Max Diehn, MD, PhD, breaks down how ctDNA testing for MRD is reshaping prognostic discussions in NSCLC.
Circulating tumor DNA (ctDNA) testing for minimal residual disease (MRD) has become one of the most active areas of biomarker development in early-stage non–small cell lung cancer (NSCLC), but a positive or negative result carries very different clinical weight than clinicians may assume, according to Max Diehn, MD, PhD.¹
"The interest in detection of MRD is driven by the fact that this could potentially help us tell apart patients who are most likely cured after standard therapy vs patients who still harbor occult [NSCLC]," Diehn said during a presentation at the
In the presentation, Diehn focused on all angles of ctDNA MRD in early-stage NSCLC, including tumor-informed MRD assays, positive vs negative MRD results and their implications, surrounding guidelines, appropriate contexts for the testing, and remaining improvements that are needed.
Diehn is the Jack, Lulu, and Sam Willson Professor, the vice chair of research, and the division chief of radiation and cancer biology in the Department of Biology at Stanford Cancer Institute in Palo Alto, California.
Diehn began by outlining MRD and its importance for lung cancer and NSCLC, noting that interest in detecting the biomarker lies in its potential to distinguish between patients who have been cured by standard therapies compared with patients who still possess less visible disease. More specifically, he mentioned how quick MRD detection following local therapy might enable better personalization of adjuvant therapy for patients with lung cancer. Regarding early-stage lung cancer, Diehn underscored how there is now strong evidence to support ctDNA as a biomarker.
“One thing that's very important in this area is to understand the importance of the assay sensitivity. There are many kinds of ctDNA assays and they have different sensitivities,” Diehn added.
ctDNA assays fall into two broad categories with different performance characteristics: tumor-naive or tumor-agnostic and tumor-informed assays, he explained. Tumor-naive assays are run directly on a plasma sample using fragmentomic, single-nucleotide variant genotyping or methylation-based approaches, with an approximate 95% limit of detection (LOD95) that ranges from approximately 1 in 102 to 1 in 103. Tumor-informed assays instead begin by sequencing a patient's tumor tissue to identify mutations, then build a personalized approach to track those specific alterations in the blood of patients. First-generation tumor-informed assays achieve an LOD95 of approximately 1 in 104, and newer ultrasensitive versions can reach into approximate LOD95 ranges of 1 in 105 to 1 in 107.
Ultimately, the higher sensitivity of tumor-informed assays gives them the edge over others, standing as the preferred method for detecting ctDNA MRD, according to Diehn. “Tumor-informed assays are the preferred approach [due to] this increased sensitivity because you're trying to detect microscopic disease, which of course sheds relatively little,” he elaborated.
Diehn also detailed the workflow associated with tumor-informed assays. “You start by sending a piece of tumor to the lab that gets sequenced. Mutations are identified, and then a custom assay is created for each patient; then the custom assay is applied to the subsequent plasma samples. [Next], reports are generated about the presence of tumor DNA and the concentration,” he explained.
Diehn presented data from six independent studies, all of which showed the same pattern: patients who are ctDNA MRD-positive after completion of definitive local therapy fared worse than those who tested negative.2-7
He cautioned that despite significant differences, the curves for ctDNA MRD-negative patients were not flat. "A negative [ctDNA MRD] result does not mean for sure the patient is cured; it just means the patient has a much lower likelihood of recurring than a patient who has not been tested or is not known to be negative,” he added.
Looking ahead, Diehn suggested perioperative time points—after neoadjuvant therapy and again after surgery—as logical places to test ctDNA MRD status to inform trials of therapy escalation or de-escalation, though he stressed these remain investigational.
Centers for Medicare and Medicaid Services requires an assay to demonstrate earlier detection of recurrence than imaging surveillance to qualify for Medicare reimbursement.1 Additionally, since 2025, tumor-informed ctDNA MRD testing has been reimbursed by Medicare for patients with early-stage NSCLC.
However, NCCN guidelines do not mention ctDNA MRD testing for early-stage NSCLC, reflecting the absence of interventional evidence that demonstrates changes in outcomes, according to Diehn. “We need trials even though we already have reimbursement,” he expressed. “We can order ctDNA MRD testing based on insurance; we need interventional trials to prove utility to push this field forward.”
A positive ctDNA MRD test carries a likelihood higher than 90% to 95% of recurrence in patients with early-stage NSCLC, he said. Moreover, up to two-thirds of patients who experience recurrence demonstrate ctDNA MRD-negative results that are initially false, further emphasizing the need for ctDNA MRD assays with higher sensitivities.
“A negative [ctDNA MRD result] does not mean the patient is cured, it means the patient is likely to do better than an average patient in that setting. [It is important to] remember that there are false negatives, we don't have proof that withholding treatment is safe [following a negative ctDNA MRD result],” he said. “In my clinic, with ctDNA MRD-positive results, I try to enroll patients on these trials we have available. And I encourage everyone who is in the position to do so to try to open [MRD-based] trials... If that's not possible, there's still potential utility, you can discuss prognosis with the patients and consider more frequent imaging, if your standard is every 6 months in these patients, maybe you want to consider every 3 months so that you could detect the likely recurrence.”
For MRD-negative patients, "it is an easier discussion because it’s a good thing, patients really like it," Diehn said. Ultimately, for those who are ctDNA MRD negative, Diehn suggested discussing prognosis with patients and considering less frequent imaging while still emphasizing that there is no evidence that de-escalation of adjuvant treatment following the result is safe.
