Supplementary Figure S2. Classification of cancer cells according to WD/DD liposarcoma signatures.

Heatmaps of expression of 29 key gene signatures in (A) SARC028 and (B) Stanford RNA sequencing data

11553 Background: Immune checkpoint inhibitors (ICIs) are an established treatment for many malignancies, with an emerging role in sarcoma. While the impact of concomitant medications (CM) on ICIs has been described in other cancers, it has not been evaluated in sarcoma. We assessed the association between CM use and ICIs outcomes in sarcoma patients. Methods: This pooled post hoc analysis included 7 investigator-initiated phase II ICI-based trials enrolling patients with metastatic sarcoma between April 1, 2017, and May 30, 2024. Patients receiving ≥1 dose of ICIs were included. CM taken within 30 days prior to treatment initiation were defined as baseline; PRN medications were excluded. CM were recorded at each cycle through end of treatment and categorized by medication class; supplements/herbal agents included vitamins, minerals, probiotics, botanicals, amino acids, and other dietary substances. For progression-free survival (PFS), baseline and on-treatment CM were analyzed as time-dependent covariates using Cox proportional hazards models. Overall survival (OS) analyses were limited to baseline CM. Logistic regression evaluated associations between baseline CM and objective response (ORR) and immune-related adverse events (irAEs). Toxicity was graded per CTCAE v5.0. Results: A total of 321 patients (median age, 58 years) were included. The most common histologies were liposarcoma (all subtypes, 22%), undifferentiated pleomorphic sarcoma (22%), and leiomyosarcoma (19%); 46% had received ≥3 prior lines of therapy. ICI was combined with cytokine/oncolytic therapies (43%), targeted agents (38%), or chemotherapy (19%). After a median follow-up of 47.4 months, ORR was 18%, median PFS was 3.9 months (95% CI, 3.09–5.32), and median OS was 19.9 months (95% CI, 17.5–22.9). Grade ≥3 irAEs occurred in 25% of patients, most commonly hematologic (6%) and hepatic (5%). On-treatment use of supplements/herbal agents was associated with longer PFS (HR 0.64, 95% CI 0.50–0.82; p<0.01), while on-treatment anti-infective use was associated with shorter PFS (HR 1.55, 95% CI 1.06–2.25; p=0.02); baseline anti-infective use was not associated with PFS. Baseline supplement/herbal use was associated with longer OS (HR 0.69, 95% CI 0.53–0.90; p<0.01), whereas baseline non-opioid (HR 1.54, 95% CI 1.10–2.16; p=0.01) and opioid analgesic use (HR 2.14, 95% CI 1.45–3.14; p<0.01) were associated with shorter OS. Baseline antihistamine use showed a trend toward increased gastrointestinal irAEs (OR 2.37, 95% CI 0.97–5.98; p=0.06). No CM class was associated with ORR. Conclusions: Concomitant medication use is associated with differential ICI outcomes in metastatic sarcoma and warrants prospective evaluation. CM Timing Outcome HR SH OT PFS 0.64 AI OT PFS 1.55 AI BL PFS 1.37* SH BL OS 0.69 Non-OA BL OS 1.54 OA BL OS 2.14 *Not statistically significant; SH = Supplements/herbal agents, AI = Anti-infectives, OA = Opioid analgesia, OT = on-treatment, BL = Baseline.

