e23554 Background: Radiographic imaging remains the standard of care for disease surveillance and response assessment in soft tissue sarcoma (STS). Circulating tumor DNA (ctDNA) has emerged as a potential biomarker in solid tumors; however, its feasibility and detectability in STS remain incompletely defined due to histologic heterogeneity and variable tumor DNA shedding. We evaluated ctDNA detectability and its relationship to radiographic disease status in the most common STS subtypes: undifferentiated pleomorphic sarcoma (UPS) and leiomyosarcoma (LMS). Methods: We retrospectively analyzed patients with UPS or LMS who underwent serial ctDNA testing using a personalized assay (Signatera) alongside routine radiographic surveillance between 2022 and 2025. Any detectable mean tumor molecules (MTM) >0 was considered positive. ctDNA results were paired with the closest imaging within ±30 days and assessed using radiology report–based determinations of disease status. Exploratory sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were estimated per paired assessment using radiographic progression or recurrence as the reference standard. Results: 13 patients were included (7 female, 6 male), with a median age of 62 years (range 28–75); 7 patients (54%) had LMS and 6 (46%) had UPS. A total of 59 ctDNA–imaging paired assessments were analyzed (median 4 per patient). Overall ctDNA detectability was low and histology-dependent. Baseline ctDNA was detectable in 2 of 13 patients (15%), both with LMS. During longitudinal follow-up, ctDNA remained persistently undetectable in 10 of 13 patients (77%), including all UPS patients and 4 of 7 LMS patients. Detectable ctDNA occurred almost exclusively in LMS and was often observed in association with radiographic disease progression. Across all paired assessments, 46 of 59 (78%) were concordant. Discordance included false-negative ctDNA (negative ctDNA with radiographic disease/progression; n=9) and transient ctDNA positivity without immediate radiographic progression (n=4). Where longitudinal follow-up was available, subsequent ctDNA or imaging assessments frequently demonstrated eventual alignment; lead-time analysis for transient positives was limited by sample size. Using radiographic progression as the reference standard, exploratory per-pair test characteristics demonstrated a sensitivity of approximately 64%, specificity of 88%, PPV of 80%, and NPV of 77%. Conclusions: In this real-world cohort, ctDNA detectability differed by histology and was observed predominantly in LMS. When detectable, ctDNA trends often aligned with radiographic disease progression or recurrence, supporting a potential complementary role in disease monitoring for selected patients. These findings require confirmation in larger, prospective studies.
Orosco-Ttamina et al. (Thu,) studied this question.