New data show that circulating tumor DNA (ctDNA) positivity after definitive treatment strongly predicts relapse in patients with locally advanced cervical cancer — with all recurrences in a real-world cohort occurring in patients who tested ctDNA-positive soon after treatment.
“What we see is that patients who fail to clear ctDNA have significantly worse relapse-free survival,” said Theresa Kuhn, MD, with AdventHealth Cancer Institute, Orlando, Florida.
Kuhn presented the results at Society of Gynecologic Oncology (SGO) Annual Meeting on Women’s Cancer 2026.
Despite the availability of preventive vaccines and screening, cervical cancer remains a leading cause of cancer deaths worldwide. Patients with locally advanced disease face particularly poor outcomes, Kuhn pointed out — with recurrence rates of 30%-50% within 2 years and 5-year survival as low as 15%-50% after recurrence.
Recurrence is typically detected only after symptoms arise, and there is currently insufficient evidence to support routine use of surveillance biomarkers — underscoring the need for better tools, Kuhn said.
To address this gap, she and her colleagues conducted a retrospective, real-world study of 58 patients with cervical cancer, the majority of whom had locally advanced disease and squamous histology.
ctDNA testing was done on plasma samples using a personalized tumor informed assay, meaning each assay was personalized based on the patient’s tumor.
“This improved sensitivity for detecting minimal residual disease compared to tumor agnostic approaches,” Kuhn said. “So we’re not just detecting ctDNA, we’re detecting patient-specific tumor DNA after definitive treatment.”
The analysis showed that ctDNA status shortly after treatment was highly prognostic.
Patients who were ctDNA-positive within 3 months of definitive treatment had significantly worse relapse-free survival (hazard ratio [HR], 225.8; P < .0001) than patients who were ctDNA-negative during that timeframe. Median relapse-free survival was 3 months in patients who were ctDNA-positive within 3 months of treatment.
Similarly, patients who were initially ctDNA-negative 0-3 months after treatment but converted to positive during that time also had significantly worse relapse-free survival than peers who remained ctDNA-negative (HR, 45.7; P = .001).
Notably, no recurrences were observed among patients who remained ctDNA-negative throughout surveillance, while 14 of 16 ctDNA-positive patients experienced recurrence. All recurrences were preceded or accompanied by ctDNA positivity.
Beyond prognostication, ctDNA results also appeared to influence clinical management.
A positive ctDNA test led to shorter time to the next ctDNA test and about a twofold shorter time to imaging. Importantly, 90% of positive ctDNA tests triggered imaging within 3 months, and all of the patients who underwent imaging had radiographic confirmation of recurrence.
“In practice, ctDNA is not only prognostic,” Kuhn said. “In this real-world data set, it appears to influence clinical decision-making with respect to radiographic imaging and earlier detection of disease.”
Whether ctDNA testing ultimately affects patients’ survival is unknown, and Kuhn said that further studies are needed to answer that question.
Study discussant Rebecca Brooks, MD, director of gynecologic oncology at the University of California, Davis, agreed.
Brooks framed ctDNA as an “emerging, promising but still investigational biomarker” in cervical cancer, with its value closely tied to disease stage and timing of assessment: ctDNA is typically not detectable in pre-invasive disease, low in early-stage disease, and higher in locally advanced disease.
And the strongest evidence has been in the post-treatment setting, where a positive ctDNA result consistently predicts a higher risk for recurrence.
“The central unanswered question is actionability,” Brooks said. “Does intervening on molecular recurrence improve outcomes, or only reveal relapse earlier?”
The study had no commercial funding. Kuhn had no disclosures. Brooks disclosed relationships with GSK, AstraZeneca, Natera, and Corcept.
Admin_Adham