Please activate JavaScript!
Please install Adobe Flash Player, click here for download


ISCB 2014 Vienna, Austria • Abstracts - Poster Presentations 93Monday, 25th August 2014 • 15:30-16:00 Monday25thAugustTuesday26thAugustThursday28thAugustAuthorIndexPostersWednesday27thAugustSunday24thAugust Med Res Methodol 2013) a bound expressed in terms of the observed test statistic at the interim is exhibited : If z ≥ b(q, V), for a certain function b where q = n/(N0 +r) and V=(N0+r)/N0, then one may use the traditional threshold zα at the final analysis without compromising the type I error rate. The current work proves these results to be equivalent. Also, further de- tails regarding the asymptotics of the bound b are derived.   P1.2.33 Use of an adaptive approach to design and evaluation of multi-regional clinical trials C Chiang1 , J-P Liu1 , C-F Hsiao2 1 National Taiwan University, Taipei, Republic of China (Taiwan), 2 National Health Research Institutes, Zhunan, Republic of China (Taiwan)   In recent years, global collaboration has become a conventional strategy for new drug development. To accelerate the development process and shorten approval time, the design of multi-regional clinical trials (MRCTs) incorporates subjects from many countries/regions around the world un- der the same protocol. At the design stage of a multi-regional clinical trial, a common treatment effect and an equal variability of the primary end- point across regions are usually assumed. However, at any data monitoring point of a MRCT, we may discover re- gional differences. In this case, modification of the planned design will be inevitable. In this paper, we develop an adaptive selection design if we discover regional differences at an interim analysis in a MRCT. More spe- cifically, we will stop recruiting subjects in regions without expected treat- ment differences. Statistical method on sample size modification will also be developed.   P1.2.35 Appropriate or inappropriate use of REMARK guidelines C Ciria1 , S Love1 , C Roberts1 , M Shanyinde1 , D Altman1 1 University of Oxford, Oxford, United Kingdom   Background: Appropriate reporting is critical for the interpretability and application of clinical research. REMARK is a guideline of reporting recom- mendations for tumour marker prognostic studies that encourages com- plete reporting of relevant information. Objective: Evaluate the appropriateness of citations of REMARK in pub- lished articles. Methods: We searched the Web of Science database in September 2013 and identified 878 articles that cited REMARK. A random sample of 100 articles was selected for review. Each article was read by two reviewers with professional experience in medical research. The citation of REMARK was classified as appropriate reporting or conduct, or inappropriate. Appropriate reporting was considered when REMARK was used as a re- porting guideline, appropriate conduct when REMARK criteria were fol- lowed throughout the study, and inappropriate use when the indication was neither of these. Results: In the random sample of 100 articles, most were oncology re- search (94%) of prognostic biomarkers (45%) reporting on a single dataset of patients (71%). The use of REMARK was considered appropriate in 67% of the articles (52% reporting; 15% conduct) and inappropriate in 33%. Appropriate uses included REMARK as a reporting guideline (38%), a tool to conduct the study (13%), and importance of reporting guidelines (7%). The most common inappropriate use was as a methodology guideline (28%). Conclusion: REMARK is commonly used as a guideline of reporting rec- ommendations. However, inappropriate use as a methodology guideline is also frequently observed with potential negative consequences over the design of the study. P1.2.47 Using longitudinal toxicity score to detect time trend in dose-finding trials. Application to 19 phase I studies M Ezzalfani1 , M Marous1 , L Collette2 , E Rizzo2 , X Paoletti1 1 Institut Curie / U900, Paris, France, 2 EORTC Headquarters, Brussels, Belgium   The risk of late and cumulative toxicities is a concern with molecularly tar- geted agents in oncology dose-finding trials, because of their long-term administration.The power to detect time trend in toxicity has been shown low using binary or ordinal endpoints. We propose a new approach to detect time trend using a quasi-continu- ous toxicity score endpoint which is defined as the Euclidean norm of the grades of toxicities experienced by a patient at each cycle (C). The asso- ciation between the toxicity score and the binary endpoint dose-limiting toxicity (DLT) was studied using phase-I data. A mixed model is being de- veloped to detect time trend with the toxicity score. This model will be compared with proportional odds mixed-effect models (POMM) for ordi- nal endpoint with longitudinal measurements. Eight different molecules were investigated for the first 6 cycles. 536 pa- tients were treated (1387 cycles, median cycle was 2). 5742 toxicities were observed during C1, 8% grade3-4 after C1.The mean toxicity score was 0.068 (min=0; max=0.31). The median and the interquartile range were 0.056 and 0.066, respectively. At cycle-1, the mean score (95%CI) of pa- tients with DLT was significantly different from that of patients with no DLT: 0.170 (0.147, 0.194) versus 0.061 (0.057-0.065). A stratified analysis on cycle showed a significant association (p-value<0.0001) between the mean score and grades3-4 group: 0.126 (0.122-0.131) versus 0.050 (0.048- 0.052). A significant time effect was found with POMM for two studies (p- value<0.05). Using repeated toxicity score data in dose-finding trials is promising to investigate cumulative effects.   P1.2.48 Standardizing safety analyses for clinical trials: a story of success in the making M Fayaz1 , M Nilsson2 1 Tehran University of Medical Sciences, Tehran, Islamic Republic of Iran, 2 Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, United States   Industry standards have evolved over time for data collection (CDASH), observed data (SDTM), and analysis datasets (ADaM). Although substan- tial progress has been made, additional standardization can improve product development. Development of standard tables and figures with associated analyses will lead to improved product life-cycle evaluation by ensuring reviewers receive the desired analyses for the evaluation of pa- tient safety. More importantly, having an organized process for shared learning of improved methodologies can lead to earlier safety signal detection and better characterization of the safety profile of our products. A cross- industry working group (the PhUSE Computational Science Symposium Development of Standard Scripts for Analysis and Reporting Working Group) is providing recommendations for analyses, tables, and figures for data that is common across therapeutic areas (laboratory measurements, vital signs, electrocardiograms, adverse events, demographics, medica- tions, disposition, hepatotoxicity, pharmacokinetics). A Script Repository has been created to contain associated code for the recommended analy- ses and displays. This poster will provide an update of this effort, and instructions for how to participate in the development and review process. We encourage all medical researchers to adopt this collaborative culture change!

Pages Overview