Virtual Library

Start Your Search

R. Stephens

Moderator of

  • +

    MS25 - Translating Research into Practice (Applied Statistics) (ID 42)

    • Event: WCLC 2013
    • Type: Mini Symposia
    • Track: Statistics
    • Presentations: 4
    • +

      MS25.1 - Defining the Gold Standard: PFS v OS (ID 579)

      14:05 - 14:25  |  Author(s): K. Hotta

      • Abstract
      • Presentation
      • Slides

      Abstract
      Approval of most new agents for advanced non-small-cell lung cancer (NSCLC) has been based on prolongation of OS, representing a direct measure of clinical benefit, in randomized clinical trials. Recently, however, as we have more opportunity to obtain trial results showing “significant improvement in PFS without any OS benefit”, most investigators feel that the assessment of OS may become of limited use and that PFS must be surrogate endpoint for assessing the efficacy of an experimental agent in NSCLC. Originally, PFS is quite different from OS since it is subjective, but not a direct measure of clinical benefit. Progression is often asymptomatic, and it is not always clinically relevant. To use PFS as surrogate endpoint, its validity should be statistically evaluated. In the pooled analysis using individual patient data of randomized trials comparing first-line docetaxel with vinca-alkaloids, Buyse et al. showed relatively weak PFS-OS correlation [BMJ open 2013;3:e001802]. So far, there seem no other large-scaled validation studies investigating the surrogacy of PFS for OS endpoint with formal statistical approach. Thus, PFS has not yet been a statistically acceptable surrogate endpoint for OS in patients with metastatic NSCLC. Now, what could we interpret “significant improvement in PFS without any OS benefit”? Also, what would be the clinically meaningful endpoint in advanced NSCLC? The literature-based study was conducted to assess the PFS-OS relationship with the data of the phase III trials investigating molecular-targeted agents in advanced NSCLC [Lung Cancer 2013;79:20]. It showed a strong PFS-OS correlation only in trials where subsequent therapy was conducted less frequently, whilst trials with a higher proportion of use of subsequent therapy had weaker association. The study concluded that 1) the relationship between PFS and OS is originally strong enough to support potential surrogacy of PFS, but that 2) given the observation that increasing use of effective salvage therapies could affect the PFS-OS association, improvement in PFS without any OS benefit does not mean the experimental agent fails to have a true clinical benefit. That is, a potentially true OS benefit by an experimental agent would have been seen without any confounding if no subsequent therapy had been given. This theory is called “explanatory approach”, supporting the use of PFS as a clinically meaningful outcome. Rather, there is an opposite opinion, suggesting the observed difference in OS would be considered the measure of clinical benefit, regardless of subsequent therapies, provided that they follow the current standard of care. This is the pragmatic approach [JCO 2011;29:2439]. The observed difference in OS, of course, will be expected smaller by the subsequent therapy than one would see if it was not available, leading to the need of a large sample size to detect such differences. But, according to the pragmatic approach, such potentially attenuated but actually observed OS difference should be used for assessing the true clinical benefit of the experimental agent in the given clinical setting under reflecting the clinical reality of available subsequent treatments. Recently, ASCO discussed what would be clinical meaningful outcome in advanced NSCLC without EGFR or EML4-ALK mutations, and provided draft recommendations as follows: 1) survival after first line therapy is relatively short, and OS is a feasible endpoint although the effects of the experimental agent on OS can be clouded by treatments administered after the period of therapy, and 2) clinical trials should aim to improve OS by a minimum of 25% as compared with standard therapy. ASCO seems to support the importance of measuring OS rather PFS, in favor of the pragmatic approach. Even in the first-line metastatic colorectal cancer, though PFS has already been established as a surrogate for OS in the 90’s, its surrogacy was reappraised using individual patient data between 1997 and 2006 because of the advance in the treatment during the last decade and recent improvement of OS [JCO 2013 (suppl;abstr 3533)]. Surprisingly, the PFS-OS relationship was not as strong as was seen in the 90’s. The study concluded that in modern metastatic colorectal cancer trials where SPP exceeds time to first progression, the treatment effects on PFS do not reliably predict those on OS. It seems that even though PFS was once accepted as a suitable surrogate for OS, its validity should be assessed repeatedly along with the advances especially in the post-progression treatment, which also supports the concept of the pragmatic approach. Regarding the regulatory considerations, the FDA stated OS should be considered the standard clinical benefit endpoint and that it should be used to establish the efficacy of a treatment in metastatic NSCLC, although it can also consider PFS for regulatory decision of drug approval based on the population. Some of the abovementioned findings potentially support the rationale of the pragmatic approach, stressing the importance of a possibly attenuated, but actually observed OS difference. However, this theory is unlikely to be applied in more recent trials investigating specific targeted therapies including EGFR-TKIs or ALK inhibitors, because no OS difference would be arguably obtained due to a quite high level of crossover inevitably for the ethical reason. Thus, under the special situation where i) an experimental agent has theoretically been considered to target specific molecules, and ii) the earlier trials showed dramatic effect, this possibly high proportion of crossover can compromise the ability to assess clinical benefit, and also lead to the unrealistic sample size to detect significant OS difference. Rather, in this condition, whether both large magnitude of PFS advantage and great OS advantage compared with historical control can be obtained would be more important than the conventional assessment of OS difference between the arms. In conclusion, the goals of any new cancer treatment are to offer patients a true clinical benefit. PFS has not yet been formally accepted as a valid surrogate for the OS endpoint in advanced NSCLC, and its surrogacy will be addressed by each drug mechanism of action and patient population. Finally, OS remains the primary endpoint of clinical trials, except in a situation where agents targeting specifically driver oncogenes are being evaluated with anticipation of high level of crossover.

      Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.

    • +

      MS25.2 - Statistical Requirements for Screening Trials (ID 580)

      14:25 - 14:45  |  Author(s): L. Billingham

      • Abstract
      • Presentation
      • Slides

      Abstract
      A screening programme is a set of procedures that can be applied to an asymptomatic population to enable early detection and treatment of disease. The success of the programme is dependent on this early intervention reducing morbidity and mortality associated with the disease. There has been much research to develop and assess potential screening programmes for lung cancer. In particular, there are a number of major clinical trials such as the National Lung Screening Trial in the United States, the United Kingdom Lung Cancer Screening Trial and the Dutch-Belgian NELSON trial, that assess the effectiveness of screening a high risk population with low dose computed tomography scans to reduce mortality from lung cancer. The key elements of any screening programme are: (i) identification of the target population for screening, (ii) the screening test that will be used to classify patients as likely or unlikely to have the disease, (iii) the frequency of applying the screening test, (iv) the diagnostic test that will be used to determine whether people are truly diseased or not, (v) the treatment that will be available for those diagnosed early. These choices will impact on the statistical design. The research question may be whether to introduce screening, to determine the frequency of screening, to identify the appropriate target population for screening or whether to add an additional screening tool to an existing screening modality. Randomised controlled trials are the gold standard for assessing a screening programme as they overcome major biases specific to screening namely lead-time bias, length bias, over-diagnosis bias and selection bias. The usual trial design parameters are important but special statistical issues arise in relation to screening trials. Statistical inferences should only be made back to the eligible population defined for the trial so the choice of eligibility criteria is important. Screening interventions can cause harm to individuals that are potentially disease-free so the target population is usually those who are at high risk of disease. Interventions such as CT scans which have both a relatively high risk and high level of harm should be targeted at a high risk population whilst those that are low risk to the individual such as sputum cytology could be targeted at a wider population. Harms also include the inconvenience and psychological effects of a false positive result. A high risk population is usually defined in terms of pack-years of smoking and time since quitting and simple patient characteristics such as age. Statistical models that predict risk of disease could enhance the identification of a high risk population but the accuracy of prediction should be considered in relation to its impact on the trial. The primary outcome measure to assess benefit of a lung cancer screening programme should be lung cancer specific mortality. Duration of follow-up is an important design parameter. It needs to be long enough to allow for the time lag before the impact of screening becomes apparent and not so long after the cessation of screening as to include a period of time when screening would have lost its impact. Appropriate statistical analysis of the primary outcome measure is essential to properly evaluate the benefits of screening. Typically the screening intervention will be compared to the control arm in terms of its ability to reduce cumulative mortality. If this is calculated over the entire period of screening and follow-up then the benefit may be underestimated. Comparing time-specific mortality rates is the recommended approach [1]. The analysis is also complicated by the problems of non-attendance and contamination but methods have been proposed to adjust for these [2]. Sample size calculations involve key design parameters including hypothesised mortality reductions, expected compliance and contamination, number of screening rounds and length of follow-up [3]. The accuracy of the screening test to predict disease in asymptomatic people is not only important for the feasibility and ethics but will also impact on sample size as inaccurate predictions will dilute the potential benefit of screening. Screening tests, such as those that involve biomarkers measured on a continuous measurement scale, should be rigorously developed and validated before assessing their clinical utility within a randomised controlled trial environment. References [1] Hanley JA; Measuring mortality reductions in cancer screening trials; Epidemiologic Reviews 2011; 33: 36-45. [2] Baker SG, Kramer BS, Prorok PC; Statistical issues in randomised trials of cancer screening; BMC Medical Research Methodology 2002; 2: 11. [3] Prorok PC, Marcus PM; Cancer screening trials: nuts and bolts; Seminars in Oncology 2010; 37(3): 216-223.

      Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.

    • +

      MS25.3 - Cost-Effectiveness of Modern Therapies (ID 581)

      14:45 - 15:05  |  Author(s): N. Leighl

      • Abstract
      • Presentation
      • Slides

      Abstract
      The growing number and rising costs of modern lung cancer therapies have brought the issue of cost and cost effectiveness to the forefront of clinical practice. While personalized medicine improves outcomes in specific subgroups, sparing others from ineffective costly treatment and toxicity, it can be challenging to incorporate into economic analyses. Defining the target population for the new treatment is key, and then evaluating the costs and benefits of the new intervention compared to the previous standard. However, the definition of molecular populations for targeted therapies has emerged as an important consideration when considering whether or not to adopt a new targeted therapy. The cost of biomarker testing can have a major impact on healthcare costs, and many countries are struggling with how to best incorporate the "hidden" costs of personalized medicine into adopting new targeted therapies. Focusing only on the target population, comparing the new treatment with standard comparators does not incorporate the costs of biomarker testing, or need for repeat biopsies for successful testing, but will be a better reflection of the benefit of the new treatment in that population. Comparing a test-and-treat strategy to a strategy without testing or the new therapy allows incorporation of the costs of testing, but has some important challenges. Biomarker frequency is a key driver in these analyses, with smaller populations as a particular challenge, such as ALK positive nonsmall cell lung cancer. Presenting the cost of both a test-and-treat strategy alongside an evaluation of the cost effectiveness of therapy in the target population may be a better way to illustrate the impact of a novel treatment, especially when the target population is small, while acknowledging the incremental financial burden of biomarker testing in cancer. This may allow new therapies to compete with current alternatives on a comparable footing, and not underestimate the impact of a new treatment in a small subgroup. It would also permit the development of more effective and cost-efficient screening methods for the desired target population. It is important to recall that technological methods and costs involved in biomarker testing and molecular analysis are rapidly changing, and should be revisited over time. The technological methods used to identify molecular abnormalities in cancer, such as sequencing and antibody development, are changing rapidly along with associated costs. Thus less expensive or more efficient methods (e.g. multiplex testing) may be more affordable compared to more labor-intensive methods used in initial clinical trials. For example, immunohistochemical techniques may replace more expensive methods, allowing more jurisdictions to take up testing and treatment of new therapies.

      Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.

