Browse Articles

XALience: A phase 3, open-label, multicenter, randomized study of xaluritamig plus abiraterone vs investigator’s choice in patients with chemotherapy-naïve metastatic castration-resistant prostate cancer.

TPS5144 Background: In the first-in-human (FIH) study, xaluritamig, a T-cell engager directed at six-transmembrane epithelial antigen of the prostate 1 (STEAP1), demonstrated potent antitumor activity with a manageable safety profile in patients with metastatic castration-resistant prostate cancer (mCRPC). 1,2 Preliminary data from this FIH study of xaluritamig in combination with abiraterone in chemotherapy-naïve patients with mCRPC 3 supports the initiation of a phase 3 study of xaluritamig plus abiraterone with the aim of improving survival in this population. Methods: XALience is a randomized, multicenter, open-label, phase 3 study (NCT07213674) of xaluritamig plus abiraterone vs investigator’s choice in patients with chemotherapy-naïve mCRPC. Approximately 750 eligible patients ≥18 years old with histologically confirmed chemotherapy-naïve mCRPC; evidence of disease progression on prior enzalutamide, apalutamide, or darolutamide; ECOG PS of 0-1; and adequate organ function and who may have previously received ≤6 cycles of docetaxel in the metastatic hormone-sensitive prostate cancer (mHSPC) setting will be randomized 1:1 to receive xaluritamig plus abiraterone or investigator’s choice of docetaxel, cabazitaxel, or abiraterone. Patients will be stratified by prior docetaxel exposure in mHSPC, presence of liver metastases, prior treatment with prostate-specific membrane antigen radioligand therapy, and planned intention-to-treat with investigator’s choice. This study consists of a 28-day screening period, treatment until radiographic progression based on Prostate Cancer Working Group 3 (PCWG3) guidelines, a safety follow-up period, and a long-term follow-up period. To mitigate the risk of cytokine release syndrome, xaluritamig will be initiated at weekly step-up dosing over four doses up to the target dose during cycle 1 followed by target dose administration every 2 weeks from cycle 2 day 1 onwards. The primary endpoint is overall survival, which will be analyzed using a stratified log-rank test at a one-sided 2.5% significance level. Key secondary endpoint is investigator-assessed radiographic progression-free survival per PCWG3. Other secondary endpoints include radiographic and prostate-specific antigen response, safety, tolerability, health-related quality of life, patient-reported outcomes, pharmacokinetics, and immunogenicity. Enrollment is ongoing and 20 patients have been enrolled as of January 19, 2026. References: Kelly WK, Danila DC, Lin CC, et al. Cancer Discov. 2024;14:76-89. Kelly WK, Appleman L, Lin C-C, et al. Ann Oncol . 2024;35: S963-S964. Data on file, Amgen; 2025. Clinical trial information: NCT07213674 .

Safety and efficacy of relacorilant and nab-paclitaxel in platinum resistant ovarian cancer: A GRADE assessed systematic review and meta-analysis of randomized controlled trials.

e17604 Background: Relacorilant, a first-in-class selective glucocorticoid receptor antagonist, is being used along with Nab-Paclitaxel for the treatment of Platinum resistant ovarian cancer and has shown better results as compared to Nab-Paclitaxel alone. To systematically evaluate the efficacy and safety of Relacorilant in combination with nab-paclitaxel compared with nab-paclitaxel alone in patients with platinum-resistant ovarian cancer using a GRADE-appraisal meta-analysis of randomized controlled trials. Methods: Embase, PubMed, Cochrane, and ClinicalTrials.gov were searched through December 2025. Four Randomised Controlled Trials assessing the effects of Relacorilant and Nab-Paclitaxel in Platinum resistant ovarian cancer. Pooled analyses of Progression Free Survival (PFS), Overall Survival (OS), Duration of Response (DOR), and Objective Response Rate (ORR), with adverse effects including anemia, neutropenia, and urinary tract infections were performed using random-effects models in R 4.5.2 using “metafor” package. Heterogeneity was quantified using I² statistics. Results: A total of randomized controlled trials evaluating Relacorilant plus nab-paclitaxel versus nab-paclitaxel monotherapy in platinum-resistant ovarian cancer were included. The meta-analysis showed a numerical improvement in progression-free survival (PFS) with combination therapy (pooled mean difference [MD] 1.26 months, 95% CI 0.15 to 2.38) with low heterogeneity (I² = 0%). Overall survival (OS) also favored Relacorilant-based treatment (MD 1.70 months, 95% CI −2.02 to 5.43), although the effect was not statistically significant. The objective response rate (ORR) was higher in the combination group (risk ratio [RR] 1.13, 95% CI 0.91 to 1.42, I² = 0%). Patients receiving Relacorilant experienced a longer duration of response (DOR) (MD 1.54 months, 95% CI 0.57 to 2.51, I² = 5%). Safety analyzes revealed no significant difference between the groups in the risk of anemia (RR 1.02, 95% CI 0.66 to 1.59, I² = 8.9%), neutropenia (RR 0.77, 95% CI 0.21 to 2.80, I² = 84.3%), or urinary tract infection (RR 1.00, 95% CI 0.06 to 0.06). Showed. 15.62). Overall variation in safety outcomes ranged from low to substantial. Conclusions: Relacorilant combined with nab-paclitaxel shows modest improvements in clinical efficacy outcomes, including progression-free survival, objective response rate, and duration of response, without a significant increase in treatment-related adverse events, supporting its potential role as a safe and effective therapeutic option for patients with platinum-resistant ovarian cancer, with the overall certainty of evidence ranging from low to moderate depending on the GRADE assessment.

