Supplemental Figure S3 shows peripheral blood and serum analysis of patients in Cohort 0 with irAEs.

Figure S7 shows clinical outcomes related to neoadjuvant RT + BO-112, related to Figure 7.

Supplementary Table S4 shows patient characteristics of surgical specimens from patients with sarcoma undergoing routine standard of care (n=44).

Supplementary Table S2 shows patient characteristics of 14 patients enrolled in an exploratory Phase 1 clinical trial to evaluate the safety of integrating intratumoral BO-112 with nivolumab and hypofractionated RT in the neoadjuvant setting for patients with resectable, high-risk soft tissue sarcoma (NCT04420975).

Supplemental Figure S5 shows impact of BO-112 and RT on intratumoral immune composition in human specimens.

Supplementary Video S1 shows study patient walking after limb salvage surgery as part of neoadjuvant RT and BO-112 therapy.

Low-grade fibromyxoid sarcoma (LGFMS) is a deceptively bland spindle cell neoplasm with malignant potential, most commonly arising in the extremities and trunk, and limited to at least110 cases in the head and neck. This report describes the first molecularly confirmed LGFMS of the floor of the mouth demonstrating a FUS::CREB3L2 gene fusion. An 18-year-old male presented with a painless floor of mouth mass of 1-year duration that was clinically suspected to represent a ranula. The mass was subjected to excisional biopsy. Histological examination was salient for an infiltrating proliferation of bland spindle cells arranged in whorled patterns within a collagenous stroma with focal myxoid change. Immunohistochemical analysis revealed diffuse MUC4 expression, narrowing the diagnosis. Targeted RNA sequencing identified a FUS::CREB3L2 fusion involving FUS exon 6 and CREB3L2 exon 5, confirming LGFMS. At 1-year follow-up, the patient showed no clinical evidence of disease and no detectable circulating tumor DNA. This case highlights the critical role of molecular testing in the diagnosis of LGFMS and emphasizes the importance of appropriate management strategies, including aggressive local control and long-term surveillance, given the tumor’s risk of late recurrence and distant metastasis.

Supplemental Figure S4 shows analysis of patient tumors and plasma, related to Figure 4.

Supplementary Data is an excel file with supplementary tables S1, S3, and S6-S9 compiled. Supplemental Table 1 shows complete adverse events and relatedness to BO-112 or nivolumab by cohort. Supplemental Table 3 shows overall treatment emergent adverse events. Supplemental Table 6 shows key genes for annotation of myeloid clusters in scRNA-seq. Supplemental Table 7 shows top 100 differentially expressed genes for each myeloid cluster. Supplemental Table 8 shows annotation of T cell subsets by top differentially expressed genes. Supplemental Table 9 shows representativeness of Study Participants

Supplemental Figure S6 shows intratumoral T cell evolution is facilitated by clonal replacement by less exhausted T cells, related to Figure 6.

Supplementary Table S10 shows individual patient level clinical outcomes

Supplemental Figure S1 shows focal RT using image guidance and treatment planning to target the tumor and spare local lymph node

Neoadjuvant immune checkpoint blockade (ICB) and radiotherapy (RT) improve disease-free survival in select patients with soft-tissue sarcoma (STS). However, most STS are myeloid-rich and lack preexisting T cells associated with ICB response. In preclinical models, we observed that intratumoral BO-112 [nanoplexed polyinosinic: polycytidylic acid (poly I:C)] engages myeloid cells that persist after RT, ultimately enhancing T cell-dependent tumor control. We evaluated BO-112 and hypofractionated RT, with or without nivolumab, in 14 patients with high-risk STS in a phase I neoadjuvant trial. Consistent with its immunologic potency, the triple combination induced rare immune-related adverse events (myositis-myocarditis-myasthenia gravis spectrum), mitigated by BO-112 and nivolumab dose adjustment. BO-112 and RT reprogrammed tumor-associated myeloid cells toward antigen-presenting states, promoted clonal replacement by less exhausted T cells, and enhanced malignant cell depletion compared with standard RT. These immunologic changes coincided with encouraging disease control in a small, high-risk cohort, supporting further clinical development. SIGNIFICANCE:Intratumoral BO-112 and hypofractionated RT activate systemic T-cell immunity in mouse models and in a phase I neoadjuvant study of high-risk, resectable sarcoma. Engaging myeloid cells with BO-112 represents a potent strategy with RT to replete T cell-deficient tumors and expand the benefits of neoadjuvant ICB.

