Friends of Cancer Research annual meeting to discuss validation of ctDNA endpoints, plus interim OS analysis and pragmatic trials
Evidence for circulating tumor DNA reduction as a new intermediate endpoint to support accelerated approvals of solid tumor therapies is nearing a tipping point. The endpoint stands to transform clinical development across many settings, but it must first be more fully validated.
For that to happen, companies need to commit to standardizing both the assays used to measure circulating tumor DNA (ctDNA) and the protocols for collecting samples and analyzing the data.
Friends of Cancer Research has taken on the challenge through a working group that aims to harmonize ctDNA analyses in trials of immunotherapies for non-small cell lung cancer (NSCLC). The goal is to conduct meta-analyses of the harmonized studies to validate reduction in ctDNA as an endpoint that predicts progression-free survival (PFS) and overall survival (OS).
Even though the analysis will focus on immunotherapies for NSCLC, due to the high level of clinical development activity in the setting, the results are likely help validate the endpoint more broadly.
Friends of Cancer Research has released a white paper on the topic, one of three that will be covered at its annual meeting to be held Tuesday. Also on the agenda is optimization of interim overall survival analyses and enhancing postmarket studies using pragmatism — a priority the Oncology Center of Excellence is addressing through the ongoing Project Pragmatica. White papers were also released on these topics.
The case for ctDNA
Circulating tumor DNA is a measurement of genetic material shed into plasma by solid tumors. Levels appear to correlate with the size of a tumor, and possibly, its response to therapy.
The endpoint is relatively low-burden for patients as it involves a simple blood draw. It can also be used when the traditional radiographic intermediate endpoints that measure the size of a tumor, such as objective response rate (ORR), fall short. For example, bone metastases can’t be imaged through traditional radiographic endpoints, tumor size can be difficult to assess in lung cancers when pleura effusions are extensive, and repeated imaging — particularly for head and neck cancers — can expose patients to excessive radiation that could contribute to future tumor formation.
Circulating tumor DNA could also be useful in perioperative settings, where the tumor is removed so imaging isn’t effective in determining whether residual cancer is present.
In effect, reduction in ctDNA could help streamline trials, open the accelerated approval pathway to new applications and encourage participation. Outside of clinical trials, the measure could provide an early indication of whether a patient is responding to a therapy, to ensure that they receive optimal care.
FDA has signaled a willingness to accept new intermediate endpoints in oncology. In April, the agency’s Oncologic Drugs Advisory Committee (ODAC) agreed to support the absence of minimal residual disease (MRD) as an intermediate endpoint for multiple myeloma. MRD measures residual cancer cells in the bone marrow after treatment.
ctDNA, in contrast, has primarily served as an internal decision-making tool in clinical development. Transitioning the measurement from research tool to clinical endpoint will require standardizing the practices around its capture and analysis, then determining whether it correlates with survival endpoints.
Harmonizing data collection
Establishing consistency in how and when ctDNA data are collected is the first step in creating a standardized relationship with clinical outcomes.
Currently, companies measure ctDNA in different ways using different assays. They also collect data on different schedules in their clinical trials.
Next-generation sequencing and digital droplet PCR assays are both used to detect ctDNA. The tests are often designed to detect the presence of particular tumor variants into the blood, but which variants and how many are included in the screen are variable. Most assays detect a set of variants found in the primary tumor sample, but some look for a single disease-driving allele.
Additionally, ctDNA methylation and fragment size are emerging measurement strategies.
Another problem companies need to solve is that DNA variants shed into blood by stem cells can obscure the ctDNA results. Here too, companies use different methods to eliminate the signals from stem cell DNA such as variants resulting from clonal hematopoiesis (CH). Some methods employ algorithms to remove CH variants, while others use tumor-informed methods to distinguish variants found in the tumor from those that likely come from CH. A third method involves removing peripheral blood mononuclear cells to identify and filter out germline variants.
In the white paper, the working group doesn’t recommend a particular assay or method to measure variant allele frequency (VAF), but it does recommend using assays that can detect ctDNA in at least 70% of cases at baseline as a minimum standard. The limit of detection should be around 0.1% of plasma DNA. And although it doesn’t recommend a method for filtering CH variants, it does stipulate that the methods used should be disclosed.
The group recommends measuring ctDNA on the first day of treatment prior to initiation, plus at least three more times: on the starting day of the second and third treatment cycles, and then when six-month PFS data are recorded.
Through a project called ctMoniTR, Friends of Cancer Research collected data from clinical trials to determine whether changes in ctDNA correlate with clinical outcomes, and indeed found correlations with survival endpoints.
Other meeting topics
The Friends of Cancer Research meeting will also discuss best practices for planning interim overall survival analyses, and applications for incorporating pragmatic trial features into postmarket studies.
In a white paper released ahead of the meeting on overall survival, a working group outlined the issues that interim overall survival readouts present. Because the data are immature at the interim analysis points, early OS data may conflict with positive outcomes on intermediate endpoints.
The group suggests prespecifying OS milestones and harm thresholds to improve interpretation of the interim data. They also propose working with regulators to harmonize simulation methods for interpretation of interim OS results, and using structured and multi-step approaches that include descriptive assessments then quantitative analyses.
In preparation for a session that will be held on pragmatism in postmarket studies, a separate working group introduced three scenarios where pragmatic elements may help answer research questions after a drug is approved. They are studies with racial and ethnic diversity matched to the real-world population, safety and toxicity studies on the cumulative effects in the real world and how to mitigate them, and evaluation of safety and efficacy in biomarker-defined populations.
In all cases, the goal is to collect more data on the generalizability of the results from premarket studies to real-world populations through studies that incorporate real-world data; minimize the burden on patients, sites and clinicians; and more closely align clinical research with routine care.
https://www.biocentury.com/article/654136/turning-ctdna-into-an-intermediate-endpoint-will-require-harmonization