Results from OCEAN II: A phase II study of encorafenib + binimetinib combination in Chinese patients with <i> BRAF ^V600E </i> mutated metastatic non-small cell lung cancer.

8640 Background: Lung cancer is the most common cause of cancer and cancer death in China. Patients (pts) with BRAF V600E mutant metastatic non-small cell lung cancer (mNSCLC) can benefit from targeted treatments such as encorafenib + binimetinib (E+B). E+B is approved in Western countries (PHAROS: NCT03915951); however, evidence of efficacy and safety in Chinese pts is lacking. OCEAN II (NCT05195632) is the 1st study to specifically evaluate the efficacy, safety and pharmacokinetic of this targeted combination in Chinese pts with BRAF V600E mutant mNSCLC. Methods: OCEAN II is an ongoing multicenter, open-label, phase 2 (P2) study with a Safety Lead-in (SLI). Eligible pts had unresectable stage IV BRAF V600E mutant NSCLC, were treatment-naive or had received systemic therapy, excluding BRAF/MEK inhibitors. Pts received E 450 mg once daily + B 45 mg twice daily. Primary endpoints were dose-limiting toxicities (DLT) during the 1st 28 days (SLI) and confirmed objective response rate (cORR) by independent central review (ICR) (P2). Secondary endpoints (P2) were disease control rate (DCR), duration of response (DoR), progression-free survival (PFS), overall survival (OS), and safety. Results: Between June 2022 and October 2023, 63 pts were enrolled (SLI = 15, P2 = 48): median age 65 yrs (range 52–82), 49% female, 75% with ECOG score = 1. There was only 1 DLT (non-serious grade 3 lipase increased) during SLI, supporting P2 initiation. At the primary analysis cut-off date (27 May 2024), the primary endpoint was met; cORR by ICR was 59.3% (95% confidence interval [CI] 45.0–72.4) in the Efficacy set, with pre-defined statistical significance criteria met. An ad hoc analysis (cut-off 27 March 2025) was conducted when all pts had ≥12 months (mo) follow up. cORR by ICR was 61.1% (95% CI 46.9–74.1), with 6 (11%) complete and 27 (50%) partial responses, and DCR of 87.0% (95% CI 75.1–94.6). Median time to response was 1.8 mo (range 1.7–13.7), median DoR 17.5 mo (95% CI 12.9–NR), median PFS 13.8 mo (95% CI 7.5–NR) and median OS 27.9 mo (95% CI 14.7–30.0). DoR and OS are not yet mature (median follow-up for OS: 21 months). Median treatment duration was 32.3 weeks, with 7 (11%) pts receiving treatment for &gt; 2 years, and 10 (15.9%) remaining on treatment. Treatment-emergent adverse events (TEAEs) ≥ grade 3 occurred in 56% of pts. Related TEAEs in ≥25% were anemia (36%), increased aspartate aminotransferase (33%), increased alanine aminotransferase (30%), increased creatine kinase (29%), vomiting (27%), increased creatinine (25%). Nine (14%) pts discontinued any drug due to AEs. Conclusions: The data confirm the clinical benefit of E+B in Chinese pts with BRAF V600E mutant mNSCLC, a population with distinct genomic and disease characteristics. No new safety concern was identified in the Chinese population. These data can support regulatory decisions and clinical practice guidelines in China. Clinical trial information: NCT03915951 .

Modular CAR T cells: Advancing flexibility, safety, and efficacy in solid cancer therapy.

e14521 Background: Chimeric antigen receptor (CAR) T cell immunotherapy has revolutionized treatment for hematological malignancies, where antigen expression is fairly ubiquitous, and the cancer cells are easily accessible. However, translating this success to solid tumors remains challenge due to unique obstacles in the tumor microenvironment (TME) including, a) metabolic insufficiency, where T cells fail to sustain effector functions and persistence within the metabolically hostile and nutrient-deprived TME; and b) T cell exhaustion and anergy, driven by the immunosuppressive milieu that dampens anti-tumor immunity. To overcome these formidable hurdles, CAR T cell therapy must evolve to incorporate the ability to finely tune CAR T cell activity by adjusting its “intensity” with time, and “flexibility” to target multiple antigens either simultaneously or sequentially. Methods: We used AVATAR and AVATAR AI incubators (Xcell Bio) to precisely control oxygen concentration and pressure to simulate TME conditions. Effector T cells including activated T cells (ATC) or bispecific antibody (BiAb) armed ATC (BATs) and bispecific antibody armed headless CAR T cells (ahCART) or unarmed hCART either grown or were acclimated under TME condition before coculturing with target cells. The effect of TME condition compared to normal incubator environment (NE) were determined on T cell proliferation, phenotypic changes and functional activity of hCARTs. Functional readouts were measured by real time cytotoxicity assay, and cytokine profiles under NE or TME-like conditions. Results: The T cells, when transduced with 4-1BB-ζ hCAR lentiviral vector show metabolically enhanced activity and are able to withstand hypoxic TME without compromising T cell effector functions. Survival was differentially affected by the co-stimulatory endodomains, hCARTs with 4-1BB-ζ (BBζ) showed superior survival under stringent hypoxia 0.5% O 2 (13% apoptosis) compared to hCARTs CD28-ζ (61% apoptosis). Specific cytotoxicity of anti-HER2 (HER2 hCARTs) and anti-EGFR BiAb armed hCARTs (EGFR hCARTs) show that MCF-7 cells are killed more effectively (~65-90%) by sequential treatment than by HER2- or EGFR-hCARTs alone (30-60% and 40-50%, respectively, using by real time cell analysis (RTCA) using xCelligence System. Targeting both EGFR and HER2 antigens sequentially induces a stronger immune response, reduces antigen escape, kills more tumor cells, and sustains a robust anti-tumor Th1 environment compared to targeting a single antigen. T cell reprogramming to adapt TME is likely to represent a novel and important aspect to enhance anti-tumor activity in solid tumor microenvironment. Conclusions: Our new metabolically enhanced TME acclimated hCAR T cells offer a new, versatile, immunotherapy platform that can be targeted "at will" to various tumor targets for prolonged killing and controlled toxicity.