ABSTRACT Background Concomitant medications (CMs) influence outcomes in patients receiving immune checkpoint inhibitors (ICIs), but their impact in sarcoma remains undefined. We assessed the association between CM use and ICI outcomes in patients with advanced or metastatic sarcoma. Methods This pooled analysis included patients from seven investigator‐initiated phase II trials of ICI‐based therapy for sarcoma. CMs within 30 days of treatment were defined as baseline; on‐treatment exposure was captured longitudinally. The primary endpoint was progression‐free survival (PFS); secondary endpoints included overall survival (OS), objective response rate (ORR), and immune‐related adverse events (irAEs). Multivariable Cox models adjusted for age, ECOG performance status, histological subtype, treatment regimen, and race. Results Among 321 patients (median follow‐up: 47.4 months), in time‐dependent analyses, exposure to anti‐infective medications was associated with shorter PFS (adjusted hazard ratio [aHR] 1.54, 95% CI 1.05–2.27). Baseline use of vitamins/minerals/supplements/herbal products was associated with longer PFS (aHR 0.73, 95% CI 0.56–0.95) and OS (aHR 0.67, 95% CI 0.51–0.89). Baseline statin use was associated with longer PFS (aHR 0.64, 95% CI 0.47–0.87) but not OS. Baseline use of opioids was associated with shorter OS (aHR 1.71, 95% CI 1.12–2.61). No CM class was significantly associated with ORR or irAE occurrence. Conclusions CM use is associated with differential efficacy outcomes in sarcoma patients receiving ICIs. These findings highlight the need for prospective studies to define the impact of commonly prescribed medications on ICI outcomes and to optimize clinical trial stratification.

TPS11594 Background: Dedifferentiated liposarcoma (DDLPS) is a rare adipocytic malignancy characterized by frequent amplification of MDM2 and CDK4 and associated with poor outcomes, with a median overall survival of approximately 15 months. A phase II trial at Memorial Sloan Kettering Cancer Center demonstrated activity of the CDK4/6 inhibitor palbociclib in advanced WD/DDLPS, with a 12-week progression-free survival (PFS) rate of 57%, exceeding historical chemotherapy benchmarks (<35%) and leading to its inclusion in the NCCN Guidelines. Despite this, durability of benefit remains limited, and outcomes following progression are poor. Preclinical work from our group demonstrated that CDK4/6 inhibition in DDLPS induces cellular geroconversion from quiescence to irreversible senescence, accompanied by MDM2 downregulation. This process requires suppression of HRAS signaling and inhibition of the MAPK pathway and is associated with activation of the senescence-associated secretory phenotype (SASP). Preclinical models further suggest that dual CDK4/6 and MEK inhibition enhances SASP activation, providing a mechanistic rationale for combination therapy. Based on these findings, we initiated a phase Ib/II clinical trial of mirdametinib, a MEK1/2 inhibitor, in combination with palbociclib in patients with DDLPS (NCT06843967). Methods: This is an ongoing phase Ib/II, single-arm, open-label, single-center study evaluating the safety, tolerability, and efficacy of mirdametinib plus palbociclib in patients with unresectable, recurrent, or metastatic DDLPS (NCT06843967). Patients may enroll at any line of therapy, including first line, consistent with contemporary palbociclib use; patients with prior CDK4/6 inhibitor exposure are eligible for the phase Ib portion. Key eligibility criteria include ECOG performance status 0–2 and RECIST v1.1–measurable disease. Phase Ib uses a Bayesian optimal interval (BOIN) design with three dose levels to determine dose-limiting toxicities, maximum tolerated dose, and recommended phase II dose (RP2D), enrolling up to 24 patients. Phase II will assess efficacy at the RP2D, with the primary endpoint of PFS at 18 weeks by RECIST v1.1 in 30 patients. Enrollment began February 19, 2025. As of January 2026, accrual is ongoing at dose level 2. Enrollment to dose level 3 is anticipated in February 2026. Clinical trial information: NCT06843967 .

Histologic distribution by immunotype and response

Table S7. Clinical responses of advanced PDGFRA-mutant GIST patients treated with first-line imatinib.

