Response-Dependent and Reduced Treatment in Lower Risk Hodgkin Lymphoma in Children and Adolescents, Results of P9426: A Report from the Children’s Oncology Group
Associated Data
Abstract
Background
Hodgkin lymphoma is highly curable but associated with significant late effects. Reduction of total treatment would be anticipated to reduce late effects. This aim of this study was to demonstrate that a reduction in treatment was possible without compromising survival outcomes.
Methods
Protocol P9426, a response-dependent and reduced treatment for low risk Hodgkin lymphoma (stages I, IIA, and IIIA1) was designed in 1994 based on a previous pilot project. Patients were enrolled from 10/15/1996 to 09/19/2000. Patients were randomized to receive or not receive Dexrazoxane and received 2 cycles of chemotherapy consisting of Doxorubicin, Bleomycin, Vincristine, and Etoposide. After 2 cycles, patients were evaluated for response. Those in complete response (CR) received 2550 cGy of involved field radiation therapy (IFRT). Patient with partial response or stable disease, received 2 more cycles of chemotherapy and IFRT at 2550 cGy.
Results
There were 294 patients enrolled, with 255 eligible for analysis. The 8 year event free survival (EFS) between the Dexrazoxane randomized groups did not differ (EFS 86.8 + 3.1% with DRZ, and 85.7 + 3.3% without DRZ (p=0.70). Forty five percent of patients demonstrated CR after two cycles of chemotherapy. There was no difference in EFS by histology, rapidity of response, or number of cycles of chemotherapy. Six of the eight secondary malignancies in this study have been previously reported.
Conclusions
Despite reduced therapy and exclusion of most patients with Lymphocyte Predominant histology, EFS and overall survival are similar to other reported studies. The protocol documents that it is safe and effective to reduce therapy in low risk Hodgkin lymphoma based on early response to chemotherapy with rapid responding patients having the same outcome as slower-responding patients when given 50% of the chemotherapy.
BACKGROUND
Hodgkin Lymphoma (HL) was one of the first malignancies for which curative chemotherapy regimens were developed.1 The therapy for lower risk patients, Ann Arbor Stages I, IIA and IIIA1 is very effective with event free survivals and overall survivals greater than 90% 2, 3. Although intensive combined modality chemotherapy and radiation therapy regimens have achieved excellent results, concern over short and long-term side-effects are substantial and many groups are currently evaluating reduced forms of therapy 4–8. In 1992 we developed a novel reduced intensity combined modality regimen for the treatment of lower risk HL. Based on the results of this study9 and our previous reports suggesting that response to chemotherapy is predictive of long term survival 10,11 a new trial, P9426, was designed for lower risk patients (Stages I, IIA, and IIIA1) which used completeness and rate of response to guide the amount of treatment. The inclusion of only Stage IIIA1 in the lower risk group was consistent with our prior studies (9–12). In this trial, overall treatment to patients with favorable chemotherapy response was reduced, with the goal of reduction in treatment related toxicity. We report herein the long term results of P9426 which demonstrate that excellent OS can be maintained when therapy is reduced for patients with rapid early response (RER).
PATIENTS AND METHODS
The study was conducted at Pediatric Oncology Group Institutions from 10/15/1996 to 09/19/2000 and at the Children’s Cancer Group Institutions from 06/25/1999 until the trial closed 09/19/2000. Written informed consent was obtained from all patients according to institutional guidelines, and participating institutions obtained approval from local Institutional Review Boards.
Eligibility
Patients ≤21 years of age with newly diagnosed, histologically proven, clinically staged IA, IIA and IIIA1 HL (according to the Ann Arbor staging criteria as modified by Desser13) were eligible for study. Nodular sclerosis, mixed cellularity and lymphocyte depleted histologic types were eligible for the study throughout the duration of the protocol. Lymphocyte predominant HL was included only during the latter two years of the study, after 10/12/98. All patients were required to have physical assessment, a postero-anterior (PA) and lateral chest x-ray, computed tomography (CT) scan of neck, chest, abdomen and pelvis, and a gallium scan. Bone marrow biopsies were not routinely required and all other studies were performed at the discretion of the treating physician. Clinical symptoms of HL were defined as unexplained, persistent or intermittent fever > 38.3°C, weight loss of >10% of weight six months prior to the diagnosis of HL, and night sweats; patients with any of these symptoms were not eligible. All histology and imaging studies were centrally reviewed. No patients had surgical staging.
