PET scan and either 75% or more reduction of the sum of the products of the perpendicular diameters of the lesions of all measurable or evaluable disease or return of nodes to their normal size. A partial response required at least a 50% reduction in the size of the measurable lesions of the original tumor volume, or gallium or PET avidity persistence in nodal masses despite 75% or more reduction in volume. Less than 50% reduction in the size of the measurable lesions was defined as stable disease, and an increase of more than 25% of the original tumor volume or appearance of new areas of disease represented progressive disease.¹⁴

There was no blinding in this study. Radiotherapy assignment was based on early response to chemotherapy. Patients with a complete response at early response evaluation completed therapy following the fourth chemotherapy cycle. Patients who achieved less than a complete response after 2 cycles received 25.5-Gy−involved field radiotherapy (IFRT) in 17 fractions of 1.5 Gy beginning 2 to 4 weeks after completion of all chemotherapy and included treatment of initially involved nodes and surrounding nodal region.

Follow-up evaluations included history and physical examination, chest radiograph, and laboratory examinations every 3 months during the first year, every 4 months during the second and third year, every 6 months up to 5 years off therapy and yearly thereafter. PET and CT scans were performed at 1 and 2 years off therapy and as clinically indicated. Ongoing assessments for late treatment complications include annual evaluations of growth and pubertal development; thyroid, pulmonary, and cardiac function; and pregnancy outcomes. Relapses were confirmed by biopsy. Following relapse, patients are observed for survival. Retrieval therapy (cytoreductive chemotherapy, additional radiotherapy with or without consolidative autologous stem cell transplant) was not specified and was administered according to investigator preference.

The phase 2 study aimed to determine whether the 2-year event-free survival is more than 90% or whether there is evidence that the true 2-year event-free survival is less than 80%. Accrual of 87 patients was required to test the hypothesis with a type I error rate of 5% and 80% power. We also report 5-year outcome estimates as post hoc analyses.

The 1-sample binomial test was used to test whether the proportion of patients who were event-free at 2 years was significantly different from the unacceptably low rate of 80% and also the desired rate of 90%. Demographic and disease characteristics were compared according to early response to therapy using a 2-tailed Fisher exact test (categorical variables) or Wilcoxon rank sum test (continuous variables).¹⁵ Event-free survival was defined as the time from study enrollment to the date of treatment failure (relapse or progressive disease) or date of last follow-up for nonrelapsing patients. Event-free and overall survival distributions were estimated using the method of Kaplan-Meier,¹⁶ and reported with 95% confidence intervals. Demographic and disease characteristics were examined as predictors of event-free survival using Cox proportional hazards regression.¹⁷ Each factor was examined univariately because there were too few events to include multiple predictors in a model. The adequacy of each model was assessed using graphical and numerical methods derived from cumulative sums of martingale residuals over follow-up times and covariate values. Early response to therapy (complete response vs less than complete response), was treated as a time-dependent covariate. These analyses were post hoc and have limited power due to the small number of events and should be considered exploratory. SAS version 9.2 (SAS Institute Inc) was used for statistical analysis. All reported P values are 2-sided and considered to be statistically significant if less than .05.

Forty-four patients (48%) were enrolled at St Jude Children's Research Hospital, 14 (15%) at Stanford University Medical Center, 19 (21%) at Dana-Farber Cancer Institute, 13 (14%) at Massachusetts General Hospital, and 1 (1%) at Maine Medical Center. Of 91 patients enrolled, 3 were ineligible and not considered in this report. Two of these patients did not meet criteria for favorable-risk disease. Pathology review of the third patient failed to confirm the diagnosis of Hodgkin lymphoma. Thus, 88 patients are the focus of this report. Fifty-nine patients (67%) were male; 77 (88%), white; and 7 (8%), African American. The median age at diagnosis was 13.9 years (range, 4.4-20.6 years). Histological subtype revealed classical Hodgkin lymphoma in 56 patients (64%) and nodular lymphocyte predominant Hodgkin lymphoma in 32 (36%). Thirty-nine patients (44%) had stage IA, 35 (40%) had a mediastinal mass, and 13 (15%) had peripheral bulk disease (≥6 cm).

Forty-seven patients (53%) achieved a complete response after 2 cycles of VAMP and finished therapy without IFRT. Among the 41 (47%) who did not achieve a complete response (39 partial responses, 2 stable diseases), 39 received radiotherapy according to protocol, 1 withdrew consent for participation and received radiotherapy elsewhere, and 1 had early disease progression prior to radiotherapy and received retrieval therapy. Demographic and disease characteristics according to early response to therapy are listed in Table 1. There was no difference in distribution according to sex, race, or age. Presence of peripheral bulk disease did not differ between groups. Patients with nodular sclerosing Hodgkin lymphoma (63%, P < .001), patients with stage IIA disease (71%, P = .01) and patients with a mediastinal mass (59%, P = .001) were less likely to achieve a complete response after 2 cycles of chemotherapy.

