Current Evidence for Circulating Tumor DNA in Sarcoma: Challenges and Opportunities for Clinical Application

ABSTRACT

Sarcomas represent a diverse group of mesenchymal tumors with high rates of recurrence after resection. While recent technical advances have enabled the detection of rare circulating tumor DNA (ctDNA) in other malignancies, the complexity and heterogeneity of sarcoma genomics have historically limited ctDNA in these cancers. This narrative review highlights the rapidly evolving evidence supporting potential clinical applications of ctDNA in common sarcoma subtypes including gastrointestinal stromal tumor, leiomyosarcoma, rhabdomyosarcoma, osteosarcoma, and Ewing sarcoma.

1. Introduction

Sarcomas are a heterogenous collection of rare bone and soft tissue tumors encompassing more than 100 recognized histological and molecular subtypes [1, 2]. Despite often having low tumor mutational burdens, sarcomas may harbor a wide array of genomic alternations [3, 4, 5]; this genetic diversity poses significant challenges for diagnosis, treatment, and disease monitoring [6, 7, 8]. While our evolving understanding of sarcoma oncogenes has facilitated the development of targeted therapies in a limited subset of these cancers [9], surgery with or without chemoradiation remains the cornerstone of treatment [10, 11, 12]. Unfortunately, local or metastatic recurrences remain common, and conventional imaging‐based surveillance methods are limited in their ability to identify early recurrences [13, 14, 15, 16, 17, 18]. Improved methods to assess real‐time treatment response, detect minimal residual disease, and efficiently identify early recurrences are urgently needed [19].

The presence of circulating cell‐free DNA (cfDNA) in malignant disease is well established, with seminal studies correlating serum cfDNA levels to tumor burden [20, 21]. The subset of cfDNA released from cancer cells following apoptosis, necrosis, or spontaneous secretion is referred to as circulating tumor DNA (ctDNA) and has increasingly garnered interest as a potential target for liquid biopsies (Figure 1A) [22, 23, 24, 25]. ctDNA fragments contain tumor‐specific somatic mutations that can be quantitatively measured, and ctDNA levels often closely correlate with tumor dynamics during treatment and recurrence across a variety of tumor types [26, 27, 28]. For instance, ctDNA levels have been found to decline following surgical resection in numerous solid tumor types including colon, breast, and non‐small‐cell lung cancer. Applications of ctDNA includes early diagnosis, monitoring of tumor burden, identification of clonal evolution throughout treatment, and surveillance for disease recurrence (Figure 2) [29, 30, 31, 32, 33, 34]. However, given the rare nature of and genomic variability across sarcoma subtypes, only a limited number of contemporary studies have included sarcoma patients [35, 36, 37, 38].

Figure 1.

ctDNA sources and detection methods. (A) Circulating tumor DNA (ctDNA) may be released from primary, metastatic, residual, or recurrent tumor cells through necrosis, apoptosis, or spontaneous secretion. (B) ctDNA, which contains the genetic information of tumor cells and thereby tumor‐specific mutations, offers the potential of a liquid biopsy. Using peripheral blood samples, ctDNA can be detected through either tumor informed or tumor agnostic assays. Tumor informed methods, such as PCR based assays, require prior knowledge of a patient's specific tumor neoantigen and custom designed primers. Alternatively, tumor agnostic methods typically use next generation sequencing to detect and quantify ctDNA without prior tumor‐specific information. Figure created with biorender.com.

Figure 2.

Schematic overview highlighting the potential clinical utility of ctDNA in sarcoma. The clinical use of ctDNA can be broadly divided into three distinct timepoints: (1) disease screening, early diagnosis, and/or establishment of preoperative baseline, (2) monitoring response following resection and/or detection of minimal residual disease (MRD) to inform the use of adjuvant therapy, and (3) disease surveillance and/or identification of potentially targetable genetic alterations in recurrent disease. Figure created with biorender.com.