Chemotherapy
Patients received DBVE chemotherapy cycles every 28 days consisting of:
- (D) Doxorubicin 25 mg/m2/day intravenously (IV) on days 1 and 15 of each cycle
- (B) Bleomycin 10 IU/m2/day IV on days 1 and 15 of each cycle
- (V) Vincristine 1.5 mg/m2/day (max 2mg) IV on days 1 and 15 of each cycle
- (E) Etoposide 100 mg/m2/day IV days 1–5
- Filgrastrim 5 mcg/kg/day sq on days 6 to 14 and 16 until absolute neutrophil count increased to >1000/mcl.
Study design
Patients were randomized to receive a regimen of DBVE with or without Dexrazoxane (DRZ). DRZ was added to the design to examine whether this agent could reduce late cardiopulmonary toxicity. Results from this randomization will be reported separately. The objective of this report is to present the results of the response-dependent therapy algorithm. The schema of therapy is shown in Figure 1. All patients were evaluated for response after two cycles of DBVE. If they had achieved a complete response, they proceeded directly to involved field radiation therapy (IFRT) with no additional chemotherapy. Patients who had not achieved a complete response after the initial two cycles of DBVE received two additional cycles of DBVE therapy prior to IFRT. The final response to treatment was evaluated in all patients after completion of radiation therapy. Response evaluation included physical assessment, PA and lateral chest x-ray, CT scan of neck, chest, abdomen and pelvis, gallium scan and any other studies positive for disease at diagnosis. Approximately 45% of patients demonstrated RER and thus received 50% of the chemotherapy otherwise intended9.
Radiation therapy
All patients were scheduled to receive IFRT after completion of chemotherapy. A dose of 2550 cGy was administered in 17 fractions of 150 cGy each. Real time central review of the treatment plan was conducted by the Quality Assurance Review Center (QARC) to ensure compliance with protocol specified treatment parameters.
Response criteria
Complete response (CR) was defined as disappearance of all clinically and imaging-evident active disease which included a negative gallium scan. Patients with a mediastinal mass were considered in CR if the mass was <50% of the original CT measurements, the gallium scan was negative and they fit all other CR criteria. This definition was developed from prior experience in which patients with negative gallium scans did not have tumor found on biopsy of residual mediastinal mass which fit the <50% criterion (14). Rapid early response (RER) was defined as a complete response (CR) after two cycles of DBVE. Slow early response (SER) was defined as less than CR after 2 cycles of DBVE. Partial response (PR) was defined as ≤50% decrease in the sum of the product of the perpendicular diameters of all lesions and resolution of any gallium positivity. The final size of a mediastinal mass had to be ≤50% of its initial size on a PA upright chest x-ray and stable for at least two CT scans taken two months apart after completion of therapy. Stable disease (SD) was defined as <50% decrease in the sum of the products of the perpendicular diameters of all measurable lesions and the development of no new lesions or increases in measurement of the disease as compared to the baseline diagnosis. No response (NR) was defined as no decrease in the sum of the product of the perpendicular diameters of all measurable lesions. Progressive disease (PD) was defined as unequivocal increase of at least 25% in the size of any measurable lesion or appearance of new lesions. Histologic documentation was required for NR and PD.
Statistical Considerations
The primary endpoint was event-free survival (EFS). Time to event was calculated as the time from study entry until the first occurrence of progression, relapse, second malignant neoplasm, or death from any cause, or until last contact if no event occurred. The secondary endpoint was overall survival (OS), where OS time was calculated as time from study entry to death from any cause, or until last contact if the patient did not die. EFS and OS estimates were computed using the Kaplan-Meier method15 and standard errors were reported according to Peto 16. The logrank test was used to test for differences between survival curves. P-values ≤0.05 were considered statistically significant. All analyses were conducted as intent-to-treat within the cohort of eligible patients.
