|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 1999 American Society for Clinical Oncology
MOPP Versus ABVD and Low-Dose Versus High-Dose Irradiation in Hodgkin's Disease at Intermediate and Advanced Stages: Analysis of a Meta-Analysis by CliniciansHôpital Laënnec, Paris, France To the Editor: Over the last three decades, the prognosis for advanced-stage Hodgkin's disease (HD) has been dramatically improved by the introduction of highly active chemotherapies (CTs), such as mechlorethamine, vincristine, procarbazine, and prednisone (MOPP) and doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD). However, the role of radiotherapy (RT) is still debated in these advanced stages. In this setting, the meta-analysis of Loeffler et al1 is particularly timely and appropriate. They gathered the individual data of 1,740 patients included in 16 randomized trials comparing a pure chemotherapeutic strategy (CT arm) and an association of CT and RT (CT + RT arm) for treating HD at intermediate or advanced stages. Loeffler et al chose to share the 16 trials they had selected in two groups of eight trials. In the first group, the initial CT was of the same type and same duration (6 to 12 months) in both arms, and RT was given as an additional treatment in the CT + RT arm (CT1 v CT1 + RT) (Table 1, trials 1 to 8). In fact, in trial 8, RT was given as an initial treatment and CT as an adjuvant one (RT + CT); moreover, in trials 3 and 6, RT was given in sandwich after an initial CT and before a final CT (CT-RT-CT). In the other group, RT was given in a parallel design (CT1 + CT2 v CT1 + RT or CT1 v CT2 + RT) (Table 1, trials 9 to 16). This "parallel design" corresponds, in fact, to three different types of specific associations: in trials 11, 12, and 13, the same type of CT was given in both arms but was longer in the CT arm than in the CT + RT arm (8 months v 6 months, 10 months v 6 months, and 10 months v 3 months; respectively); trial 14 was a variant of the above-mentioned trials, as six courses of MOPP plus six courses of a MOPP-like regimen were compared with six courses of MOPP plus RT. A second type of association was represented by trials 9, 15, and 16; in these trials, the CT arm was composed of six courses of MOPP or MOPP-like regimens combined with three (trial 9) or six (trials 15 and 16) courses of ABVD, whereas the CT + RT arm combined six courses of MOPP or MOPP-like regimens and RT. Trial 10 did not belong strictly to the parallel design because the CT arm was composed of 4 to 6 months of MOPP combined with 4 to 6 months of ABVD (for a total duration of 8 to 12 months), whereas the CT of the CT + RT arm was composed only of MOPP given for 8 to 12 months also.
Whether RT was performed in "additional" or parallel design, Loeffler et al1 found a statistical advantage for the CT arm in terms of 10-year overall survival; in regard to disease control, the CT + RT arm was found to be better than the CT arm when RT was added, whereas the CT arm was found to be equivalent to the CT + RT arm when RT was given in parallel. Loeffler et al concluded their meta-analysis with the following statement: "despite its effectiveness in local disease control, RT induced long-term hazards and might compromise possibilities to rescue patients from relapse. Hence, RT in this setting should be limited to specific indications." As clinicians involved for years in HD management, we believe that the results of these 16 trials should be interpreted on the basis of some cardinal, well-demonstrated features: (i) six cycles of MOPP, when given alone, allow almost half of the patients with advanced-stage HD to enjoy a permanent complete remission,2 and thus far, there has been no demonstration that MOPP-like regimens are better than MOPP, nor that increasing the number of cycles of MOPP or MOPP-like regimens beyond six could give better disease control or survival rates. (ii) In contrast, six cycles of ABVD or ABVD combined in different manners with MOPP or MOPP-like regimens give better disease control and better survival rates than MOPP2 or MOPP-like3 regimens. (iii) When RT is given alone, there is a strong correlation between the dose of RT and the percentage of in-field sterilization4,5; in this setting, it has been shown that when RT is given as an adjuvant therapy after CT, a dose of more than 20 Gy prevents the occurrence of relapses significantly better than lower doses.6
We thus decided to examine the 16 trials presented by Loeffler et al1 (i) in light of the dose of RT (
