Browse Articles

Scalable green synthesis of Ag/Cu bimetallic nanoparticles via electron beam irradiation for synergistic antimicrobial protection against plant pathogens

Reactive Laser Additive Manufacturing of Hierarchically Structured Aerogels

ABSTRACT As demands for sustainable and scalable energy materials manufacturing accelerate, additive manufacturing (AM) remains largely limited to passive shaping of predefined precursors. Here, we introduce reactive laser AM, in which precursor composition is designed to transform the printing step itself into a chemically active stage of materials synthesis. Incorporating eutectic alkali halide salts into protein‐based powders converts localized laser heating into transient reaction environments that drive vapor‐phase chemistry, surface etching, and in situ hierarchical growth without external reagents or solvents. This internally activated reactivity enables the rapid formation of graphitic aerogel monoliths with multilevel architecture—macroporous frameworks decorated with microtubular arrays and nanoscale features—within seconds in a single process. As energy storage electrodes, these hierarchically structured aerogels exhibit a tenfold enhancement in gravimetric capacitance (∼162 F g −1 ) relative to salt‐free counterparts. By engineering reactivity through feedstock design, this work reframes laser AM as a dynamic platform for reaction‐driven materials‐by‐design.

Machine learning models for predicting postpartum convulsions using clinical indicators from PMA Ethiopia data

Autoinhibition is Not a Universal Feature of Transcription-Repair Coupling Factors

Cell-free DNA fragmentomics and methylomics for thymic epithelial tumors.

8118 Background: Development of circulating cell-free (cf) DNA assays for cancer detection and monitoring is an area of active research. Limited data are available on the utility of cfDNA for thymic epithelial tumors (TETs). Evaluation of cfDNA fragmentation patterns (fragmentomics) offers advantages over conventional cfDNA analysis. Studies have shown that concurrent generation of fragmentomics and methylomics features via enzymatic methyl-sequencing (EM-Seq) are highly predictive at classifying cancer vs. non-cancer and capture key cancer-related biological features. We conducted a study to determine the feasibility of fragmentomics and methylomics for evaluation of TETs. Methods: cfDNA was extracted from 1-2 mL of plasma EDTA derived from patients with advanced TETs using the QIAamp Circulating Nucleic Acid Kit, underwent enzymatic conversion via the NEBNext Enzymatic Methyl-Seq v2 Kit, and sequenced on a NovaSeq X 10B flowcell. Estimated tumor fractions were generated using ichorCNA with low tumor settings and via fragle . Recurrent copy number alterations (CNAs) were plotted and compared against external TET copy number datasets. Short fragment percentage (90-150 bps) was calculated and average DELFI scores (DNA evaluation of fragments for early interception) were generated via FinaleToolkit . Results: Plasma samples from 47 of 50 participants (thymic carcinoma: 26, thymoma: 24; median age: 55 years (range 23-77); females 22; stage distributions: I/II/III/IV = 1/0/1/48) yielded high-quality cfDNA which was sufficient for sequencing. Individuals with thymic carcinoma displayed higher estimated tumor fractions via ichorCNA (P<0.05) and Fragle (P=0.05). Moreover, although not statistically significant, individuals with thymic carcinoma exhibited a general trend of higher short fragment percentage and higher average DELFI scores. When evaluating recurrent CNAs, individuals with thymic carcinoma displayed greater enrichments in CNAs in general with noted amplifications/gains at chr 1, 5, 12, and 17 and deletions at chr 16, whereas individuals with thymoma showed deletions at chr 3. When grouped and compared to external thymoma tissue CNAs, recurrent CNAs detected in plasma showed partial concordance. Large alterations in fragmentation profiles were also identified across both groups, which showed increased variability near recurrent CNA regions. Conclusions: This proof-of-concept study demonstrates that plasma cfDNA is detectable and fragmentomics is feasible among individuals with TETs, thus providing a non-invasive means of capturing cancer-related aberrant signals. Validation of this result in larger studies would establish cfDNA fragmentomics as a novel non-invasive means for disease monitoring and biomarker identification in patients with TETs.

Beyond-line immunotherapy in ES-SCLC: Efficacy and safety of second-line PD-1/PD-L1 inhibitors after progression on first-line PD-1/PD-L1 inhibitors plus platinum-etoposide chemotherapy.

e20149 Background: While Programmed Death-1/Programmed Death-Ligand 1(PD-1/PD-L1) inhibitors combined with platinum-etoposide are the standard first-line treatment for extensive-stage small cell lung cancer (ES-SCLC), optimal second-line strategies remain poorly defined. Evidence regarding the clinical benefit of continuing immunotherapy beyond progression (IBP) is lacking. This study evaluated the efficacy and safety of second-line PD-1/PD-L1 inhibition following progression on first-line immunochemotherapy. Methods: We retrospectively analyzed 96 patients with ES-SCLC who progressed after first-line PD-1/PD-L1 inhibitors plus EP/EC chemotherapy at The First Affiliated Hospital of Dalian Medical University between January 2020 and March 2025. Patients were categorized into two groups: the IBP group (n = 55), receiving continuous PD-1/PD-L1 inhibitors as second-line therapy, and the Non-IBP group (n = 41). Survival outcomes were compared using Kaplan-Meier curves and Cox proportional hazards models. Results: The IBP group demonstrated numerically higher objective response rates (ORR: 28.30% vs. 16.28%, p = 0.2231) and disease control rates (DCR: 69.81% vs. 58.14%, p = 0.2854) compared to the non-IBP group. Notably, IBP was associated with significantly prolonged median progression-free survival (mPFS: 4.3 vs. 2.2 months; HR = 0.59; 95% CI, 0.38–0.94; p = 0.0118) and median overall survival (mOS: 12.5 vs. 6.8 months; HR = 0.56; 95% CI, 0.34–0.92; p = 0.0120). Subgroup analyses revealed significant PFS benefits for IBP in patients who were male, aged < 65 years, had ECOG PS 1, involvement of > 3 metastatic organs, liver/brain metastases, absence of bone metastases, new lesions at progression, > 4 cycles of first-line therapy, first-line clinical benefit (PR/CR/SD), first-line PFS > 3 months, and the use of irinotecan-based second-line regimens. Furthermore, the OS benefit remained robust in patients who were male, had ECOG PS 1, involvement of > 3 metastatic organs, liver/brain metastases, > 4 cycles of first-line therapy, first-line clinical benefit (PR/CR/SD) and first-line PFS > 3 months. The safety profile of IBP was manageable, with no new safety signals. Conclusions: Second-line PD-1/PD-L1 inhibitors beyond progression significantly improved PFS and OS in ES-SCLC. These findings suggest that continuing PD-1/PD-L1 inhibition is a promising strategy for specific patient populations, especially those with high metastatic burden or favorable response to initial immunotherapy.

