BACKGROUND:To speed up the evaluation of new therapies, the multi-arm, multi-stage trial design was suggested previously by the authors.METHODS:In this paper, we evaluate the performance of the two-stage, multi-arm design using four cancer trials conducted at the MRC CTU. The performance of the design at fictitious interim analyses is assessed using a conditional bootstrap approach.RESULTS:Two main aims are addressed: the error rate of correctly carrying on/stopping the trial at an interim analysis as well as quantifying the gains in terms of resources by employing this design. Furthermore, we make suggestions for the best timing of this interim analysis.CONCLUSION:Multi-arm, multi-stage trials are an effective way of speeding up the therapy evaluation process. The design performs well in terms of the type I and II error rates.

Ovarian cancer is a relatively rare disease, and patients are infrequently diagnosed with early stages of the disease.Among those with high-risk stage I disease, the probability of disease recurrence within 10 years is approximately 27%, and for those with stage II disease, 44%. 1 Most

Summary of end point recommendations With all the above-mentioned considerations working group Bformulated the recommendations listed in Table 5. Acknowledgements This document summarizes the discussions held in the group Bsection during the 3rd Ovarian Cancer Consensus Conference,organized by the GCIG and held in Baden-Baden, Germany,3–5 September 2004. None of the participants in this sectionindicated any conflict of interest related to the sponsor of thisevent (Bristol-Myers Squibb). References 1. Sylvester R. Planning cancer clinical trials. In Buyse ME, Staquet MJ,Sylvester RJ (eds): Oxford Cancer Clinical Trials – Methods andPractice. 47–63. Oxford University Press 1982.2. Brady MF, Thigpen JT, Vermorken JB, Parmar MKB. Randomisedtrials in ovarian cancer: trial design considerations. Ann Oncol 1999;10 (Suppl 1): S75–S82.3. Cannistra SA. Cancer of the ovary. N Engl J Med 2004; 351:2519–2529.4. Berek JS, Bertelsen K, du Bois A et al. Advanced epithelial ovariancancer: 1998 consensus statements. Ann Oncol 1999; 10 (Suppl 1):S87–S92.5. Vergote I, De Brabanter J, Fyles A et al. Prognostic importance ofdegree of differentiation and cyst rupture in stage I invasive epithelialovarian carcinoma. Lancet 2001; 357: 176–182.6. Zanetta G. The accuracy of staging: an important prognostic determi-nator in stage I ovarian carcinoma. A multivariate analysis. Ann Oncol1998; 9: 1097–1101.7. Eisenhauer EA, Gore M, Neijt JP. Ovarian cancer: should we bemanaging patients with good and bad prognostic factors in the samemanner? Ann Oncol 1999; 10 (Suppl 1): S9–S15.8. Trope´ C, Kaern J, Hogberg T et al. Randomized study on adjuvantchemotherapy in stage I high-risk ovarian cancer with evaluation ofDNA-ploidy as prognostic instrument. Ann Oncol 2000; 11: 281–288.9. Eisenhauer EA, Vermorken JB, van Glabbeke M. Predictors of responseto subsequent chemotherapy in platinum pretreated ovarian cancer: ananalysis of 704 patients. Ann Oncol 1997; 8: 963–968.10. van Glabbeke M, Vermorken J, Eisenhauer E. Predictors of survival(OS) and time to progression (TTP) following 2

Formal monitoring of data from randomised controlled trials (RCTs) is becoming more common. Wide variation exists in the structure and organisation of data monitoring committees (DMCs), with little guidance on how they should operate. We used various strategies to consider the behavioural, procedural, and organisational aspects of data monitoring in RCTs: systematic reviews of DMCs and small group processes in decision making; surveys of reports of RCTs, recently completed and ongoing RCTs, and the policies of major organisations connected with RCTs; detailed case studies of four DMCs that faced difficult decisions; and interviews with experienced DMC members. The findings aided the development of a template for a charter for DMCs. We summarise the findings and outline the key considerations at every stage of the data monitoring process. Widespread use of a charter for the structure and organisation of DMCs would promote a systematic and transparent approach, and enable them to operate more effectively and efficiently.

