Unblinded Data Research Articles

PL‐03‐02: Evaluating semagacestat, a gamma‐secretase inhibitor, in a phase III trial

PL‐03‐02: Evaluating semagacestat, a gamma‐secretase inhibitor, in a phase III trial

Continuous safety monitoring for randomized controlled clinical trials with blinded treatment information: Part 4: One method

Standard methods for testing safety data are needed to ensure the safe conduct of clinical trials. In particular, objective rules for reliably identifying unsafe treatments need to be put into place to help protect patients from unnecessary harm. DMCs are uniquely qualified to evaluate accumulating unblinded data and make recommendations about the continuing safe conduct of a trial. However, it is the trial leadership who must make the tough ethical decision about stopping a trial, and they could benefit from objective statistical rules that help them judge the strength of evidence contained in the blinded data. We design objective early stopping rules for harm that act as continuous safety screens for randomized controlled clinical trials using the accumulating data with blinded treatment information, which could be used by anyone, including trial investigators and trial leadership. A Bayesian framework, with emphasis on the likelihood function, is used to allow for continuous monitoring without adjusting for multiple comparisons. Close collaboration between the statistician and the clinical investigators will be needed in order to design safety screens with good operating characteristics. Though the math underlying this procedure may be computationally intensive, implementation of the statistical rules will be easy and the continuous screening provided will give suitabley early warning when real problems were to emerge. Trial investigators and trial leadership need these safety screens to help them to effectively monitor the ongoing safe conduct of clinical trials with blinded data.

Accelerated clinical discovery using self-reported patient data collected online and a patient-matching algorithm

Patients with serious diseases may experiment with drugs that have not received regulatory approval. Online patient communities structured around quantitative outcome data have the potential to provide an observational environment to monitor such drug usage and its consequences. Here we describe an analysis of data reported on the website PatientsLikeMe by patients with amyotrophic lateral sclerosis (ALS) who experimented with lithium carbonate treatment. To reduce potential bias owing to lack of randomization, we developed an algorithm to match 149 treated patients to multiple controls (447 total) based on the progression of their disease course. At 12 months after treatment, we found no effect of lithium on disease progression. Although observational studies using unblinded data are not a substitute for double-blind randomized control trials, this study reached the same conclusion as subsequent randomized trials, suggesting that data reported by patients over the internet may be useful for accelerating clinical discovery and evaluating the effectiveness of drugs already in use.

Reaction to the recent publication by Rosemary Basson entitled ‘Testosterone therapy for reduced libido in women’

Reaction to the recent publication by Rosemary Basson entitled ‘Testosterone therapy for reduced libido in women’

Improvements in haemolysis and indicators of erythrocyte survival do not correlate with acute vaso‐occlusive crises in patients with sickle cell disease: a phase III randomized, placebo‐controlled, double‐blind study of the gardos channel blocker senicapoc (ICA‐17043)

Red blood cell (RBC) hydration is regulated in part by the Ca(2+) -activated K(+) efflux (Gardos) channel. Senicapoc selectively blocks potassium efflux through the Gardos channel, reducing RBC dehydration and haemolysis, and increasing haemoglobin levels in sickle cell disease (SCD). This randomized, placebo-controlled trial was designed to determine the safety and clinical efficacy of senicapoc in SCD patients. One hundred and forty-five patients were randomized to receive senicapoc and 144 patients to receive placebo for 52 weeks. Consistent with a previous study, patients in the senicapoc group had significantly increased haematocrit, haemoglobin, and decreased numbers of both dense erythrocytes and reticulocytes when compared to the placebo group. The unblinded Data Monitoring Committee terminated this study early due to a lack of efficacy when it determined that, despite improvements in anaemia and haemolysis, no significant improvement in the rate of sickle cell painful crises was observed in patients treated with senicapoc compared to those on placebo (0·38 vs. 0·31, respectively). Comparisons of the times to first, second and third crises between the senicapoc and placebo groups were not statistically significant. Nausea and urinary tract infections occurred more frequently in the senicapoc group than placebo. Serious adverse events were similar in the two groups.

