Accurate detection of minimal residual disease remains a major challenge in early breast cancer. Current risk assessment after neoadjuvant therapy relies largely on clinicopathologic factors and pathologic complete response (pCR), which is associated with improved outcomes, particularly in triple-negative and HER2-positive disease. However, pCR is an imperfect surrogate marker: some patients without pCR remain disease-free long term, while others eventually experience relapse. Because conventional imaging and pathology cannot reliably detect microscopic systemic disease, circulating tumor DNA (ctDNA) has emerged as a promising biomarker for identifying molecular residual disease and refining postoperative risk stratification. The Pathologic Response Evaluation and Detection in Circulating Tumor-DNA (PREDICT-DNA) study was designed to evaluate whether ultrasensitive ctDNA monitoring could improve risk assessment in this setting.
Study Design and Objectives
The PREDICT-DNA study is a prospective, multicenter study evaluating circulating tumor DNA as a biomarker of minimal residual disease in patients with stage II–III HER2-positive or triple-negative breast cancer receiving neoadjuvant therapy. Using a tumor-informed ultrasensitive ctDNA assay capable of detecting variants at levels below 100 parts per million, investigators analyzed plasma samples collected at baseline, after completion of neoadjuvant therapy before surgery, and after surgery. The primary objective was to determine whether ctDNA negativity after neoadjuvant therapy could predict pathologic complete response with a negative predictive value of at least 90%. Among the pre-specified secondary endpoints, the study evaluated the relationship between ctDNA detection and 5-year invasive disease–free survival (IDFS) in the TNBC cohort.
Patient Population
In the PREDICT-DNA study, a total of 227 patients were enrolled, of whom 220 were evaluable for pathologic complete response (pCR). The study population included stage II–III breast cancer, comprising 48% HER2-positive disease and 52% triple-negative breast cancer (TNBC). Overall, 41% achieved pCR following neoadjuvant therapy.
Primary Endpoint
In the PREDICT-DNA study, the primary endpoint was not met. While all patients who achieved pathologic complete response (pCR) were ctDNA-negative after completion of neoadjuvant therapy, ctDNA negativity was also observed in a substantial proportion of patients with residual disease. Specifically, 40% of patients without pCR were ctDNA-negative after neoadjuvant therapy, resulting in a negative predictive value of 60%, which did not reach the prespecified threshold of ≥90%required for the primary endpoint.
Prognostic Value of ctDNA
Despite the primary endpoint not being met, the PREDICT-DNA study demonstrated strong prognostic value for ctDNA monitoring. Detectable ctDNA after completion of neoadjuvant therapy was significantly associated with an increased risk of recurrence (HR 8.9; 95% CI, 2.4–33; P = .001), and this association remained independent of pathologic complete response (pCR). Importantly, ctDNA detection after surgery identified patients at extremely high risk of relapse (HR 128; 95% CI, 15–1083; P < .001). In contrast, patients who were ctDNA-negative after surgery had excellent outcomes, with 94% 5-year invasive disease-free survival.
Investigator Perspective
Commenting on the findings, Ben Ho Park, MD, PhD, Director of the Vanderbilt-Ingram Cancer Center, highlighted the prognostic implications of ctDNA monitoring in early breast cancer.
“PREDICT-DNA is the first prospective multi-institutional study with pre-specified endpoints that validates ctDNA as a diagnostic tool for detecting microscopic minimal residual disease using an ultrasensitive assay in patients with Stage II and III triple-negative breast cancer (TNBC) receiving neoadjuvant therapy. The results suggest that clearance or persistence of ctDNA after neoadjuvant therapy is an improved prognostic marker compared to pathologic complete response, and additionally, ctDNA MRD assessment after surgery accurately reflects patients who still have disease that will recur (ctDNA positive), versus those that have a high likelihood of cure (ctDNA negative).”
Ben Ho Park, MD, PhD
Director, Vanderbilt-Ingram Cancer Center
Dr. Park also clarified that although the primary endpoint was not met, the PREDICT-DNA study included pre-specified and statistically powered analyses evaluating the association between ctDNA status and outcomes. He emphasized that clearance of ctDNA after completion of therapy was associated with excellent long-term outcomes, whereas persistent ctDNA after treatment was associated with recurrence within three years. He also noted that the SCANDARE trial, using the same assay, reported a similar signal supporting ctDNA as a potentially stronger prognostic marker than pathologic complete response, and added:
“Together with similar studies, validation of ctDNA MRD assessment in early-stage TNBC now opens the possibility of designing future clinical utility studies to safely de-escalate therapies, as well as escalate additional treatments for patients who will otherwise recur.”
Clinical Implications
These findings highlight the potential clinical value of molecular residual disease assessment in early breast cancer. Although ctDNA-guided treatment strategies are not yet part of routine clinical practice, the results of the PREDICT-DNA study support the development of prospective trials evaluating treatment escalation or de-escalation strategies guided by ctDNA status.
Written by Marine Rushanyan, MD