PURPOSEThe Big Lung Trial (BLT) was a large, pragmatic trial to evaluate the addition of chemotherapy to primary treatment (ie, surgery, radical radiotherapy, or supportive care) in non-small-cell lung cancer (NSCLC). In the supportive care group, there was a small but significant survival benefit in patients treated with chemotherapy compared with supportive care alone (no chemotherapy). A substudy was undertaken to evaluate the quality of life (QoL) implications of the treatment options. QoL was assessed using European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaires C30 (QLQ-C30) and LC17, and daily diary cards.PATIENTS AND METHODSEORTC QLQ-C30 and LC17 were collected at 0, 6 to 8, 12, 18, and 24 weeks. Diary cards were completed during the first 12 weeks of the study. The primary end point was global QoL at 12 weeks.RESULTSA total of 273 patients were randomly assigned: 138 to no chemotherapy and 135 to chemotherapy. There was no evidence of a large detrimental effect on QoL of chemotherapy. No statistically significant differences in global QoL or physical/emotional functioning, fatigue and dyspnea, and pain were detected at 12 weeks. Higher rates of palliative radiotherapy in the no chemotherapy arm may have lessened differences in QoL. Global QoL, role functioning, fatigue, appetite loss, and constipation were prognostic indicators of survival at 12 weeks.CONCLUSIONThere were no important adverse effects of chemotherapy on QoL.

OBJECTIVESTo address issues about data monitoring committees (DMCs) for randomised controlled trials (RCTs).DATA SOURCESElectronic databases. Handsearching of selected books. Personal contacts with experts in the field.REVIEW METHODSSystematic literature reviews of DMCs and small group processes in decision-making; sample surveys of: reports of RCTs, recently completed and ongoing RCTs and policies of major organisations involved in RCTs; case studies of four DMCs; and interviews with experienced DMC members. All focused on 23 prestated questions.RESULTSAlthough still a minority, RCTs increasingly have DMCs. There is wide agreement that nearly all trials need some form of data monitoring. Central to the role of the DMC is monitoring accumulating evidence related to benefit and toxicity; variation in emphasis has been reflected in the plethora of names. DMCs for trials performed for regulatory purposes should be aware of any special requirements and regulatory consequences. Advantages were identified for both larger and smaller DMCs. There is general agreement that a DMC should be independent and multidisciplinary. Consumer and ethicist membership is controversial. The chair is recognised as being particularly influential, and likely to be most effective if he or she is experienced, understands both statistical and clinical issues, and is facilitating in style and impartial. There is no evidence available to judge suggested approaches to training. The review suggested that costs should be covered, but other rewards must be so minimal as to not affect decision-making. It is usual to have a minimum frequency of DMC meetings, with evidence that face-to-face meetings are preferable. It is common to have open sessions and a closed session. A report to a DMC should cover benefits and risks in a balanced way, summarised in an accessible style, avoiding excessive detail, and be as current as possible. Disadvantages of blinded analyses seem to outweigh advantages. Information about comparable studies should be included, although interaction with the DMCs of similar ongoing trials is controversial. A range of formal statistical approaches can be used, although this is only one of a number of considerations. DMCs usually reach decisions by consensus, but other approaches are sometimes used. The general, but not unanimous, view is that DMCs should be advisory rather than executive on the basis that it is the trial organisers who are ultimately responsible for the conduct of the trial.CONCLUSIONSSome form of data monitoring should be considered for all RCTs, with reasons given where there is no DMC or when any member is not independent. An early DMC meeting is helpful, determining roles and responsibilities; planned operations can be agreed with investigators and sponsors/funders. A template for a DMC charter is suggested. Competing interests should be declared. DMC size (commonly three to eight people) is chosen to optimise performance. Members are usually independent and drawn from appropriate backgrounds, and some, particularly the chair, are experienced. A minimum frequency of meetings is usually agreed, with flexibility for more if needed. The DMC should understand and agree the statistical approach (and guidelines) chosen, with both the DMC statistician and analysis statistician competent to apply the method. A DMC's primary purpose is to ensure that continuing a trial according to its protocol is ethical, taking account of both individual and collective ethics. A broader remit in respect of wider ethical issues is controversial; arguably, these are primarily the responsibility of research ethics committees, trial steering committees and investigators. The DMC should know the range of recommendations or decisions open to it, in advance. A record should be kept describing the key issues discussed and the rationale for decisions taken. Errors are likely to be reduced if a DMC makes a thorough review of the evidence and has a clear understanding of how it should function, there is active participation by all members, differences are resolved through discussion and there is systematic consideration of the various decision options. DMCs should be encouraged to comment on draft final trial reports. These should include information about the data monitoring process and detail the DMC membership. It is recommended that groups responsible for data monitoring be given the standard name 'Data Monitoring Committee' (DMC). Areas for further research include: widening DMC membership beyond clinicians, trialists and statisticians; initiatives to train DMC members; methods of DMC decision-making; 'open' data monitoring; DMCs covering a portfolio of trials rather than single trials; DMC size and membership, incorporating issues of group dynamics; empirical study of the workings of DMCs and their decision-making, and which trials should or should not have a DMC.