Supplementary Table S5 shows patient characteristics of surgical specimens from patients with sarcoma undergoing pre-operative, hypofractionated RT alone as part of a separate Phase 2 clinical trial (NCT02701153).

Supplemental Figure S2 shows intratumoral administration is necessary for optimal anti-tumor efficacy of BO-112

BACKGROUND:Surgery is potentially curative for solid cancers; however, in cases of incomplete surgery, the impact of surgery on immune surveillance in the residual tumor microenvironment is not known. We sought to understand how surgery impacts immune populations in a residual tumor and correlates with overall survival in patients with primary pleomorphic liposarcoma. METHODS:This retrospective cohort study was conducted by searching the UCLA Sarcoma Program database for all patients with a histologic diagnosis of primary pleomorphic liposarcoma from 1995 to 2015. Patient follow-up was carried out through 2021. Patients were stratified by completeness of initial surgery: microscopically complete (R0), microscopically incomplete (R1), and grossly incomplete (R2). Six out of seven patients with an initial R2 resection underwent short-interval re-resection to negative margins within 120 days (R2-to-R0). We used immunofluorescence microscopy to characterize changes in immune populations of the tumor microenvironment. RESULTS:On multivariate analysis of this 32-patient cohort, age, tumor size, and R2-to-R0 resection were significantly associated with mortality. The hazard ratio for mortality after R2-to-R0 resection was 109 (p value < 0.01). The median overall survival for patients with R2-to-R0 resection was 2.0 years compared to 8.5 years for an upfront R0 resection (p value < 0.001). Immunofluorescence on four pairs of initial and re-resected tumors revealed a postoperative accumulation of suppressive myeloid and T regulatory immune populations in the residual microenvironment. DISCUSSION:We found that an initial incomplete surgery correlated with the accumulation of suppressive immune populations in the residual tumor microenvironment and mortality-a phenomenon we call hyper-progression of disease. Our findings have implications for therapeutically targeting immunosuppressive populations in the perioperative period to improve patient survival.

BACKGROUND:Chordomas are rare ectodermal bone malignancies derived from transformed notochordal remnants. Histologic variants include conventional (80%-90%), chondroid (5%-15%), and dedifferentiated (2%-8%). Because chordomas are relatively resistant to chemotherapy and radiotherapy, novel targeted agents are needed to expand treatment approaches and improve outcomes. This study analyzes the genomic landscape of chordoma and identifies potential pathogenic and druggable targets. METHODS:Eighty-six tumor samples derived from chordoma patients treated at Massachusetts General Hospital, University of California, Los Angeles, and the University of Miami were included. Tumor specimens were sent for comprehensive molecular profiling using next-generation sequencing. The most frequently mutated genes were identified and categorized by subtype, and microsatellite instability and programmed death ligand-1 (PD-L1) staining were assessed. RESULTS:Histologic subtypes included 70 conventional (81.4%), nine chondroid (10.5%), and seven dedifferentiated chordomas (8.1%). The most common mutations were cyclin-dependent kinase inhibitor 2A/B (CDKN2A/B) (28 of 86, 33%), low-density lipoprotein receptor-related protein 1B (10 of 86, 12%), polybromo-1 (9 of 86, 11%), and epidermal growth factor receptor (EGFR) (8 of 86, 9%). By subtype, CDKN2A/B mutation was most common in conventional chordoma (24 of 70, 34%), and chondroid chordoma (3 of 9, 33%). CDKN2A/B and EGFR mutations were most common in dedifferentiated chordoma (2/7, 29%). Microsatellite instability was not detected in seven of 69 (10.1%) samples. PD-L1 staining of tumor and immune cells was scarce, with scores <1 in 38 of 41 (92.7%) and 22 of 25 (88%) patients, respectively. CONCLUSIONS:This study provides a robust, high-dimensional sequencing assessment from 86 chordoma tissue samples and a descriptive overview of the genomic landscape of this rare, difficult to treat malignancy. Future studies should include in vitro assessment of gain and loss of function of frequently altered pathways to validate these findings.

