The clinical value of tumor-informed minimal residual disease detection in sarcoma

11510 Background: The high recurrence and metastasis rates lead to poor survival of patients with sarcoma. At present, circulating tumor DNA-based minimal residual disease (MRD) has been studied in varies solid tumors and proved to be valuable in guiding treatment and predicting recurrence, but its application in sarcomas is rarely reported. Methods: Tumor tissue samples were used for whole exome sequencing (WES) and clonal mutations were selected for the customized MRD panel. Blood samples were collected for MRD detection. Experiments were conducted at OrigiMed, a College of American Pathologists accredited and Clinical Laboratory Improvement Amendments certified laboratory. Results: A total of 84 sarcoma patients including 47 bone tumor (BT), 32 soft-tissue sarcoma (STS), and 5 undifferentiated small round cell sarcoma (USRCS) patients were enrolled. Patients consisted of 49 (58.3%) males and 35 (41.7%) females, with a median age of 23 years old (1-85 years old). Based on WES, we found that 64% (4936/7715) of the mutated genes were unique to each patient, and 70.4% (859/1220) of selected tumor-informed single nucleotide variants (SNV) were variants with unknown significance, suggesting that tumor-informed MRD is superior to panel-based MRD in sarcomas. In this cohort, 78.6% (66/84) of patients successfully constructed MRD panels, including 85.1% (40/47) of BT patients, 75% (24/32) of STS patients, and 40% (2/5) of USRCS patients. The main reason for failure was the scarcity of SNVs detected. Preoperative MRD detection was performed on 6 cases and revealed a positive rate of 100%. MRD detection was performed twice on 2 patients during preoperative neoadjuvant therapy. For the patient with constant allele frequencies detected throughout the treatment course, imaging showed tumor enlargement; While for another patient with a trend of reduction in allelic frequency, imaging showed tumor shrinkage and necrotic tumor tissue was observed in surgery one month later. These two patients had not recurred after surgery at the time of analysis. Post-surgery MRD detection were performed on 27 patients, of which 1 STS and 7 BTs were positive (30%, 8/27). Three of them confirmed recurrence by imaging at 12, 30, and 101 days after detection of positive MRD, respectively. However, there were still no imaging recurrence in the other patients so far, including the 5 remaining patients with positive MRD and 19 patients with negative MRD. Follow-up is still ongoing. Conclusions: MRD tests had been successfully performed in most of the sarcoma patients. Due to the heterogeneity of sarcomas, about 64% of mutated genes of sarcomas are unique individually, which may be more suitable for monitoring with tumor-informed MRD panel. Tumor-informed MRD could be used to predict the efficacy of preoperative neoadjuvant therapy and discover recurrence earlier than imaging in sarcoma.