BACKGROUND:Aggressive angiomyxoma (AA) is a rare, locally infiltrative mesenchymal tumor with frequent local recurrence. Optimal management remains poorly defined, particularly the roles of surgery, endocrine therapy, and observation. METHODS:We retrospectively analyzed patients with histologically confirmed AA treated at a single institution from 2000 to 2024. Clinical, pathologic, and treatment data were collected. Overall survival (OS), relapse-free survival (RFS), duration of treatment (DOT), and radiographic response by quasi-RECIST 1.1 were assessed descriptively. RESULTS:Forty patients were included; median age was 42 years, 87.5% were female, and median tumor size was 9.9 cm. Initial management was surgery in 30 patients, systemic therapy in 8, and observation in 2. Among patients initially managed with surgery, median OS from the time of first resection was not reached at a median follow-up of 68.6 months; estimated 10-year OS was 95% (95% CI, 68%-99%). Among patients with R0/R1 resection, median RFS was 49.8 months (95% CI, 21.0-64.7), with 3-year RFS of 56% (95% CI, 31%-75%). Tumor size and margin status were not significantly associated with recurrence. Thirteen patients received systemic therapy, predominantly endocrine-based; ORR was 25% and median DOT was 15.4 months. Two patients were initially observed, and three with residual macroscopic disease after R2 resection remained free from further intervention for prolonged periods. Radiation therapy was rarely used. CONCLUSIONS:AA was associated with excellent long-term survival despite frequent recurrence or subsequent intervention after surgery. Endocrine therapy showed objective activity in a subset of patients. Observation or delayed intervention may be feasible in selected, clinically stable patients with untreated or residual macroscopic disease, supporting individualized, function-preserving management focused on symptoms, morbidity, and patient preferences.

Risk stratification across the three main clinical subsets of malignant peripheral nerve sheath tumor (MPNST), neurofibromatosis type I (NF1-related), sporadic, and prior radiation therapy (RT), is based mainly on clinicopathologic parameters, such as size, grade, stage, and NF1 status. Moreover, no prior study investigated the additional impact of genomic alterations in the prognosis of high-grade MPNST using clinically validated DNA targeted next-generation sequencing (NGS) panels. Our goal was to integrate clinicopathologic and genomic parameters using an elastic-net penalized Cox proportional hazards machine learning model using OncoCast for risk prediction. Herein we perform comprehensive mutational and copy number profiling using Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT) on 81 primary localized high-grade MPNSTs (51% NF1-related, 38% sporadic, 11% RT-associated). The most common genomic alterations included NF1 (51% germline, 59% somatic), CDKN2A/B (62%), PRC2 components (SUZ12, EED) (53%), and TP53 (25%). Variables selected by OncoCast as significantly associated with survival were used to construct a three-tier risk stratification model for progression-free survival (PFS) and disease-specific survival (DSS). For PFS, patients with chr16 deletion were classified as high-risk, those with concurrent germline and somatic NF1 alterations as low-risk, and the remainder was assigned to the intermediate-risk group. For DSS, cases with fraction genome altered (FGA) > 50% were defined as high-risk, those with PRC2 abnormalities, CDKN2A deletion, TERT promoter mutation, or chr16 deletion as intermediate-risk, and cases lacking all the aforementioned alterations as low-risk. The high-risk group showed significantly inferior survival compared to the low-risk group (both PFS and DSS p < 0.001). Subgroup analysis showed that among NF1-related MPNST, co-occurring somatic NF1 mutation or WT TP53 was associated with superior PFS. Collectively, genomic alterations detected by clinical NGS panels provide potential new biomarkers for risk stratification that can be integrated with conventional parameters to provide improved prognostication and guide therapeutic strategies. © 2026 The Pathological Society of Great Britain and Ireland.

