ctDNA MRD Testing in Lymphoma: Clinical Utility and Applications

Overview: In this discussion, Dr David Russler-Germain explains how circulating tumor DNA (ctDNA)–based minimal residual disease (MRD) testing is helping refine post-treatment risk stratification, detect relapse earlier, and guide more personalized treatment decisions in diffuse large B-cell lymphoma (DLBCL), particularly as immunotherapies complicate interpretation of traditional PET/CT imaging.

David Russler-Germain, MD, PhD: When we think about the clinical rationale for using ctDNA testing after frontline therapy in DLBCL, three main areas come to mind. The first is that simply we can help avoid over-treating low-risk patients. Confirming that patients with complete metabolic response by PETs are indeed MRD-undetectable conveys that they have deep, deep remissions with incredibly low risks of relapse. We should be running clinical trials to study de-escalation of therapy in these patients, and we should also not over-surveil them with unnecessary scans, radiation, or biopsies in that context. 

Second, we want to improve our post-treatment risk stratification beyond standard PET CT imaging. As we've discussed, multiple studies have shown that if you compare ctDNA-based MRD assessment to PET scans, every time the outcome looks the same, that you get a greater hazard ratio and better prognostic capacity from the MRD testing compared to end-of-treatment PET scans in frontline DLBCL. Thus, being highly specific for residual disease, we need to better understand who are our patients likely to be cured versus those who aren't, especially with the advent and introduction of second-line CAR T-cell therapies, another curative but unfortunately somewhat toxic therapy that we need to judiciously apply to patients with true residual disease, either those with clear progression, biopsy-proven residual disease, or MRD-detectable disease with some perhaps signal residual on PET. 

And finally, thinking about how do we detect residual disease before clinical relapse, many of our patients will be in the end-of-treatment complete metabolic response after frontline chemoimmunotherapy, but unfortunately the low sensitivity of PET scans leaves their significant opportunity for earlier detection of relapse. And we know that patients do better both in terms of efficacy and toxicity when they receive second-line therapy with lower burdens of disease. There are ongoing studies examining surveillance MRD by ctDNA sequencing at structured time points after frontline therapy to catch relapses as soon as possible and not wait for patients to have clinical symptoms or unfortunately become ill from their recurrent lymphoma. 

In recent studies that have performed serial MRD testing by ctDNA sequencing on patients in remission after frontline treatment for DLBCL, we've seen over and over that ctDNA is able to identify recurrent lymphoma with a lead time advantage compared to clinical or imaging-based assessments. What we've seen is that perhaps you achieve in the order of two to three months of additional lead time through the MRD surveillance analyses, which is actually surprisingly important considering our common second-line therapy can and should be CAR T-cells in many patient circumstances. While these therapies are fantastic, they're actually surprisingly convoluted to deliver to patients, including referrals to specialized centers, multiple rounds of pretesting, arranging collection for apheresis production and administration, and thus every week or every month counts as far as preventing our patients from becoming ill from recurrent lymphomas and delivering them the most effective therapy in the safest way possible. 

To date, there has actually not been a study yet to show that ctDNA or MRD testing does not outperform PET scans as far as prognostication after frontline treatment for DLBCL, with most studies showing a 6 to 10 times stronger relapse prediction by the MRD approach compared to PET scans alone. What you can see here is that roughly half of patients with end of treatment positive PETs don't go on to relapse in the next two to three years after frontline treatment, whereas essentially 80 to 90, if not all, depending on the cohort you look at, of patients with positive ctDNA at end of treatment do unfortunately go on to relapse compared to fewer than 10 to 15% of patients with undetectable MRD having subsequent progression based on the ctDNA assay. 

And importantly, the cTDNA testing shouldn't be restricted based on whether a patient's outcome by PET is one result or the other. The studies that have adjudicated PETs versus CTs from the same time point have repeatedly shown that the MRD result is highly informative, irrespective of whether patients have positive or negative end-of-treatment PETs. What you can see here is on the left is that patients with negative end-of-treatment PETs who unfortunately are MRD-positive have a, in this small cohort, 100% chance of recurrence in the first one to two years post-treatment versus a 90% or better chance of being cured if they were on MRD-undetectable. And conversely, the same almost applies with the PET-positive patients where those who are MRD-undetectable have less than a quarter odds of subsequent recurrence, where all of those with MRD-positive disease did unfortunately go on to relapse. This highly impacts how we implement MRD testing, where it's not a sequential analysis to adjudicate PETs after the fact, but rather it is optimally likely implemented concurrently or even instead of in certain circumstances PET-based imaging, depending on the resources of your practice. 

As novel immunotherapies have been introduced into the DLBCL treatment armamentarium, such as bispecific antibodies and CAR T-cells, what we've unfortunately seen is that the risk of false-positive PET signals due to induced inflammation from these T-cell-based therapies interferes with our ability to assess for responses. Thus, as immunotherapies move earlier into lines of therapy, more recently in the second-line setting for CAR-T cells, for DLBCL, and there are ongoing trials of bispecific combinations in the frontline setting, we're going to need to move beyond PET CTs for a response assessment and use more reliable and more prognostic response adjudication, namely MRD-based assessment, to help make the best choices for our patients. 

It's extremely exciting to see ctDNA MRD testing in DLBCL move from the research sphere to the clinical sphere. Our colleagues in the care of patients with CLL, mantle cell, and multiple myeloma have had MRD testing, typically by the immunoglobulin sequencing assays, integrated into their treatment algorithms for several years now. Given that DLBCL is an aggressive, yet curable malignancy where the goal is to achieve complete disease-free state, the NCCN now recommends consideration of ctDNA-based MRD testing at end of treatment for patients with residual signals seen on their end-of-treatment PET scans. In a typical scenario, our patients are receiving three to four cycles of chemoimmunotherapy for limited-stage disease and six cycles of therapy for patients with advanced-stage disease in undergoing an end-of-treatment PET scan, hopefully showing a complete metabolic response. For those with a Deauville 4 or 5, particularly those with a partial response where this is likely most important, confirmatory testing of residual refractory lymphoma by tissue biopsy or MRD testing is highly recommended before moving on to second-line therapy such as CAR T-cells. 

In many situations, we've gotten used to deep biopsies or invasive procedures to assess for residual lymphoma, or if the procedure isn't amenable from a safety perspective, we say that we'll just get another PET scan in two to four months and see where the disease goes from there. But more recently, the NCCN and the field is moving towards the concept of using a blood-based biomarker, namely ctDNA sequencing, to be that adjudication for end-of-treatment PET scans, saving patients both the time getting to the biopsy and the potential risks of the procedures, as well as the false-negative risk as some procedures may find only necrotic lymphoma, when in fact residual disease is left at a different focus.