A meta-analysis of stages I and II Hodgkin's disease

Abstract

To compare radiotherapy alone to chemotherapy plus radiotherapy in the treatment of early stage Hodgkin's disease, the English language medical literature was searched for reports on randomized clinical trials in Stages I and I1 Hodgkin's disease from 1975 through 1986. Twenty-three reports with 2999 patients were entered into matched study analysis. Data on extended-field radiotherapy (EF), involved-field (IF), chemotherapy alone, combination chemotherapy and radiotherapy (CM), disease stage, laparotomy staging, and complications were gathered. A proportional hazard rate was used to estimate and compare relapsefree (RFS) and overall survival rates (S). Iteratively reweighted least square analysis was used to estimate survival curves. Twelve-year RFS for CM (889 patients) was significantly superior to EF (1350 patients) (P < 0.01). Twenty-year RFS in EF was better than IF (760 patients) (P < 0.01). Twelve-year S for CM was not significantly different than for EF but was better than for IF (P < 0.05).Cancer 65:1155-1160, 1990.HE MANAGEMENT of Stages 1 and 11 Hodgkin's dis-T ease is controversial. 1-4 Proponents of radiotherapy alone (after laparotomy staging) cite a decreased risk of secondary hematologic malignancies and chemotherapyrelated toxicities with excellent results of chemotherapybased salvage if necessary. Proponents of initial chemotherapy avoid laparotomy-related sequelae as well as the even higher rate of secondary leukemia in patients subjected to initial radiotherapy and subsequent salvage chemotherapy. In addition, it is apparent that salvage chemotherapy is not always entirely effective in achieving cure.5To improve our understanding of the optimal therapeutic modality for Stage I and I1 Hodgkin's disease, we chose to do a meta-analysis of the English language literature addressing this issue. Many randomized controlled trials on Hodgkin's disease management are too small to detect clinically important differences. However, metaanalysis allowed us to summarize results of different randomized studies, thus increasing the statistical power of study endpoints and resolving the differences between studies.6-8The objectives of our study were as follows: (1) to determine relapse-free survival (RFS) rates based on treatment modality (involved-field radiotherapy [IF], extended-field radiotherapy [EF], combined modality therapy [CM], i.e., radiation and chemotherapy, or chemotherapy alone; (2) to determine overall survival (S) rates based on treatment modality; (3) to determine the effect on RFS and S of laparotomy staging or no laparotomy staging; and (4) to determine complications and long-term toxicities related to treatment modalities. Materials and Methods Literature SearchThe English language literature was searched for reports on clinical trials in Stages I and I1 Hodgkin's disease from January 1975 to December 1986. The search was conducted from Index Medicus, from a Med Line (National Library of Medicine MEDLARS Retrieval System, Bethesda, Maryland) computerized search, from reviewing the refere...

Meta‐analysis of summary survival curve data

“…All of the above-mentioned methods deal with one summary measure for the treatment effect. A number of methods that focus on estimating the whole survival curve have been proposed [11,[18][19][20][21]. In these meta-analysis methods, data from entire survival curves are combined, instead of artificially compromising the data to just one single survival statistic or to a survival estimate on just one fixed point in time.…”

Hodgkin's Disease

“…25, 31-33), A meta-analysis of 23 studies involving 3000 patients treated between 1975 and I986 shows that the risk for relapse is clearly higher following radiotherapy, especially following IF radiotherapy versus combination therapy. No difference in survival was found ( 80%) between EF radiotherapy and combination therapy (18). Likewise, another meta-analysis of 22 studies involving slightly more than 2000 patients could not show clearly better survival for combination therapy compared to radiotherapy alone (34).…”

“…The studies suggest a relapse frequency around 50% and longterm survival of 60% to 70%. More recent controlled and non-controlled studies and a meta-analysis show that radiation therapy based on E F principles yields better relapsefree survival (55% to 85%) than therapy based on I F principles (30%) t o 50'%1), but no difference appears in total long-term survival, 80%) to 90%) (8,17,18). When I F is defined as a volume greater than the involved field alone (usually meaning the mantle volume), studies show that EF methods yield longer relapse-free survival, but d o not improve overall survival (19).…”

A non‐parametric procedure for evaluating treatment effect in the meta‐analysis of survival data

“…In all experiments, e ective sample sizes were computed for the interval [0; 2] years using the deÿnition of Cutler and Ederer in (9). In addition, e ective sample sizes for the interval [0; 2] years using Peto's deÿnition in (10) were used in two experiments. With estimatesŜ gi (t) and e ective sample sizes, N 1i ; N 2i , thus generated for Groups 1 and 2 for a trial, a weighting factor, w i , based on e ective sample sizes was calculated for the trial, where w i = (N 1i N 2i )=(N 1i + N 2i ).…”

“…This contrast index, denoted as Â, is shown in Section 2 to be interpretable as a log-transformed weighted average of the hazard ratios in the corresponding interval for the cumulative survival rates. The log hazard ratio is an attractive contrast index for the meta-analysis of survival data [5,9,10]. When death is the outcome for survival, the exponentiated value of this contrast index expresses how many times as fast patients are dying in one treatment regimen relative to a reference treatment.…”