In this narrative review, we provide an overview of the opportunities and challenges associated with the use of ctDNA in sarcoma for monitoring treatment response, detecting minimal residual disease, and informing the (de‐)escalation of therapy. We first spotlight the status of ctDNA in other solid tumors to provide a framework for potential clinical applications of this tool in sarcoma. We then provide a high‐level overview of sarcoma genomics and the technological considerations that have posed unique challenges for the application of ctDNA across major disease subtypes. With this perspective, we summarize the current state of ctDNA in soft tissue and bone sarcomas, focusing primarily on gastrointestinal stromal tumors (GIST), leiomyosarcoma (LMS), rhabdomyosarcoma (RMS), osteosarcoma, and Ewing sarcoma. While the use of ctDNA remains in the early investigational stages with current support stemming largely from small, retrospective studies and correlative analyses, the existing literature suggests that ctDNA in sarcoma may become a useful tool for guiding clinical decision making. Finally, we identify future directions and areas for further development in leveraging ctDNA in sarcoma.

2. Current Status of ctDNA in Non‐Sarcoma Malignancies

The use of liquid biopsy has long been appealing in cancer with the goal of identifying and applying disease biomarkers such as circulating tumor cells, cell free nucleotides, cell free DNA, and circulating tumor DNA to diagnose, characterize, and monitor disease [39]. In early‐stage disease, the use of ctDNA has been explored for establishing pre‐treatment tumor burden baseline; monitoring response and detecting persistent or recurrent disease after neoadjuvant therapy and/or surgical intervention; and enhancing radiographic or clinical surveillance after completion of standard of care treatment [40]. Measurements of ctDNA may also enable risk stratification after surgical resection to determine which patients require intensive monitoring or additional therapy. For instance, patients who remain ctDNA positive after surgery may be stratified as higher risk, leading to more intensive surveillance and/or treatment escalation, whereas patients with complete ctDNA clearance after resection might be stratified as lower risk and/or candidates for treatment de‐escalation such as omission of adjuvant therapy. In later‐stage disease, the use of ctDNA typically centers around monitoring response to systemic treatments or identifying genetic targets for directed therapy. The following overview of ctDNA in selected solid tumor types highlights the potential applications of ctDNA in cancer management.

The use of ctDNA has been explored in almost all stages of colorectal cancer (CRC) [41]. Prospective studies have demonstrated that ctDNA corresponds to tumor burden to enable real‐time monitoring, and postoperative ctDNA positivity has demonstrated strong prognostic value for recurrence in the adjuvant setting, potentially identifying recurrence up to a year before radiographic detection [42, 43, 44, 45, 46]. Several large studies have additionally shown that ctDNA monitoring may help to identify a subset of CRC patients who would benefit from adjuvant chemotherapy and to differentiate low‐risk patients in whom adjuvant therapy can be avoided without increasing recurrence [47, 48, 49, 50].

Similarly, in breast cancer, ctDNA represents an appealing biomarker for monitoring disease response to therapy and post‐treatment surveillance given its generally robust baseline detection [51, 52]. In early‐stage disease, favorable dynamics in ctDNA during neoadjuvant therapy are associated with pathological response and improved outcomes [53]. Postoperatively, patients with positive ctDNA on serial assessment demonstrate shorter relapse‐free and overall survival. Furthermore, the median lead time between ctDNA detection and radiographic evidence of recurrence following surgery has been reported in the range of 10–15 months. In more advanced breast cancer, large studies including have demonstrated the utility of ctDNA as a tool for profiling and identification of targetable genetic alterations [54, 55].

Finally, ctDNA has long been thought to correlate with disease burden and potentially clinical outcomes in lung cancer patients [56, 57, 58]. In non‐small cell lung cancer (NSCLC), specifically, there are multiple FDA approved liquid biopsy assays capable of detecting conserved oncogenic drivers that can inform targeted therapy [59, 60, 61].

