Phased Variant-MRD ctDNA Assay Outperforms PET in Predicting LBCL Outcomes

Findings from the prospective DIRECT (NCT04226937) study demonstrated that an ultrasensitive circulating tumor DNA (ctDNA) assay tracking tumor-specific phased variants (PVs) to detect minimal residual disease (MRD) at the end of first-line therapy is a powerful and independent predictor of patient outcomes for patients with large B-cell lymphoma (LBCL).^1,2

A key finding is that PV-MRD assessment provides more powerful prognostic information than conventional radiological response assessment using PET. Specifically, detectable PV-MRD at the end of therapy was strongly associated with inferior outcomes.

The study established a clear and statistically significant link between end of first-line therapy PV-MRD status and the likelihood of tumor progression. After a median of 24.5 months, 2-year time to tumor progression (TTP) for patients with detectable vs undetectable PV-MRD was 42% vs 95%, respectively (HR, 13.7; P <.001). Among those receiving full-dose anthracycline-based immunochemotherapy, the 2-year TTP was 45% vs 96%, respectively (HR, 15.4; P <.001), outperforming PET (HRs, 6.9 for PET vs 16.9 for PV-MRD).

When directly compared, the HR for predicting progression was significantly higher for PV-MRD (HR, 16.9) than for PET assessment (HR, 6.9).

These figures indicate that PV-MRD status is a more robust predictor of relapse than the current imaging standard of care at the end of treatment.

Methodological Considerations and Caveats

The study also identified important technical and clinical nuances that must be considered when interpreting PV-MRD results.

Firstly, patient-specific LoD95 was not uniform. It varied by "more than 2 orders of magnitude across patients,” according to Krupka et al, authors of the study. This finding underscores the critical need to report a patient-specific LoD95 to accurately contextualize an undetectable result.

The analysis also noted cases of "persistent PV-MRD in the absence of relapse." This highlights a potential caveat in interpreting a positive result, as it may not always signal imminent clinical progression. This was particularly observed in patients with transformed follicular lymphoma, where 3 out of 4 patients exhibited persistent PV-MRD without relapsing. This suggests that the biological context of the lymphoma may influence the interpretation of ctDNA results.

Study Overview

The DIRECT study was a prospective, multisite study designed to evaluate the clinical utility of ctDNA monitoring in patients with LBCL. The research aimed to overcome the limitations of previous studies, which were often constrained by small cohort sizes, retrospective designs, or reliance on single commercial assay providers.

The study developed and utilized a lymphoma-customized, open-source ctDNA assay and pipeline. This approach was capable of capturing hundreds of PVs per patient, enabling ultrasensitive detection.

The analysis included 155 patients for whom end of first-line therapy PV-MRD status was available.

Clinical Implications

The DIRECT study concludes that PV-MRD analysis at the end of therapy is a sensitive and clinically meaningful method for response assessment in LBCL. The evidence demonstrates that it offers independent prognostic information that significantly enhances response assessment beyond what is possible with conventional radiologic methods alone.

Based on these findings, the authors advocate for the integration of PV-MRD monitoring into clinical trials and routine management.

REFERENCES

1.Krupka J, Moutsopoulos I, Cutmore N et al. Phased variant-supported circulating tumor DNA as a prognostic biomarker after first-line treatment in large B-cell lymphoma: Findings from the DIRECT study. Published December 22, 2025. Accessed January 2, 2026. J Clin Oncol. doi: 10.1200/JCO-25-01587.

2.DLBCL interim response evaluation for customized therapy (DIRECT). ClinicalTrials.gov. Updated August 5, 2022. Accessed January 2, 2026. https://clinicaltrials.gov/study/NCT04226937