All but 1 patient were alive at the time of the analysis. The median follow-up for survivors was 6.9 years (range, 2.5-11.4 years) excluding the patient who withdrew consent prior to radiotherapy. At the time of analysis (data current to March 12, 2012), 80% of patients had been seen or contacted within the past year, and 91% within the past 18 months. Figure 1 shows overall and event-free survival distributions for the study cohort. The estimated 2-year event-free survival is 90.8% (95% CI, 84.7%-96.9%), which is significantly different from 80% (P = .01) but is not significantly different from 90% (P = .98). Figure 2 shows event-free survival distributions according to early response to therapy. The 2-year event-free survival for patients who achieved early complete response and did not receive IFRT is 89.4% (95% CI, 80.8%- 98.0%) compared with 92.5% (95% CI, 84.5%-100%) for those who did not achieve complete response.

Excluding the patient who withdrew consent early, 56 of 87 patients (64%) had at least 5 years of follow-up. The 5-year event-free survival is 88.5% (95% CI, 80.7%-96.3%), and the 5-year overall survival 100%; 1 patient died approximately 7.5 years after study enrollment. The clinical characteristics of the 11 patients who experienced treatment failure (median, 17 months; range, 4-79 months) are summarized in Table 2. Patients who did not undergo irradiation had an estimated 5-year event-free survival of 89.4% (95% CI, 79.0%-99.8%), which is similar to patients who did (5-year event-free survival, 87.5%; 95% CI, 75.7%-99.3%). There was no evidence that early complete response was a significant predictor of event-free survival (P = .61). All patients who were not irradiated who experienced recurrence did so at a previously involved site and were successfully retrieved with chemotherapy and IFRT without stem cell transplant (Table 2). Sites of failure in 5 of the 6 patients with less than complete response at early response evaluation also included previously irradiated sites. Four of these patients received high-dose chemotherapy and autologous stem cell transplant; 1 patient with a localized late relapse (30 months after original diagnosis) was retrieved with chemotherapy and IFRT. The remaining patient developed recurrent nodular lymphocyte predominant Hodgkin lymphoma concurrent with transformation to diffuse large B-cell lymphoma and succumbed to refractory disease.

Prognostic factors for treatment failure are presented in Table 3. Neither patient characteristics (sex, race, and age) nor tumor characteristics (histology, stage, presence of a mediastinal mass or peripheral lymph node bulk) predicted failure, although the power to detect such prognostic factors is limited by the relatively small number of patients and events.

The estimated 5-year event-free survival for patients with nodular lymphocyte predominant Hodgkin lymphoma is 87.5% (95% CI, 74.0%-100%), compared with 89.1% (95% CI, 79.7%- 98.5%; P = .50) for classical Hodgkin lymphoma (Figure 3).

Therapy was well tolerated without major complications. Delay or dose modifications due to adverse toxic effects were rare. The most common adverse effects were neuropathic pain (2% of patients) and nausea and vomiting (3% of patients), all of which are readily managed with supportive care. Neutropenia (absolute neutrophil count <1000/dL) was observed in 60% of patients (32% of cycles), and febrile neutropenia in 2% of patients (0.9% of cycles). Nine patients (10%) were hospitalized 11 times (3% of cycles) for febrile neutropenia (3 hospitalizations) or nonneutropenic infection. Granulocyte colony–stimulating factors were not used.

No second malignant neoplasm occurred except for a histology transformation in a patient with nodular lymphocyte predominant Hodgkin lymphoma. Nine of 88 patients (10%) developed asymptomatic compensated hypothyroidism 16 to 30 months from study enrollment. All of them had received IFRT to the neck. Two patients developed osteonecrosis, and 2 others developed moderate osteopenia. Twelve patients (14%) developed subclinical pulmonary dysfunction following thoracic radiation, most commonly asymptomatic mild to moderate pulmonary diffusion deficits (n = 8; DLCO, 57%-74% of what is predicted for their sex and age). Four patients (5%; 2 irradiated and 2 nonirradiated) experienced asymptomatic left ventricular dysfunction (shortening fractions, <30%; range, 24%-29%) at a median of 35 months after study enrollment (range, 8- 86 months). Two of them (1 irradiated and 1 nonirradiated) recovered shortening fraction to 30% or more, and the other 2 patients to 29% and 28%, respectively. All remained asymptomatic. Among patients older than 18 years (23 women and 38 men), 3 couples involving 2 women and 1 man in the cohort conceived 6 times, resulting in 4 live births.

