Browse Articles

Impact of baseline demographics on therapy management in patients (pts) with metastatic castration-resistant prostate cancer (mCRPC) treated with talazoparib (TALA) + enzalutamide (ENZA) in the TALAPRO-2 study: Extended follow-up.

5043 Background: The randomized, Phase 3 TALAPRO-2 trial showed improved radiographic progression-free survival (rPFS) and overall survival (OS) in pts with mCRPC treated with TALA + ENZA vs placebo + ENZA in two cohorts: unselected and homologous recombination repair (HRR)-deficient pts. Median TALA duration of treatment (DoT) was 19.7 mo (unselected) and 20.3 mo (HRR-deficient). Study protocol required dose reduction for pts with reported Grade 3 or 4 hematologic adverse events (AEs); 55% of pts in both cohorts had a TALA dose reduction due to AEs at final analysis. We analyze the impact of baseline demographics on TALA dose reductions and clinical outcomes in the TALA + ENZA arm. Methods: Pts in the TALA + ENZA arm received TALA 0.5 mg (in case of moderate renal impairment [30–59 mL/min/1.73 m 2 ] 0.35 mg) + ENZA 160 mg PO QD. Time-to-event endpoints were summarized by the Kaplan–Meier method and compared by log-rank test. Results: In both cohorts, pts with a TALA dose reduction due to any cause (AEs and other reasons) were more likely to be older, be of Asian ethnicity, have lower bodyweight (BW), and have mild renal impairment vs pts without a TALA dose reduction (Table). Fewer pts with a TALA dose reduction than without discontinued treatment before wk 16 in the unselected (6.1% vs 20.0%) and HRR-deficient (2.7% vs 13.8%) cohorts. Median TALA DoT was longer in pts with a TALA dose reduction in the unselected cohort and similar between groups in the HRR-deficient cohort (Table). No clinically meaningful differences were observed in rPFS or OS in pts with a TALA dose reduction vs without in both cohorts (unselected: rPFS HR 0.94; 95% CI 0.70–1.25; OS HR 0.90; 95% CI 0.68–1.20; HRR-deficient: rPFS HR 0.96; 95% CI 0.64–1.44; OS HR 0.98; 95% CI 0.64–1.49; Table). Conclusions: Pts with specific baseline demographics (older, Asian ethnicity, lower BW, mild renal impairment) were more likely to have a TALA dose reduction. TALA dose reduction did not negatively impact TALA DoT or efficacy in pts who received TALA + ENZA in the TALAPRO-2 trial. Clinical trial information: NCT03395197 . Unselected with dose reduction(n=228) Unselected without dose reduction(n=170) HRR-deficient with dose reduction(n=111) HRR-deficient without dose reduction(n=87) Median age (range), y 73.0 (52–90) 70.0 (41–87) 73.0 (53–90) 68.0 (41–88) White, % 53.1 70.0 60.4 78.2 Black/African American, % 2.6 2.9 3.6 2.3 Asian, % 39.5 21.2 31.5 11.5 Median weight (range), kg 75.0 (45–169) 85.0 (59–155) 78.0 (45–118) 85.0 (60–135) Mild renal impairment (60–89 mL/min/1.73 m 2 ), % 52.2 36.5 59.5 34.5 Median TALA DoT (range), mo 23.0 (1.9–67.4) 15.2 (0.1–58.6) 20.3 (2.6–61.1) 20.8 (0.3–58.6) Median rPFS (95% CI), mo 33.1 (27.4–41.4) 33.0 (22.3–41.4) 33.1 (24.3–38.5) 30.2 (18.1–46.7) Median OS (95% CI), mo 46.9 (40–53.3) 45.1 (30.7–57.2) 41.9 (35.4–not reached [NR]) 48.4 (31.5–NR)

Integrating artificial intelligence to evaluate outcomes and disparities in young-onset breast cancer: A SEER-based retrospective cohort study.