RESULTS
Study Population
A total of 294 patients were enrolled from 10/15/96 to 09/19/2000. Thirty-nine patients were found to be ineligible and excluded, leaving 255 eligible patients for analysis. The reasons for exclusion were mediastinal/thoracic ratio greater than 1:3 (n=22), stage IV (n=6) or IIIA2 (n=3) disease, lymphocyte predominant histology in the early phase of the study before eligibility of this histologic subtype (n=2) and other miscellaneous reasons (n=6). One hundred and twenty seven of the 255 eligible patients were randomized to receive DBVE with dexrazoxane and 128 without dexrazoxane. There were 150 (59%) White, 52 (20%) Hispanic, 38 (15%) African-American and 15 (6%) other patients (Table I). The age ranged from 2 to 20 years, with a median of 13 years. The histology is known for 239 patients and included 156 (65.3%) patients with nodular sclerosis, 54 (23%) mixed cellularity, 26 (11%) lymphocyte predominant and 3 (1%) lymphocyte depleted. Distribution of stage was as follows: stage IA 61 (24%), stage IIA 179 (70%) and stage IIIA1 14 (6%) (Table I).
Table I
Patient characteristics by histological subtype of Hodgkin disease
| Characteristics | Nodular sclerosis | Mixed cellularity | Lymphocyte predominant | Lymphocyte depleted | Overall |
|---|---|---|---|---|---|
| Overall | 156 | 54 | 26 | 3 | 255 |
| Age (years) | |||||
| <=10 | 29 | 22 | 9 | 2 | 67 |
| 11–15 | 72 | 23 | 13 | 1 | 117 |
| 16–20 | 55 | 9 | 4 | 0 | 71 |
| Sex | |||||
| Male | 76 | 42 | 21 | 2 | 150 |
| Female | 80 | 12 | 5 | 1 | 105 |
| Race | |||||
| White | 105 | 18 | 15 | 3 | 150 |
| Hispanic | 25 | 21 | 3 | 0 | 52 |
| Black | 18 | 11 | 5 | 0 | 38 |
| Other | 8 | 4 | 3 | 0 | 15 |
| Stage | |||||
| Stage I | 24 | 19 | 13 | 0 | 61 |
| Stage II | 119 | 34 | 12 | 3 | 179 |
| Stage III1 | 12 | 1 | 1 | 0 | 14 |
| Unknown | 1 | 0 | 0 | 0 | 1 |
| WBC (x103/µL) | |||||
| <=5 | 14 | 16 | 5 | 2 | 39 |
| 6–7 | 38 | 18 | 8 | 0 | 68 |
| 8–9 | 45 | 13 | 8 | 1 | 72 |
| ≥10 | 59 | 7 | 5 | 0 | 76 |
| Response after 2 cycles | |||||
| CR | 61 | 26 | 17 | 2 | 112 |
| PR | 90 | 28 | 8 | 1 | 134 |
| NR/PD | 1 | 0 | 0 | 0 | 1 |
| Unknown | 4 | 0 | 1 | 0 | 8 |
Response
Two hundred and forty-seven of the 255 eligible patients were evaluated for early response as shown in Table II. One-hundred twelve (45%) patients were considered RER. Conversely, 135 (55%) were considered SER because they did not achieve a CR after 2 cycles. Initial response was not reported in 8 patients who are included in all analyses other than RER/SER calculations, based upon actual treatment received. Patients with RER proceeded to radiation therapy after the 2 cycles of DBVE, except for 18 who were treated as if they had had SER with two more cycles of chemotherapy and IFRT. Patients with SER received a total of 4 cycles of DBVE before proceeding to radiation therapy, except for nine patients who were treated like RER’s with two cycles of DBVE and IFRT. All patients who were incorrectly allocated to treatment arm by the institution are included in analysis as they should have been treated based on response status. Exclusion of these patients does not alter the overall findings or conclusions (data not shown). Patient registration and response evaluations are shown in the CONSORT diagram (Supplemental Figure). Upon completion of chemotherapy 169 (68%) patients achieved a complete response and 78 (32%) a partial response. After completing radiation therapy, 226 (91%) patients achieved a complete response and 22 (9%) a partial response. The CR rates after radiation were similar for Stage I, II and III1 patients (91%, 91% and 93%, respectively) (Table 2). The early response rate was lower for patients with nodular sclerosis (40%) as compared to patients with all other histologies (55%) (p=0.04).
