This blog post provides an overview of Friends’ recent ctMoniTR publication. Please see graphical abstract below that provides an overview of the manuscript approach, findings, and context. 

Background 

In oncology clinical trials, the gold standard approach for measuring efficacy of cancer therapies is by determining overall survival (OS), or the length of time a patient lives after starting treatment in a clinical trial. Thanks to advances in cancer treatments, patients are living longer, but this also means trials can take more time to complete. Long timelines can slow the approval and availability of new therapies that may have better efficacy and safety compared to existing treatments. To address this, drug sponsors can measure early endpoints (i.e., measurements that are reasonably likely to predict OS) to demonstrate efficacy using the U.S. Food and Drug Administration’s (FDA) Accelerated Approval pathway, enabling earlier assessments of treatment efficacy, regulatory approval, and patient access.  

One promising early endpoint for solid tumors is circulating tumor DNA (ctDNA), fragments of DNA shed by tumors into the bloodstream. The amount of ctDNA is measured from a blood sample and reductions in the ctDNA levels present at baseline compared to an on-treatment levels correlate with the tumor shrinking. Since ctDNA can be measured through blood tests, it allows for more frequent and potentially earlier monitoring than traditional tumor imaging. While the potential of ctDNA is recognized, additional data from multiple clinical trials are needed before it can be routinely used as an early endpoint.

Key challenges in ctDNA-based trial designs 

Despite data being developed over time, several important questions remain regarding the optimal use of ctDNA in clinical trial design: 

• How soon after treatment should blood be collected to measure ctDNA? 

• How should a molecular response (MR) be defined? What level of ctDNA reduction is clinically meaningful?  

How our study contributes 

In our analysis, we combined data from four randomized clinical trials of patients with advanced non-small cell lung cancer (aNSCLC) treated with anti-PD-(L)1 immunotherapy and/or chemotherapy. To address questions of timing and response definitions, we evaluated ctDNA changes from baseline (start of treatment) at two follow-up timepoints, either: 

• Timepoint 1 (T1; an “early” timepoint): Within 7 weeks of treatment initiation OR 

• Timepoint 2 (T2; a “later” timepoint): Between 7–13 weeks of treatment initiation 

Additionally, we defined molecular response (MR) by the percent reduction in ctDNA levels from baseline, and assessed the following categories: 

• MR50: ≥50% reduction  

• MR90: ≥90% reduction  

• MR100: 100% clearance (non-detected ctDNA on treatment) 

Results 

We assessed whether reductions in ctDNA were associated with improved OS by calculating adjusted hazard ratios (aHRs) for each MR level at both timepoints. In this analysis, higher aHRs reflect a greater risk of death for patients who did not achieve a MR, meaning larger values indicate stronger associations between ctDNA reduction and improved OS. 

Anti-PD-(L)1 Group: 

• Patients in the clinical trials were treated with anti-PD-(L)1 with or without chemotherapy. 

• At both T1 and T2, reduced ctDNA was associated with improved OS and associations at T2 were slightly stronger. 

• For both timepoints, MR100 had a higher aHR than MR50, suggesting greater ctDNA clearance is more strongly linked to improved outcomes, though fewer patients were able to achieve MR100. 

Chemotherapy Group: 

• Patients in the clinical trials were treated with chemotherapy alone. 

• Reduced ctDNA was associated with improved OS at T2. 

• Associations were slightly lower than the anti-PD-(L)1 group. 

• MR100 showed the strongest association, similar to findings in the anti-PD-(L)1 group. 

Interpretation 

These findings suggest that ctDNA assessment at a later timepoint (T2) may be a stronger predictor of overall survival than earlier measurements. However, ctDNA changes at an earlier timepoint (T1) were still predictive of OS, indicating early measurements provide valuable information—especially given the benefit of obtaining early insights on treatment efficacy. 

Associations between ctDNA response and OS were stronger in the anti-PD-(L)1 group compared to the chemotherapy group, suggesting a stronger relationship between ctDNA response and OS in patients receiving anti-PD(L)1 immunotherapy. These findings should be interpreted with caution, as the anti-PD(L)1 group had a larger sample size, which may have contributed to stronger statistical signals. Overall, the results demonstrate that decreases in ctDNA are associated with improved OS for patients with aNSCLC treated with either anti-PD-(L)1 and/or chemotherapy. 

Remaining Questions 

While this study offers valuable insights, several important questions remain: 

1. How do different treatments influence ctDNA dynamics? Unequal group sizes in this study limited our ability to directly compare anti-PD(L)1 vs. chemotherapy. 

1. Is ctDNA predictive of OS in other cancer types? This analysis focused solely on aNSCLC. 

1. What are the clinical advantages of later ctDNA timepoints, and how might they affect treatment decision-making? This study suggested later measurements may provide more information, but the practical implications remain unclear. 

1. Which ctDNA cutoff (MR50, MR90, MR100) is most appropriate for applications in clinical trials? Further research is needed to validate these thresholds. 

This study is part of our ongoing effort to generate the evidence needed to determine how ctDNA can serve as a reliable tool to accelerate the development of effective cancer therapies. More work is needed to standardize approaches across trials, validate meaningful thresholds, and understand how ctDNA dynamics vary across treatments and cancer types. Addressing these questions will be critical to advancing ctDNA as a trustworthy measure of treatment benefit and helping bring new therapies to patients faster.