In light of this evidence, current National Comprehensive Cancer Network (NCCN) guidelines support the use of ctDNA in specific circumstances for these tumor types. In colon cancer, ctDNA is recognized as a prognostic marker for disease recurrence, although neither treatment decisions nor surveillance based on ctDNA are currently recommended pending further confirmatory studies. In breast cancer, 2026 NCCN guidelines recommended the use of ctDNA assays to identify biomarkers for targeted therapies (e.g., evaluation for PIK3C1, AKT1, PTEN, ESR1 mutations in recurrent, unresectable, or metastatic hormone receptor positive, HER‐2 negative tumors). Similarly, NSCLC guidelines recommend restricting the use of ctDNA assays in advanced disease as a complementary approach to tissue‐based testing to perform molecular profiling; ctDNA testing is also recognized as a surrogate in cases where tissue is insufficient or for patients who are unable to undergo tissue sampling. These guidelines continue to rapidly evolve, with numerous active or recruiting phase III trials investigating the use of ctDNA to guide administration, de‐escalation, or escalation of adjuvant therapy (NCT04089631, NCT04068103, NCT03803553, NCT05174169, (NCT04093167, NCT06564844).

3. Technical Considerations for Sarcoma ctDNA

When considering the application of ctDNA to sarcoma, numerous advances in the understanding of molecular oncogenesis have highlighted the impressive diversity of tumor genomics, with sarcomas generally classified as either “simple” or “complex” according to their karyotypes [62]. Sarcomas with simple karyotypes (e.g., GIST, Ewing sarcoma) occur most commonly in young patients with relatively low overall tumor mutational burdens (TMB) and are often due to the presence of a specific oncogenic driver. Conversely, complex karyotype sarcomas (e.g., leiomyosarcoma, osteosarcoma) are characterized by many copy number variations (CNVs) and aneuploidy. Oncogenesis typically occurs in the setting of many genomic alterations, often arising from the loss of tumor suppressor gene function. These alterations are less likely to be conserved across patients and may make these tumors more prone to clonal evolution over time [63, 64]. Potentially related to their higher TMB, complex karyotype sarcomas tend to have more immune infiltration into the tumor microenvironment, which may correlate with better responses to immune checkpoint inhibition or other immune‐based approaches [62, 65]. The diversity across sarcoma subtypes creates unique requirements for the development and validation of ctDNA assays.

The use of ctDNA as a surrogate marker of tumor dynamics is predicated on the ability to detect tumor‐specific genomic alterations within all circulating cfDNA. This represents a fundamental technical challenge given that ctDNA may represent as little as 0.01% of cfDNA [26, 27]. Early work in ctDNA utilized digital PCR (dPCR) to identify rare mutations, relying on careful dilution of samples to probabilistically isolate single molecules followed by amplification and fluorescent tagging to delineate mutational status [66]. Throughput was subsequently improved with BEAMing (beads, emulsions, amplification, and magnetics) and droplet digital PCR (ddPCR) techniques [67, 68, 69]. Despite these advances in detection, these PCR based techniques are designed to determine whether a pre‐specified genetic marker is present within a sample and such approaches are therefore limited by the requirement that they target one or more specific, previously known mutations [70]. This is a particularly important limitation for primary bone sarcomas, which require decalcification for morphologic analysis, and therefore specialized tissue preparation to obtain DNA for molecular assays [71].

Alternatively, next generation sequencing (NGS) assays have the potential to address some of these shortcomings but have historically been limited by the difficulty of differentiating rare, meaningful variants from technical noise. More recent advances combining NGS with specialized bioinformatic pipelines have significantly increased the fidelity of NGS assays, allowing for the robust detection of mutant allele frequencies (MAF) as low as ~0.02% without prior knowledge of tumor mutational status [72, 73]. In contrast to PCR based methods, these NGS methods sacrifice some sensitivity in exchange for allowing de novo detection of genomic abnormalities (i.e., NGS approaches do not necessarily require knowledge of a pre‐specified mutation).

Overall, PCR and NGS assays have several important trade‐offs related to cost, throughput, sensitivity, generalizability, and need for prior knowledge regarding a particular tumor's genetic profile (Table 1, Figure 1B). The heterogeneity of sarcoma genomic alterations, particularly between simple and complex karyotype sarcomas, present a diverse set of technical problems for ctDNA detection and have consequently led to a variety of disease‐specific approaches. Similarly, clonal evolution in recurrent or metastatic disease may lead to false negatives during longitudinal monitoring and will require specific consideration as approaches to ctDNA in sarcomas are standardized. In this review, we highlight a subset of the early, small‐scale studies that are designed to quantify ctDNA during treatment and seek to primarily address questions related to using ctDNA for clinical decision making across sarcoma subtypes. These studies employ PCR, NGS, and combination assays in a range of different sarcoma subtypes. Notably, some investigators are assessing ctDNA across soft tissue sarcomas broadly, while other trials are focused on tumor subtypes.