Therapeutically targetable mutation changes in lung adenocarcinoma metastasis.

e20719 Background: Use of targeted therapies was reported to contribute to the sharp decline of annual mortality from NSCLC in the U.S. between 2013 and 2016. However, it only benefits patients who harbor specific mutations. At present, NCCN clinical practice guidelines only recommend that when a patient's cancer progresses after initial targeted treatment, plasma or tissue-based mutation testing should be considered to understand the resistance mechanisms. Patients with prior negative findings are rarely re-biopsied and re-profiled genomically when metastases develop. Methods: We conducted analyses on applying 449 paired samples from 197 patients to examine targetable mutation changes in therapy–actionable genes during metastasis. We assessed targetable mutation changes in therapy–actionable genes, including KRAS, ALK, ROS1, BRAF, MET, RET, ERBB2 and NTRK, during metastasis. Results: During metastasis, 38.3% of patients receiving prior targeted therapy exhibited targetable mutation change in therapy–actionable genes, whereas 16.9% patients who never received targeted therapy also developed those changes. Stratification analysis showed that during spread from the primary tumor to metastases, 16.9% of patients exhibited such changes and 16.7% patients with paired samples of metastases across different sites had such changes. During metastasis, targeted therapy significantly increased targetable mutations in EGFR (P = 0.00006), but had no significant effect on other therapy–actionable genes. Conclusions: Among patients not receiving targeted therapy, a substantial percentage of therapeutic targetable mutations can change during metastasis. Therefore, in clinical practice, re-biopsy of metastases with genomic re-profiling should be recommended to patients when metastases develop, even if they never had prior positive results of targetable mutation testing.

Utility of hemoglobin A1c and blood glucose monitoring for early detection of pancreatic cancer in a high-risk cohort.

e16466 Background: Pancreatic ductal adenocarcinoma (PDAC) has a poor prognosis largely due to late-stage diagnosis, even among high-risk individuals under surveillance with MRI/MRCP and/or EUS. However current surveillance methods can be invasive, costly, and are imperfect in their ability to detect PDAC early. PDAC can be associated with early metabolic changes, including hyperglycemia and diabetes, and therefore changes in HgbA1c and glucose have been hypothesized to potentially serve as a non-invasive adjunctive measure for PDAC early detection in high-risk populations. Inspired by the Meziani et al study of HgbA1c and glucose trends in a Dutch cohort of high-risk individuals, the aim of this study is to examine HgbA1c and glucose trends in a more heterogenous population with a broader genetic profile and risk factors, greater racial diversity, and longer mean length of follow-up. Methods: Retrospective analysis of high-risk individuals undergoing PDAC surveillance who were enrolled in the CAPS5 study at a single study center between 1/1/2015 and 12/31/2025 (mean follow-up of 56 months). Those included in the analysis had at least one measured glucose and HgbA1c level, without a prior history of diabetes. Individuals were separated into cohorts of those who developed PDAC during the surveillance period and those who did not. Percent change in glucose and HgbA1c was assessed in each cohort from first visit under surveillance to date of last follow up. Results: Of the 268 included individuals, 6 (2.2%) developed PDAC during the surveillance period. Two (33%) developed diabetes in the PDAC cohort and 15 (5.72%) developed diabetes in the surveillance cohort who did not develop PDAC. The mean values of glucose (109.5 in non-PDAC vs 123.2 in PDAC, p = 0.015) and HgbA1c (5.8 in non-PDAC vs 6.2 in PDAC, p = 0.019) did differ significantly between cohorts. However, neither glucose trends nor HgbA1c trends differed significantly between the control and PDAC cohorts. Conclusions: In this study, measures of glucose and HgbA1c did not meaningfully contribute to the early detection of PDAC, nor provide a clear pre-diagnostic signal in a high-risk cohort. However, the sample size of this study is small, and larger prospective studies of glucose/HgbA1c monitoring need to be performed to confirm these results. Overall, this data highlights the potential limitations of glycemic monitoring as a non-invasive blood- based tool for early PDAC detection, and emphasizes the ongoing need for more sensitive biomarkers as an adjunct to imaging surveillance strategies. Changes in Glucose and A1c from initial visit to last follow-up. Non-PDAC (n=262) PDAC (n =6) p-value Glucose Range of change (%) 3.81% 7.29% 0.67 Mean (SD) 2.20 (21.96) 4.83 (22.76) 0.79 HbA1c Range of change (%) 1.86% 1.82% 0.98 Mean (SD) 0.09 (0.36) 0.12 (0.27) 0.75

Real-world treatment patterns and overall survival in recurrent/metastatic head and neck squamous cell carcinoma following treatment with immune checkpoint inhibitor and platinum-based chemotherapy.

