Congruence between circulating tumor DNA and radiographic response during treatment of advanced sarcoma patients.

e23531 Background: Imaging assessments remain the standard of care to monitor therapy response or disease recurrence in advanced sarcoma. It remains unclear whether tumor- specific ctDNA (circulating tumor DNA, Signatera by Natera) adds predictive value to radiography, given the histologic diversity and limited understanding of tumor DNA shedding in sarcoma. Methods: We retrospectively assessed the congruence of ctDNA and imaging at serial timepoints in recurrent/metastatic sarcoma patients 1) undergoing regular interval imaging, with 2) ctDNA assessed +/- 30 days of scans, and 3) ≥ 2 ctDNA assessments. Congruence was achieved if A) ctDNA levels decreased and scans showed clinical benefit (stable disease or partial/complete response (PR/CR)) or if B) ctDNA levels increased and scans showed progression. A negative ctDNA with scans showing post-treatment CR was considered congruent and each ctDNA/scan pairing was evaluated independently for congruence. Results: 13 patients (7 male/6 female; median age 52yrs, range 33-73) treated from 5/2022-1/2025 with recurrent/metastatic leiomyosarcoma (n = 5), angiosarcoma (n = 3), and alveolar rhabdomyosarcoma, alveolar soft part sarcoma, Ewing’s sarcoma, osteosarcoma and sclerosing epithelioid fibrosarcoma (n = 1 each) met evaluation criteria. Treatment included chemotherapy (11/13), immunotherapy (4/13), radiation (2/13), and/or ablation (2/13). The median follow-up time was 370 days (52-909) with 4 median ctDNA assessments (2-14) and 2 median treatments (1-5) per patient. 69 ctDNA draws met criteria to assess congruence. CtDNA was congruent in 55/69 (86%) of assessments (sensitivity 69.2% (52.4-83.0%); specificity 93.3% (77.9-99.2%)). PPV and NPV were 93.1% (77.7-98.1%) and 70.0% (59.1-79.1%), respectively. Of the 14 incongruent values, 2 were positive ctDNA values with radiographic CR. One showed disease on follow-up scans, and the other ctDNA converted to negative with CR seen on scans. Amongst the remaining 12/14 values, 9 (5 while on active treatment) showed a negative ctDNA with active disease/progression and 3 showed decreasing ctDNA with progression on scans. In 8/12 cases follow-up ctDNA increased in congruence with scans and in 2 cases follow-up scans showed CR, congruent with negative ctDNA. In 2 cases, follow-up assessments are pending. Conclusions: In this analysis, we mimicked real-world clinical conditions by treating each timepoint as an independent assessment of ctDNA vs scans. The findings suggest that tumor-specific ctDNA may be a valuable tool for disease monitoring in recurrent/metastatic sarcoma on treatment, and efforts in additional ctDNA follow-up, further analysis of ctDNA vs scan timing, and stratification by histology are ongoing. The incongruency between ctDNA and radiographic response supports further studies on DNA shedding kinetics, therapy impacts on ctDNA, and variations by sarcoma histology.