Fitbit-derived activity and heart-rate patterns for predicting 12-month hospitalization in an NIH All of Us cancer cohort.

1629 Background: Hospitalizations in oncology are common and diverse, and are often limited in their predictive accuracy with routine variables. Wearable devices capture functional and physiological patterns that may indicate vulnerability. We examined whether Fitbit-derived activity and heart rate features enhance the prediction of all-cause hospitalization beyond demographic and clinical factors like frailty. Methods: Using data from the NIH All of Us Research Program, we conducted a retrospective analysis of participants aged ≥50 years with a history of cancer, linked electronic health records, and short-term Fitbit data (with imputed missing days). Baseline predictors included demographic variables, clinical factors such as the All of Us frailty index, and indicators of cancer subtype. The outcome was any hospitalization within 12 months of the index date, relative to the last Fitbit data day. Features included step count, step variability, activity distribution, rhythmicity, heart rate, and heart rate variability. Missing data were imputed using predictive mean matching. Models were optimized via nested stratified cross-validation and evaluated with out-of-fold predictions. Discrimination was measured by AUC and average precision with bootstrap confidence intervals. Predictive value for the top 10% at-risk group was also assessed. Results: The study cohort consisted of 278 participants (average age 64.6). Twelve percent (35) experienced hospitalization within a year. Time to hospitalization ranged up to 342 days. Hospitalization diagnoses were heterogeneous. Models using demographic and clinical data alone had limited discrimination (AUC 0.41–0.50). Incorporating Fitbit activity features improved logistic performance: average steps yielded an AUC of 0.522, step-pattern quartiles yielded 0.559, and distributional rhythmic features yielded 0.628 (CI 0.520–0.733; average precision 0.283). Heart rate and variability did not materially enhance logistic prediction. Gradient-boosted trees combining activity and heart rate data achieved the highest discrimination (AUC 0.704; CI 0.621–0.786; average precision 0.286). Using the best-performing model, the top 10% predicted-risk group had a hospitalization rate of 28.6% (8/28), compared with the overall cohort hospitalization rate of 12.6%, capturing 8 of 35 events. Conclusions: Fitbit-derived activity patterns and heart rate patterns, when modeled using machine learning, improved near-term hospitalization prediction in an NIH cancer cohort beyond demographic and clinical models, supporting their potential for practical risk assessment. Ongoing analyses will address missingness, calibration, and subgroup robustness to inform prospective validation.

AI-driven virtual transcriptome screening to identify repurposable combination partners for JIN-A02 in NSCLC.

e20528 Background: Combination therapy is a key strategy to overcome drug resistance and improve therapeutic efficacy in non-small cell lung cancer (NSCLC). However, experimentally exploring the full landscape of combination partners for a new therapeutic agent is costly and time-consuming, limiting the pace of rational combination discovery. Transcriptome-guided, scalable prioritization approaches may accelerate identification of clinically actionable combinations and drug repurposing opportunities. Methods: We conducted a large-scale in silico screening to identify potential combination treatment candidates in the context of JIN-A02 treatment. We used RNA-seq data generated from JIN-A02-treated Ba/F3 cells engineered to express EGFR 19del/T790M/C797S to define the JIN-A02-induced transcriptional response signature. Starting from this transcriptomic profile, we virtually generated transcriptomes representing co-inhibition conditions of 15,089 individual targets. These synthetic co-inhibition transcriptomes were systematically compared to assess their ability to enhance or modulate the JIN-A02-associated transcriptional program, including amplification of desired response pathways and suppression of compensatory signaling patterns. Candidate conditions were ranked based on transcriptome-level impact and prioritization of pharmacologically actionable targets. Results: The screening produced a ranked list of putative co-inhibition conditions, and we identified 12 promising combination candidates within the top 100 conditions. Top-ranked candidates showed consistent transcriptomic modulation patterns suggestive of strengthened anti-tumor signaling and attenuation of potential adaptive responses relative to the JIN-A02 signature alone. Notably, one of the top-ranked candidates corresponded to an approved drug, highlighting a drug repurposing opportunity with potential translational advantages. Conclusions: Our results demonstrate the feasibility of virtual transcriptome generation as a scalable framework for discovering rational combination therapies in NSCLC. This strategy enables systematic exploration of a large target space, prioritizes clinically actionable combination hypotheses for JIN-A02, and reveals opportunities for drug repurposing. Follow-up studies will focus on experimental validation of predicted combinations and identification of biomarkers associated with combination sensitivity.