Table II
Response rates by stage of disease
| Stage of Disease | N | Early responders* N (%) | CR after chemotherapy N (%) | CR after radiation N (%) |
|---|---|---|---|---|
| Stage I | 61 | 40/58 (69%) | 47 (81%) | 53 (91%) |
| Stage II | 179 | 69/174 (40%) | 112 (64%) | 159 (91%) |
| Stage III1 | 14 | 3/14 (21%) | 10 (71%) | 13 (93%) |
| Unknown | 1 | 0 | 0 | 1 (100%) |
| Total | 255 | 112 (45%) | 169 (68%) | 226 (91%) |
Note: Early response percentages were calculated based on the 248 patients with known response data;
Toxicity
Treatment-related side effects were relatively mild and most patients tolerated treatment as outpatients, with rare requirement for admission. Main toxicities reported in both arms of the study, i.e. with and without dexrazoxane, are shown in Table III. There was increased bone marrow suppression in the dexrazoxane arm manifested by increased suppression in white blood-cell count (WBC) (p<0.01), ANC (p=0.03); hemoglobin and platelet counts did not reach statistical significance. This is similar to the toxicities seen in P942512 and did not interfere with the scheduled delivery of chemotherapy.
Table III
Toxicity rates by treatment
| TOXICITY | WITHOUT DRZ | WITH DRZ | ||
|---|---|---|---|---|
| Count (n=113) | Incidence (%) | Count (n=109) | Incidence (%) | |
| WBC* | 30 | 26.5 | 54 | 49.5 |
| ANC* | 61 | 54.0 | 75 | 68.8 |
| Lymphs | 1 | 0.9 | 1 | 0.9 |
| Platelets | 10 | 8.8 | 18 | 16.5 |
| Hemoglobin | 7 | 6.2 | 20 | 18.3 |
| Infection | 3 | 2.7 | 1 | 0.9 |
| Sepsis Bacteria | 1 | 0.9 | 1 | 0.9 |
| Allergic Reaction | 4 | 3.5 | 1 | 0.9 |
| Nausea | 2 | 1.8 | 2 | 1.8 |
| Vomiting | 5 | 4.4 | 3 | 2.8 |
Survival
The median follow-up time was 7.5 years in patients alive without an event, and 98 patients had more than 8 years of follow-up. The 5-year EFS and OS for the overall cohort were 88.3 ± 2.1% and 97.6 ± 1.0% and the 8-year EFS and OS were 86.3 ± 1.4% and 96.5 ± 3.5%, respectively (Figure 2). There were 34 events, with first events due to either relapses/progression (n=29) or second malignant neoplasm (n=5). There were a total of 8 deaths in this study. Four of these deaths were related to disease and 4 to second malignancies.
EFS rates by treatment did not differ: 86.8 + 3.1% at 8-years with DRZ, and 85.7 + 3.3% at 8-years without DRZ (p=0.70) (Table IV). Survival rates were not different by stage of disease, rapidity of response, number of cycles received, LP histology, or by rapidity of response within each stage (Table IV). The small group of Stage III patients had lower EFS, but OS was equal to the entire group. There was a slight trend for better EFS and OS among patients with LP histology.