Table 1.

Summary of relevant techniques for detection of ctDNA in sarcoma.

Technique	Method	Tumor informed or tumor agnostic?	Advantages	Disadvantages	Example application in sarcoma
Droplet Digital PCR (ddPCR)	PCR‐based	Tumor informed	Sensitivity and quantification; useful for serial monitoring	Requires known mutation or fusion target	KIT mutations in GIST [74]; monitoring patient‐specific EWSR1 fusion breakpoints in ES [75, 76, 77]; methylation‐based assay in OS [78]
Targeted NGS panels (e.g., hybrid capture panels)	NGS	Tumor agnostic (disease panel, not patient specific)	Detection of multiple mutations; scalable for clinical use	Limited to panel genes	Detection of KIT and PDGFRA mutations in GIST (e.g., Guardant360 CDx) [79]
Cancer Personalized Profiling by Deep Sequencing (CAPP‐Seq)	NGS	Tumor agnostic	Sensitivity to detect single nucleotide variants, insertions/deletions, copy number changes	Requires advanced bioinformatics workflow to distinguish signal from noise	Detection of translocations in RMS [80]
Tumor‐informed sequencing assays (e.g., WES‐informed personalized panels)	NGS‐based personalized assay	Tumor informed	Highly patient‐specific; useful for MRD detection	Requires tumor sequencing and customized assay	WES to identify tumor specific mutations to develop MRD panel in OS [81]
Ultra‐low pass whole genome sequencing (ULP‐WGS)	Genome‐wide sequencing	Tumor agnostic	Detects copy number variations without prior mutation knowledge; useful for complex karyotypes	Lower sensitivity for point mutations	Quantification of CNV in LMS [82] and OS [83]

Abbreviations: ES, Ewing Sarcoma; GIST, gastrointestinal stromal tumor; LMS, Leiomyosarcoma; MRD, minimal residual disease; NGS, next generation sequencing; OS, Osteosarcoma; PCR, polymerase chain reaction; WES, whole exome sequencing.

4. Gastrointestinal Stromal Tumors

GISTs represent the most common gastrointestinal sarcoma subtype and predominantly arise from activating point mutations in KIT and/or PDGFRA [84, 85, 86]. The tyrosine kinase inhibitor (TKI) imatinib revolutionized the clinical care of GISTs, with partial responses seen in more than half of patients with unresectable or metastatic disease [87]. Genetic alterations have important implications for treatment, as the locations of KIT mutations predict response to imatinib [88]. Secondary KIT and PDGFRA mutations directly lead to imatinib resistance [89, 90, 91], which has enabled the development of novel multikinase inhibitors for the treatment of imatinib‐resistant GISTs [92, 93, 94]. Several features make GISTs a promising early candidate for the clinical implementation of ctDNA [95, 96, 97, 98, 99, 100, 101, 102, 103, 104]: genomically, they have simple karyotypes with highly‐conserved and response‐predictive mutations, adjuvant therapy has been shown to reduce the risk of relapse in a subset of high‐risk patients who undergo upfront resection but identification of those patients likely to benefit remains challenging, and early detection of recurrent oligometastatic disease has been demonstrated to improve overall survival [105, 106, 107, 108].

In a phase Ib trial focused on the tolerability of rapidly alternating sunitinib and regorafenib, a combined ddPCR and targeted error correction sequencing (TEC‐seq) approach with serially collected ctDNA in 12 patients suggested that secondary KIT mutations mediated resistance to the combination therapy [74]. Secondary analysis of ctDNA collected during the phase I NAVIGATOR trial further demonstrated that patients with KIT activating loop mutations in exons 17 or 18, but without ATP binding pocket mutations in exons 13 or 14, had better outcomes when treated with avapritinib [109, 110]. More recently, the phase III VOYAGER trial compared avapritinib versus regorafenib as third‐line therapy in unresectable GIST. While the trial did not meet the primary end point of improving progression free survival [111], secondary analysis using ctDNA prospectively collected during the trial and the commercially available 74‐gene Guardant360 CDx panel revealed KIT and PDGFRA mutations in 75% and 5% of 386 enrolled patients, respectively, emphasizing the feasibility of clinical ctDNA monitoring with available technology [79]. Resistance mutations in the KIT ATP binding pocket in exons 13 or 14 were interestingly associated with decreased median progression free survival (PFS) in patients treated with avapritinib but not regorafenib, with this correlative analysis in a relatively large cohort suggesting that ctDNA can potentially guide personalized TKI selection for a subset of GIST patients [79].