Many models for clinical prediction (prognosis or diagnosis) are published in the medical literature every year but few such models find their way into clinical practice. The reason may be that since in most cases models have not been validated in independent data, they lack generality and/or credibility. In this paper we consider the situation in which several compatible, independent data sets relating to a given disease with a time‐to‐event endpoint are available for analysis. The aim is to construct and evaluate a single prognostic model. Building a multivariable model from the available prognostic factors is accomplished within the Cox proportional hazards framework, stratifying by study. Non‐linear relationships with continuous predictors are modelled by using fractional polynomials. To assess the discrimination or separation of a survival model, we use the D statistic of Royston and Sauerbrei. D may be interpreted as the separation (log hazard ratio) between the survival distributions for two independent prognostic groups. To evaluate the generality of a prognostic model across the data sets, we propose ‘internal–external cross‐validation’ on D : each study is omitted in turn, the model parameters are estimated from the remaining studies and D is evaluated in the omitted study. Because the linear predictor of a survival model tells only part of the story, we also suggest a method for investigating heterogeneity in the baseline distribution function across studies which involves fitting completely specified, flexible parametric survival models (Royston and Parmar). Our final models combine the prognostic index (obtained with stratification by study) with the pooled baseline survival distribution (estimated parametrically). By applying this methodology, we construct two prognostic scores in superficial bladder cancer. The simpler of the two scores is more suited to clinical application. We show that a three‐group prognostic classification scheme based on either score produces well‐separated survival curves for each of the data sets, despite identifiable heterogeneity among the baseline distribution functions and to a lesser extent among the prognostic indexes for the individual studies. Copyright © 2004 John Wiley & Sons, Ltd.

BACKGROUND:The incidence of mesothelioma is rising rapidly in the UK. There is no generally accepted standard treatment. The BTS recommends active symptom control (ASC). It is not known whether chemotherapy in addition prolongs survival or provides worthwhile palliation with acceptable toxicity. Palliation as recorded by patients has been fully reported for only two regimens: mitomycin, vinblastine, and cisplatin (MVP), and vinorelbine (N). The BTS and collaborators planned to conduct a phase III randomised trial comparing ASC only, ASC+MVP, and ASC+N in 840 patients with survival as the primary outcome measure. The aim of the present study was to assess the acceptability of the trial design to patients and the suitability of two standard quality of life (QL) questionnaires for mesothelioma.METHODS:Collaborating centres registered all new patients with mesothelioma. Those eligible and giving informed consent completed EORTC QLQ-C30+LC13 and FACT-L QL questionnaires and were randomised between all three or any two of (1) ASC only, (2) ASC+4 cycles of MVP, and (3) ASC+12 weekly doses of N.RESULTS:During 1 year, 242 patients were registered of whom 109 (45%) were randomised (55% of the 197 eligible patients). Fifty two patients from 20 centres were randomised to an option including ASC only. This translates into a rate of 312 per year from 60 centres interested in collaborating in the phase III trial. The EORTC QL questionnaire was superior to FACT-L in terms of completeness of data and patient preference. Clinically relevant palliation was achieved with ASC.CONCLUSION:The planned phase III trial is feasible.