To our knowledge, this is the first trial in which a select group of children with favorable–risk Hodgkin lymphoma experienced a high rate of 2- and 5-year event-free survival without exposure to radiotherapy, alkylating agent, epipodophyllotoxin, or bleomycin chemotherapy and a relatively low cumulative dose of anthracyclines. The desire to avoid late treatment complications—particularly those resulting from high doses of irradiation—has motivated most treatment modifications for pediatric Hodgkin lymphoma. Early trials established the effectiveness of combined-modality therapy featuring multiagent chemotherapy and lower cumulative doses of radiation to involved sites of disease.^18- 20 To avoid radiation complications altogether, chemotherapy-only trials were developed prescribing multiple courses of nitrogen mustard, vincristine, procarbazine, and prednisone (MOPP) or derivatives.^21- 23 These regimens proved to be effective in achieving long-term remissions.^24- 26 However, most trials featured high cumulative doses of alkylating agents, anthracyclines, or bleomycin leading to increased morbidity and toxic injury to their heart, lungs, and sexual organs, and secondary leukemia. As a result, contemporary combined-modality trials focus on balancing efficacy and toxic effects of therapy. Investigators have also sought to identify patients with favorable features who would be candidates for therapy reductions. These efforts led to the evaluation of a response-based radiation approach, which was first undertaken in Stanford pediatric protocols prescribing lower doses of radiation to patients with good response to MOPP.²⁷ This experience has shaped our consortium trials since.^6,28- 29

The Children's Cancer Group study CCG5942³⁰ and the German Pediatric Oncology and Hematology (GPOH-HD95) study⁴ (confirmed in the GPOH-HD2002 study⁵) pursued response-based trials in the mid-1990s aimed to omit radiation. The GPOH trials had comparable outcomes for nonirradiated favorable-risk patients who achieved complete response after 2 cycles of vincristine, procarbazine, prednisone, and doxorubicin (OPPA for girls) or vincristine, etoposide, prednisone, and doxorubicin (OEPA for boys) compared with those who achieved less than complete response and received radiotherapy. A marked event-free survival advantage with combined-modality therapy was only appreciated in intermediate- and high-risk groups, while overall survival remained comparable.⁴ In the Children's Cancer Group study, patients who achieved a complete response after all chemotherapy were randomized to receive 21 Gy IFRT or no additional treatment (eTable). There was a 3-year event-free survival advantage for the group that underwent irradiation (100% vs 89%); however, survival in the 2 groups was identical (3-year overall survival 100%), raising the concern of the number of children needed to be irradiated to prevent 1 relapse. This is particularly pertinent in our study in which patients' developing relapse after treated only with chemotherapy were successfully retrieved with standard multiagent chemotherapy and IFRT without high-dose chemotherapy or stem cell transplant. As a result, more than 50% of patients in our study were spared radiotherapy, and the majority could be cured without exposure to leukemogenic agents.

We previously reported the results using VAMP chemotherapy and low-dose IFRT.^6,11 The 5-year event-free and overall survival for the entire cohort were 93% and 99%, respectively. Results of the current study are similar with a 5-year event-free and overall survival of 89% and 100%, respectively. In the present study, the 5-year event-free survival for patients with classical Hodgkin lymphoma was 89%, with no difference in outcome between those treated with and without radiotherapy (P = .61).

Historically, patients with nodular lymphocyte predominant Hodgkin lymphoma have a favorable outcome and have been treated on regimens suitable for classical Hodgkin lymphoma; however, there are several reports in the adult and pediatric literature suggesting that such patients can be cured with less therapy. In adults, radiotherapy-only approaches are favored^31- 32; however, the required doses of 30 to 36 Gy would result in significant musculoskeletal toxic effects in children. Most children with nodular lymphocyte predominant Hodgkin lymphoma do well regardless of therapy chosen; adverse toxic effects remain the main concern for more involved treatment approaches.^33- 35 Remarkably, a substantial proportion of children with limited-stage, completely-resected nodular lymphocyte predominant Hodgkin lymphoma achieve long-term remission with surgical resection alone without additional therapy.³⁷ In view of these results, the outcome for patients with nodular lymphocyte predominant Hodgkin lymphoma in our study who were treated without irradiation was disappointing. Twenty-six of 32 patients were early responders and thus treated without irradiation. Four patients (15%) relapsed compared with no treatment failures in our prior experience with a combined-modality approach wherein all such patients were irradiated.⁶ In the current study, none of the 10 patients with nodular lymphocyte predominant Hodgkin lymphoma who had undergone complete resection relapsed; thus, it is possible that many of them could have been spared chemotherapy altogether. In contrast, patients with stage II (and unresected) nodular lymphocyte predominant Hodgkin lymphoma who did not receive radiotherapy are at increased risk of relapse. Whether the omission of alkylating agents from the VAMP regimen can account for the less favorable result is speculative. However, it appears that even low-dose irradiation may be beneficial for children with nodular lymphocyte predominant Hodgkin lymphoma who are treated with chemotherapy regimens that omit alkylating agents.

A limitation of this study is its relatively small sample size limiting the power to assess differences between study sites and limiting subgroup analyses. Thus, it would be important to confirm the results in a larger cohort. Our results suggest that a risk-adapted response-based approach may be very effective and well tolerated for a selected group of patients with favorable risk Hodgkin lymphoma. Such patients can achieve high 2-year event-free survival without alkylating agent, bleomycin, or epipodophyllotoxin chemotherapy, and more than half without radiotherapy. Future studies should consider further tailoring of radiotherapy reserving irradiation for patients who remain PET positive at early response evaluation.