    • +

      MS25.4 - Assessing New Treatments for Lung Cancer - Regulatory and Cost-Effectiveness Implications (ID 582)

      15:05 - 15:25  |  Author(s): S.R. Hill

      • Abstract
      • Presentation
      • Slides

      Abstract
      Over the last several years, there have been a number of new classes of drugs approved for the management of lung cancer. Regulatory pathways are evolving in many ways – for example, to recognize that targeted therapies, such as the EGFR inhibitors, will require evaluation of the drug alongside a diagnostic test; or that the increasing specificity of targets may result in smaller and shorter trials with different endpoints; or that regulatory authorities might identify potential significant advances in therapy by the use of special evaluation pathways, such as the ‘breakthrough’ therapy designation from the US FDA. Sometimes the trials now also measure quality of life outcomes or patient preferences. But these trials are still fundamentally designed to ask the ‘regulatory’ question – can the new product work – as well as assessing the risk-benefit ratio. But where does this changing regulatory environment leave payers? The information set that is available to guide assessments of value of money generally seems to becomes more limited as the regulators increase the pace of their decision-making. Trials are truncated or treatment groups are crossed over to the new treatment at the earliest possible opportunity, often before there is a confident estimate of the effect size in terms of final clinical outcomes. The ethical imperative to offer access to possibly effective treatments outweighs ensuring adequate trial design to be confident in the estimate of effect. Statistical techniques to ‘adjust’ for limitations in design become more and more complex, and more prone to uncertainty. And at the same time, most countries are still struggling with faltering economies and diminishing health budgets. Patients still want access to the latest treatments, but are less and less willing to pay increasing prices. Payers then have to compare the limited information set available for new drugs with what is known about current treatments. Arguments over the value of differences such as a 1.4 months gain in overall survival compared to existing treatment become conflated with the cost of this gain. That is assuming, of course, that there is a gain in overall survival, which has not been often shown to date in the trials of the new drugs for lung cancer. When high prices and high prevalence are combined, the value-for-money question is rightly raised – why pay so much more for so much uncertainty? Arguments about incremental advances in therapy, innovation and technological developments then dominate the discussion, rather than the appropriate focus on health gain for communities and individuals. So what is the solution? Options suggested include ‘managed entry’ of new products with additional data collection as a condition of price negotiation, ‘paying for performance’ with outcome data collection, or ‘value-based pricing’ that allows a premium for ‘innovation’. However, given the high clinical need for effective treatments, the emphasis should be on getting the best health outcomes for an affordable price, to the community and for individuals.

      Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.



Author of

  • +

    MS13 - Statistics of Personalised Medicine (ID 30)

    • Event: WCLC 2013
    • Type: Mini Symposia
    • Track: Statistics
    • Presentations: 1
    • +

      MS13.1 - A Review of Randomised Trials in Lung Cancer (ID 516)