T cell subpopulations associated with shorter survival after accounting for histologic subtype and treatment

Mutation analysis of our PDGFRA-mutant GIST cohort reveals clustering around the exon 18, 842-codon position. A, The protein domains of PDGFRA (AL, activation loop; AP, ATP-binding domain; EC, extracellular domain) and the exon distribution of unique primary mutations seen in PDGFRA-mutant GIST (n = 1,379 cases). B, Breakdown of mutations seen in exon 18 PDGFRA-mutant GIST cases (n = 1,122/1,379). The gray-colored proportion in the chart indicates mutations with five or fewer reported cases (n = 90/1,122). C, Number of cases with exon 18 in/del mutations and the net number of deleted residues in the final protein sequence, along with the breakdown of the type of mutations observed with a 4-residue deletion. Red “X” denotes the number of amino acids inserted (X = 1, XX = 2). D, Proportion of cases with mutations that directly alter the 842 residue. E, Distribution of the amino acid occupying the 842-position in exon 18–mutant cases (n = 1,122). Colors in the legend correspond to the amino acid class of the 842-position residue (hydrophobic, polar uncharged, special case, positively charged, and negatively charged). [Portions of A were Created in BioRender. Khosroyani, H. (2026) https://BioRender.com/l88qy7p.]

Abstract EP300 and CREBBP encode p300 and CBP respectively, which are H3K27 acetylases. Both EP300 and CREBBP mutated in a subset of urothelial cancers (∼15% for EP300 or CREBBP). The EP300 and CREBBP genes have ∼60% sequence similarity and thus have been presumed to have largely overlapping functional roles in cell homeostasis and cancer pathogenesis. To characterize the role of EP300 mutations in urothelial cancer pathogenesis and to identify non-redundant roles of these paralogues, we generated EP300 and CREBBP isogenic knockout urothelial cancer cell lines and characterized low passage mutant and wild type for EP300 and CREBBP urothelial cancer-derived patient derived organoids. EP300 KO and loss-of-function mutation was associated with enhanced cell growth in soft agar, increased invasive potential in vitro and altered cellular metabolism. These gain-of-function phenotypes were mediated by enhanced JAK-STAT3 activation resulting from IL-6 trans-signaling, the proximal driver of which was increased transcription and production of IL-1á. Notably, isogenic BLCA cells with CREBBP knockout did not confer IL-1á hypersecretion or hyperactivation of the IL-6/JAK1/STAT3 signaling axis indicating that this is a phenomenon was specific to EP300 loss-of-function. Transcriptomic analysis of Parental, EP300 KO, and CREBBP KO RT112 clones revealed that CREBBP KO significantly depressed IL1A transcript levels which EP300 KO significantly elevated. Additionally, CREBBP inactivated BLCA lines could not upregulate IL1A expression following genotoxic stress in contrast to parental and EP300 null cell lines. Using an inducible short hairpin RNA construct targeting CREBBP, we also find that CREBBP knockdown rescued IL1A upregulation in EP300 ko clones that coincided with a significant growth defect. Pharmacologic inhibition via the p300/CBP specific inhibitor A485 also abolished IL1A upregulation and caused significant growth defect in EP300 KO cells. EP300 null clones were also significantly more sensitive to A485 treatment than either CREBBP null urothelial cell or parental cell lines. In sum, our results identify regulation of IL-1á-JAK-STAT3 signaling as a novel non-redundancy between EP300 and CREBBP that could be exploited therapeutically in patient with EP300 loss-of-function mutations. Citation Format: James August Rodrigues, Hikmat A. Al-Ahmadie, Sizhi P. Gao, Jiaqian Luo, Jacob Tallman, Fengshen Kuo, Merve Basar, Cansu Yol, Jordan Eichholz, Alejandra Lopez Rojas, Ecenur Turkay, Jonathan E. Rosenberg, Gopa Iyer, Eugene J. Pietzak, Dan Li, Dana Schoeps, Shipra Shukla, Zoe Jacobs, Chen Khuan Wong, Woo Hyun Cho, Ping Chi, David B. Solit, Yu Chen, . CREBBP drives tumorigenicity via aberrant IL-1α signaling in EP300 altered bladder cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 4757.