Statistical Inference in Adaptive Group Sequential Trials with the Standardized Mean Difference as Effect Size

We consider the standardized difference of normal means as the outcome measure for comparing two independent groups, say experimental E and control C. This scale- and translation-invariant effect measure enables a convenient specification of a noninferiority margin in applications. Starting with a family of null hypotheses, we derive adaptive group sequential confidence intervals keeping the predefined confidence coefficient. The interval at the end of the trial determines whether and which null hypotheses can be rejected. During the course of the trial, the sample size can be calculated in a completely adaptive way based on the unblinded data of previously performed stages. Concrete rules for sample size updating are provided. Moreover, in each interim analysis, it is possible to change the planning from showing noninferiority to showing superiority or vice versa without affecting the overall type I error. A real data example is worked out in detail and the change in the sample size planning from showing noninferiority to showing superiority is considered during the ongoing trial. In Monte Carlo simulations, we investigate the practical properties of the proposed intervals.

Data Monitoring in Adaptive Dose-Ranging Trials

In clinical studies that address dose-finding objectives, the use of trial designs that allow adaptations based upon accruing unblinded data have much potential to introduce added efficiency and to improve the decision making process, compared to traditional designs. In this paper we consider the processes by which such adaptations may be made. Conventions for interim data monitoring and decision making, particularly with regard to restricting access to data, have largely evolved based upon the needs in Phase III nonadaptive studies; as such, they may not be an ideal fit for adaptive dose-ranging studies, because of the different natures of the trials, their placement in the drug development process, and the objectives of the monitoring. In this paper, we specifically discuss what processes for interim monitoring, analysis, decision making and implementation, might be advantageous and appropriate for adaptive dose-ranging studies, and then illustrate how processes were implemented and carried out in a particular case study.

Adaptive group sequential confidence intervals for the ratio of normal means

In controlled group sequential trials, we consider the ratio of normal means as the effect measure of interest and derive group sequential nested confidence intervals on this parameter. In an interim analysis, the sample sizes of the following stages can be determined in a completely adaptive way using the unblinded data from all previously performed stages. Despite the data-dependent adaptation, the nested confidence intervals always keep the predefined confidence coefficient.

P02-135 - A pattern recognition matrix for placebo-response in schizophrenia

ObjectivesThere is a need for real-time capability to detect inconsistencies in efficacy outcome measures and to predict individual-level or group-level placebo response. Training and calibration to improve reliability is one method; though this can be undermined by the enrollment of placebo responders.MethodsA pattern recognition matrix for placebo-response in schizophrenia was developed based on a Phase II study of schizophrenia conducted in the US. A data monitoring algorithm based on the Positive and Negative Syndrome Scale (PANSS) was retrospectively applied to the unblinded data and score patterns for the placebo responders versus placebo non-responders were analyzed.ResultsA total of 35 placebo responders and 35 randomly selected placebo non-responders were compared. For specific patterns during the first 2 visits, patients were significantly likely to be placebo responders. Within this sample of patients who were randomized to placebo, those who demonstrated the pattern within the initial study visits were approximately three times more likely (OR=2.9, p=0.027) to demonstrate a placebo response.ConclusionThis initial finding suggests that this method could aid in the detection of placebo response early in a trial. In concert with a data-monitoring process we would expect more robust signal detection. The use of the same system, coupled with ongoing training has demonstrated significant improvements in reliability, increasing the ICC by 19% in a 3-month period. If deployed at startup, this method provides a cost-effective way of managing the data quality in RCTs.