6038 Background: Treatments for recurrent/metastatic head and neck squamous cell carcinoma (R/M HNSCC) include immune checkpoint inhibitors (ICI) and/or platinum-based chemotherapy, yet most patients (pts) will experience disease progression, with poor survival and limited therapeutic options in later-line settings. Real-world evidence in R/M HNSCC after ICI and platinum-based chemotherapy is limited. We characterize real-world treatments and survival outcomes in pts with heavily pretreated R/M HNSCC. Methods: This retrospective analysis used electronic medical records from the US Flatiron Health Advanced EDM (71% community centers, 22% academic centers, 7% both). Study period was 1 Jan 2015 to 31 Mar 2025 (date of data cutoff). All identified pts had R/M HNSCC and had received prior PD-1 inhibitor and platinum-based chemotherapy. Demographics, baseline disease characteristics, and treatment patterns were analyzed descriptively. Kaplan-Meier method was used to evaluate real-world overall survival (rwOS), defined as time from index date (first date of treatment after PD-1 inhibitor and platinum-based chemotherapy) until date of recorded death. Subgroups by HPV-unrelated (oral cavity, larynx, hypopharynx, HPV-negative oropharyngeal) and HPV-related (HPV-positive oropharyngeal) HNSCC were evaluated. Results: In total, 2105 pts with R/M HNSCC were identified from the Flatiron database as previously treated with PD-1 inhibitor and platinum-based chemotherapy. There were 1271 pts who did not receive any subsequent treatment, while 834 pts received an index line of therapy after PD-1 inhibitor and platinum-based chemotherapy and were eligible for analysis. Among those eligible, primary tumor site distribution was 55% oropharynx (81% HPV-positive / 19% HPV-negative), 21% larynx, 18% oral cavity, and 6% hypopharynx. Most pts had an ECOG score of 0 or 1 (77%). Most common regimens in the index line of therapy included chemotherapy, anti-EGFR, and/or PD-(L)1 agents. Among the 834 pts identified, rwOS was 7.8 months (95% CI, 7.0–8.4). There were 373 and 461 pts identified with HPV-related and HPV-unrelated disease, respectively. Median rwOS was significantly longer for the HPV-related vs HPV-unrelated group (9.3 months [95% CI, 8.2–10.9] vs 6.8 months [95% CI, 6.1–7.3]; HR, 0.85 [95% CI, 0.73–0.99]; P =0.036). The 6-month rwOS rate was 65% vs 55%, respectively. Of note, a substantial portion of pts died prior to receiving another line of therapy (39% for HPV-related and 49% for HPV-unrelated). Conclusions: Median rwOS is poor in R/M HNSCC following treatment with PD-1 inhibitor and platinum-based chemotherapy, and particularly poor for HPV-unrelated disease, with nearly half dying by 6 months. These results highlight the pressing need for innovative treatments to improve survival in R/M HNSCC.

Impact of neoadjuvant durvalumab (D) on tumor microenvironment (TME) features and their association with event-free survival (EFS) in patients with resectable NSCLC (R-NSCLC) from the phase 3 AEGEAN trial.

8015 Background: In AEGEAN, perioperative D + neoadj CT improved EFS and pathological complete response vs neoadj CT alone in pts with R-NSCLC. Here, we report exploratory transcriptomic analyses of the TME in tumor samples collected at baseline (BL) and surgery (Sx) to investigate the impact of neoadj D on TME features and their association with EFS. Methods: AEGEAN is a double-blind PBO-controlled study (NCT03800134). Adults with Tx-naïve R-NSCLC (stage II–IIIB[N2]) and ECOG PS 0/1 were randomized 1:1 to neoadj platinum-based CT + D or PBO IV (Q3W, 4 cycles) before Sx followed by D or PBO IV (Q4W, 12 cycles) after Sx. EFS was evaluated by BICR (RECIST v1.1) in the modified ITT (mITT) population, which excluded pts with known EGFR / ALK aberrations. Total RNA was extracted from BL and Sx tumor samples and sequenced (Illumina NovaSeq X Plus). Unsupervised hierarchical clustering of samples was conducted based on previously reported gene signatures reflective of tumor and TME components. Results: Transcriptomic data were available from 366 samples in 292 mITT pts across both arms (at BL, 257 pts; at Sx, 109 pts) whose characteristics and outcomes were broadly representative of the mITT population (74 pts with paired samples). At BL, 3 distinct phenotypic clusters (C) based on TME features were identified: an immune desert (C1, 24.9% of pts), characterized by a predominance of proliferating tumor cells; immune suppressed (C2, 39.3%), by elevated levels of suppressive myeloid cells, angiogenesis, and fibroblasts; and immune activated (C3, 35.8%), by high levels of effector T cells. A higher proportion of pts with squamous vs non-squamous tumors had phenotype C1 (37.5% vs 13.9%, respectively) while a lower proportion had C3 (22.5% vs 47.4%). The addition of perioperative D improved EFS across all BL C (C1: HR, 0.43; 95% CI, 0.19–0.94; C2: HR, 0.90; 95% CI, 0.50–1.63; C3: HR, 0.41; 95% CI, 0.20–0.81), with least improvement in C2. The same clusters were detected at Sx (C1, 32.1%; C2, 33.9%; C3, 33.9%) and pts with C1 tumors had the highest risk of progression; 36-month EFS rates (95% CI) were 27.8% (12.1–46.0), 55.7% (33.7–73.0), and 77.2% (59.3–88.0) for C1, C2, and C3, respectively. Neoadj D was associated with higher proportions of pts with C2 and C3 phenotypes at Sx, such that the proportion with poor prognosis C1 tumors at Sx was 17.9% vs 47.2% in the D vs PBO arms, respectively. Among pts with C3 tumors at Sx, EFS benefit in the D vs PBO arm was striking (HR, 0.16; 95% CI, 0.03–0.79). Conclusions: The TME before and after neoadj Tx impacts EFS in pts with R-NSCLC, with an immune suppressed phenotype associated with reduced perioperative D benefit. The TME differs between squamous and non-squamous tumors and is influenced by neoadj D, which may promote an immune-activated phenotype associated with prolonged EFS benefit with perioperative D. Clinical trial information: NCT03800134 .