A comparative analysis of the performance of leading large language models on the endodontics section of the dentistry specialization exam in Türkiye

Objective This study aimed to evaluate and compare the performance of eight contemporary LLMs on the endodontics section of the DUS, assessing their accuracy in both theoretical knowledge and simulated clinical scenarios from historical exam data. Methods The performance of eight different large language models (Claude 4, DeepSeek V3, Gemini 2.5 Pro, ChatGPT-4o, ChatGPT-5, Grok 4, LLaMA 4, and Perplexity) was evaluated using 127 multiple-choice endodontics questions from the Specialization Exam in Dentistry (DUS) administered by the Student Selection and Placement Center (ÖSYM) between 2012 and 2021. The models’ responses were compared against the official answer keys. Statistical analyses were performed using Pearson’s chi-square and McNemar tests, with a significance level of α = 0.05. Results Significant differences existed among LLMs in overall accuracy (p &lt; 0.001). Gemini 2.5 Pro achieved the highest accuracy (90.6%), outperforming ChatGPT-4o (61.4%) and LLaMA 4 (71.7%). In Clinical Practice Questions (CPQ), Gemini 2.5 Pro (93.9%) surpassed ChatGPT-4o (57.6%; p = 0.019). For General Knowledge and Concept Questions (GKCQ), Gemini 2.5 Pro (89.4%), Grok 4 (85.1%), and DeepSeek V3 (84.0%) exceeded ChatGPT-4o (62.8%; p &lt; 0.001). No significant intra-model differences emerged between CPQ and GKCQ performance (p &gt; 0.05). Conclusion Contemporary LLMs demonstrate substantial competence in endodontic knowledge, with Gemini 2.5 Pro excelling in both theoretical and clinical queries. However, significant performance variability across models (61.4%−90.6%) and the complexity of retrieving and resolving clinical exam queries necessitate domain-specific optimization and expert oversight for reliable integration into dental education and practice.

Light‐Induced Dispersion of Pd Single Atoms Provides Steady‐State‐Stabilized Activity for Photocatalytic H ₂ Generation

ABSTRACT In recent years, the use of single atoms (SAs) has become increasingly significant in photocatalysis. Particularly, SA noble metals of Pt and Pd can act as highly effective co‐catalysts for the hydrogen production reaction. The main challenge in the use of SA catalysts is agglomeration, and most importantly, for photocatalysts, light‐induced agglomeration that rapidly degrades the catalytic performance. Here, we show, however, that for Pd SAs on an ideal flat TiO 2 thin‐film platform, under a constant, sufficiently high photon flux, a highly dispersed and highly active SA configuration can be maintained, i.e., agglomeration to nanoparticles and activity loss can be prevented. We ascribe this, supported by DFT, to the destabilizing effect of a high hydrogen loading on Pd 2D rafts consisting of few atoms and the high mobility of the Pd‐H intermediates. This, in contrast to Pt, provides an energetics that allows to maintain a dispersion‐aggregation equilibrium, which results in a steady‐state that macroscopically can provide a remarkably high SA‐stability during photocatalytic experiments, maintaining the photocatalyst in a state of highest activity.

Very simple sunlight-assisted precipitation of hierarchical urchin-like zinc oxide nanostructures and evaluation of their photocatalytic activity

Sea Urchin‐Inspired Immuno‐Instructive Ionic Flow Drives MSCs‐Macrophage Communication to Promote Bone Regeneration

ABSTRACT Bone tissue regeneration is a complex physiological process dependent on the spatiotemporal coordination of immune cells and stem cells. Conventional research primarily elucidates the mechanism through which materials facilitate bone formation by initially modulating macrophages and subsequently encouraging the osteogenic differentiation of stem cells, while largely overlooking the proactive regulatory influence of stem cells on immune cells. Consequently, investigating the capacity of materials to concurrently attract stem cells and modulate innate immune infiltration, while establishing a tissue microenvironment response mechanism, holds substantial importance for the development of next‐generation tissue regeneration materials. This study introduces a sea urchin‐inspired immune instructional ionic flux (SUIF) platform, characterized by a radial mesoporous structure that employs B─O bond dissociation/dissolution rates and Sr 2 + slow‐release mechanisms to create a “fast‐slow biphasic” ionic flow release, thereby establishing a dynamic alkaline ionic flow microenvironment. Within this microenvironment, 5B5Sr‐SUIF recruits MSCs and stimulates MSCs to enhance the expression of miR‐466m‐5p, thereby obstructing the nuclear translocation of NF‐κB in macrophages. This facilitates immunological regulatory communication between mesenchymal stem cells and macrophages, offering novel material design principles and molecular mechanism support for the exact regeneration of complicated bone deformities.

