Laying the Groundwork to Use ctDNA to Catch Progression in DLBCL

Among the most captivating areas of cancer research is the use of circulating tumor DNA to shift cancer treatment midcourse. Can markers in the blood “catch” the recurrence of cancer before it presents as tumor progression?

This is already the concept of SERENA-6,¹ the groundbreaking phase 3 study in breast cancer used circulating tumor DNA (ctDNA) monitoring to determine which patients should switch from a CDK4/6 inhibitor to an oral selective estrogen receptor degrader camizestrant at the first sign of an emergent ESR1 mutation. Those who switched had significantly improved outcomes. The trial led to an approval for camizestrant in Europe, but the approach has been a source of debate in the United States, where an FDA advisory committee recommended against approval.²

Now, a group of researchers from Limerick, Ireland, are deploying this concept in diffuse large B-cell lymphoma (DLBCL), a disease in which 40% of patients relapse or prove refractory to treatment. Writing in Hematological Oncology,³ the authors note that existing tools—the International Prognostic Index and interim/end-of-treatment PET/CT scans—often fail to flag high-risk patients before a recurrence.

In a rapidly advancing disease such as DLBCL, the stakes are high. “Early identification of patients at high risk of relapsed or refractory disease remains an unmet clinical need,” the authors write. “This could facilitate earlier consideration of alternative treatments such as bispecific T-cell engagers (BiTEs) or chimeric antigen receptor T-cells (CAR-T).”

Their scoping review, conducted to inform an Irish feasibility study, mapped the evidence on ctDNA and circulating immune markers as noninvasive predictors of treatment response in previously untreated DLBCL. Following PRISMA-ScR methodology, the authors searched PubMed, Embase, CINAHL, and Cochrane from inception through May 2025, ultimately including 61 sources (42 journal articles, 18 conference abstracts, and 1 guideline from the National Comprehensive Cancer Network) out of 1791 initially retrieved records.

In Key ctDNA Findings, Timing Matters

Across 42 ctDNA studies included, sampling consistently occurred at baseline, with 23 studies also collecting at end of treatment (EOT) and 24 during treatment. Plasma was separated from EDTA blood via double centrifugation and frozen within 2–4 hours to preserve sample integrity. Laboratory approaches varied widely: 14 studies used custom NGS panels (11–413 genes), while 10 relied on commercial panels such as Roche Avenio or the Euroclonality lymphoma panel; 9 studies used the ultrasensitive PhasED-Seq method, capable of detecting 1 variant in 1 million bases.

Detection rates varied substantially by platform and methodology, with studies using hybrid capture or PhasED-Seq generally achieving detection in over 90% of patients, compared with lower rates (60%–80%) for other approaches. Concordance between tissue and ctDNA mutation detection averaged 72.3%.

The strongest evidence centered on timing: Kurtz et al. (2018) defined “early molecular response” at cycle 2 day 1 and “major molecular response” at cycle 3 day 1 as thresholds predicting superior progression-free survival, concepts now widely referenced in the field. Failure to clear ctDNA by EOT was consistently linked to inferior outcomes. Notably, combining ctDNA with PET/CT substantially improved risk stratification; several studies found that patients with a positive EOT PET scan but undetectable ctDNA did not progress, suggesting ctDNA could reduce unnecessary confirmatory biopsies or repeat scans. One study found ctDNA-based molecular clustering (the "LymphGen" tool) could reassign risk even among patients with high baseline ctDNA, identifying a favorable-outcome subgroup (BN2/ST2) with 4-year survival rates far exceeding unclassified patients.

Recurrently mutated genes associated with inferior outcomes included TP53, KMT2D, EP300, CD79b, PIM1, MYC, and B2M, while CREBBP mutations were linked to longer progression-free survival.

Key Immune Marker Findings

Eighteen studies examined circulating immune markers, nearly all using pre-treatment samples only, which limited insight into dynamic changes over treatment. ELISA was the dominant method. The most consistently prognostic markers were CXCL9, CXCL10, IL-10, TNF-α, and monocytic myeloid-derived suppressor cells (M-MDSCs)—all elevated at diagnosis in patients with inferior survival. Elevated CXCL10 was particularly notable for identifying relapse risk even among germinal-center B-cell (GCB) subtype patients, who typically have favorable prognoses. One study identified a 3-protein "inflammation score" (IL-10, IL-18, CXCL9) that stratified outcomes and correlated with an exhausted, checkpoint-marker-rich tumor microenvironment.

A single study directly combined both data types. It found that inflammatory serum protein clusters correlated with higher baseline ctDNA and with specific molecular subtypes (e.g., elevated TACI and IL-16 with the MCD/C5 subtype), but did not find a relationship between inflammatory markers and MRD positivity at later timepoints. The authors describe this integration as a largely unexplored area.

“Questions remain open regarding which immune markers are significantly associated with high ctDNA at diagnosis or with failure to achieve [major molecular response] after 2 cycles of treatment,” they write. “With the identification of molecular subgroups and the potential for personalized medicine, much work remains to integrate the patient's immune status into the prognostic models.”

Limitations and Conclusions

The review's authors acknowledge that inclusion of conference abstracts—necessary given how fast the ctDNA field is evolving—meant some included sources lacked complete methodological detail, complicating quality comparisons. They also note that too few immune marker studies extended beyond baseline sampling to draw firm conclusions about their value during or after treatment.

The review concludes that ctDNA shows clear potential to improve DLBCL outcome prediction, particularly for interpreting ambiguous PET/CT results, and that ctDNA-based MRD assessment is now recommended in NCCN guidelines when EOT PET is positive and repeat biopsy is not feasible. However, they emphasize that “Assay standardization for ctDNA analysis is currently lacking,” given the heterogeneity of panels, platforms, and sensitivity thresholds used across studies—the recommended limit of detection (1 part per million) may itself be a barrier to widespread clinical adoption due to cost and expertise requirements. Immune markers offer complementary pretreatment prognostic value, but their role in monitoring treatment response and their relationship to ctDNA-defined molecular subgroups remain open questions for future research.

References

1. Bidard FC, Mayer EL, Park YH, et al, for the SERENA-6 Study Group. First-line camizestrant for emerging ESR1-mutated advanced breast cancer. N Engl J Med. 2025;393(6):569-580. doi: 10.1056/NEJMoa2502929
2. Update on FDA Advisory Committee vote on camizestrant in combination with a CDK4/6 inhibitor for advanced HR-positive breast cancer. News release. AstraZeneca. April 30, 2026. Accessed August 26, 2026. https://www.astrazeneca-us.com/media/press-releases/2026/Update-on-FDA-Advisory-Committee-vote-on-camizestrant-in-combination-with-a-CDK-4-6-inhibitor-for-advanced-HR-positive-breast-cancer.html
3. McMahon A, Ryan EJ, Dillon S, Clifford R. Circulating tumor DNA and immune response markers for improved treatment outcome prediction in diffuse large B‐cell lymphoma: A scoping review Hematol Oncol 2026;44(5):e70242. doi:10.1002/hon.70242