Dipeptidyl Peptidase-IV Inhibitor Use Associated With Increased Risk of ACE Inhibitor-Associated Angioedema

Dipeptidyl peptidase-IV (DPP-IV) inhibitors decrease degradation of the incretins. DPP-IV inhibitors also decrease degradation of peptides, such as substance P, that may be involved in the pathogenesis of angiotensin-converting enzyme (ACE) inhibitor-associated angioedema. This study tested the hypothesis that DPP-IV inhibition affects risk of clinical angioedema, by comparing the incidence of angioedema in patients treated with the DPP-IV inhibitor vildagliptin versus those treated with comparator in Phase III randomized clinical trials. Prospectively defined angioedema-related events were adjudicated in a blinded fashion by an internal medicine adjudication committee and expert reviewer. Concurrent ACE inhibitor or angiotensin receptor blocker exposure was ascertained from case report forms. Study drug exposure was ascertained from unblinded data from phase III studies. Odds ratios and 95% confidence intervals comparing angioedema risk in vildagliptin-treated and comparator-treated patients were calculated for the overall population and for patients taking ACE inhibitors or angiotensin receptor blockers, using both an analysis of pooled data and a meta-analysis (Peto method). Overall, there was no association between vildagliptin use and angioedema. Among individuals taking an ACE inhibitor, however, vildagliptin use was associated with an increased risk of angioedema (14 confirmed cases among 2754 vildagliptin users versus 1 case among 1819 comparator users: odds ratio 4.57 [95% confidence interval 1.57 to 13.28]) in the meta-analysis. Vildagliptin use may be associated with increased risk of angioedema among patients taking ACE inhibitors, although absolute risk is small. Physicians confronted with angioedema in a patient taking an ACE inhibitor and DPP-IV inhibitor should consider this possible drug-drug interaction.

The influence of unscheduled biopsies on prostate cancer detection in the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) trial

1504^ Background: REDUCE is an ongoing 4-year, international, randomized, placebo-controlled study investigating the effect of dutasteride on the incidence of biopsy-detectable prostate cancer (PCa). We assessed the influence of prostate specific antigen (PSA)-driven, unscheduled biopsies on PCa detection. Methods: Eligible subjects were aged 50–75 years with serum PSA 2.5–10 ng/mL (50–60 years) or 3–10 ng/mL (>60 years) and a negative 6–12-core biopsy ≤6 months before enrollment. Protocol-mandated transrectal ultrasound directed biopsies (10 cores) were taken at 2 and 4 years; we report unblinded 2-year data. Results: Of the 8108 subjects in the efficacy population (subjects with no PCa in baseline biopsy as assessed by Central lab and who took at least one dose of study medication), 356 subjects (4.4%) underwent an unscheduled biopsy (0–18 months), and 6460 subjects (79.7%) underwent a protocol-mandated (scheduled) biopsy (19–24 months). Rising PSA was the primary reason for an unscheduled biopsy. PCa was diagnosed in 60 subjects (16.9%) at an unscheduled biopsy compared with 950 subjects (14.7%) at scheduled biopsy. Subjects with PCa on unscheduled, compared with scheduled, biopsy had higher baseline PSA (6.63 vs. 6.06 ng/mL), lower % free PSA (13.2% vs. 15.5%); there was no difference in the mean number of baseline biopsy cores (8.6 vs. 8.4). Conclusions: Although unscheduled biopsies in REDUCE were slightly more likely to result in a cancer diagnosis and have a higher Gleason score, the small number of such biopsies (356/6591, 5.4% of subjects biopsied) would not cause a significant ascertainment bias. This contrasts the prostate cancer prevention trial (PCPT), where close to 40% of subjects had for-cause biopsies (most often PSA-driven), creating a substantial detection bias for high-grade cancers. [Table: see text] [Table: see text] ASCO Conflict of Interest Policy and Exceptions In compliance with the guidelines established by the ASCO Conflict of Interest Policy (J Clin Oncol. 2006 Jan 20;24[3]:519–521) and the Accreditation Council for Continuing Medical Education (ACCME), ASCO strives to promote balance, independence, objectivity, and scientific rigor through disclosure of financial and other interests, and identification and management of potential conflicts. According to the ASCO Conflict of Interest Policy, the following financial and other relationships must be disclosed: employment or leadership position, consultant or advisory role, stock ownership, honoraria, research funding, expert testimony, and other remuneration (J Clin Oncol. 2006 Jan 20;24[3]:520). The ASCO Conflict of Interest Policy disclosure requirements apply to all authors who submit abstracts to the Annual Meeting. For clinical trials that began accrual on or after April 29, 2004, ASCO's Policy places some restrictions on the financial relationships of principal investigators (J Clin Oncol. 2006 Jan 20;24[3]:521). If a principal investigator holds any restricted relationships, his or her abstract will be ineligible for placement in the 2009 Annual Meeting unless the ASCO Ethics Committee grants an exception. Among the circumstances that might justify an exception are that the principal investigator (1) is a widely acknowledged expert in a particular therapeutic area; (2) is the inventor of a unique technology or treatment being evaluated in the clinical trial; or (3) is involved in international clinical oncology research and has acted consistently with recognized international standards of ethics in the conduct of clinical research. NIH-sponsored trials are exempt from the Policy restrictions. Abstracts for which authors requested and have been granted an exception in accordance with ASCO's Policy are designated with a caret symbol (^) in the Annual Meeting Proceedings. For more information about the ASCO Conflict of Interest Policy and the exceptions process, please visit www.asco.org/conflictofinterest .