Mechanical and rheological behaviour of acrylonitrile butadiene styrene/polyolefin elastomers composites: Experimental analysis and numerical optimization

The role of phospho-ubiquitin in mitochondrial health and diseases

A phase 1b study of SYS6002 in advanced solid tumors.

4579 Background: SYS6002 is a next-generation Nectin-4-targeting antibody-drug conjugate (ADC) utilizing a novel site-specific enzymatic conjugation technology to link monomethyl auristatin E (MMAE) with a uniform drug-antibody ratio of 2. This design aims to enhance linker stability, minimize free payload release, and improve the therapeutic window. Methods: This phase 1 study (ChiCTR2200066256) included dose escalation (0.2–4.5 mg/kg Q3W; 2.4–2.7 mg/kg Q2W), pharmacokinetic (PK) expansion, and cohort expansion phases. Eligible patients had advanced solid tumors refractory to standard therapies. The primary endpoints were safety and recommended phase 2 dose (RP2D). Secondary endpoints included PK profile and efficacy, with the latter reported herein were specific to the pretreated advanced urothelial carcinoma (aUC) cohort. Results: As of Dec 19, 2025, 150 patients were enrolled (dose escalation n=32; PK expansion n=70; cohort expansion n=48). The maximum tolerated dose was not reached up to 4.5 mg/kg. The RP2D in aUC patients was determined as 3.6 mg/kg Q3W. SYS6002 exhibited a manageable safety profile in all 150 patients. With a median follow-up of 11.5 months (IQR 6.5-15.5), 2 (1.3%) patients discontinued treatment due to treatment-related adverse events (TRAEs) and no TRAE led to death. The most common TRAEs were ocular disorders (e.g., corneal disorder, dry eye), which were predominantly Grade 1-2. While ocular events were frequent (consistent with on-target effects), they were reversible and effectively managed with prophylaxis, resulting in no treatment discontinuations due to ocular toxicity. Notably, TRAEs typically associated with payload shedding were characterized by low incidence, with peripheral neuropathy reported in 11.3% of patients (no Grade ≥3) and rash in 20.7% (2.0% Grade ≥3), which compared favorably to historical data of other MMAE-based ADCs. Furthermore, no Stevens-Johnson Syndrome or Toxic Epidermal Necrolysis reported. In 85 efficacy-evaluable aUC patients across all tested doses, the objective response rate (ORR) was 37.6% (95%CI 27.4-48.8) and disease control rate (DCR) was 77.6% (95%CI 67.3-86.0), with a median duration of response of 6.2 months (95%CI 4.3-10.4) and median overall survival of 13.2 months (95%CI 10.6-14.9). At the RP2D (n=39), the ORR reached 41.0% (95%CI 25.6-57.9). Encouragingly, in heavily pretreated patients with prior MMAE-based ADC exposure (n=25), SYS6002 maintained activity with an ORR of 24.0% (95%CI 9.4-45.1) and DCR of 72.0% (95%CI 50.6-87.9), suggesting potential to overcome resistance. Conclusions: SYS6002 demonstrated a distinct safety profile from other Nectin-4 ADCs. Its promising efficacy in aUC patients, extending to those progressing on prior MMAE-based ADCs, supports further phase 3 trials. Clinical trial information: ChiCTR2200066256.

Treatment holiday for patients with metastatic hormone sensitive prostate cancer: Real-world experience at a single center.

e17105 Background: Continuous androgen deprivation therapy (ADT) is the standard for patients with hormone sensitive prostate cancer that is metastatic on conventional imaging (mHSPC). In the current era many patients receive more potent therapy with both ADT and an ARPI without (doublet therapy) or with docetaxel (triplet therapy). Several trials are examining de-escalation by stopping the ARPI, introducing a treatment holiday, or treating for a finite duration of systemic therapy. Reports on real world experience of treatment holidays for patients with mHSPC are limited. Methods: Following IRB approval, we identified patients in our hospital system with mHSPC, which was M1b or M1c based on CT or MRI and bone scan, and who received ADT and an ARPI for at least 6 months. We examined stage at diagnosis and prior to doublet therapy, treatment history, and testosterone recovery, and recorded if there was PSA or radiologic progression while on treatment holiday. PSA progression was defined stringently as PSA &gt; 0.2 for those who had undergone prostatectomy, and PSA &gt; 2.0 for those who had not. Radiologic progression was on either PET or conventional imaging. Results: Twenty-seven patients with mHSPC went on treatment holiday from doublet therapy. Median age at diagnosis was 67 (range 52-84). Twenty-one of 27 (78%) patients had synchronous metastatic disease, 6 of 27 (22%) had metachronous metastatic disease, and 17 of 27 (63%) underwent radiotherapy to metastatic sites. Median PSA before starting doublet therapy was 19.72 (range 1.03-176), and all achieved undetectable PSA prior to treatment holiday. Twenty-five of 27 (93%) had bone metastases, and 3 of 27 (11%) had lung metastases. Nineteen of 27 (70%) patients continued off therapy without evidence of progression, with a median follow-up of 27 months (range 4.0-57). Eight of 27 (30%) patients had progression (2 with PSA progression, 4 radiologic progression, and 2 with both), and 72% (95% CI: 47%-86%) remained progression-free at 36 months. Median time to testosterone recovery to &gt; 150 ng/mL was 26 months (95% CI: 8.7, NR). Conclusions: In this real-world data, a subset of patients with metastatic prostate cancer and on treatment holiday from doublet therapy had durable time off therapy and without progression. This included patients with synchronous metastatic disease and with high volume disease. Prospective studies are warranted to evaluate which patients are most likely to have durable time off therapy. Of note, a quarter of those whose cancer progressed during treatment holiday had radiologic progression without PSA progression.

