(P 04) EARLY ON-TREATMENT CIRCULATING TUMOR DNA DYNAMICS TO PREDICT RESPONSE IN PATIENTS WITH SARCOMA

Objective: ctDNA (circulating tumor DNA) holds promise as an emerging biomarker for risk stratification and treatment response monitoring across multiple cancers. We explored early on-treatment ctDNA dynamics to predict response in patients (pts) with soft tissue and bone sarcomas receiving cytotoxic chemotherapy.

Methods: 22 pts with sarcoma undergoing chemotherapy at Wilmot Cancer Institute through 12/31/24 with serial tumor-informed ctDNA testing (SignateraTM, Natera, Inc.) were retrospectively evaluated; 4 had a second set of treatment (Tx). We additionally prospectively enrolled 12 patients for banked ctDNA analysis at early timepoints on chemotherapy and interrogated these with Natera; 6 met inclusion criteria for a total of 32 Tx. Clinical responders (CR) were classified as those achieving a complete or partial response on first follow up imaging. Pts with progressive disease were classified as non-responders (NR), and those with mixed response or stable disease were considered intermediate responders (IR). ctDNA clearance (ND) was approximated to 0.01 MTM/mL for log-fold calculations.

Results: Of the 32 Tx, 14 led to CR, 10 NR, and 8 IR. Early on-Tx ctDNA dynamics were highly associated with response, with CRs showing a statistically significant decrease in ctDNA levels compared to NRs (p=0.001). CRs achieved a median of -1.9 (ND to -0.02) and -2.6 (ND to -1.3) log10 fold decrease in ctDNA levels by cycle 1 & 2, respectively. Conversely, NRs were associated with a median of -0.3 (-1.0 to 0.16) and -0.3 (-1.2 to 0.21) fold decrease by cycle 1 & 2, respectively. A value of -1.25 log fold fall in ctDNA by cycle 2 separated CR from NR. Of 13 pts with ctDNA clearance, none were in NR group (100% NPV).

Conclusion: Early ctDNA dynamics, particularly log-fold reduction of -1.25 or greater after 1-2 cycles predicts Tx response. ctDNA may offer clinical utility for interpreting equivocal scans, bony lesions or infiltrative subtypes such as angiosarcoma. Prospective validation in larger cohorts is warranted.