Carotid Endarterectomy: Technical Practices of Surgeons Participating in the GALA Trial

ObjectiveRecent meta-analyses confirm an advantage to patch angioplasty during carotid endarterectomy (CEA) and suggest a benefit from routine shunting. GALA Trial (RCT: general [GA] versus local [LA] anaesthesia for CEA) collaborators (non-UK [European] and UK) were surveyed to assess current practice techniques. Materials and MethodsPostal questionnaires determined: shunt usage, monitoring techniques dictating shunt deployment, criteria for patching and the influence of anaesthetic technique upon these decisions. Results157/216 surgeons (73%) replied. For UK surgeons (n=76) performing GA CEA a shunt was always, never, or selectively used by 73.6%, 4.2% and 22.2% respectively. Figures for non-UK surgeons (n=77) were 20.8% (p<0.0001), 26% (p<0.0002) and 53.2% (p<0.0001). When shunting selectively, fewer UK surgeons relied on stump pressure (26.4% v 48.1%; p<0.0064) with TCD more widely used (38.9% v 11.7%; p<0.0001). Shunting criteria during LA CEA were the same for both groups (impaired awake-testing). Routine patching was commoner amongst UK surgeons (GA: 76.4% v 34.2%, p<0.0001; LA: 70.1% v 31.9%, p<0.0001). ConclusionsThese results indicate that more UK surgeons have adopted current suggestions for improving CEA outcomes. Future analysis of unblinded GALA Trial data may provide further information about the impact of different policies for shunting and patching.

First limits on WIMP dark matter from the XENON10 experiment

. XENON10 is a dual phase liquid/gas xenon Time Projection Chamber (TPC) for the direct search of Weakly Interacting Massive Particle (WIMP) Dark Matter (DM). The detector, located in the Gran Sasso National Laboratory in Italy, aims at measuring the recoil energy from elastic WIMP-nucleus scattering. It distinguishes electronic interactions (gamma and beta backgrounds) from nuclear recoils based on a different ratio in yields of ionization charge and scintillation light. This discrimination technique reduces background by a factor of 200 – 1000 in the energy range of interest, at a nuclear recoil acceptance of 50%. Background is further reduced by the self-shielding properties of liquid xenon, as the 3D position resolution of the TPC is used to define an inner fiducial volume of low background. XENON10 was extensively calibrated with gamma and neutron sources. A blind search, with data cuts fixed a priori based on calibration data and 16 days of unblinded background data, had an exposure of 136 kg days at a fiducial mass of 5.4 kg. We report our first upper limit on WIMP DM interactions of 8.8 × 10-44 cm2 for a WIMP mass of 100 GeV/c2, which is a factor of 2.3 below the best previous limit.