BACKGROUND:The cyclin-dependent kinase 4 and 6 (CDK4/6) inhibitor palbociclib delays disease progression in dedifferentiated liposarcoma (DDLPS) by inducing tumor cell quiescence or senescence, though most tumors ultimately progress. Preclinical data suggest CDK4/6 inhibition enhances intratumoral inflammation and may synergize with immune checkpoint inhibitors. METHODS:This non-randomized, open-label, phase 2 study was conducted at Memorial Sloan Kettering Cancer Center. Patients with metastatic or unresectable DDLPS, or those expected to benefit from systemic therapy prior to surgery, were eligible. Patients received palbociclib (125 mg orally, days 1-21 of a 28-day cycle) plus retifanlimab (500 mg intravenous flat dose every 4 weeks). The primary endpoint was to assess the best objective response rate (ORR) by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. RESULTS:30 patients were treated and evaluable. Median age was 61 years (range 36-81), 67% were male, and 63% were treatment-naive. The median follow-up was 14.8 months. ORR was 20% (95% CI 8% to 39%) and the clinical benefit rate was 53% (95% CI 34% to 72%). Median progression-free survival and overall survival were 4.8 (95% CI 1.71 to 15.7) and 27 months (95% CI 21.7 to not reached (NR)), respectively. Median duration of response was 18.4 months (95% CI 9.4 months to NR). Grade ≥3 treatment-related adverse events occurred in 61% of patients, including one Grade 4 neutropenia. There were no Grade 5 events. Retifanlimab and palbociclib were discontinued due to toxicity in 23% and 17% of patients, respectively. Tumor sequencing identified JUN amplification in 6 of 14 patients with progressive disease versus 2 patients with stable disease or partial response. Patients with progressive disease had a higher median copy number alteration burden compared with those with stable disease or partial response. CONCLUSIONS:Palbociclib plus retifanlimab demonstrated deep and durable responses in a subset of patients with advanced DDLPS, with an ORR exceeding that historically observed with either agent alone. Adverse events were generally manageable. Copy number alteration burden and JUN amplification merit further evaluation as potential biomarkers of resistance. TRIAL REGISTRATION NUMBER:https://clinicaltrials.gov/study/NCT04438824.

2661 Background: CRD3874-SI is a first in class, systemically administered, allosteric STING agonist that blocks STING’s proton channel activity differentiating it from previously developed STING agonists. The drug has demonstrated pre-clinical anti-cancer activity in several murine tumor models and has pharmacological properties distinct from earlier generation STING agonists. Methods: This is a single institution, open-label, phase I study of CRD3874-SI in patients with advanced solid tumors. The dose escalation study follows a standard 3+3 design. CRD3874-SI is administered intravenously once per week for 2 cycles. Cycle duration is 28 days. From cycle 3 onwards, continuous weekly treatment +/- one week break (week 4) may be considered. The primary objective is to assess the safety of CRD3874-SI by determining the maximum tolerated dose, recommended phase 2 dose and schedule of administration. Secondary objectives include examining the pharmacokinetics and pharmacodynamics (IP10 analysis) of CRD3874-SI and evaluating the efficacy of CRD3874-SI as determined by best objective response rate per RECIST v 1.1. Clinical trial information: NCT06021626. Research sponsor: Curadev Pharma, Inc. Results: As of January 5th 2026, 21 patients (sarcoma n=20, adenoid cystic carcinoma n=1) received treatment at four escalating dose levels (0.1-1.8mg/kg). The median number of prior lines of treatment was 4 (1-11). The median duration of treatment was 7 weeks (range: 1-32 weeks). 2 patients continue treatment. Reasons for treatment discontinuation include: progression of disease (n=16), toxicity (n=2), patient withdrawal (n=1). Treatment emergent adverse events (TEAEs) were manageable and reversible. Only low-grade cytokine related symptoms were reported. TEAEs possibly related to study treatment reported in >20% of participants and were mostly low grade include: fatigue (43%), chills (38%), nausea (38%), diarrhea (33% (G3 (10%)), headache (29%) and flu-like symptoms (24%). Low grade colitis not typical of auto-immune mechanism, responding well to temporary treatment pause +/- oral budesonide, were reported in 4 patients. One DLT at dose level 4 (G3 dyspnea) was observed. 19 patients are evaluable for efficacy. The best objective response per RECIST v1.1: confirmed partial response, n=1 (malignant phyllodes tumor); stable disease, n=9; progressive disease, n=9. Dose level 3 (0.9mg/kg) represents a biologically active dose (one confirmed PR and a second case with -27% tumor regression was observed). Dose proportional increase in AUC with concomitant increase in plasma CXCL10 levels were observed. Conclusions: CRD-3874-SI has demonstrated clinical activity with manageable safety. Dose level 0.9mg/kg is biologically and clinically active. A dose expansion phase at this dose level in is planned in >5 histology specific cohorts. Clinical trial information: NCT06021626 .