Table IV
Survival rates by patient subsets
| Factors | # of patients | 8-year Event-free survival ± S.E* | 8-year Overall survival ±S.E.* |
|---|---|---|---|
| Overall | 255 | 86.3 ± 2.1% | 96.5 ± 3.5% |
| By treatment | |||
| No DRZ | 128 | 85.7 ± 3.3% | 96.4 ± 1.8% |
| Yes DRZ | 127 | 86.8 ± 3.1% | 96.6 ± 1.7% |
| p = 0.70 | p = 0.99 | ||
| By stage | |||
| Stage I | 61 | 91.7 ± 3.6% | 96.4 ± 2.4% |
| Stage II | 179 | 85.6 ±2.7% | 97.1 ± 1.3% |
| Stage III1 | 14 | 69.8 ± 12.8% | 90.9 ± 8.7% |
| p = 0.12 | p = 0.71 | ||
| By rapidity of response | |||
| Early response | 112 | 87.3 ± 3.3% | 97.1 ± 1.7% |
| Slow response | 135 | 85.4 ± 3.1% | 95.9 ± 1.8% |
| p = 0.40 | p = 0.65 | ||
| By total # cycles received | |||
| 2 cycles DBVE | 103 | 86.7 ± 3.5% | 96.8 ± 1.8% |
| 4 cycles DBVE | 144 | 85.8 ± 3.1% | 96.2 ± 1.7% |
| p = 0.78 | p = 0.82 | ||
| By LP Histology | |||
| Non-LPs | 213 | 85.1 ± 2.5% | 95.9 ± 1.5% |
| LPs | 26 | 94.7 ± 5.1% | 100% |
| p = 0.13 | p=0.31 | ||
| By stage & rapidity of response | |||
| Stage I | |||
| Early response | 40 | 92.4 ± 4.2% | 94.6 ± 3.7% |
| Slow response | 18 | 94.4 ± 5.4% | 100.0% |
| p = 0.82 | p = 0.36 | ||
| Stage II | |||
| Early response | 69 | 83.6 ± 4.8% | 98.5 ± 1.5% |
| Slow response | 105 | 86.2 ± 3.5% | 96.1 ± 1.9% |
| p = 0.98 | p = 0.37 | ||
| Stage III1 | |||
| Early response | 3 | 100% | 100% |
| Slow response | 11 | 63.6 ± 14.5% | 90.0 ± 9.5% |
| p = 0.31 | p = 0.75 |
LP histology has been considered favorable in early stage patients. As noted in Table IV, due to eligibility restrictions we enrolled only 26 LP patients on this study. In spite of an OS of 100%, it was not possible to demonstrate statistically superior results with this number and an OS of 95.9 % for the other patients. COG has subsequently conducted a separate trial for early stage LP patients, the results of which are not yet available.
There have been eight secondary malignancies (SMs) reported in this study: five occurred as first events (three acute myeloid leukemia (AML), one thyroid carcinoma, one osteosarcoma) and three occurred after relapse (one AML, one myelodysplastic syndrome, one Ewing sarcoma). Only the thyroid carcinoma was within a primary radiation field; all four AML cases occurred less than 2.5 years from original diagnosis, The carcinoma of the thyroid was 3 years from diagnosis and the myelodysplastic syndrome 4 years post diagnosis. The Ewing sarcoma was 6 years and the osteosarcoma 9 years after diagnosis. Four of the five SMs which occurred as primary events were in patients randomized to DRZ. The rates of second malignancy in P9426 and P9425, the simultaneous study for higher risk HL12, have been previously reported17. While we found higher SMs in patients receiving DRZ, we are aware of numerous reports that found no such association(18, 19, 20) and have suggested that our study differs from others in administering etoposide (a third topoisomerase inhibitor) simultaneously (on the same days) with doxorubicin and DRZ(17,21). We therefore speculate that the difference was not simply a result of administration of DRZ.
DISCUSSION
Based on our previous studies documenting effectiveness and improved EFS and OS in patients with low risk (Ann Arbor stage I, IIA and IIIA1) HL with DBVE chemotherapy and IFRT, the 9426 trial was designed to determine whether the amount of chemotherapy could be reduced based on the extent and rapidity of chemotherapy response without loss of effectiveness. We report here the successful results of P9426, which demonstrate the ability to reduce DBVE chemotherapy to only two cycles, followed by IFRT to 2550 cGy, in patients who obtain a complete response by the end of the second cycle of chemotherapy. While the evaluation of RER has evolved to having PET scans replace Gallium scans, this study supports prior publications from our group which have noted superior outcomes for RER patients when treated identically to SER patients10,11 and at least equivalent outcomes when treated with reductions in therapy12. It is not possible, even in this time frame, to show reduced late effects, although it can be assumed that less therapy will not increase late effects. A complete analysis of late effects will involve the COG late effects studies and will be reported separately and combined with a detailed analysis of HD treatment protocols of the COG over several decades.