The statistical analysis of a clinical trial when a protocol amendment changed the inclusion criteria

BackgroundSometimes, protocol amendments that change the inclusion and exclusion criteria are required in clinical trials. Then, the patient populations before and after the amendment may differ.MethodsWe propose to perform separate statistical tests for the different phases, i.e. for the patients recruited before and after the amendment, and to combine the tests using Fisher's combination test. After a significant combination test a multiple testing procedure can be applied to identify the phase(s) to which a proof of efficacy refers. We assume that the amendment(s) are not based on any type of unblinded data. The proposed method is investigated within a simulation study.ResultsThe proposed combination approach is superior to the 'naïve' strategy to ignore the differences between the phases and pooling the data to perform just one statistical test. This superiority disappears when there are hardly any differences between the two phases.ConclusionWhen one or more protocol amendments change the inclusion and exclusion criteria, one should realize that the populations may differ. In this case, separate tests for the different phases together with a combination test are a powerful method that can be applied in a variety of settings. The (first) amendment should specify the combination test to be applied in order to combine the different phases.

Muscle Tenderness in Men With Chronic Prostatitis/Chronic Pelvic Pain Syndrome: The Chronic Prostatitis Cohort Study

Myofascial pain is a possible etiology for category III chronic prostatitis/chronic pelvic pain syndrome, either secondary to infection/inflammation or as the primary cause. We documented tenderness on physical examination in a large multicenter cohort of patients with chronic pelvic pain syndrome and compared to controls. Data were reviewed from the National Institutes of Health Chronic Prostatitis Cohort study on 384 men with chronic pelvic pain syndrome and 121 asymptomatic controls who had complete unblinded physical examination data from 7 clinical centers between October 1998 and August 2001. Tenderness in 11 sites including prostate, genitals, abdomen and pelvic floor together with prostate size and consistency was evaluated. Data were correlated with cultures and symptoms. Overall 51% of patients with chronic pelvic pain syndrome and 7% of controls had any tenderness. The most common site was prostate (41% chronic pelvic pain syndrome, 5% controls), followed by external and internal pelvic floor (13% and 14% chronic pelvic pain syndrome, 0 controls) and suprapubic area (9% chronic pelvic pain syndrome, 0 controls). Of patients with chronic pelvic pain syndrome 25% had 1 tender site, 11% had 2 and 6% had 3 tender sites. Tenderness did not correlate with inflammation or infection in the prostate fluid. Prostate consistency was normal in 79% of patients with chronic pelvic pain syndrome and in 95% of controls, and did not correlate with symptom severity. Patients with chronic pelvic pain syndrome with any tenderness had significantly higher Chronic Prostatitis Symptom Index scores at baseline and at 1 year (24.1 vs 21.2 and 20.2 vs 17.5, p <0.0001) compared to patients without tenderness. Abdominal/pelvic tenderness is present in half of the patients with chronic pelvic pain syndrome but only 7% of controls. Extraprostatic tenderness may identify a cohort of patients with a neuromuscular source of pain.

BiovaxID Vaccine Therapy of Follicular Lymphoma in First Remission: Phase III Blinded Safety Update.

Background: A previous single-arm Phase 2 study (Nature Med 5:1171–7, 1999), evaluated the ability of tumor-specific idiotype (ID) conjugated to keyhole limpet hemocyanin (KLH) administered with granulocyte-monocyte colony-stimulating factor - GM-CSF (BiovaxID vaccine) to induce complete remission (CR) and molecular remission (MR) in follicular NHL patients in first CR after chemotherapy. We reported (ASH 2005, Abstr 2441) that at 9.2 years follow-up, disease free survival (DFS) and overall survival (OS) were 45% and 95%, respectively; median DFS was 8.0 years. To date, there have been no additional reported mortalities in this cohort. Although these data are promising, improvements in the treatments and increasing survival for patients with follicular lymphoma have necessitated additional rigor in the evaluation of overall safety and overall risk-benefit. The ongoing Phase 3 study that opened in 2000 by the NCI is designed to evaluate the impact of this vaccine on DFS in previously untreated subjects who have attained CR/Cru. As of August 2007, 231 subjects have been enrolled to receive chemotherapy and 176 have been randomized to receive vaccine (ID-KLH or KLH-KLH).Methods: Since Phase 3 trial inception in 2000, SAEs have been reported through a pharmacoviligance group and identified according to time period on the study (i.e. during chemotherapy, during vaccine, after vaccine, or unknown), causality, and relatedness to study agent.Results: As of August 2007, 32 SAEs have been reported in the BV301 trial, 7 of which occurred during the vaccination segment of the study. At a blinded interim analysis of the Phase 3 safety data, of the 7 SAEs reported during the vaccine segment of the study, 3 are considered unlikely to be related to the study agent (chest pain, compression fracture, herpes zoster), 2 possibly related (arrhythmia supraventricular, chest pain), 1 unrelated (vomiting), and 1 unknown (febrile neutropenia). The 1 SAE, cerebral ischaemia, that occurred post vaccine is considered unrelated to the study agent.Conclusions: Additional follow-up on the Phase 2 patients is needed, as well as similar analyses in other cohorts. Blinded Phase 3 safety data shows a positive safety profile for BiovaxID. Unblinded safety data will be needed to determine the final efficacy and safety profile.

