What Is ctDNA MRD Testing and How Does It Work?

This podcast explores the principles and clinical applications of circulating ctDNA-based MRD testing in diffuse large B-cell lymphoma, including how personalized sequencing assays can detect microscopic residual disease with greater sensitivity than imaging. Dr David Russler-Germain discusses the growing evidence supporting ctDNA MRD testing for prognostication and treatment monitoring, while also addressing current limitations, such as variability across lymphoma subtypes and factors that can affect assay performance and interpretation.

Kelly Conger: Hello, and welcome to the Oncology Learning Network. I'm Kelly Conger, and on today's podcast, we are joined by Dr David Russler-Germain, who will be discussing ctDNA MRD testing in lymphoma. So to kick us off, can you define ctDNA at the most basic level before we dive in to how it can be harnessed for residual disease detection?  

Dr Russler-Germain: All dividing cells in our body are at risk of releasing some degree of their DNA contents into the bloodstream as cell-free DNA. But when those cell-free DNA molecules, typically in the 100 to 150 base pair length, have mutations due to being derived from malignant or tumor cells somewhere in the body, we call that circulating tumor DNA. Whether those mutations do or don't relate to the pathogenesis of the disease is a separate question. They can be driver mutations in genes such as MYC or p53 in some contexts. But alternatively, they can be passenger mutations. But in the end, the bottom line is that they're markers of the diseased cells being present in the body and releasing those mutated DNA molecules for the detection for one purpose or another.  

Kelly Conger: Now, how is ctDNA implicated in a disease like diffuse large B-cell lymphoma? Don't we already have tools to detect residual disease in this setting?  

Dr Russler-Germain: In the end, response assessment in DLBCL boils down to detecting some degree of residual cells somewhere in the body. And often that is a residual mass in patients who unfortunately have refractory disease where it can be seen on a scan, either by anatomic measurements by CT or by PET signal, which reflects glucose uptake by the cancer cells. But as our treatments get better and better and patients have a higher likelihood of achieving a complete response by imaging, We're going to be eventually below the limit of detection of PET scans where we have microscopic amounts of disease, just a few cubic millimeters left in the patient that still unfortunately puts them at risk for recurrence, even within months to the year, few years' timeframe.  

Compared to PET scans, however, ctDNA-based MRD testing is able to identify very low levels of disease with high degrees of sensitivity. Because of the highly proliferative nature of the residual cancer cells, even if the residual lesions, so to speak, are smaller than the resolution of the imaging tests, they are still going to release some degree of cell-free DNA into the plasma, which can be identified using our highly sensitive sequencing assays as ctDNA, and give us that MRD-detectable or undetectable analysis, hence the improved prognostic value of ctDNA testing compared to PET scans in end-of-treatment DLBCL management.  

Kelly Conger: I want to dig into some of the practical considerations of ctDNA testing. Can you walk us through how these tests are conducted and how the technology has evolved?  

Dr Russler-Germain: One of the major benefits of MRD testing by ctDNA sequencing is that it's done typically by something as simple as a peripheral blood draw or more, if a little more invasively, a bone marrow biopsy or CSF testing by lumbar puncture. In essence, we're looking for a body fluid, whether that's blood, marrow, or CSF, and isolating the cell-free DNA fraction from that, which can then undergo next-generation sequencing to look for specific mutations that were previously identified in either the same sample when the patient had active disease pretreatment or from their original tumor biopsy that led to the original diagnosis.  

So how does applying ctDNA and MRD testing actually work? We need to first identify a patient's specific mutational fingerprint from their tumor to identify the mutations that we're then going to look for, for MRD testing down the line. Using a patient's archival tumor tissue, it'll either undergo a targeted panel sequencing, whole exome or whole genome sequencing to obtain that mutational fingerprint. And then a personalized assay or a panel-based assay might be applied to MRD testing samples down the line. Typically, we are comparing the tumor tissue to a matched germline sample so that we know exactly which are the somatic alterations we should track over time. And then more often than not, a personalized assay will be looking for the specific mutations from that tumor sample in the future blood samples where the ctDNA is isolated down the line.  