This nonrandomized clinical trial evaluates the long-term toxic effects and clinical outcomes associated with a shorter 5-day, dose-equivalent preoperative radiation therapy (RT) regimen among patients with soft tissue sarcoma. QuestionWhat safety and clinical outcomes are associated with a shorter 5-day preoperative radiotherapy regimen for extremity and truncal soft tissue sarcoma?FindingsIn this nonrandomized phase 2 trial of 110 patients, median follow-up for the initial and expansion cohorts was 64.2 and 30.0 months, respectively. Major wound complications occurred in 33 of 110 patients (30.0%); at 2 years, 14 of 74 evaluable patients (18.9%) developed grade 2 or higher toxic effects. Two-year local control was 92.4% (95% CI, 86.3%-96.5%), and time to wound closure exceeded 6 months for 15 of 110 patients (13.6%).MeaningThese findings suggest 5-day preoperative radiotherapy was associated with favorable long-term clinical outcomes, but its potential impact on time-to-wound closure should be assessed in a randomized study. ImportanceStandard preoperative radiotherapy (RT) for high-risk soft tissue sarcoma (STS) is delivered over 5 weeks, which can be a logistical challenge for patients.ObjectiveTo evaluate the long-term toxic effects and clinical outcomes associated with a shorter 5-day, dose-equivalent preoperative RT regimen.Design, Setting, and ParticipantsPhase 2, single-group nonrandomized trial with an initial cohort (April 2016 to May 2018) and expansion cohort (October 2018 to May 2023) at a single academic center in the US. Participants were patients with histologically confirmed extremity or trunk STS recommended to undergo standard preoperative RT and surgery. Patients with planned neoadjuvant systemic therapy who were enrolled in the expansion group were excluded from this analysis. Analysis was conducted September 2024 to August 2025.InterventionA total of 30 Gy in 5 fractions were delivered preoperatively.Main Outcomes and MeasuresThe primary end point was 2-year grade 2 or higher radiation toxic effects. Secondary end points included major wound complications (MWC), local failure, distant progression, and overall survival.ResultsA total of 110 patients were treated with preoperative RT and surgery (42 patients [38%] were aged 65-79 years; 64 [58%] were male; 75 had tumors of the lower extremity [68%], and 64 patients [58%] had high-grade disease). The initial cohort accrued 50 patients who underwent surgery. The expansion cohort accrued 83 patients; 60 of 83 were treated without neoadjuvant chemotherapy and were included. Median (IQR) follow-up was 37.3 (20.1-60.6) months, including 64.2 (36.3-74.1) months for the initial cohort and 30.0 (13.5-40.2) months for the expansion cohort. At 2 years, 14 of 74 evaluable patients (18.9%) developed grade 2 or higher toxic effects (10 patients [25.0%] for the initial cohort and 4 patients [11.8%] for the expansion cohort). MWCs occurred in 33 of 110 patients (30.0%); (17 [34.0%] for the initial cohort and 16 [26.7%] for the expansion cohort). Time to wound closure exceeded 6 months for 15 patients (13.6%), including 12 of 29 patients (41.4%) who underwent local tissue advancement flaps. Two-year local control adjusting for competing risk of death was 92.4% (95% CI, 86.3%-96.5%). There were 3 (2.7%) bone fractures and 5 (4.5%) amputations.Conclusions and RelevanceThis nonrandomized clinical trial of ultrahypofractionated preoperative RT identified durable local control with MWC and favorable late grade 2 or higher toxic effects rates. Randomized data are necessary to differentiate the safety profiles of various fractionation regimens, especially duration of wound healing.Trial RegistrationClinicalTrials.gov Identifier: NCT02701153