Group A was made up of nine trials of a total of 786 randomized patients (trials 1, 2, 3, 6, 7, 8, 12, 13, and 14); in group A, patients on both arms received the same type of initial CT, followed in the CT + RT arm by a dose of RT that was Group B was made up of three trials with a total of 596 patients (trials 4, 5, and 11); in group B, patients also received the same type of CT in both arms; however, the dose of RT given in the CT + RT arm was only 20 Gy (trials 4 and 11) or even less (between 15 and 20 Gy, trial 5). Again, when the number of patients and the log hazards ratio of each of these three trials were taken into account, we found that the CT arm was better for overall survival, whereas the CT + RT arm remained better for disease control. Group C was made up of four trials with a total of 560 patients (trials 9, 10, 15, and 16). In group C, the initial CT of the CT arm was an ABVD-containing regimen, whereas the CT + RT arm was composed of MOPP-like regimens plus RT. In this group, the CT arm was clearly better than the CT + RT arm for survival as well as for disease control (although the CT + RT arm was better than the CT arm when RT was given at 40 to 44 Gy). Altogether, we do not come to the conclusion of Loeffler et al. Our conclusions are as follows: (i) six or more cycles of CT, whatever its type (MOPP, MOPP-like, or ABVD-containing regimens), when combined with nodal irradiation delivered at doses of more than 20 Gy, gave better long-term survival and disease control than similar CTs (even given for longer periods of time) given without irradiation (group A); (ii) MOPP or MOPP-like regimens (six cycles) combined with ABVD (three to six cycles) gave better survival and disease control than the same MOPP or MOPP-like CT (six to 12 cycles) combined with RT. These two conclusions are in keeping with the demonstration (which took 20 years!) that ABVD was better than MOPP and that doses of more than 20 Gy were more effective than doses of less than 20 Gy in the treatment of HD. REFERENCES 1. Loeffler M, Brosteanu O, Hasenclever D, et al: Meta-analysis of chemotherapy versus combined modality treatment trials in Hodgkin's disease. J Clin Oncol 16:818-829, 1998[Abstract] 2. Canellos GP, Anderson JR, Propert KJ, et al: Chemotherapy of advanced Hodgkin's disease with MOPP, ABVD, or MOPP alternating with ABVD. N Engl J Med 327:1478-1484, 1992[Abstract] 3. Urba WJ, Longo DL: Hodgkin's disease. N Engl J Med 326:678-687, 1992[Medline]
4.
Kaplan HS: Evidence for a tumoricidal dose level in the radiotherapy of Hodgkin's disease. Cancer Res 26:1221-1224, 1966 5. Vijaykumar S, Myrianthopoulos LC: An updated dose-response analysis in Hodgkin's disease. Radiother Oncol 24:1-13, 1992[Medline]
6.
Yelle L, Bergsagel D, Basco T, et al: Combined modality therapy of Hodgkin's disease: 10-year results of the National Cancer Institute of Canada Clinical Trials Group multicenter clinical trial. J Clin Oncol 9:1983-1993, 1991
Institute for Medical Informatics, Statistics and Epidemiology, University of Leipzig, Leipzig, Germany In Reply: We thank Drs Andrieu, Yilmaz and Colonna for the critical comments on a meta-analysis about the role of radiotherapy in advanced-stage HD recently published by us and others.1 We'd like to take the opportunity to supplement our analysis by the major aspects suggested by Andrieu et al. This analysis, however, demonstrates that our original conclusions are supported, contrary to what Andrieu et al expected. The 16 comparisons of CT alone versus CT plus RT included in the meta-analysis are heterogeneous with respect to CT regimen (MOPP, MOPP-like, doxorubicin-containing) and number of CT cycles given in the two arms, as well as extension and dose of RT in the combined modality arm. We presented strong arguments for grouping of trials according to additional RT, respectively, parallel RT/CT trial design. Nevertheless, we are aware that our grouping is open to discussion. In this letter, we'd like to show that our conclusions do not change with the grouping proposed by Andrieu et al. In addition, we will take the opportunity to critically discuss the assumptions leading to the new grouping proposed. The results presented by Andrieu et al in their Table 1 are potentially misleading because they do not take into account the magnitude and precision (reflected in the 95% confidence intervals) of the log hazards ratio estimates (as given in our Table 11). Thus, we reanalyzed the data in the sense of an auxilliary sensitivity analysis, using exactly the groups A, B, and C proposed by Andrieu et al. The results are shown in Fig 1.