Genomic profiling of Burkitt lymphoma in adults: Novel molecular insights and implications for outcomes.

7085 Background: The real-world treatment outcomes of Burkitt lymphoma (BL) from low-middle income countries appear to be inferior (<60%). Delays in diagnosis, treatment interruptions and toxicities are often attributed as the reasons for poor outcomes. However, biological differences could also explain the observed differences. However, there is limited information on the molecular profile of adult BL from developing countries. Methods: We evaluated a cohort of adult BL (histopathologically confirmed) who received treatment at our centre between 2012-2019. The data on demography, stage, HIV status, treatment type, response and survival was recorded. The formalin-fixed paraffin blocks were subjected to targeted next-generation sequencing using the hybrid capture based targeted panel- SOPHiA Lymphoma Solution which identifies single nucleotide variants (SNVs), indels from 73 genes and gene amplification events in 68 genes. Results: Among a total of 165 patients, the adequate quality DNA was extracted from 76 patients(83% males). Median age was 34 years ;67% patients had stage III/IV disease,10% were HIV positive. Treatment regimen were high-dose methotrexate based-74 %, 24%- DA-EPOCH. Rituximab was given to 83% of patients. Complete response was seen in 69% of patients. 67 cases (88.2%) harboured at least one somatic variant, while no variants were detected in 9 cases (11.8%) within the genes covered by the targeted panel. A total of 712 somatic variants were identified across the cohort. Based on ACMG/AMP classification, 88 variants (12.4%) were categorized as oncogenic and 318 variants (44.7%) as likely oncogenic. This indicates a substantial mutational burden in Burkitt lymphoma, despite its classical characterization as a MYC-driven malignancy. Considering only oncogenic and likely oncogenic variants, the most recurrently mutated genes were KMT2D, CREBBP, ARID1A, EP300 and TP53 . TP53 showed the highest number of confirmed oncogenic variants (22 variants) and also accounted for the largest number of missense mutations (16 out of 158 missense variants). Mutations in PTPRD and ITPKB were mutually exclusive (p = 0.051), as were mutations in CCND3 and KMT2D (p = 0.053). These patterns suggest possible distinct molecular subsets of Burkitt lymphoma, potentially reflecting alternative oncogenic pathways driven either by aberrant cell cycle regulation ( CCND3 ) or by epigenetic dysregulation (KMT2D ). Conclusions: Our data suggests that histopathologically confirmed BL have complex mutational landscape characterized by frequent alterations in epigenetic regulators, tumour suppressor genes, and key signalling molecules which can lead to inferior outcomes.

Preclinical investigations and first-in-human phase I trial of KP-483 in solid tumors: Safety, antitumor activity, and preliminary efficacy.

2659 Background: KP-483 is a novel small-molecule antagonist of the E-type prostanoid receptor 4 (EP4) that potentially exerts antitumor effects by modulating the tumor microenvironment and restoring antitumor immunity. Based on preclinical and phase I studies, this report presents the pharmacological properties, safety, and preliminary efficacy of KP-483. Methods: In preclinical studies, the pharmacodynamic profile of KP-483 was characterized in vitro and in vivo . Antitumor activity was evaluated in tumor-bearing mice, and tumor infiltrating CD8⁺ and CD163⁺ cells were quantified. A first-in-human phase I study (jRCT2031220311) using a 3+3 dose-escalation design enrolled patients with solid tumors who had progressed after standard treatments or for whom no appropriate standard treatment was available. KP-483 was administrated orally once daily, with dose escalation across five cohorts (50, 100, 200, 400, and 800 mg). The primary endpoints were dose-limiting toxicities (DLTs) and safety profiles. Plasma concentrations of KP-483 were also measured. Preliminary efficacy was evaluated according to RECIST criteria, and the antitumor mechanisms were explored by flow cytometry analysis of tumor tissue. Results: KP-483 exhibited greater EP4 antagonist activity than existing antagonists (IC 50 : 1.0 nmol/L in human) and had higher binding affinity to the EP4 receptor and a longer dissociation half-life than PGE 2 in vitro . In mouse models, KP-483 exhibited dose-dependent antitumor effects following 14 days of oral administration at doses of 3, 30, and 300 mg/kg once daily. The number of CD8 + T cells in tumor tissue increased while that of CD163 + cells decreased in a dose-dependent manner. In addition, combination treatment with KP-483 (15 mg/kg, twice daily) and either anti-PD-1 or anti-PD-L1 antibodies enhanced antitumor effects compared with either monotherapy. In the phase I study, safety was assessed in a total of 19 patients. The median treatment duration was 43 days (range: 24–713 days), and no DLTs were observed. The most common treatment-emergent adverse events were anemia and nausea (each 26.3%), with only one study-drug-related grade ≥3 event (anemia). Systemic exposure to KP-483 increased with dose, and the plasma half-life of KP-483 ranged from 7.9 to 12.8 hours on treatment day 15. Best overall responses included stable disease in four patients and partial response in one patient. Flow cytometry analysis of tumor tissue showed a trend toward increased activated cytotoxic T cells, dendritic cells, and M1-like macrophages. Conclusions: KP-483 is a potent EP4 antagonist that exerts antitumor effects through modulation of the tumor immune microenvironment. Given its favorable safety and pharmacokinetic profiles, KP-483 represents a promising EP4 receptor–targeted therapeutic strategy with potential for combination with immune checkpoint inhibitors. Clinical trial information: jRCT2031220311.