Similarly, the phase III INTRIGUE trial compared the multitargeted TKI sunitinib with ripretinib, a switch control TKI that forces an inactive kinase conformation shown to have efficacy across a broad range of resistance mutations [112, 113], without a demonstrable improvement in PFS for patients with advanced GIST previously treated with imatinib [114]. However, exploratory analysis using a NGS assay with ctDNA from 362 patients prior to treatment in the study demonstrated significant differences in PFS for TKI sunitinib and ripretinib depending on the location of the KIT mutations‐ with KIT exon 11 + 13/14 mutations responding better to sunitinib and KIT exon 11 + 17/18 mutations responding better to ripretinib [115]. These findings are under further investigation in the INSIGHT phase III clinical trial designed to compare ripretinib and sunitinib for patients with KIT exon 11 + 17/18 mutations (NCT05734105) [116]. Together, these studies demonstrate the promise of ctDNA as a clinical decision aid for patients with early‐ and late‐stage GIST, including selecting candidates for (neo)adjuvant therapy, detecting resistance mutations, and guiding therapy selection. While these studies exploring ctDNA in GISTs represent some of the most robust evidence to date for ctDNA in sarcoma, it is important to note that these findings are largely correlative. Future prospective and randomized studies will therefore be critical to confirming the clinical utility of ctDNA in GISTs in guiding both treatment selection and surveillance.

4.1. Leiomyosarcoma

Leiomyosarcomas (LMS) are soft tissue sarcomas arising from dedifferentiation along the smooth muscle lineage and often occur in the extremities, retroperitoneum, or uterus. Survival depends on the primary site and tumor grade, with varied responses to chemotherapy [11, 117]. LMS is a complex karyotype sarcoma and is characterized by widespread genomic instability, including frequent structural rearrangements and inactivation of tumor suppressor genes such as TP53 and RB1 [118, 119, 120]. LMS tumors also have significant genetic diversity both within the primary tumor and between metastatic sites, with evidence that clonal seeding for metastasis may arise years prior to diagnosis [121]. The heterogeneity of LMS genomic alterations have historically made assessment of ctDNA challenging, though several recent NGS approaches have shown promise [95].

Early approaches for examining ctDNA in LMS included the application of a commercial ctDNA NGS approach to examine a predefined gene panel in 73 patients with metastatic LMS [122]. Alternatively, tumor informed ctDNA assays using personalized multiplex PCR primers for up to 16 patient‐specific single nucleotide variations (SNVs) identified from tumor whole‐exome sequencing, have demonstrated that detection of ctDNA correlated with disease progression in a cohort of 34 patients [123]. Combination approaches, such as mutational profiling using cancer personalized profiling by deep sequencing (CAPP‐Seq) in addition to CNV identification using genome representation profiling (GRP), further improved detection of ctDNA with mutant allele frequencies as low as 0.01% [124]. The development of ultra‐low pass whole genome sequencing (ULP‐WGS) has further enabled quantification of tumor mutations and CNVs without requiring prior knowledge of patient mutations, with one study detecting LMS ctDNA in 11/16 patients with progressive disease [125, 126]. This ULP‐WGS was applied to 167 samples from the SARC021 phase III clinical trial examining doxorubicin plus evofosfamide versus doxorubicin alone [127]. Secondary analysis of ctDNA samples was detectable in approximately half of 98 examined pre‐treatment samples, with detectable ctDNA levels correlating with lower likelihood of treatment response and poor survival [128]. Elsewhere, a retrospective analysis including 15 cases of LMS demonstrated strong concordance between ctDNA and radiographic findings, highlighting a potential complementary role for tumor‐informed assays in disease surveillance for soft tissue sarcomas such as LMS [129].