Group A summarizes all trials that used MOPP or MOPP-like CT regimens in both trial arms and RT delivered at doses of more than 20 Gy in the combined modality arm. None of the CT regimens contains doxorubicin. Here, the combined modality arm shows a 17% improvement in disease control rate after 10 years (P < .0001; 95% confidence interval [CI], 9% to 25%). For overall survival, however, there is only a weak nonsignificant trend toward superiority of the combined modality arm (difference at 10 years, 3%; 95% CI, -4% to 10%; P = .22). Thus, the improvement in disease control does not translate into a similar improvement in overall survival. Group B summarizes trials with RT delivered at doses of 20 Gy or lower in the combined modality arm and either MOPP-like or doxorubicin-containing CT in both arms. There is identical disease control with both treatment modalities (difference at 10 years, 0%; 95% CI, -9% to 9%; P = .8), whereas the CT-alone arm is significantly superior in overall survival (difference at 10 years, 13%; 95% CI, 1% to 24%; P = .03). Group C summarizes trials comparing doxorubicin-containing regimens with MOPP-like regimens combined with RT. For disease control, there is no significant difference between the treatment modalities, although there is a weak trend in favor of CT alone (difference at 10 years, 5%; 95% CI, -4% to 14%; P = .19). With respect to overall survival, the CT-alone arm is clearly superior (difference at 10 years, 12%; 95% CI, 3% to 21%; P = .01). It should be noted that the absolute failure rates between the groups should not be compared. The three trial groups have different patient compositions (more intermediate stages in group A and more advanced stages in group C). The above results support two conclusions. First, combined modality is an effective treatment principle in advanced HD as far as disease control is concerned. The advantage of adding RT, however, is lost if doxorubicin-containing regimens are used or if the CT regimens are sufficiently long. Second, with respect to overall survival, in all three groups, the benefit of RT on disease control dimishes or disappears. This is in accordance with our previous finding that the effect of combined modality treatment in disease control is compromised by an increased rate of fatal events in continuous complete remission and possibly by a decreased salvageability after treatment failure. Thus, the results of our reanalysis are fully consistent with our previous results1 and are clearly in disagreement with the first conclusion of Andrieu et al, that "six or more cycles of CT, whatever its type (MOPP, MOPP-like, or ABVD-containing regimens), when combined with nodal irradiation delivered at doses of more than 20 Gy, gave better long-term survival and disease control than similar CTs (even given for longer periods of time) given without irradiation." Better disease control is only obtained in group A, which did not receive doxorubicin-containing regimens. However, in overall survival, the difference between the two arms diminishes to a nonsignificant trend. Furthermore, in group B, in which longer and/or doxorubicin-containing regimens are used in both arms, there is no effect of RT with respect to disease control. There is also additional information supporting the above conclusions. First, there are indications for a beneficial role of more than six cycles. Björkholm et al2 have shown in a randomized trial of four cycles (MOPP-ABVD) versus response-adapted two to four cycles (MOPP-ABVD) a significant superiority of the fixed arm with respect to disease control. Further evidence for the effect of the number of cycles given stems from the HD3 trial of the German Hodgkin's Study Group3 (trial 11). In that study, a cohort of patients in complete remission refused randomization for consolidation treatment after three cycles (cyclophosphamide, vincristine, procarbazine, and prednisone [COPP]-ABVD) and received no further therapy. In this cohort, disease control and overall survival were significantly reduced as compared with patients who received consolidation treatment. Furthermore, there is no demonstration of a dose-response relationship if RT is given after prolonged CT. The randomized HD1 trial of the German Hodgkin's Lymphoma Study Group4 compared 20 Gy with 40 Gy extended-field RT after two cycles (COPP-ABVD) in intermediate stages. No difference could be detected with respect to disease control or overall survival. Similar results were shown by Schellong.5 In a trial randomizing 36 to 40 Gy versus 18 to 20 Gy extended-field RT in a combined modality setting for children with advanced-stage HD, the two RT doses were