Adaptive designs: Looking for a needle in the haystack—A new challenge in medical research

The statistical principles of fully adaptive designs are outlined. The options of flexibility and the price to be paid in terms of statistical properties of the test procedures are discussed. It is stressed that controlled inference after major design modifications (changing hypotheses) will include a penalty: Intersections among all the hypotheses considered throughout the trial have to be rejected before testing individual hypotheses. Moreover, feasibility in terms of integrity and persuasiveness of the results achieved after adaptations based on unblinded data is considered as the crucial issue in practice. In the second part, sample size adaptive procedures are considered testing a large number of hypotheses under constraints on total sample size as in genetic studies. The advantage of sequential procedures is sketched for the example of two-stage designs with a pilot phase for screening promising hypotheses (markers) and controlling the false discovery rate. Finally, we turn to the clinical problem how to select markers and estimate a score from limited samples, e.g. for predicting the response to therapy of a future patient. The predictive ability of such scores will be rather poor when investigating a large number of hypotheses and truly large marker effects are lacking. An obvious dilemma will show up: More optimistic selection rules may be superior if in fact effective markers exist, but will produce more nuisance prediction if no effective markers exist compared with more cautious strategies, e.g. aiming at some control of type I error probabilities.

Lapatinib (L) with paclitaxel compared to paclitaxel as first-line treatment for patients with metastatic breast cancer: A phase III randomized, double-blind study of 580 patients

1011 Background: L is an oral tyrosine kinase inhibitor of EGFR/HER2, active as monotherapy and in combination for HER2-overexpressing advanced/metastatic breast cancer (BC). A Phase I study of L with paclitaxel (P) indicated no unexpected adverse events (AEs). PK profile indicated no relationship between peak plasma concentration of P+L and neuropathy, neutropenia, diarrhea, rash or myalgia. We report here blinded efficacy and safety data for patients (pts) with incurable Stage IIIb/IIIc/IV BC at first diagnosis or relapse, untested or negative (0/1+ IHC or FISH neg) for HER2. Unblinded data will be presented at ASCO 2007. Methods: Between Jan 2004 and Jul 2005, 580 pts from 24 countries were stratified by metastatic site and randomized 1:1 to L 1500 mg QD + 175 mg/m2 P q3w or placebo QD + 175 mg/m2 P q3w. Primary endpoint was TTP; secondary endpoints were AEs, ORR, PFS, CBR, RFS, and OS. Tumor tissue was obtained from the most recent biopsy of 451 (78%) pts and was centrally analyzed in blinded fashion for biomarker patterns. Serum samples were collected for central EGFR and HER2 ECD analysis. Results and Conclusions: 579 pts were analyzed; 87% presented with Stage IV BC. 55% received prior adjuvant chemotherapy or anti-hormonal therapy. No pts received previous trastuzumab. At the time of analysis, 561 (97%) pts progressed or otherwise withdrew. Most common AEs were alopecia (58%), neurological (55%, gr=3:8%), diarrhea (42%, gr=3:9%), nausea (32%), and rash (32%, gr=3:2%). Neutropenia and thrombocytopenia AEs related to study treatment were 18% and &lt;1%, respectively. LVEF decrease of 20% relative to baseline and below LLN was reported 15 times. 12% of AEs led to treatment withdrawal. Blinded data analysis revealed a median TTP of 25 wks and ORR of 30%. CNS relapse was reported in 11 pts (2%). Enrollment predominantly came from countries with limited HER2 testing capacity thus a subgroup of pts is expected to be HER2+ve. Blinded analyses of HER2, ER and PR are ongoing at this time and final biomarker evaluations will be presented with unblinded efficacy data. [Table: see text]