Extrapulmonary small cell carcinoma treated with conventional or tumor-targeted topoisomerase I inhibition plus ATR blockade: Clinical outcomes and molecular correlates.

3027 Background: Extrapulmonary Small Cell Carcinoma (EPSCC) is a rare (0.1–0.4% of all cancers), aggressive neuroendocrine cancer with poor outcomes (Median PFS 2-4 months), no established second-line standard, and limited prospective data. We evaluated ATR inhibition with berzosertib combined with topoisomerase I (TOP1) inhibition with topotecan or sacituzumab govitecan (SG) and explored molecular correlates of response. Methods: Given the rare nature of EPSCC, patients were pooled from 3 prospective phase II trials evaluating topotecan or SG with berzosertib (NCT03896503, NCT02487095, NCT04826341) without intent for direct regimen comparison. Eligible patients had platinum-refractory EPSCC and received treatment at recommended phase II doses. Endpoints included efficacy and safety, with exploratory molecular correlates including somatic mutations, copy number alterations, circulating tumor DNA, tumor RNA sequencing, and pharmacogenomic analyses. Results: Of the 41 patients enrolled, 34 were evaluable (topotecan+berzosertib, n=20; SG+berzosertib, n=14). ORR based on RECIST was 10% (95% CI, 2.8–30.1) with topotecan+berzosertib and 21.4% (95% CI, 7.6–47.6) with SG+berzosertib; disease-control rates were 45% and 86%, respectively. Durable clinical benefit was observed in subsets of patients, with PFS ≥6.8 months and OS up to 19.7 months in the SG cohort, and OS >18 months in patients treated with topotecan+berzosertib. Patients with prolonged PFS represented multiple primary sites, including bladder, prostate and laryngeal. Grade 3–4 hematologic adverse events were more frequent with topotecan+berzosertib, whereas gastrointestinal toxicity and alopecia were more common with SG+berzosertib; no unexpected safety signals were observed. In the SG cohort, UGT1A1 intermediate and poor metabolizers experienced higher rates (90%) of grade 3–4 toxicities than normal metabolizers (10%). Early ctDNA declines correlated with radiographic response, while rising ctDNA preceded progression; baseline ctDNA correlated with tumor burden (r = 0.83; P = 0.006). Transcriptomic analyses demonstrated enrichment of E2F-driven and neuroendocrine lineage programs in responders (NES = 1.64–1.67; P < 0.001). Conclusions: In relapsed or refractory EPSCC targeting tumor replication stress through combined ATR and TOP1 pathway inhibition yielded clinically meaningful activity. Tumor-targeted TOP1 delivery with SG plus berzosertib achieved disease-control rate and a manageable safety profile. UGT1A1 phenotype was associated with toxicity in the SG cohort, supporting the relevance of pharmacogenomic-guided risk stratification. Correlative analyses identified E2F-driven transcriptional programs, neuroendocrine lineage signatures and early ctDNA suppression as promising biomarkers for future trials. Clinical trial information: NCT03896503 , NCT02487095 , NCT04826341 .

Clinical utility of repeat next-generation sequencing in patients with driver-negative non-small cell lung cancer: A report from the LC-SCRUM-TRY screening platform.

8626 Background: While precision medicine targeting driver oncogenes has improved clinical outcome of non-small cell lung cancer (NSCLC), the clinical utility of repeat or late-line multiplex genomic profiling remains unclear for patients whose tumors were initially identified as driver-negative. We evaluated the clinical impact of genomic screening for treatment-resistant NSCLC within the nationwide framework (LC-SCRUM-TRY). Methods: Between September 2020 and November 2025, a total of 2,088 patients with NSCLC were enrolled into LC-SCRUM-TRY. Of these, this analysis focused on 840 patients who were diagnosed as driver-negative at the time of enrollment. Genomic profiling was performed using Oncomine Precision Assay (ThermoFisher Scientific) for tumor tissue sample, or liquid biopsy (Guardant360 [Guardant Health] or liquid OPA). We compared overall survival (OS) among three groups: (A) Driver-negative, (B) Driver-positive without targeted therapy, and (C) Driver-positive with subsequent corresponding targeted therapy. Results: Among the 840 patients, median age at enrollment was 68 years (range 34-86); the majority were male (70%), ever-smokers (79%), and had adenocarcinoma (74%), with good ECOG-PS of 0-1 (91%). Of these, 89% underwent tissue-based NGS using OPA, while 11% underwent liquid biopsy. Actionable driver alterations were identified in 186 patients (22%), without significant difference between tissue- and liquid-based NGS. The detected drivers included EGFR mutation (mut) (n = 60), HER2 mut (n = 33), KRAS G12C mut (n = 31), BRAF V600E mut (n = 7), MET exon 14 skipping (n = 17), RET fusion (fus) (n = 17), ALK fus (n = 9), ROS1 fus (n = 7), and NRG1 fus (n = 5). Of the 186 patients with detected drivers, 74 patients (40%) received subsequent matched targeted therapies (Group C), including 13 patients enrolled in clinical trials. The median OS was 36.2 months in Group A, 34.1 months in Group B, and 70.2 months in Group C, respectively. Patients who received matched targeted therapies (Group C) had significantly longer OS compared to Groups A and B (p < 0.0001). Conclusions: Multiplex genomic profiling effectively identified actionable driver oncogenes in 22% of NSCLC patients previously considered driver-negative. Access to matched targeted therapies based on these results more than doubled the median OS (70.2 vs. 34.1 months), highlighting the critical importance of a multiplex genomic re-screening approach in this population. Clinical trial information: UMIN000041957.

