TOPLINE

In an analysis of patients with early-stage triple-negative breast cancer (TNBC), detection of tissue-free circulating tumor DNA (ctDNA) during surveillance strongly predicted recurrence, with a hazard ratio (HR) of 27. The median lead time from detection to recurrence was comparable to that of multivariant tumor-informed assays.

METHODOLOGY

• Molecular residual disease (MRD) detection has the potential to transform adjuvant therapy for patients with early TNBC, but methods that require sequencing the primary tumor pose logistical challenges. Tissue-free ctDNA detection could offer an alternative. 
• Researchers conducted a prognostic and exploratory analysis of 159 patients with TNBC at moderate to high risk for recurrence participating in the c-TRAK phase 2 clinical trial in the UK. They looked at whether ctDNA detection accurately estimated recurrence and how it compared to tumor-informed approaches. 
• A total of 1026 plasma samples were collected from patients every 3 months for up to 2 years after completion of adjuvant therapy. Analyses compared tissue-free ctDNA detection results with both digital polymerase chain reaction (PCR) tracking 1 to 2 patient-specific variants and whole-exome sequencing (WES)-powered multivariant tumor-informed assays.
• The primary outcome was recurrence-free survival by tissue-free ctDNA detection status, with secondary outcomes comparing detection timing and lead times between tissue-free and tumor-informed approaches. 

TAKEAWAY

• Tissue-free ctDNA detection during surveillance was strongly associated with risk for recurrence (HR, 27.2; P < .001), with a sensitivity of 84.4% and specificity of 87.8% for identifying recurrence by 24 months. 
• Among 42 patients with ctDNA detected by both tissue-free and digital PCR assays, tissue-free detection occurred earlier in 14 patients (33.3%), with no cases of earlier PCR detection. Median lead time to recurrence was longer with tissue-free ctDNA compared with digital PCR (7.9 months vs 5.8 months; P = .03). 
• Among 41 patients with ctDNA detected by both tissue-free and multivariant tumor-informed assays, 12 patients (29.3%) had earlier detection by multivariant assay vs one patient (2.4%) by tissue-free assay. There was no significant difference in median lead times between the tissue-free and tumor-informed approaches (7.6 months vs 7.1 months; P = .15). 
• Concordance between tissue-free and multivariant tumor-informed assays was 95.2% at the sample level and 92.5% at patient level. 

IN PRACTICE

“These findings support the use of tissue-free assays as a practical alternative in patients for which a tumor-informed assay is not possible,” the study authors wrote. They stressed, however, that “the clinical benefit of early MRD detection and intervention remains uncertain and prospective ctDNA-guided intervention studies are warranted to assess clinical utility of tissue-free assays.”

SOURCE

The study, led by Niamh Cunningham, MBBS, of Royal Marsden Hospital in London, England, was published online in JAMA Oncology.

LIMITATIONS

The study did not compare the tissue-free assay to whole genome sequencing-powered tumor-informed assays, which may have longer median lead times than the WES-based multivariant approach used. Patients did not undergo routine surveillance imaging for metastatic disease, potentially introducing bias in lead time assessment as occult metastatic disease at detection time could not be ruled out. Detection of ctDNA in patients without subsequent clinical relapse raises questions about whether these represent false positives, immune surveillance of micrometastatic disease, indolent disease biology, or insufficient follow-up duration.

DISCLOSURES

The study had numerous funders, including Le Cure, the Royal Marsden Cancer Charity, Breast Cancer Now, Cancer Research UK, and Guardant Health. Several co-authors disclosed financial relationships with various companies, including AstraZeneca, Pfizer, Roche/Genentech, and Merck Sharp & Dohme. Full disclosures are noted in the original article.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.