Objectives: The non-small cell lung cancer (NSCLC) meta-analysis suggested a survival benefit for cisplatin-based chemotherapy when given in addition to surgery. radical radiotherapy or 'best supportive care'. However, it included many small trials and trials with differing eligibility criteria and chemotherapy regimens. The aim of the Big Lung Trial was therefore to run a large pragmatic trial to confirm the survival benefits seen in the meta-analysis. Methods: In the surgery setting, a total of 381 patients were randomised to chemotherapy (C, 192 patients) or no chemotherapy (NoC, 189 patients). C was three 3-weekly cycles of cisplatin/vindesine, mitomycin/ifosfamide/cisplatin, mitomycin/vinblastine/cisplatin or vinorelbine/cisplatin. Results: Chemotherapy was given before surgery in 3% of patients whilst 97% received adjuvant chemotherapy. Baseline characteristics were: median age 61 years, 69% male, 48% squamous cell, 93% WHO PS 0-1, 27% stage I, 38% stage II, and 34% stage III. Complete resection was achieved in approximately 95% of patients. In the C group, 13% received no chemotherapy, 21% one or two cycles, and 64% all three cycles of their prescribed chemotherapy (60% of the latter with no delays or modification). 30% had grade 3/4 toxicity, mainly haematological, nausea/vomiting and neutropenic fever, and six patients were reported as having a treatment-related death. 198 (52%) of patients have died, but there is currently no evidence of a benefit in overall survival to the C group: HR 1.02 (95% CI 0.77-1.35), P = 0.90). Conclusions: This trial has failed to observe a survival benefit with adjuvant chemotherapy following complete resection of stage I-III NSCLC. However, the hazard ratio and 95% confidence intervals are consistent with the previously reported meta-analysis and two large recently reported trials, which suggest a small survival benefit with cisplatin-based chemotherapy. (C) 2004 Elsevier B.V. All rights reserved.

Background: In 1995 a meta-analysis of randomised trials investigating the value of adding chemotherapy to primary treatment for non-small cell lung cancer (NSCLC) suggested a small survival benefit for cisplatin-based chemotherapy in each of the primary treatment settings. However, the meta-analysis included many small trials and trials with differing eligibility criteria and chemotherapy regimens. Methods: The aim of the Big Lung Trial was to confirm the survival benefits seen in the meta-analysis and to assess quality of life and cost in the supportive care setting. A total of 725 patients were randomised to receive supportive care alone (n = 361) or supportive care plus cisplatin-based chemotherapy (n = 364). Results: 65% of patients allocated chemotherapy (C) received all three cycles of treatment and a further 27% received one or two cycles. 74% of patients allocated no chemotherapy (NoC) received thoracic radiotherapy compared with 47% of the C group. Patients allocated C had a significantly better survival than those allocated NoC: HR 0.77 (95% CI 0.66 to 0.89, p = 0.0006), median survival 8.0 months for the C group v 5.7 months for the NoC group, a difference of 9 weeks. There were 19 (5%) treatment related deaths in the C group. There was no evidence that any subgroup benefited more or less from chemotherapy. No significant differences were observed between the two groups in terms of the pre-defined primary and secondary quality of life end points, although large negative effects of chemotherapy were ruled out. The regimens used proved to be cost effective, the extra cost of chemotherapy being offset by longer survival. Conclusions: The survival benefit seen in this trial was entirely consistent with the NSCLC meta-analysis and subsequent similarly designed large trials. The information on quality of life and cost should enable patients and their clinicians to make more informed treatment choices.