Despite their heterogeneity, together these studies demonstrate that detecting ctDNA in complex karyotype sarcomas is feasible and provide preliminary data to suggest that ctDNA assays may be used to track LMS disease progression, Further work is needed to standardize the methods used to detect ctDNA in complex karyotype sarcomas and determine whether such methods can be used to prospectively guide treatment decisions in the clinic.

4.2. Rhabdomyosarcoma

Rhabdomyosarcoma (RMS) is the most common type of soft tissue sarcoma in children and adolescents, arising from skeletal muscle precursors and generally divided into embryonal, alveolar, pleomorphic, and spindle‐cell subtypes [130, 131]. These subtypes of RMS include examples of both simple and complex karyotype. For example, alveolar RMS notably can be classified according to the presence of a PAX3 or PAX7 gene fusion with FOXO1 that is associated with poor overall survival [132, 133, 134]. The presence of these conserved fusion genes has enabled targeted PCR‐based assays to detect and quantify fusion‐positive ctDNA, which has been correlated with radiographic tumor volume during treatment in a case report [135]. Among RMS lacking the fusion gene, mutations in tyrosine kinase/RAS/PIK3CA signaling are commonly found despite low tumor mutational burdens; however, these mutations are less conserved and require alternative detection approaches [131, 132]. For example, ULP‐WGS was able to detect CNVs in pediatric sarcomas including alveolar RMS, which correlated to tumor dynamics in a small cohort [82]. Similarly, application of CAPP‐seq detected translocations in 4/4 RMS patients [80] and detection of cancer‐specific chromatin changes in 16 patients have shown potential in RMS [136, 137]. Alternatively, pleomorphic RMS frequently demonstrate TP53 mutations, CNVs, and an overall complex karyotype that pose distinct challenges for the development of ctDNA assays [138].

In a proof‐of‐concept pilot study, a ddPCR approach combined with whole exome sequencing (WES) was used to track tumor burden in a RMS patient derived xenograft (PDX) mouse model and successfully detected ctDNA in 21 out of 25 pretreatment patients with serial measurements correlating to treatment response [139]. Furthermore, a combination of ULP‐WGS and an RMS‐specific hybrid capture assay detected ctDNA in a subset of patients with both fusion‐positive (n = 49) and fusion‐negative (n = 75) RMS prior to treatment [140]. A similar approach using ULP‐WGS and a custom 36‐gene RMS panel revealed that secondary mutations in fusion‐positive RMS correlate with decreased overall survival across 18 patients [141]. Taken as a whole, these pilot studies provide preliminary evidence to suggest that ctDNA can be reliably detected in RMS and is potentially correlated with clinically meaningful outcomes, although these applications will require standardization and validation in prospective clinical trials.

4.3. Osteosarcoma

Osteosarcoma (OS) is the most common primary malignant bone tumor and disproportionally occurs in pediatric and young adult patients. A complex karyotype sarcoma with significant heterogeneity across patients, OS is often complicated by different genomic profiles across metastatic sites and recurrences even in the same patient [142, 143, 144]. Therefore, most ctDNA studies in OS rely on NGS approaches, with early analyses detecting ctDNA in approximately half of cases, using a combination of tumor informed and tumor agnostic assays (combined n = 92 OS patients) [145, 146, 147]. More recently, secondary analysis of plasma samples from 183 patients in the phase III OS2006 trial examined the ability of ULP‐WGS to detect CNVs in ctDNA and to risk stratify OS patients [83]. In multivariable risk modeling, the level of ctDNA at diagnosis was an independent risk factor for worse progression free and overall survival, which enabled the development of a prognosis prediction tool [83].