Systematic safety information gaps in AI-generated melanoma therapy information: An exploratory multi-platform evaluation.

e13704 Background: Patients and clinicians increasingly use large language models (LLMs) for pharmaceutical information. Prior studies have evaluated AI chatbot accuracy for drug information and oncology applications, finding variable performance with notable hallucinations. However, whether consumer AI platforms adequately communicate critical safety information for oncology therapies has not been systematically characterized. Methods: We conducted an exploratory pilot evaluation of AI-generated responses for 8 FDA-approved melanoma therapies across 3 platforms (ChatGPT/GPT-4-Turbo, Claude/Sonnet, Gemini/2.0-Flash). A total of 6,941 responses were collected and scored against FDA prescribing information using automated LLM-based concordance assessment (Claude Opus 4.5). This pilot used automated scoring without human validation. Results: Our primary finding was systematic safety information gaps across all evaluated drugs. Responses were flagged for missing safety information in 80-85% of cases for established therapies and 99% for the recently-approved TIL therapy Amtagvi. Overall concordance was 83.5% (range: 78.9%-86.3%). We identified 1,137 responses with critical severity flags, including instances where platforms failed to mention serious adverse events. The recently-approved Amtagvi showed the poorest performance, with 98.7% of responses requiring review. We did not assess whether platforms communicated equivalent risk using alternative terminology. Conclusions: This exploratory study identifies systematic safety information gaps as a consistent pattern in AI-generated melanoma therapy information across all drugs and platforms. Findings warrant validation through human expert review to characterize the clinical significance of these omissions.

FAPI-CUP: A prospective cohort study of patients with cancer of unknown primary (CUP) to evaluate a novel radiotracer, fibroblast activated protein (FAPI), targeting cancer-associated fibroblasts for primary site detection.

3072 Background: Cancer of unknown primary (CUP) represents a heterogeneous group of metastatic tumours for which standardised diagnostic work-up fails to identify the site of origin at diagnosis. 18 F-FDG-PET/CT aids in identifying a putative primary site in ~30-50% of CUP patient (pts); however, has limited sensitivity for detecting cancers with high stromal content. Fibroblast activation protein (FAP) is a type II transmembrane serine protease and highly expressed by cancer-associated fibroblasts abundant in desmoplastic tumours such as CUP. 68 Ga-FAPI-46 is a FAP-targeting PET tracer but has not been directly compared to 18 F-FDG-PET/CT in CUP. Aims: We aimed to determine if the addition of 68 Ga-FAPI-46 detects more primary sites compared with 18 F-FDG-PET/CT and CT CAP (standard of care; SoC) in CUP pts and if there is an association between the level of FAPI avidity and response to systemic treatment. Methods: A prospective cohort study recruiting CUP pts across four sites in Australia. Key inclusion criteria: 1) pts considered CUP after diagnostic work-up, pathological review and gender appropriate tests; 2) adequate haematologic and organ function; 3) not commenced current line of systemic treatment (exception palliative radiotherapy for symptom control); 4) ECOG ≤ 2 and life expectancy &gt; 3months. 5) ≤1 prior line of systemic treatment. Pts undergo SoC imaging (CT CAP, 18 F-FDG-PET/CT) and 68 Ga-FAPI-46 PET/CT scan which are reviewed at a multidisciplinary meeting to determine primary site. Clinicopathological review and genomic analysis (if available) were used to determine final primary site. Pts start systemic treatment and have SoC follow-up with data regarding RECIST 1.1 response, survival outcome and further lines of treatment for 12 months. Results: At interim analysis 60 pts were recruited from 03/2022 to 06/2025. Median age was 68 years [30-83] with 89% being ECOG 0-1 and 80% classified as having an unfavourable CUP subtype. A primary site was detected in 37/60 (62%) and 33/60 (55%) pts with 68 Ga-FAPI-46 PET/CT and SoC and SoC alone, respectively. Cholangiocarcinoma (9/37; 24%) and lung (7/37; 19%) were the commonest primary sites detected. 46/60 (77%) pts commenced systemic treatment with a best overall response rate of 9/38 (24%) in RECIST evaluable pts. There was no association between RECIST response and baseline FAPI avidity for either average SUVmax top 3 lesions (p=0.653) or highest SUVmax (p=0.416). Conclusions: 68 Ga-FAPI-46 PET/CT in addition to SoC detected a primary site in more patients compared to SoC imaging alone. There was no association between RECIST response and baseline FAPI avidity. Clinical trial information: NCT05263700 .

Surufatinib for advanced or metastatic chemotherapy-refractory thymic epithelial tumor: A single-arm, single-center, phase II study.