Real-world incidence and clinical outcomes of trastuzumab/pertuzumab-induced thrombocytopenia in adjuvant HER2-positive breast cancer: 8-year retrospective multicenter study.

e23379 Background: Trastuzumab, alone or with pertuzumab, is standard adjuvant therapy for HER2-positive breast cancer. Real-world evidence describing the incidence, clinical course, and management impact of trastuzumab-induced thrombocytopenia (TIT) is limited. Methods: Multicenter retrospective cohort study across three UAE infusion centers (2016–2024). Electronic records of 87 women receiving adjuvant trastuzumab (IV/SC; originator or biosimilar) with or without pertuzumab were reviewed. All patients had platelet recovery after (neo)adjuvant chemotherapy (baseline > 150×10^9/L). TIT was defined as any platelet count < 150×10^9/L after HER2-targeted therapy initiation and graded by CTCAE v4.0. We assessed onset cycle, platelet nadir, recurrence (multi-cycle drops), transfusion requirement, and management actions (delay, discontinuation, regimen change). Predictors were explored using univariable and multivariable logistic regression. Results: TIT occurred in 35.6% and was typically early (median onset cycle 3). Events were predominantly mild (Grade 1: 29), with rare higher-grade events (Grade 2: 1; Grade 4: 1; no Grade 3). Mean platelet nadir was 140.1×10^9/L (SD 28.5; range 17–276). Recurrent/multi-cycle platelet drops occurred in 16/87 (18.4%). Platelet transfusion was required in 8% of TIT cases. Treatment impact was limited: > 90% continued HER2-targeted therapy without interruption; cycle postponement occurred in 2.5%, pertuzumab discontinuation in 2.5%, and regimen change in 1.3%. TIT incidence was similar with trastuzumab+pertuzumab vs trastuzumab alone (36.8% vs 31.6%; p = 0.68). No baseline demographic, stage, comorbidity, formulation, or regimen variables were associated with TIT risk. Conclusions: In this real-world multicenter UAE cohort, adjuvant trastuzumab ± pertuzumab was associated with a moderate rate of laboratory-defined thrombocytopenia that was usually early-onset and low-grade, with infrequent transfusion or treatment modification, although recurrence occurred in a meaningful minority. Routine platelet monitoring is appropriate; severe or treatment-limiting events appear rare. Logistic regression predicting thrombocytopenia. Predictor Adjusted OR 95% CI p Stage I (vs. II) 0.00 0.00 – ∞ .949 Stage III (vs. II) 1.09 0.32 – 3.72 .890 Regimen: Phesgo SC (vs. Trastuzumab alone) 0.65 0.10 – 4.17 .646 Regimen: T+P IV (vs. Trastuzumab alone) 0.74 0.17 – 3.17 .686 Age (per year) 1.01 0.94 – 1.07 .837 Weight (per kg) 0.99 0.95 – 1.03 .540 Number of cycles 1.02 0.93 – 1.12 .711 Chronic disease (yes vs. no) 0.87 0.24 – 3.10 .828 Baseline platelets <150 3.94 0.29 – 54.35 .305 Note. OR = odds ratio; CI = confidence interval. Reference groups: Stage II, Trastuzumab alone.

CV CARE: Cardiovascular care of androgen-related effects in prostate cancer patients.

e13581 Background: Cardiovascular disease (CVD) is a leading cause of death among prostate cancer (PC) survivors, and treatment with androgen deprivation therapy (ADT) negatively impacts metabolic factors and CVD risk. A standardized approach to delivering patient education regarding CVD risk remains a major unmet need. To address this, we implemented CV CARE (CardioVascular Care of Androgen Related Effects) to assess the feasibility of delivering advanced practice provider (APP)-driven personalized CV risk assessment and patient education in an oncology clinic at ADT initiation. Methods: Patients with PC receiving ≥ 24 weeks of ADT were enrolled in a DFCI IRB approved study within 12 weeks of starting ADT. CV risk labs (lipid panel, HbgA1c) were collected at enrollment, and patients received education from APPs at study start and 24 weeks later. Atherosclerotic Cardiovascular Disease (ASCVD) risk scores were calculated to determine personalized CVD risk, and the ABCDE (Awareness, Blood pressure, Cholesterol/Cigarette cessation, Diabetes/Diet, and Exercise) algorithm was used to educate on reversible CVD risk factors. Patients identified as high risk (ASCVD score of > 20%) were referred to cardio-oncology. Lipid levels, HgbA1c, ASCVD risk, and ADT effects were communicated via electronic medical record to patients’ care teams. Following study completion, patients and APPs were interviewed to provide feedback that will be used to revise the program. The primary endpoint is the feasibility of integrating the revised CV CARE program into the clinic workflow, as determined by continued participation rates (if ≥75% of patients remain in the program at week 24). Secondary endpoints include medication changes between ADT initiation and week 24, and clinician and patient reported satisfaction. Results: In total, 60 patients were enrolled into CV CARE (Table 1). 43.3% (26) completed the study, and 26.7% (16) remain in follow-up. Qualitative interviews assessing clinician and patient satisfaction with CV CARE are ongoing. Conclusions: We have completed enrollment in a study to explore the feasibility of integrating APP-delivered CVD risk assessment, patient education, and standardized multi-disciplinary communication for PC patients starting ADT. Ongoing assessments include feasibility, clinician and patient satisfaction, and changes in CVD-associated medications. This study was approved and funded by the National Comprehensive Cancer Network (NCCN) Oncology Research Program (ORP) from general research support provided by Pfizer Inc. and Sumitomo Pharma America, Inc. Clinical trial information: NCT06202820 . Baseline patient characteristics. Characteristics Overall N=60 (%) Age, Median, IQR 69 (64-73) Ethnicity (non-Hispanic) 95% (57) Localized Disease 70% (42) Current or Former Smoker 30% (18) Patients on Anti-HTN Medication 58.3% (35) Patients on Statin Medication 61.7% (37) Hypertension 71.7% (43) Hyperlipidemia 70% (42) Diabetes 20% (12)