Tumor‐informed WES has also been leveraged to enable detection of minimal residual disease after surgery. In one study, WES was utilized to identify tumor specific mutations in 83 patients that were used to generate personalized ctDNA panel for MRD detection, enabling ctDNA detection in 13/59 post‐operative patients [81]. Serial monitoring with these personalized assays revealed that postoperative ctDNA detection was associated with worse event‐free survival. Furthermore, ctDNA was detected on average more than 90 days prior to cross‐sectional imaging recurrence in 5 patients with disease relapse [81]. Notably, there was significant mutational heterogeneity in this cohort. Alternative methods for detecting OS ctDNA have been proposed, including a customized methylation‐based ddPCR assay that successfully detected ctDNA in 50/72 pre‐operative and 5/17 post‐operative OS samples [78]. While these studies show that ctDNA can be detected in OS patients and potentially provide clinically useful information, further standardization and validation of these methods is needed. Furthermore, considering the genetic heterogeneity of OS and varying rates of detection across ctDNA techniques, further work is needed to optimize the reliability and scalability of these assays for clinical applications.

4.4. Ewing Sarcoma

Ewing sarcoma represents a relatively rare but aggressive sarcoma that primarily arises from bone in children and adolescents. Historically, a lack of serum biomarkers and difficulty ascertaining treatment response based on cross‐sectional imaging has limited the ability to risk stratify patients for adjuvant therapy. Surveillance strategies relying on frequent cross‐sectional imaging can also expose a predominantly adolescent population to repeated radiation, further emphasizing the need for serum‐based disease assessment. Genomic alterations involving EWSR1 gene rearrangements are the hallmark of Ewing sarcoma, most commonly resulting in the EWSR1‐FLI1 or, less frequently, EWSR1‐ERG or other fusion genes [148, 149]. Despite the prevalence of EWSR1 rearrangements in this disease, detection of these conserved fusion genes in ctDNA is complicated by intronic breakpoints unique to each patient. Therefore, approaches for detecting ctDNA in Ewing sarcoma often focus on the use of PCR‐based assays using patient‐specific primers [150]. While these breakpoints are unique to each patient, they are typically conserved between primary and recurrent tumors in an individual patient, potentially increasing the feasibility of longitudinal ctDNA monitoring [151].

ctDNA has been preliminarily evaluated in multiple contexts in Ewing sarcoma. ctDNA levels measured by ddPCR have been correlated with radiographic response during systemic therapy in a total of 40 patients across 2 cohorts [75, 76]. In a separate 20 patient cohort, ddPCR‐based detection of patient‐specific breakpoints in combination with PET imaging demonstrated a strong positive predictive value (88%) for disease remission and all recurrences were associated with increased ctDNA levels [77]. In a recent phase I clinical trial for recurrent/refractory Ewing sarcoma, ctDNA assays targeting patient‐specific fusion breakpoints similarly demonstrated a relationship between ctDNA levels and radiographic disease burden in 8 patients [152]. Likewise, in a 10‐patient phase II trial examining palbociclib and ganitumab in relapsed Ewing sarcoma, ctDNA levels declined with treatment [153].

In cases where patient‐specific fusions are not known or where somatic mutations in STAG2 or TP53 can be targeted, alternative NGS assays have been utilized in this population. For example, a sarcoma‐focused NGS assay identified ctDNA in more than half of patients and found that the detection of ctDNA at diagnosis was associated with worse 3‐year event free survival (n = 94 patients with Ewing sarcoma) [147]. Elsewhere, a hybrid capture sequencing assay was more sensitive than ULP‐WGS due to the low rates of CNVs in a small cohort of Ewing sarcoma patients, highlighting the ongoing work to optimize detection of ctDNA in this disease [146, 147]. More recent application of CAPP‐seq to detect common translocations and CNVs in pediatric sarcoma patient plasma has further improved the sensitivity of the assay and allowed for detection of complex rearrangements from ctDNA; in this study of 8 patients with Ewing sarcoma, ctDNA levels correlated with clinical response and relapse [80]. These studies show that a variety of approaches may be capable of detecting ctDNA in Ewing sarcoma patients and provide correlative evidence that the presence of such ctDNA may carry prognostic information. While promising, additional work is needed to optimize and standardize these assays. Finally, the clinical utility of ctDNA in Ewing sarcoma will require extensive validation in future prospective trials.