Much of the innovation of ctDNA testing over the last five to 10 years has come from both molecular biology and technology or informatics approaches to really get that level of sensitivity down to one in one million mutated molecules in essence. Part of it involves whether those mutations being tracked are required to be on the same DNA fragment, sometimes called phased variants. Other times it is regarding the depth of sequencing as well as breadth of sequencing across the number of mutations being tracked. In essence, the idea is that each assay has its own specific parameters for whether a sample will be called positive or negative based on very stringent requirements that have been optimized internally.  

Kelly Conger: So it sounds like both the enhanced sensitivity, and the ability to personalize these assays makes them powerful tools for lymphoma residual disease detection. Moving into the clinical setting, can you speak to how ctDNA MRD testing is being implemented and at what stages of the patient's journey this technology can be leveraged?  

Dr Russler-Germain: One of the landmark analyses by Mark Roschewski and others applying phased variant ctDNA sequencing in a frontline cohort of DLBCL patients looked at the serial samples with each cycle during treatment, showing that subsequent clearance with each cycle of therapy was highly prognostic for PFS. A substantial fraction of patients even convert to undetectable ctDNA after just a single cycle of therapy. The majority of patients who remain progression-free after treatment have more often than not converted to undetectable MRD by cycle three, day one of six cycles of frontline R-CHOP or an alternative therapy, indicative that the kinetics of response are highly dynamic and strongly prognostic, informing our patient care and their subsequent outcomes.  

So how is the utility of ctDNA testing and lymphoma evolving? First, we're learning that earlier looks may actually be incredibly important. Waiting till end of treatment to decide whether a patient has residual disease or not has its own value, of course, but earlier looks after one, two, or even only three cycles of treatment have shown to be highly prognostic for subsequent outcomes and are likely to move forward in the field to guide treatment escalation or de-escalation strategies. Beyond that, though, while we might think that applying ctDNA testing to our patients at highest risk of relapse at the end of treatment is where we should focus our effort, several studies actually contradict that and say that regardless of IPI, ctDNA testing A) outperforms PET scans for end of treatment adjudication and B) remains highly prognostic. While obviously patients at low risk from diagnosis tend to do better than those who are high risk, we shouldn't be restricting our optimal adjudication of end of treatment response assessment just to those at high risk from baseline. And in fact, we don't do that with PET scans versus CT scans. So why should we with MRD testing?  

Kelly Conger: Are there any limitations to ctDNA MRD testing? You're certainly making a case for its strengths, but I'm curious if there are any drawbacks to this technology.  

Dr Russler-Germain: Despite everything we've talked about today, there are some limitations to ctDNA testing and MRD evaluations that we need to acknowledge and work to overcome. First, there are some sensitivity restraints. Some histologies, especially low-grade lymphomas in particular, will release less ctDNA into the bloodstream, making them harder to detect at baseline when patients have active disease and thus even harder to assess MRD down the line. So while MRD testing for DLBCL is approaching real-world clinical utilization today, the application to other histologies, especially follicular lymphoma and marginal zone, remains on the research side of the spectrum. Part of the nuances I've been thinking about sensitivity and specificity for our ctDNA-based MRD testing relates to actually how we're doing the testing and when we draw the samples. Factors such as recent G-CSF administration, surgeries, and infections can all factor in, and thus we're learning a lot year by year as to when the optimal time points are for specific blood draws and how to adjudicate these in the context of concurrent PET scans or concurrent illnesses.  

Kelly Conger: Dr Russler-Germain, thank you so much for your time and thoughtful insights today. And thank you to our audience for listening to this Oncology Learning Network segment, where we did a deep dive into ctDNA MRD testing, including how it works and how it can be applied in clinical practice. Thank you.