8119 Background: Thymic epithelial tumors (TETs), though relatively rare, represent the most common anterior mediastinal malignancy in adults. Platinum-based regimens are the standard first-line therapy; however, treatment options for subsequent lines remain poorly defined. Surufatinib, an oral tyrosine kinase inhibitor targeting VEGFR1/2/3, FGFR1, and CSF-1R, exerts both anti-angiogenic and immune-modulating effects. This study aimed to evaluate the efficacy and safety of surufatinib in patients (pts) with type B2/B3 thymoma (TM) or thymic carcinoma (TC), providing a potential new therapeutic strategy for TETs following anthracycline/taxane-based therapy. Here, we report the initial results. Methods: This was a single-arm, prospective phase II study. Eligible pts had histologically or cytologically confirmed, unresectable or radiotherapy-ineligible, advanced recurrent or metastatic B2/B3 TM or TC, with progression following at least one prior platinum-based chemotherapy regimen. Pts were aged &gt;18 years and had at least one measurable lesion per RECIST 1.1 criteria. Treatment consisted of surufatinib 300 mg orally once daily in a 4-week cycle. Tumor assessments were conducted every 8 weeks. The primary endpoint was the objective response rate (ORR). Secondary endpoints included progression-free survival (PFS), disease control rate (DCR), and overall survival (OS). Results: As of December 25, 2025, 20 pts were enrolled and evaluated. The median age was 55.5 years (range: 35–69), with 55% being male. Pathological subtypes included B2 TM (30%), B3 TM (15%), and thymic squamous cell carcinoma (TSCC, 55%). Eighteen pts (90%) presented with &gt;5 metastatic sites. Ten pts (50%) achieved a partial response, and ten pts (50%) achieved stable disease. The ORR was 50%, and the DCR was 100%. With a median follow-up of 9.50 months, the median PFS was 11.9 months, and the 9-month PFS rate was 73.5%. Median OS was not reached, with an 18-month OS rate of 91.7%. The most common treatment-related adverse events (TRAEs) of any grade were hypertension (70%), diarrhea (40%), elevated bilirubin (40%), and elevated transaminases (25%). Grade 3 TRAEs included hypertension (15%), diarrhea (5%), and elevated transaminases (5%). No treatment-related deaths were reported. Conclusions: Surufatinib monotherapy demonstrated promising antitumor activity and a manageable safety profile in patients with previously treated, advanced recurrent or metastatic TETs, supporting further investigation in this population. Clinical trial information: ChiCTR2600116776.

Empowering pathologists as cancer precision medicine leaders: The Florida Society of Pathologists (FSP) Precision Medicine Academy (PMA) initiative.

e21013 Background: As cancer care increasingly relies on molecular and genomic biomarkers, pathologists play a central role in precision oncology. However, variability in access to structured education, emerging technologies, and interdisciplinary collaboration can limit consistent implementation. The Florida Society of Pathologists (FSP) launched the Precision Medicine Academy (PMA) to develop and implement a scalable, statewide strategy to educate and equip pathologists with the knowledge, skills, and confidence needed to lead precision medicine efforts. Methods: FSP designed a phased educational strategy informed by member needs, national guidelines, and stakeholder input. Phase 1 included a comprehensive continuing medical education (CME) curriculum delivered through live, virtual webinars. Academic pathologists, together with multidisciplinary care team members, delivered evidence-based information addressing foundational and emerging topics in precision oncology, including molecular and genomic biomarker testing, hereditary genetics and tumor genomics, multidisciplinary communication and collaboration strategies, digital pathology, bioinformatics, and AI in precision cancer diagnostics and treatment. Pre/post activity assessments measured knowledge changes and qualitative feedback captured perceived practice relevance. Results: Phase 1 successfully engaged pathologists from diverse practice settings across Florida. Learners demonstrated knowledge gains of 38.3%, 44.3%, 21.4%, 52.8%, and 50.5%, across key domains of cancer genetics, solid tumor biomarkers, hematologic malignancies, multidisciplinary collaboration, and digital pathology and AI, respectively. Qualitative feedback highlighted increased confidence in refining diagnostic workflows, improving communication with oncology and genetics colleagues, and leading the laboratory workforce to adopt precision medicine strategies. Participants emphasized the value of a pathology-focused, practice-relevant curriculum that addressed both technical and operational challenges. Conclusions: The FSP Precision Medicine Academy represents a society-led approach to building statewide capacity in precision oncology among pathologists. Phase 1 demonstrated that a CME-based curriculum can produce measurable knowledge gains and meaningful perceived practice benefits. Phase 2, launching in 2026, will expand this model through visiting rotations, virtual tumor boards, a multidisciplinary symposium, and facilitated peer exchange focusing on challenges and solutions in precision oncology. This initiative represents a significant step toward empowering pathologists as leaders in precision oncology and fostering a sustainable statewide learning community.

Efficacy and safety of vebreltinib in advanced clear cell sarcoma (VEBrant): A multicenter, phase II study.