5. Conclusions

Improved understanding of the complex genomic landscape of sarcomas, paired with technological advances, have significantly expanded the potential clinical applications of ctDNA in sarcoma. Both PCR and NGS based assays now allow for detection of ctDNA across heterogenous sarcomas, ranging from widely conserved mutations in simple karyotype sarcomas such as GIST and alveolar rhabdomyosarcoma to tumor‐specific structural alterations in complex karyotype sarcomas [154]. Primarily through small early phase clinical trials and secondary analyses of larger sarcoma trials, these techniques have provided early evidence for the potential role of ctDNA in monitoring tumor response to therapy, tracking clonal tumor evolution, detecting minimal residual disease, and identifying evidence of disease recurrence in multiple sarcoma subtypes. These largely correlative studies together highlight how ctDNA monitoring can provide useful insight into disease dynamics and may ultimately help guide clinical decisions in real time.

While the evidence supporting ctDNA in sarcomas is not yet clinically actionable, it does provide an intriguing basis for future studies to assess the role of ctDNA in guiding sarcoma prognostication, utilization of adjuvant therapy, and personalized selection of therapeutic regimens in advanced disease. However, there is significant heterogeneity in ctDNA detection methodologies, quantitative thresholds, and interpretation that will need to be addressed in future studies. Standardization of these approaches within and across sarcoma subtypes will be crucial to future prospective trials seeking to validate the findings of the early, investigational studies outlined in this manuscript regarding the possible clinical utility of ctDNA in sarcoma. Critically these future trials will need to define consensus guidelines for what it means for ctDNA to be considered detectable and to identify specific use cases where ctDNA detection correlates with meaningful improvements in patient outcomes, which may be hampered by the current lack of effective therapies for some sarcoma subtypes when residual, recurrent, or metastatic disease is identified via ctDNA.

While ongoing studies (Table 2) will contribute a valuable understanding of assay validity, technical optimization to enable highly sensitive and specific ctDNA detection/quantification and establishing guidelines for ctDNA in various disease subtypes is needed. Future liquid biopsies for sarcoma may additionally include non‐plasma sources, including cerebrospinal and peritoneal fluid, which have shown promise in other cancers [155, 156, 157, 158, 159, 160, 161]. In summary, early clinical studies demonstrate a range of important applications to directly improve the way we manage both soft‐tissue and bone sarcomas. The use of ctDNA in sarcoma represents a rapidly advancing field that requires validation and refinement to reach more standard clinical utility.

Table 2.

Selected active or recruiting sarcoma trials with primary outcome involving ctDNA, identified as of May 2026 from https://clinicaltrials.gov.

National clinical trial number	Sarcoma subtype	Primary outcome(s)
NCT03896620	Soft tissue sarcoma; Retroperitoneal sarcoma	Measurement of ctDNA as variant allele fraction percentage at intervention and treatment timepoints in 50 adult patients across stage II–IV disease
NCT06958107	Bone and soft tissue sarcoma	Number of patients with ctDNA detection by low pass whole genome sequencing (LP‐WGS), estimated enrollment of 300 adults and children
NCT06068075	Ewing Sarcoma; Osteosarcoma	Prognostic value of ctDNA detection at prespecified timepoints, including a comparison of a novel research assay versus commercially available ctDNA testing (FoundationOne Liquid CDx), in 340 adults and children ≥ 12 years old (LEOPARD)
NCT04925089	Leiomyosarcoma	(1) Association between ctDNA and tumor imaging characteristics and (2) impact of treatment on detection of ctDNA in 40 patients with localized disease
NCT05653388	Leiomyosarcoma	(1) Association between ctDNA and tumor imaging responses (RECIST) and (2) correlation of ctDNA with progression free survival in 200 patients with metastatic disease

Synopsis

• ctDNA may offer a minimally invasive tool to monitor treatment response, detect minimal residual disease, and identify recurrence in cancer.

• Sarcoma genomic heterogeneity has historically limited ctDNA detection, though newer technologies are improving assay sensitivity.

• While early studies suggest ctDNA has potential applications across multiple sarcoma subtypes, more robust evidence is needed to inform practice guidelines and clinical application.

Goodsell K. E., Carter J. A., Abrams H. R., Mogal H., Loggers E. T., and Sharib J., “Current Evidence for Circulating Tumor DNA in Sarcoma: Challenges and Opportunities for Clinical Application,” Journal of Surgical Oncology 134 (2026): 490‐501, 10.1002/jso.70326.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.