      14:05 - 14:25  |  Author(s): R. Stephens

      • Abstract
      • Presentation
      • Slides

      Abstract
      There is a long and distinguished history of clinical trials investigating the treatment of lung cancer dating back to the 1950’s. One of the first such trials to be published looked at the use of nitrogen mustard and DON (6-diazo-5-oxo-L-norleucine) in patients with various cancers including bronchogenic carcinoma[1], and probably the first published randomized trial specifically for lung cancer patients compared surgery and supervoltage radiotherapy[2]. Continued progress in treatment relies on randomized clinical trials and meta-analyses, and a review was undertaken in 2003 to assess the quality and quantity of all published randomized lung cancer treatment trials[3]. This showed that, of nearly 1000 trials, only 4 accrued more than 1000 patients, and that clinicians (and patients) could only rely on a few meta-analyses to guide treatment decision making. The review called for greater global collaboration to design and run large trials to improve the survival rates and quality of life of lung cancer patients. Ten years on, an update and extension of the 2003 review has been undertaken, to see what changes have occurred, what patterns have emerged, and whether the data will help the design of future trials. The Cochrane Library was used as the e-library, as it contains some abstracts from major meetings, as Song et al[4] estimate that up to 50% of all trials are not fully published. In addition to sample size, information was collected regarding the histological subgroup studied, the treatment modality investigated, the trial design, the outcome and the country of affiliation of the first author. A search of ‘(lung OR bronchus) AND (cancer OR carcinoma)’ in Sept 2012 produced 7130 reports (5218 full papers and 1912 meeting abstracts) classified by the Cochrane Library as ‘trials’. However, as the inclusion of meeting abstracts was found to be inconsistent, the results presented here are based only on the full publications. A total of 1792 randomized clinical trials of lung cancer were identified, of which 1677 included the sample size in the title or abstract, 49 did not, and in the remaining 66 the abstract could be accessed. An analysis of the 1677 trials with sample size stated showed that: There was an increasing number of randomized trials, mainly due to an increasing number of randomized phase II trials[5] Only 20 trials included >1000 patients, and 48% had less than 100. The median sample size has remained unchanged for 40 years (~100 patients) although the median sample size for those designated as randomized phase III trials has increased from ~200 to ~400 patients over the last 10 years In nearly half of the trials the primary research question addressed chemotherapy, 15% supportive care, 8% radiotherapy and only 1% surgery Of those that indicated the histology, trials of NSCLC cancer accounted for 73% The country producing the most trials (397) was the USA, but China is now the second country, and published the most trials over the last 10 years. Although there are limitations of this review (it is necessarily several years out of date, is only looking at the title and abstract of trials on the Cochrane Library, is complicated by multiple reporting of trials, etc), its strengths are the longitudinal overview and comprehensive inclusion criteria which substantially extend the results reported by Subramanian et al[6] who compared ongoing trials of medical treatment for NSCLC in 2012 with those in 2009. The need to improve the quantity and quality of trials in lung cancer has been highlighted for more than 20 years[7-10 ]as trials have not been sufficiently large, or poorly designed, or unnecessarily duplicated previous work[11-12]. In contrast, well designed trials make best use of researchers’ time, funders’ money and patients’ goodwill[13], and may inform future work. The results of the current review suggest that greater global collaboration is still required to run large trials that will produce reliable results and influence practice worldwide. Updated and more detailed results will be presented. References 1. Krantz S, et al for the Veterans Administration Cancer Chemotherapy Study Group. A Clinical Study of the Comparative Effect of Nitrogen Mustard and DON in Patients With Bronchogenic Carcinoma, Hodgkin's Disease, Lymphosarcoma, and Melanoma. JNCI 1959, 22, 433-9 2. Morrison R. The treatment of carcinoma of the bronchus. A clinical trial to compare surgery and supervoltage radiotherapy. Lancet 1963, 1, 683-4 3. Stephens R. The need for a world strategy for clinical trials. Lung Ca 2003, 41 (suppl 3), S96 (abs E-77) 4. Song F, et al. Dissemination and publication of research findings: an updated review of related biases. Health technology Assessment 2010, 14, No. 8 5. Turrisi AT. Creeping phase II-ism and the Medical Pharmaceutical Complex: Weapons of Mass Distraction in the War against Lung Cancer. JCO 2005, 23, 4827-9 6. Subramanian J, et al. Review of Ongoing Clinical Trials in NSCLC - a status report from the CinicalTrials.gov Web Site. J Thorac Oncol 2013, 8, 860-5 7. Nicolucci A, et al. Quality, evolution, and clinical implications of randomized, controlled trials on the treatment of lung cancer. A lost opportunity for meta-analysis. JAMA 1989, 262, 2101-7 8. Brundage MD, Mackillop WJ. Locally advanced non-small cell lung cancer: Do we know the questions? A survey of randomized trials from 1966-1993. J Clin Epidemiol 1996, 49, 183-92 9. Breathnach OS, et al. Twenty-two Years of Phase III trials for Patients With Advanced Non-Small Cell Lung Cancer: Sobering Results. J Clin Oncol 2001, 19, 1734-42 10. Macbeth F, et al. An open letter to all members of the IASLC. Lung Cancer 2004, 45, 119-20 11. Chalmers I. The lethal consequences of failing to make full use of all relevant evidence about the effects of medical treatments: the importance of systematic reviews. Rothwell P, ed. Treating individuals: from randomised trials to personalised medicine. London: Lancet, 2007: 37-58. 12. Young C, Horton R. Putting clinical trials into context. Lancet 2005, 366, 107-8 13. Chalmers I, Glasziou P. Avoidable waste in the production and reporting of research evidence. Lancet 2009, 374, 86-9

      Only Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login, select "Add to Cart" and proceed to checkout. If you would like to become a member of IASLC, please click here.

      Only Active Members that have purchased this event or have registered via an access code will be able to view this content. To view this presentation, please login or select "Add to Cart" and proceed to checkout.