Pretreatment ctDNA Status Is Associated With Distant Recurrent-Free Survival in Early Breast Cancer

Breast Cancer

| Image Credit: © Sebastian Kaulitzki – stock.adobe.com

Circulating tumor DNA (ctDNA) concentration at diagnosis was significantly associated with distant recurrence-free survival (DRFS) in patients with breast cancer, regardless of whether it was hormone receptor–positive, HER2-negative, triple-negative, or HER2-positive, according to findings from a retrospective analysis of the I-SPY 2 trial that were presented at the 2025 ESMO Breast Cancer Congress.1

With a median follow-up of 4.66 years ctDNA status at diagnosis was shown to be a significant prognostic biomarker of DRFS (HR, 5.5; 95% CI, 2.4-13; P < .001). The 36-month DRFS rate in the overall population (n = 701) was 83.9% (95% CI, 81.1%-86.7%) and rose drastically in patients with negative baseline ctDNA (n = 131), at 96.8% (95% CI, 93.8%-99.9%). In patients with positive baseline ctDNA (n = 570) the 36-month DRFS rate was 80.9% (95% CI, 77.6%-84.2%).

“Among the prognostic risk factors available at diagnosis, a machine learning model revealed that pretreatment ctDNA concentration is the most important predictor of DRFS,” Mark Magbanua, PhD, a professional researcher in the Department of Laboratory Medicine at the University of California San Francisco, said in a presentation of the data.

Accurate prediction of metastatic recurrence at the time early breast cancer is diagnosed can inform optimal patient care. With respect to prognostication, clinical features have been shown to correlate with DRFS and ctDNA positivity during and after neoadjuvant therapy is associated with increased risk of metastatic recurrence.

As such, investigators evaluated the correlation between ctDNA concentration at diagnosis quantified by mean tumor molecules (MTM/mL) and DRFS in patients with high-risk early breast cancer enrolled in the I-SPY 2 trial.

The analysis included data from 712 patients with early-stage high-risk breast cancer who were treated with neoadjuvant therapy in I-SPY 2.

All patients had ctDNA results available at the time of diagnosis courtesy of the Signatera personalized, tumor-informed ctDNA assay. The correlation between ctDNA status and concentration at diagnosis vs DRFS was evaluated using Cox regressions. Machine learning was applied to determine the significance of ctDNA concentration relative to other variables that were available at the time of diagnosis for prediction of DRFS.

Patient Characteristics

The median age was 49 (range, 20-80). A total of 134 DRFS events occurred. Patients had either hormone receptor–positive, HER2-negative (n = 293); triple-negative (n = 236); or HER2-positive (n = 183) breast cancer that was grade I/II (n = 149), III (n = 368), or unknown (n = 195).

Clinical T stage was T1/T2 (n = 489), T3/T4 (n = 210), or unknown (n = 13). Clinical N stage was node negative (n = 299), node positive (n = 392), or unknown (n = 34), and MammaPrint was either H1 (n = 319) or H2 (n = 393). ctDNA status at diagnosis was predominantly positive (n = 579) vs negative (n = 133), and the median ctDNA concentration at diagnosis was 1.35 MTM/mL, 9.25 MTM/mL, and 0.87 MTM/mL in the hormone receptor–positive, HER2-negative; triple-negative; and HER2-positive subgroups, respectively.

Additional Results and Next Steps

ctDNA concentration at diagnosis was also shown to be associated with DRFS regardless of receptor subtype. When looking at ctDNA as a continuous variable a near-linear relationship was seen between the log of the hazard rate and ctDNA concentration, Magbanua said. Moreover, median MTM/mL was significantly higher in patients with recurrence, with P values of less than .0001 in the hormone receptor–positive, HER2-negative and triple-negative populations and equal to .0046 in the HER2-positive population.

Investigators also showed that ctDNA concentration at the time of diagnosis correlates with the level of recurrence risk. The ctDNA negative–population’s 36-month DRFS rate served as the reference for comparison against the ctDNA (n = 195, > 0 to ≤2 MTM/mL; n = 192, > 2 to ≤ 20 MTM/mL; and n = 183, > 20 MTM/mL) subgroups, for which the respective rates were 94.0% (95% CI, 90.7%-97.5%), 78.1% (95% CI, 72.3%-84.4%), and 69.6% (95% CI, 63.2%-76.7%). The respective HRs for DRFS were 2.4 (95% CI, 0.96-5.9; P = .062), 6.0 (95% CI, 2.6-14; P < .001), and 9.1 (95% CI, 3.9-21; P < .001).

“In comparing the survival curves we see that there’s a difference, indicating that further subgrouping could help refine risk stratification in ctDNA-positive patients,” Magbanua said.

Magbanua added that machine learning confirmed that ctDNA concentration at diagnosis was the most important variable for predicting metastatic recurrence. “Our model performed very well in the I-SPY 2 test set with high area under the curves as well as in a validation cohort composed of 131 real-world patients with breast cancer. It also performed very well with high accuracy in different subtypes in the I-SPY and real-world cohorts,” Magbanua explained.

“Ongoing analyses are modeling combinations of variables, including tumor response and ctDNA dynamics to optimize the prediction of DRFS,” Magbanua concluded.

Disclosures: Magbanua had nothing to disclose.

Reference

Magbanua MJ, Wolf D, Sayaman R, et al. ctDNA at diagnosis predicts metastatic recurrence in patients with high-risk early breast cancer. Presented at: 2025 ESMO Breast Congress; May 14-17, 2025; Munich, Germany. Abstract 5MO.