Summary of ASA BIOP Section’s Discussion on Statistical Design Considerations in Estimating Contribution of Each Sequential Treatment Effect to the Overall Effect of a Sequence of Treatments in RCTs
Rajeshwari Sridhara (FDA), Olga Marchenko (Bayer), Qi Jiang (Pfizer), Elizabeth Barksdale (LUNGevity Foundation), Yiyi Chen (Pfizer), Marc Theoret (FDA)
On April 23rd, 2024, the American Statistical Association (ASA) Biopharmaceutical Section (BIOP) and LUNGevity Foundation hosted a virtual forum to discuss Statistical Design Considerations in Estimating Contribution of Each Sequential Treatment Effect to the Overall Effect of a Sequence of Treatments in Randomized Cancer Clinical Trials. This forum was part of a series conducted under the guidance of the U.S. FDA Oncology Center of Excellence’s Project SignifiCanT (Statistics in Cancer Trials). The goal of Project SignifiCanT is to advance cancer drug development through collaboration and engagement among various stakeholders in the design and analysis of cancer clinical trials. The discussion was organized jointly by the ASA BIOP Statistical Methods in Oncology Scientific Working Group, the FDA Oncology Center of Excellence (OCE), and LUNGevity Foundation.
Anti-cancer therapies, or regimens, increasingly comprise sequences of treatments, such as (1) neoadjuvant therapy, surgery, and adjuvant treatment in early-stage solid tumors; or (2) induction therapy, consolidation (with or without hematopoietic stem cell transplant), and maintenance therapy in some hematological malignancies. In clinical trials evaluating sequential treatments, patients are typically randomized at the start and the treatment effect is assessed as a treatment policy encompassing all sequential treatments without re-randomization. However, this approach does not enable evaluation of the contribution of each phase of the sequence to the overall combined effect. Re-randomization at initiation of each phase can help assess their contribution to treatment effect, but may add complexity to the trial design, conduct, and analyses due to the diminishing number of eligible patients for subsequent randomizations. Recent advances in innovative clinical trial designs, such as Sequential Multi-arm Randomized Trials (SMART), allow for evaluation of a sequence of treatments and are currently being used in oncologic and non-oncologic diseases. This open forum discussion among multi-disciplinary experts examined the potential use of SMART and other innovative clinical trial designs to better understand the contribution of each phase of sequential treatments to the overall treatment effect in cancer trials.
The speakers/panelists* for the discussion included members of the BIOP Statistical Methods in Oncology Scientific Working Group representing pharmaceutical companies; representatives from international regulatory agencies (Food and Drug Administration (FDA), Health Canada (HC), Medicines and Healthcare products Regulatory Agency (MHRA), and Therapeutic Goods Administration (TGA)); clinicians; academicians; and expert statisticians. In addition, over 100 participants attended the virtual meeting, including representatives from other international regulatory agencies (Brazilian Health Regulatory Agency (ANVISA), Health Sciences Authority (HAS), European Medicines Agency (EMA), Singapore; Ministry of Health, Israel; Pharmaceuticals and Medical Devices Agency (PMDA), Japan). The discussions were moderated by the BIOP Statistical Methods in Oncology Scientific Working Group co-chairs, Dr. Olga Marchenko from Bayer and Dr. Qi Jiang from Pfizer; Dr. Elizabeth Barksdale from LUNGevity Foundation; and Dr. Rajeshwari Sridhara, consultant from OCE, FDA.