11503 Background: Clear cell sarcoma (CCS) is an ultra rare cancer with no established standard systemic therapy of proven efficacy. The c-MET pathway plays an important role in CCS, with MET overexpression frequently observed. Targeted therapy by type Ia MET inhibitor crizotinib revealed a disease control rate (DCR) at 69.2% and objective response rate (ORR) at 3.8%. Vebreltinib is a highly selective type Ib MET inhibitor with verified efficacy in solid tumor harboring MET alterations, such as NSCLC with MET amplification or overexpression. This phase II study aimed to assess the efficacy and safety of Vebreltinib for advanced CCS. Methods: Patients (pts) with histopathologically confirmed, unresectable or metastatic CCS were enrolled (NCT07153887), regardless of prior therapy except for any MET inhibitors. Eligible pts received Vebreltinib (200 mg BID) until disease progression, intolerable toxicity, or death occurs. Prior PD1/PDL1 treatment for at least 4 months without tumor reduction are allowed to continue using the same PD1/PDL1 medicine after thorough evaluation by the investigator. Evaluation of tumor response was performed every 2 months. Primary endpoints was ORR per RECIST v1.1. Secondary endpoints included DCR, PFS, OS, 12m-OSR, DoR and treatment safety. MET expression by immunohistochemistry was collected when possible. A Simon’s two-stage design was applied: if no responses were seen in the first 13 pts, the study would stop; otherwise, 14 additional pts would be enrolled (total N=27). Results: At data cutoff on January 23, 2026, 23 pts were enrolled. Median age was 29 years (range: 16–59), 47.8% were male, 91.3% had prior surgery, and 95.7% received prior systemic therapy. All had metastatic disease, primarily to lung (78.3%), lymph nodes (56.5%), and bone (30.4%). With median follow-up of 134 days and median treatment duration of 103 days, 17 pts were evaluable for response. ORR was 41.2% (7 partial responses) and DCR was 70.6% (7 PR, 5 stable disease), meeting the pre-specified threshold for stage I continuation. Median PFS was 4.14 months (95% CI: 2.3–NA); 6-month PFS rate was 42.8%. Median OS was not reached. Treatment-related adverse events (TRAEs) occurred in 87.0% of pts, mostly grade 1-2. Grade 3-4 TRAEs occurred in 21.7% (n=5; rash n=4, fever n=1). Notably, 4 pts discontinued due to grade 4 rash within 2 weeks of starting vebreltinib combined with immunotherapy. Conclusions: Vebreltinib demonstrated notable antitumor activity for advanced CCS, representing a marked improvement over existing therapeutic options. Attention should be paid to the risk of severe rash, particularly when combined with immunotherapy. Longer follow-up is required to further confirm the efficacy and safety. Clinical trial information: NCT07153887 .

DZD6008, a fourth-generation EGFR TKI, in pretreated NSCLC patients with <i>EGFR</i> C797X mutations: Results from phase 1/2 studies.

8520 Background: Patients with EGFR-mutant NSCLC whose disease progresses on or after 3 rd generation EGFR TKI treatment often acquire resistance mutations, among which EGFR C797X is one of the most frequently reported. DZD6008 is a 4 th generation EGFR TKI, designed to target classical EGFR sensitizing mutations (L858R/19del), as well as resistant double (T790M and L858R/19del) and triple mutations (C797X, T790M and L858R/19del), with preclinical data showing high selectivity over wild-type EGFR and other kinases and full blood-brain-barrier (BBB) penetration. Here we report results in pretreated NSCLC patients with EGFR C797X mutations (C797X+) from phase 1/2 studies. Methods: TIAN-SHAN1 (NCT06905197) and TIAN-SHAN2 (NCT06813365; CTR20241790) are ongoing, multicenter phase 1/2 studies evaluating the safety, tolerability, and anti-tumor activity of DZD6008 in EGFR-mutant NSCLC patients, conducted in the US/Australia and China, respectively. The efficacy endpoints include objective response rate (ORR), duration of response (DoR) and progression-free survival (PFS) by investigator per RECIST v1.1, and safety endpoints include treatment-related adverse events (TRAEs) per CTCAE 5.0. Results: As of December 19, 2025, a total of 24 patients with confirmed C797X+ NSCLC were treated with once daily (QD) oral DZD6008 (20 mg, n=1; 40 mg, n=13; 60 mg, n=10), and had at least 1 post-baseline tumor assessment. The median age was 66.5 years, 58.3% were female, 91.7% were Asian, 58.3% had ECOG PS of 1. All patients had metastatic disease upon study entry and received median 2 (range 1 - 6) lines of prior therapies. Across all dose levels, tumor shrinkage was observed in 75% of patients, with an ORR of 41.7%. Intracranial anti-tumor activity was observed in patients with baseline brain metastasis. The doses of 40 mg and 60 mg QD were defined as the recommended phase 2 doses (RP2Ds). The ORRs were 23.1% and 60.0% at these two dose levels, respectively. The median DoR and median PFS were not reached for either dose. The 9-month PFS rates were 54.5% and 64.8%, respectively. DZD6008 was well tolerated at the doses evaluated, with no dose limiting toxicities observed. The majority of TRAEs were grade 1 or 2. The TRAEs with grade ≥3 included lymphocyte count decreased (8.3%), anemia, malaise, fatigue and amylase increased (all 4.2%). There were no Grade 5 TRAEs. Conclusions: DZD6008 demonstrated promising and durable anti-tumor activity in patients with EGFR C797X+ NSCLC with a manageable safety profile, supporting its potential use as a later line treatment option after 3 rd generation EGFR TKI failure. The updated data will be presented at the meeting. Clinical trial information: NCT06905197 , NCT06813365 .