Anglo-Celtic IV: First results of a UK National Cancer Research Network randomized phase III pharmacogenetic trial of weekly compared to 3 weekly paclitaxel in patients with locally advanced or metastatic breast cancer (ABC)

LBA1005 Background: Phase II studies of weekly paclitaxel (P) confirm this to be a well tolerated regimen in ABC with potentially superior efficacy to standard q3w schedules. This was tested in the phase III CALGB 9840 trial, which used a “patient resource conserving” design, combining randomised patients with imported control data from a previous CALGB study of predominantly 2nd line treatment. This and the higher total dose of P in the weekly arm, left many unsure if there really was superior activity for the weekly schedule. Methods: The primary efficacy objective was to detect a 2 month improvement (6 to 8 months, HR = 1.33) in time to progression (TTP) by the giving the same total dose of P weekly (wP) vs. 3 weekly (3wP). 560 randomised patients were required to give 90% power to detect this difference at the 0.05 significance level. A single analysis was planned six months after study closure. Secondary endpoints include overall survival, toxicity and quality of life. The primary translational endpoint is a pharmacogenetic (PG) study of the effect of MDR and CYP single nucleotide polymorphisms. Between Sept. 2002 and July 2006, 569 patients were randomized to receive either wP 90 mg/m2 for 12 weeks or 3wP 175 mg/m2 for 6 cycles. Patients were stratified for measurable disease, prior treatment for ABC and Trastuzumab use. Results: The investigator reported response rates were 27% (3wP) and 42% (wP), p=0.002. Median TTP was 22.0 weeks for 3wP (95% CI 19.7–24.6) and 23.9 weeks for wP (95% CI 20.7–26.7), HR=0.92, p=0.06. Both schedules were well tolerated. PG samples from 325 patients are being matched to the unblinded efficacy data. Conclusions: This phase III trial shows that, for a matched total dose, wP produces a higher response rate than 3wP. This confirms the result from CALGB 9840. The mismatch in treatment duration in this trial may explain why TTP with wP was superior in CALGB 9840 while in the current study we have failed to detect superiority despite a higher response rate. Combining the trial results supports the hypothesis that total dose and schedule of P are important. Our results support the widespread adoption of wP as the standard schedule. No significant financial relationships to disclose.

Flexible Designs by Adaptive Plans of Generalized Pocock‐ and O'Brien‐Fleming‐Type and by Self‐Designing Clinical Trials

Flexible designs are provided by adaptive planning of sample sizes as well as by introducing the weighted inverse normal combining method and the generalized inverse chi-square combining method in the context of conducting trials consecutively step by step. These general combining methods allow quite different weighting of sequential study parts, also in a completely adaptive way, based on full information from unblinded data in previously performed stages. So, in reviewing some basic developments of flexible designing, we consider a generalizing approach to group sequentially performed clinical trials of Pocock-type, of O'Brien-Fleming-type, and of Self-designing-type. A clinical trial may be originally planned either to show non-inferiority or superiority. The proposed flexible designs, however, allow in each interim analysis to change the planning from showing non-inferiority to showing superiority and vice versa. Several examples of clinical trials with normal and binary outcomes are worked out in detail. We demonstrate the practicable performance of the discussed approaches, confirmed in an extensive simulation study. Our flexible designing is a useful tool, provided that a priori information about parameters involved in the trial is not available or subject to uncertainty.