Approximately 10% of patients with superficial bladder cancer (pTa/pT1) recur with life‐threatening muscle‐invasive disease. Identification of these patients has been a major goal of bladder cancer research. In 1994, it was suggested that p53 immunostaining could identify the cancers that would progress and it was proposed that tumours that stain for p53 should be treated aggressively with radiotherapy or cystectomy. Despite the hundreds of studies published since on the relationship between p53 and progression in superficial bladder cancer, the clinical utility of p53 immunostaining has not been resolved because of limitations concerning the numbers of patients and the length of follow‐up. This study set out to overcome these limitations by using tissue from a large multicentre trial that recruited 502 patients with a median follow‐up of 10 years. Each of 34 patients that had progressed with ≥ pT2 disease or had distant metastases or had died from bladder cancer was compared with one or two matched controls. Sections were stained with a mouse monoclonal antibody to p53, pAb1801. In agreement with many of the earlier studies, p53 immunostaining had prognostic significance. The adjusted hazard ratio for time to progression for the pAb1801‐positive versus negative group was 2.5, with 95% confidence intervals of 1.05–5.98 (p = 0.039). The other major risk factor that is associated with progression of superficial bladder cancer is pT1G3 disease. Of the 42 pT1G3 cancers, 14 (33%) progressed. The proportion of cancers with p53 staining that progressed was similar to the proportion of pT1G3 cancers that progressed, but neither the sensitivity nor the specificity of association of p53 staining with progression is sufficient to recommend cystectomy in individual patients. Copyright © 2003 John Wiley & Sons, Ltd.

Background Despite improvements in the treatment of ovarian cancer, most patients develop recurrent disease within 3 years of diagnosis. There is no agreed second-line treatment at relapse. We assessed paclitaxel plus platinum chemotherapy as such treatment.Methods In parallel international, multicentre, randomised trials, between January, 1996, and March, 2002, 802 patients with platinum-sensitive ovarian cancer relapsing after 6 months of being treatment-free were enrolled from 119 hospitals in five countries. Patients were randomly assigned paclitaxel plus platinum chemotherapy or conventional platinum-based chemotherapy. Analysis was by intention to treat, except for toxic effects.Findings With a median follow-up of 42 months, 530 patients have died. Survival curves showed a difference in favour of paclitaxel plus platinum (hazard ratio 0.82 [95% Cl 0.69-0.97], p=0.02), corresponding to an absolute difference in 2-year survival of 7% between the paclitaxel and conventional treatment groups (57 vs 50% [95% Cl for difference 1-12]), and median survival of 5 months (29 vs 24 months [1-11). 717 patients developed progressive disease or died. The progress ion-free survival curves show a difference in favour of paclitaxel plus platinum (hazard ratio 0.76 [0.66-0.89], p=0.0004), corresponding to an absolute difference in 1-year progression-free survival of 10% (50 vs 40% [4-15]) and in median progress ion-free survival of 3 months (13 vs 10 months [1-5]).Interpretation Paclitaxel plus platinum chemotherapy seems to improve survival and progression-free survival among patients with relapsed platinum-sensitive ovarian cancer compared with conventional platinum-based chemotherapy.

Despite the enrollment of more than 3000 women in randomised trials, the benefits and risks of neoadjuvant chemotherapy in the treatment of locally advanced cervical cancer remain uncertain. We carried out a systematic review and meta-analysis of individual patient data to assess the effect of neoadjuvant chemotherapy in two comparisons. In the first comparison, of neoadjuvant chemotherapy followed by radical radiotherapy compared with the same radiotherapy alone, we obtained data from 18 trials and 2074 patients. When all trials were considered together, a high level of statistical heterogeneity suggested that the results could not be combined indiscriminately. A substantial amount of heterogeneity was explained by separate analyses of groups of trials. Trials using chemotherapy cycle lengths of 14 days and shorter (Hazard Ratio (HR))=0.83, 95% Confidence Interval (CI)=0.69-1.00, P=0.046) or cisplatin dose intensities greater than or equal to 25 mg/m2 per week (HR=0.91, 95% CI=0.78-1.05, P=0.20) tended to show an advantage for neoadjuvant chemotherapy on survival. In contrast, trials using cycle lengths longer than 14 days (HR=1.25, 95% CI=1.07-1.46, P=0.005) or cisplatin dose intensities lower than 25mg/m2 per week (HR=1.35, 95% CI=1.11-1.14, P=0.002) tended to show a detrimental effect of neoadjuvant chemotherapy on survival. In the second comparison, of neoadjuvant chemotherapy followed by surgery compared with radical radiotherapy alone, data from 5 trials and 872 patients were obtained. The combined results from all trials (HR=0.65, 95% CI=0.53-0.80, P=0.0004) indicated a highly significant reduction in the risk of death with neoadjuvant chemotherapy, but there were some differences between the trials in their design and results. Despite some unexplained heterogeneity, the timing and dose intensity of cisplatin-based neoadjuvant chemotherapy appears to have an important impact on whether or not it benefits women with locally advanced cervical cancer and warrants further exploration.