HL is a very treatable disease even in advanced stages.2 However, long term effects of therapy have prompted several groups to attempt reduction of therapy, particularly in children with low risk HL. These attempts have varied in their approach, including elimination of radiation therapy16, the use of non-alkylator chemotherapy,4, 5 or response based reduction of chemotherapy and/or radiation therapy.23 In a previous report of P8625 by Kung et al., the 5-year EFS (92.7%) of low risk HL patients who achieved early response to chemotherapy was better than that of patients who did not (76.7%, p=0.006).10 Nachman et al. reported the results of the Children’s Cancer Group protocol CCG5942 which randomized patients who achieved a complete response to chemotherapy to receive or not to receive IFRT. The patients receiving IFRT had better 3-year EFS than those that did not receive IFRT (92 vs. 87%) but the OS did not differ between the groups.23 Donaldson et al. achieved excellent results with 5-year and 10-year EFS of 92.7% and 89.4% and OS of 99.1% and 96.1% using a non-alkylator chemotherapy regimen of four cycles of chemotherapy with vinblastine, doxorubicin, methotrexate and prednisone (VAMP) plus IFRT. In these studies, patients who achieved a CR with two cycles of chemotherapy had 5 and 10 year EFS (100% and 95%) as did patients with lymphocyte predominant or nodular lymphocyte predominant histology (100%)4,5. When patients with lymphocyte predominant histology HL were excluded, patients with classic HL treated with VAMP and IFRT had 5-year and 10 year EFS of 86% and 85.4% respectively. The German Collaborative Group protocols 85, 90, 95 and 2003 have tried to reduce toxicity by adaptation of chemotherapeutic drugs (e.g. avoiding procarbazine in boys) and by response based reduction of chemotherapy and /or radiation therapy25. Low risk patients (Ann Arbor stages I, IIA) receiving two cycles of chemotherapy plus IFRT in DAL-HD-90 had a 5-year EFS of 94% and OS of 99%.7 In the GPOH-HD-2002 trial18, low risk patients who responded to two cycles of chemotherapy (31.8% of Treatment Group I patients) did not receive any further therapy, while those with a partial response received IFRT; both groups had similar 5-year EFS (93.2% and 91.7%). COG has evaluated the concept of omission of radiotherapy in RER patients of intermediate risk; preliminary data indicate that this can safely be done and we certainly plan to further evaluate this concept in future studies25.
All previously discussed protocols other than the GPOH-HD report included patients with lymphocyte predominant HL. Our trial was not open to this very favorable subset of HL patients until the end of the trial, potentially negatively biasing the results compared with other trials by reducing the proportion of the lymphocyte predominant subset. The overall 5-year and 8-year EFS for P9426 was 88.3 ± 2.1% and 86.3 ± 2.1% respectively, similar to that observed when lymphocyte predominant patients were excluded from the Donaldson study discussed above. The 8-year OS for this study was 96.5 ± 3.5% which is comparable to the results obtained by other groups. The randomization to receive or not the cardioprotectant DRZ did not appear to affect the outcome in this trial. Five secondary malignant neoplasms occurred after primary chemotherapy in this trial and three further secondary malignancies occurred after retrieval therapy. As discussed above, secondary malignant neoplasms occurred most frequently in the dexrazoxane arms of this study and the similar randomized study for intermediate and high risk HL 15 which we attribute to concomitant administration of three topoisomerase II inhibitors (dexrazoxane, etoposide, and doxorubicin) rather than any single agent. Overall this regimen was well tolerated and provided the ability to reduce therapy in 45% of the patients while retaining an excellent outcome with 8-year EFS and OS of 86.3 ± 1.4% and 96.5 ± 3.5%, respectively. While it is not possible to choose a single regimen as “best” among all of the treatments discussed above, this study further documents the effectiveness and importance of response based therapy in HL.
Acknowledgements
The authors thank all physicians, nurses and clinical research associates at treating institutions who contributed to patient care and data submission. The authors also thank Roxanne Escobedo of the COG publications office for her contributions in organizing the manuscript submission. The senior author also thanks Martha Merrill for her outstanding editorial assistance and Dr. Matt Thomas, pediatric resident at Charleston Area Medical Center, for superlative technical expertise.
Research is supported by QARC GRANT NIH/NCI CA29511, the Chair’s Grant U10 CA98543-08 and Statistics and Data Center Grant U10 CA98413-08 of the Children’s Oncology Group from the National Cancer Institute, National Institutes of Health, Bethesda, MD, USA. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NCI or the NIH.
A complete listing of grant support for research conducted by CCG and POG before initiation of the COG grant in 2003 is available online at: http://www.childrensoncologygroup.org/admin/grantinfo.htm
Footnotes
Conflict of Interest: The authors have nothing to declare