Deep learning identification of high-risk lung adenocarcinoma histologic subtypes: A decision-support tool for treatment planning.

e20022 Background: Lung adenocarcinoma (LUAD) is the most common lung cancer subtype. Solid and micropapillary patterns represent high-grade histology with significantly worse outcomes: 5-year survival of approximately 60% and 40% respectively, compared to over 90% for lepidic-predominant tumors. Accurate identification is critical for adjuvant therapy decisions, as recent evidence suggests survival benefit from chemotherapy in patients with these high-risk components. However, subtyping suffers from interobserver variability (kappa 0.38-0.55), affecting treatment consistency. We developed an AI tool to identify high-risk subtypes for treatment stratification. Methods: We analyzed 143 resected LUAD whole slide images with pathologist-confirmed subtypes: acinar (n = 60, 42%), solid (n = 55, 38%), lepidic (n = 16, 11%), micropapillary (n = 10, 7%), and papillary (n = 5, 4%). High-risk was defined as solid or micropapillary predominant (n = 65, 45%) per WHO/IASLC guidelines. Tissue patches (224×224 pixels) at 20x magnification were processed using Virchow2 foundation model, selected based on superior five-class subtyping performance. ABMIL classifiers with gated attention mechanism were trained using 5-fold stratified cross-validation with class weighting and Youden index threshold optimization. Results: The model achieved AUC of 0.951±0.06 and balanced accuracy of 86.1%±4.2% for binary high-risk classification (Table 1). At the optimized operating threshold, sensitivity was 81.7%±13.7% with specificity of 90.5%±14.7%, yielding F1 score of 0.86. Positive predictive value of 90% indicates patients flagged high-risk are likely true positives, supporting chemotherapy consideration. Among AI-classified low-risk patients, 89% were confirmed true low-risk (NPV 88.6%±8.0%), identifying candidates for observation. These operating characteristics support use as a standardized second read to reduce variability in high-risk identification. Conclusions: This tool addresses a specific clinical dilemma: which resected LUAD patients warrant adjuvant therapy intensification. The 90% PPV means patients flagged high-risk can be confidently considered for chemotherapy; the 89% NPV identifies patients where observation may be appropriate - potentially sparing treatment toxicity without compromising outcomes. In tumor boards, AI-derived risk stratification provides objective data for treatment decisions, particularly in borderline cases where pathologist interpretation varies. External validation on multi-site cohorts (including biopsies) and outcome linkage (recurrence-free survival) are needed to confirm clinical utility. Binary high-risk detection performance. Metric Value AUC 0.951 ± 0.06 Balanced Accuracy 86.1% ± 4.2% Sensitivity 81.7% ± 13.7% Specificity 90.5% ± 14.7% PPV 90% NPV 88.6% ± 8.0% F1 Score 0.86

Targeting macrophage migration inhibitory factor: A phase 2 and pharmacodynamic study of sitagliptin in patients with progressive grade 4 gliomas.

TPS2099 Background: One mechanism of immunosuppression in the glioblastoma (GBM) microenvironment involves systemic and local accumulation of myeloid-derived suppressor cells (MDSCs) that inhibit cytotoxic immune cell populations and contribute to immune suppression. GBM patients have increased circulating MDSCs compared to lower grade glioma patients, and GBM patients with a better prognosis have reduced MDSCs in their tumors as well as in their peripheral circulation. A trial (NCT02669173) performed at the Cleveland Clinic demonstrated that pre-surgical anti-MDSC therapy (capecitabine) was associated with reduced circulating MDSCs and increased cytotoxic immune infiltration in tumor tissue. This proof-of-principle pilot study demonstrated that targeting MDSCs in patients can attenuate tumor-induced immunosuppression. Subsequent work at the Cleveland Clinic demonstrated that MDSCs require dipeptidyl peptidase 4 (DPP-4) for entry into the brain and overall MDSC function. Screening for a DPP-4 inhibitor identified sitagliptin as a good inhibitor with limited toxicity with efficacy in pre-clinical models. Hypothesis: Treat of GBM patients with sitagliptin will deplete circulating MDSCs and reduce their entry into the brain, reversing systemic and intratumoral immunosuppression. To test this hypothesis, we plan a “window of opportunity” clinical trial to evaluate the safety and biological impact of sitagliptin treatment in patients with recurrent grade 4 glioma undergoing clinically indicated surgical resection. Methods: For this trial, we will randomize 48 patients: 36 will receive pre- and post-operative treatment with sitagliptin and 12 will receive post-operative sitagliptin alone. All patients will receive post-operative sitagliptin and chemotherapy until disease progression. Primary endpoint: Difference in tumor CD8+ T cell count between the participants randomized to pre-surgical sitagliptin versus the participants randomized to no pre-surgical treatment. Secondary endpoints: PFS 6 , OS 12 , and safety. Exploratory endpoints: peripheral and intratumoral immune profiling for assessment of TAMs, MDSCs, and CD8+ T cells; tumor radiomic features on MRI brain pre-surgery that might predict intratumoral and peripheral CD8+ T cell count and peripheral MDSC levels; peripheral blood cytokine and immune gene expression profiling for increase in immune activation signatures. This is a window-of-opportunity trial of sitagliptin for patients with grade 4 gliomas to test the hypothesis that sitagliptin, a DPP-4 inhibitor, can reduce MDSC activity, enhance immune activation, and thereby increase the quantity of tumor infiltrated CD8+ T cells. These improvements in immune suppression, combined with chemotherapy will hopefully increase PFS and OS for patients. Clinical trial information: NCT07003542 .