Modelling of censored survival data is almost always done by Cox proportional-hazards regression. However, use of parametric models for such data may have some advantages. For example, non-proportional hazards, a potential difficulty with Cox models, may sometimes be handled in a simple way, and visualization of the hazard function is much easier. Extensions of the Weibull and log-logistic models are proposed in which natural cubic splines are used to smooth the baseline log cumulative hazard and log cumulative odds of failure functions. Further extensions to allow non-proportional effects of some or all of the covariates are introduced. A hypothesis test of the appropriateness of the scale chosen for covariate effects (such as of treatment) is proposed. The new models are applied to two data sets in cancer. The results throw interesting light on the behaviour of both the hazard function and the hazard ratio over time. The tools described here may be a step towards providing greater insight into the natural history of the disease and into possible underlying causes of clinical events. We illustrate these aspects by using the two examples in cancer.

Background Previously, we have shown that the combination of cyclophosphamide, doxorubicin, and cisplatin (CAP) and single-agent carboplatin produce similar survival and progression-free survival rates in women with ovarian cancer. Subsequently, paclitaxel combined with platinum has become a widely accepted treatment for the disease. We aimed to compare the safety and efficacy of paclitaxel plus carboplatin with a control of either CAP or carboplatin alone.Methods Between February, 1995, and October, 1998, we enrolled 2074 patients from 130 centres in eight countries. Women were randomly assigned paclitaxel plus carboplatin or control, the control (CAP or single-agent carboplatin) being chosen by the patient and clinician before randomisation. The primary outcome measure was overall survival. Secondary outcomes were progression-free survival and toxicity. Analysis was by intention to treat.Findings With a median follow-up of 51 months, 1265 patients had died, and survival curves showed no evidence of a difference in overall survival between paclitaxel plus carboplatin and control (hazard ratio 0.98, 95% Cl 0.87-1.10, p=0.74). The median overall survival was 36.1 months on paclitaxel plus carboplatin and 35.4 months on control (difference 0.7 months, 95% CI -3.6 to 4.7). 1538 patients had progressive disease or died, and again, Kaplan-Meier curves showed no evidence of a difference between the groups (hazard ratio 0.93, 95% Cl 0.84-1.03, p=0.16). Median progression-free survival was 17.3 months on paclitaxel plus carboplatin and 16.1 months on control (difference 1.2 months, 95% Cl -0.5 to 2.8). Paclitaxel plus carboplatin caused more alopecia, fever, and sensory neuropathy than carboplatin alone, and more sensory neuropathy than CAP. CAP was associated with more fever than paclitaxel plus carboplatin.Interpretation Single-agent carboplatin and CAP are as effective as paclitaxel plus carboplatin as first-line treatment for women requiring chemotherapy for ovarian cancer. The favourable toxicity profile of single-agent carboplatin suggests that this drug is a reasonable option as first-line chemotherapy for ovarian cancer.

BACKGROUND:Trials on the effect of systemic chemotherapy on survival and recurrence in adults with high-grade glioma have had inconclusive results. We undertook a systematic review and meta-analysis to assess the effects of such treatment on survival and recurrence. METHODS:We did a systematic review and meta-analysis using updated data on individual patients from all available randomised trials that compared radiotherapy alone with radiotherapy plus chemotherapy. Data for 3004 patients from 12 randomised controlled trials were included (11 published and one unpublished). FINDINGS:Overall, the results showed significant prolongation of survival associated with chemotherapy, with a hazard ratio of 0.85 (95% CI 0.78-0.91, p<0.0001) or a 15% relative decrease in the risk of death. This effect is equivalent to an absolute increase in 1-year survival of 6% (95% CI 3-9) from 40% to 46% and a 2-month increase in median survival time (1-3). There was no evidence that the effect of chemotherapy differed in any group of patients defined by age, sex, histology, performance status, or extent of resection. INTERPRETATION:This small but clear improvement in survival from chemotherapy encourages further study of drug treatment of these tumours.