Detailed histologic diagnoses according to TME subtype

Supplementary Figure S3. T cell state contour plots.

Abstract Introduction: Identifying novel approaches that harness and augment anti-tumor immune response is an important area of ongoing research and drug development. Based on their performance in pre-clinical models, orthosteric activators of the Stimulator of Interferon Genes (STING) have been investigated in human trials. The activation of STING by its ligand cGAMP leads to both IFN dependent innate immune signaling as well as IFN independent pharmacology associated with STING's proton transport activity that leads to autophagy and pyroptosis. CRD3874 is a first in class small molecule allosteric STING agonist that binds to an allosteric site within STING's proton channel and decouples the IFN and non-IFN consequences of STING activation by blocking STING's non-IFN actions. CRD3874-SI, is a proprietary intravenous formulation of CRD3874, potent activator of all five human STING variants that has demonstrated promising pre-clinical anti-cancer activity in serval mouse tumor models when systemically administered as mono therapy or in combination with checkpoint therapy. Drug administration led to elevated plasma levels of CXCL10 and IFNβ in mice and monkeys. High IV doses up to 75mg/kg were systemically tolerated in cynomolgus monkeys. Methods: This is a single institution, phase Ia/b study of CRD3874-SI in patients with advanced solid tumors who received at least one line of prior therapy. The dose escalation phase will explore the safety and tolerability of CRD3874-SI following standard 3+3 design. CRD3874-SI is administered by intravenous infusion once per week for two cycles. From cycle 3 onwards, participants will received 3 or 4 consecutive weekly infusions, over a 28-day treatment cycle. The primary objective is to assess the safety and tolerability of CRD3874-SI by determining the maximum tolerated dose, recommended phase 2 dose and schedule of administration. Secondary objectives include further defining the safety profile, examining the pharmacokinetics and pharmacodynamics (CXCL10 analysis) of CRD3874-SI and evaluating the efficacy of CRD3874-SI as determined by the best objective response rate and clinical benefit rate per RECIST v1.1. Treatment will continue until disease progression or unacceptable toxicity. This study is currently open to enrollment. 21 patients have received treatment across 4 dose levels to date. Expansion cohorts are planned after determining the appropriate dose (RP2D). Clinical trial information: NCT 06021626. Research sponsor: Curadev Pharma, Inc. Citation Format: Ciara M. Kelly, Reinhard von Roemeling, Monali Banerjee, Viswatej Avutu, Olayade Babatunde, Lauren Banks, Ping Chi, Mark A. Dickson, Mrinal M. Gounder, Grace Gray, Camron Clark, Mary Louise Keohan, Robert G. Maki, Sujana Movva, Damon Reed, Kelly Schroeder, Li-Xuan Qin, Phillip Wong, Sandip Middya, Ritesh Shrivastava, Debjani Chakraborty, Rajib Ghosh, Sourav Basu, Arjun Surya, William D. Tap, Sandra P. D'Angelo. Phase I trial of CRD3874-SI, a systemically administered third generation allosteric STING agonist, in patients with advanced solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(8_Suppl):Abstract nr CT295.

Supplementary Table S3. Representativeness of Study Participants