KEY POINTS
- Investigators analyzed 313 plasma samples from 106 patients enrolled in the randomized phase II SU2C-SARC032 trial of high-risk localized soft tissue sarcoma. Patients received neoadjuvant radiotherapy and surgery with or without perioperative pembrolizumab.
- Instead of a fixed mutation panel, each patient received a tumor-informed personalized assay based on tumor and germline sequencing. A median of 46 somatic variants per patient was tracked.
- Baseline ctDNA was detectable in 85% of patients, despite the generally low mutational burden of soft tissue sarcoma. Median allele fraction among ctDNA-positive patients was only 0.13%.
- Higher baseline ctDNA correlated with greater tumor volume and grade 3 disease. Pretreatment ctDNA remained independently associated with disease-free survival after adjustment for treatment arm and established clinical factors (HR 1.74 per log10 increase, 95% CI 1.25–2.42; p=0.001).
- ctDNA also provided an early treatment-response signal. Post-radiotherapy/preoperative ctDNA level independently predicted DFS (HR 1.75, 95% CI 1.27–2.41; p=0.0006). Conversion from detectable to undetectable ctDNA occurred in 71% with pembrolizumab versus 53% in controls, although this difference was not statistically significant (p=0.16).
- At 3 months after surgery, ctDNA remained independently associated with DFS (HR 1.40 per log10 increase, 95% CI 1.08–1.80; p=0.01). Detection of multiple circulating tumor clones at this point was also associated with poorer DFS, suggesting that ctDNA architecture may add information beyond absolute level.
- These were exploratory correlative analyses: SU2C-SARC032 was not powered for the biomarker endpoints, sample availability decreased longitudinally, and statistical significance thresholds were not prespecified. The data establish prognostic association, not evidence that changing treatment according to ctDNA improves outcomes.
CLINICAL TAKEAWAY
This is one of the more convincing ctDNA datasets in localized soft tissue sarcoma because the signal persists before treatment, after neoadjuvant therapy and after surgery, while adding information beyond conventional clinical variables. The next step is not simply more prognostic validation: it is an interventional trial testing whether escalation or de-escalation based on ctDNA actually improves patient outcomes.