A phase 3, randomized, double-arm, open-label, controlled study of ASP-1929 photoimmunotherapy (PIT) versus physician’s choice standard of care (SOC) for patients with locoregional, recurrent head and neck squamous cell carcinoma (HNSCC).

6080 Background: Locoregional recurrence is the main cause of morbidity and mortality in HNSCC, yet therapeutic choices are limited by sequelae of previous treatment and the potential for significant loss of function. ASP-1929 PIT is a novel cancer-targeted technology, utilizing an anti-EGFR monoclonal antibody conjugated to the dye IR700, that is activated by 690 nm light to induce rapid selective tumor cell destruction and trigger immune response. Methods: This global phase 3 study was conducted at 40 study centers located in the US, Taiwan, Japan, India, and Ukraine. Patients with locoregionally recurrent HNSCC who had failed or progressed on or after at least 2 lines of therapy were randomized 2:1 to the PIT arm or SOC. The planned sample size was 275. In the PIT arm, each cycle consisted of ASP-1929 infusion (640 mg/m 2 ) on Day 1, followed 24 ± 4 hours later by illumination (50 J/cm 2 superficial and/or 100 J/cm interstitial). Retreatment occurred ≥4 weeks apart, based on tumor response, for up to 8 cycles. In the control arm, patients received the physician’s choice of standard of care (docetaxel, cetuximab, methotrexate, or paclitaxel) until disease progression, intolerable adverse effects, or discontinuation of study treatment. Safety and efficacy outcomes were evaluated, with a data cutoff of 30 April 2025. Results: Active patient enrollment was discontinued due to challenges in patient recruitment and changes in the SOC landscape. As of 17 December 2024, 135 patients had been enrolled, with 68 patients experiencing progression or death and 36 patients in the ongoing long-term survival follow-up phase of the study. Median age was 62.0 years; 80.7% were male. In PIT and SOC arms, 64.0% and 63.0% of patients had received ≥ 3 prior therapy lines. Despite similar progression-free survival (HR 0.91; 95% CI 0.24 - 3.45), median overall survival was 15.7 months in the PIT arm compared to 9.6 months in the SOC arm (HR 0.83; 95% CI 0.50 - 1.36). Objective response rate was 25.8% in the PIT arm and 15.2% in the SOC arm, and disease control rate was 68.5% and 43.5%, respectively. Grade ≥3 treatment-emergent adverse events (TEAEs) occurred in 58.3% of patients in the PIT arm and 36.4% of patients in SOC; however, fewer TEAEs led to dose modification, delay, or interruption in the PIT arm (15.5% vs 30.3%). Conclusions: Considering the limitations of the study data, these results support ASP-1929 PIT as a tolerable and clinically active treatment option for locoregional, recurrent HNSCC and its ongoing evaluation in the current randomized global Phase 3 ASP-1929-381 study. Clinical trial information: NCT03769506 .

Interrogating acquired PD-1 resistance through BET bromodomain inhibition in hepatocellular carcinoma murine models.

e16154 Background: Immune checkpoint inhibitors targeting the PD-1 pathway have become the standard of care for hepatocellular carcinoma (HCC), yet many patients develop resistance. Combination strategies to overcome PD-1 resistance are needed. Bromodomain and extra-terminal (BET) family proteins are epigenetic readers that regulate transcription of immune and cancer-related gene programs in both tumor cells and immune cells. While preclinical studies suggest that targeting BET-mediated transcription may modulate the tumor immune microenvironment, its impact on PD-1 responsiveness in HCC remains poorly defined. Here, we test the hypothesis that combining BET inhibition with PD-1 blockade favorably remodels the HCC tumor microenvironment and enhances antitumor immune responses, thereby improving sensitivity to PD-1 directed therapy. Methods: Tumor-bearing syngeneic mouse model of HCC (Hep53.4 cell line, n = 100) were randomized and treated with mouse anti-PD-1 for 14 days to model early exposure to immune checkpoint blockade, followed by re-randomization of tumors into anti-PD-1, BET inhibition (molibresib), or combination therapy. Tumor samples were subsequently analyzed by Bulk RNA sequencing using Illumina TruSeq libraries and NovaSeq sequencing. Differential gene expression and pathway enrichment analyses (GSEA) were performed using DESeq2 and FGSEA. In parallel, tumors were harvested for immune profiling and tumor infiltrating lymphocytes were isolated and analyzed by Cytometry by Time-Of-Flight (CyTOF) using a 37-marker panel. Results: The combination of the BET inhibitor molibresib plus anti-PD-1 prolongs survival versus PD1 alone ( p = 0.047). RNA-sequencing analysis revealed marked upregulation of genes indicative of transcriptionally active B cells (CXCL13) and lymphoid-associated stromal features (BANK1, and LYVE1). In the GSEA ontology gene set, adaptive immune response, lymphocyte activation, T cell activation and immune response activation were enriched, suggesting enhancement of adaptive immunity. This immune-rich phenotype was confirmed by (CyTOF) profiling which similarly demonstrated a marked upregulation of B cells and M1-like macrophages. Conclusions: In this preclinical study, BET inhibition in combination with anti–PD-1 was associated with improved survival and may influence immune activation within the tumor microenvironment when combined with PD-1 blockade, possibly via upregulation of transcriptionally active B cells within the TME. Further studies are warranted to better define the underlying mechanisms and to further explore BET inhibition as a potential clinical target.