In the introductory presentation, the OCE leadership discussed two examples of planned sequential treatments in oncology trials. The first was a trial for Rituxan in diffuse large B cell lymphoma (DLBCL), where patients were randomized to standard of care in the first phase, and responders from both arms were re-randomized to receive either Rituxan or placebo in the second phase. In the second example, patients with resectable non-small cell lung cancer (NSCLC) were randomized to platinum-based chemotherapy with or without an immune checkpoint inhibitor (ICI) prior to surgery. After surgery, patients in the experimental arm (i.e., chemotherapy plus ICI) continued to receive the immune checkpoint inhibitor for up to one year while those on the control arm received placebo. The concern highlighted by these examples is whether all treatments in a sequence are necessary to achieve the overall treatment effect, and how best to design or analyze trials to address this question
The speaker, from academia, provided an overview of the SMART approach: utilizing a multi-stage randomized design that allows for the evaluation of treatment effects of sequential treatments and their components, and leads to the development of dynamic treatment regimens (DTRs). These trials are helpful when there may be delayed, prescriptive, or sample selection effects. SMARTs involve at least two randomizations of all trial participants at critical decision points and provide robust evidence for effective DTRs: evidence-based guidelines for clinical practice that account for ongoing treatment decisions based on factors such as progress, side effects, and patient preferences. The speaker also discussed design and analysis of SMART, including sample size considerations, the use of SMART-specific methods like inverse probability of treatment weighting, and the availability of resources and workshops for those interested in learning more about SMART.
The key points raised in the panel discussion following the presentation were:
The FDA expressed concerns about potential overtreatment in perioperative trials, especially for early-stage cancer patients, and iterated that exploration of innovative designs to evaluate the contribution of each phase of a sequence to the overall treatment effect is important. SMART designs can efficiently evaluate treatment sequences by comparing treatment regimens, but potential efficiency gains should be weighed against feasibility burdens, such as patient anxiety, data management challenges, and varying dropout rates.
Companies are increasingly recognizing the importance of demonstrating the efficacy of each treatment in a sequence. However, using surrogate endpoints, such as pathological complete response (pCR), for re-randomization in SMART designs can be challenging due to their questionable predictive value for ultimate endpoints like overall survival (OS) and event-free survival (EFS), making it difficult to draw definitive conclusions about the efficacy of individual treatment components.
Alternative designs, such as factorial design, four-arm trials, or three-arm trials comparing control, neoadjuvant only, and neoadjuvant plus adjuvant arms, could provide insights into the contribution of adjuvant therapy to the total treatment effect.
SMART designs typically involve a single consent at the beginning, which can include the potential for multiple randomizations and can be reassuring for patients; they do not inherently introduce additional bias compared to other trial designs, but they are not commonly proposed by sponsors.
SMART approach will not be appropriate in all situations. Trial design must take into account the main objective(s) of the development program, potential for dropout, and other operational challenges (e.g., number of sites, sample size, etc.).
This forum provided an opportunity to have open scientific discussion among a diverse multidisciplinary stakeholder group – clinicians, epidemiologists, and statisticians from academia and pharmaceutical companies, patient advocates, and international regulators- focused on emerging statistical issues in cancer drug development. This collaborative forum highlighted the promise of SMART designs to improve patient outcomes, while recognizing the logistical and analytical complexities with such designs.
Acknowledgement: Authors thank Joan Todd (FDA) and Syed Shah (FDA) for technical support.
* Speakers/ Panelists:
Dr. Keaven Anderson (Merck), Dr. Elizabeth Barksdale (LUNGevity Foundation), Dr. Alex Bliu (Health Canada), Dr. Michael Coory (TGA, AU), Prof. Marie Davidian (North Carolina State University), Dr. Boris Freidlin (National Cancer Institute), Prof. Tim Friede (University Medical Center Göttingen), Dr. Qi Jiang (Pfizer), Prof. Kelley M. Kidwell (University of Michigan), Dr. Rong Liu (Regeneron), Dr. Olga Marchenko (Bayer), Dr. Pallavi Mishra-Kalya (FDA), Dr. Richard Pazdur (FDA), Dr. Khadija Rantell (MHRA, UK), Dr. Kapil Sen (BMS), Dr. Harpreet Singh (FDA), Dr. Rajeshwari Sridhara (FDA), Dr. Marc Theoret (FDA), Dr. Jonathon Vallejo (FDA), Dr. Xian Zhou (AstraZeneca)