similar with respect to disease control and survival. Andrieu et al base their groupings on RT doses according to Yelle et al6 (which is, in fact, a trial included in our meta-analysis, 8 and 12). Originally, all patients in this trial were to receive 20 to 30 Gy of extended-field irradiation. For unknown reasons, some patients received lower doses of RT. Their outcome was compared with that of patients treated according to protocol, and showed a reduced disease control with RT doses of less than 20 Gy. Because of the explorative and potentially biased nature of this comparison, the results cannot be considered "well-demonstrated features," particularly if considered in light of the two randomized trials mentioned before. In addition, we would like to clarify two further points in regard to the letter by Andrieu et al. In our meta-analysis, we found that overall survival was superior with CT alone in the parallel RT/CT design setting and equivalent for both treatment modalities in the additional RT design setting. We never stated a superiority of CT alone with respect to overall survival in both design settings. Because our meta-analysis was performed according to treatment intended, all randomized patients were included in the analysis. This may be easily checked by comparing patient numbers given in our Table 11 with patient numbers in the Results. In summary, we doubt the assumptions on which the alternative grouping of the trials proposed by Andrieu et al is based. Nevertheless, we hope we have shown that even with their grouping, our results and conclusions do not change. The role of RT after sufficiently extensive modern doxorubicin-containing chemotherapy must be judged as rather limited. The recently started HD12 trial of the German Hodgkin's Lymphoma Study Group (Chairman, Prof. V. Diehl) will eventually lead to a clarification of one remaining question. The trial randomizes no RT versus RT to initial bulky disease sites and to sites of partial remission after CT following two variants of the dose-escalated carmustine, etoposide, cytarabine, cyclophosphamide, vincristine, procarbazine, and prednisone regimen given in eight cycles every 3 weeks.7 The trial is designed to randomize about 700 patients into the RT question. REFERENCES 1. Loeffler M, Brosteanu O, Hasenclever D, et al: Meta-analysis of chemotherapy versus combined modality treatment trials in Hodgkin's disease. J Clin Oncol 16:818-829, 1998
2.
Bjorkholm M, Axdorph U, Grimfors G, et al: Fixed versus response-adapted MOPP/ABVD chemotherapy in Hodgkin's disease: A prospective randomized trial. Ann Oncol 6:895-899, 1995
3.
Diehl V, Loeffler M, Pfreundschuh M, et al: Further chemotherapy versus low-dose involved-field radiotherapy as consolidation of complete remission after six cycles of alternating chemotherapy in patients with advance Hodgkin's disease. German Hodgkin's Study Group (GHSG). Ann Oncol 6:901-910, 1995
4.
Loeffler M, Diehl V, Pfreundschuh M, et al: Dose-response relationship of complementary radiotherapy following four cycles of combination chemotherapy in intermediate-stage Hodgkin's disease. J Clin Oncol 15:2275-2287, 1997 5. Schellong GM The German cooperative therapy studies: An approach to minimize treatment modalities and invasive staging procedures. Cancer Treat Res 4:277-289, 1989 6. Yelle L, Bergsagel D, Basco V, et al: Combined modality therapy of Hodgkin's disease: 10-year results of National Cancer Institute of Canada Clinical Trials Group multicenter clinical trial. J Clin Oncol 9:1983-1993, 1991
7.
Diehl V, Sieber M, Rueffer U, et al: BEACOPP: An intensified chemotherapy regimen in advanced Hodgkin's disease. Ann Oncol 8:143-148, 1997
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 1999 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
20 Gy v > 20 Gy)—whether RT was given in an additional or parallel design—and (ii) in light of the type of CT (MOPP or MOPP-like v ABVD-containing regimens)—whether the CT of the CT arm was longer (parallel design) or not (additional design). For this purpose, we pooled the 16 trials into three groups 
30 Gy in seven trials and ranged between 20 and 30 Gy in the two remaining ones (trials 8 and 12). Interestingly, the CT + RT arm was better than the CT arm in eight of nine trials for overall survival and in six of seven trials for disease control (no data were available for trials 8 and 12). When the number of patients randomized in each trial (because the number of patients analyzed in each trial was not given) and the log hazards ratio of each trial were taken into account, it emerged that, in group A, the CT + RT arm was better for both survival and disease control. 