1029 Background: Young-onset breast cancer (YOBC), defined as breast cancer diagnosed in women <40 years, is associated with aggressive tumor biology and affects patients differently depending on race, ethnicity, and socioeconomic status (SES). Traditional risk models may insufficiently predict disease severity and survival in this population. Artificial intelligence (AI) can integrate multidimensional clinical and demographic data to improve outcome prediction and identify high-risk subgroups. We evaluated AI-based prediction of advanced-stage disease and 5-year survival in a large U.S. cohort over two decades. Methods: We conducted a retrospective study using the SEER database, identifying women aged <40 years diagnosed with invasive breast cancer from 2000–2021. Variables included age, race, ethnicity, marital status, tumor stage, grade, receptor status, treatment, and county-level SES. A machine learning model was trained to predict advanced-stage disease (stage III–IV) and 5-year overall survival. Model performance was compared with multivariable logistic regression using area under the receiver operating characteristic curve (AUC). We also explored differences by race, ethnicity, and SES to understand disparities. Results: Our cohort included 97,462 women--38.9 percent non-Hispanic White, 26.7 percent non-Hispanic Black, 22.3 percent Hispanic, 11.1 percent Asian/Pacific Islander, and 1 percent American Indian/Alaska Native. Overall, 33 percent presented with advanced-stage disease, with higher rates among non-Hispanic Black (42 percent) and Hispanic (37 percent) patients compared with non-Hispanic White patients (26 percent, p<0.001). Women living in the lowest SES areas were 30 percent more likely to have advanced-stage disease than those in the highest SES areas. The AI model predicted advanced-stage disease (AUC 0.81) and 5-year survival (AUC 0.84) better than traditional regression (AUC 0.67 and 0.71, respectively). Importantly, 23 percent of high-risk patients identified by AI would have been missed by conventional criteria. Survival disparities persisted: 5-year survival was 77 percent for non-Hispanic Black women versus 89 percent for non-Hispanic White women, and 75 percent for women in the lowest SES areas versus 91 percent in the highest. The AI model’s predictions were consistent across all subgroups. Conclusions: As per the study AI-based models improved prediction of advanced-stage disease and survival while highlighting persistent racial, ethnic and socioeconomic disparities. These findings suggest AI-driven risk assessment can help identify high-risk young women and support equity-focused interventions to improve outcomes in YOBC.

Perioperative or adjuvant PD-1/PD-L1 blockade with curative-intent multimodality therapy for locally advanced head and neck squamous cell carcinoma: A systematic review and meta-analysis of randomized trials.

e18074 Background: Multiple phase III trials have evaluated the addition of perioperative or adjuvant/postoperative PD-1/PD-L1 blockade to curative-intent multimodality therapy in locally advanced head and neck squamous cell carcinoma (LA-HNSCC), yielding heterogeneous results across treatment strategies. This study aims to assess the efficacy and safety of immune checkpoint inhibitors in LA-HNSCC. Methods: PubMed, Embase, the Cochrane Library, and ClinicalTrials.gov were searched from inception through January 2026 for randomized controlled trials in LA-HNSCC comparing standard curative-intent multimodality therapy with perioperative (neoadjuvant ± adjuvant) or adjuvant/postoperative PD-1/PD-L1 blockade (including maintenance after definitive therapy) versus standard therapy alone; trials evaluating concurrent PD-1/PD-L1 blockade during definitive chemoradiation were excluded. The primary outcome was time-to-event efficacy, pooling event-free survival and disease-free survival as hazard ratios (HRs). Secondary outcomes included treatment-related serious adverse events and treatment-related deaths (grade 5 events). Results: Three phase III randomized trials were included (n=1,786). Pooled analysis demonstrated improved event-free/disease-free survival with checkpoint inhibition: HR 0.79 (95% CI 0.68-0.91, I²=0%). Treatment-related serious adverse events were more frequent in the checkpoint inhibitors group RR 1.79 (95% CI 1.43-2.24, I²=0%). Rates of treatment-related deaths (grade 5 events) were similar between the two groups, RR 1.90 (95% CI 0.50-7.14, I²=0). Conclusions: The addition of perioperative or adjuvant/postoperative PD-1/PD-L1 blockade to curative-intent multimodality therapy in LA-HNSCC was associated with a significant improvement in event-free/disease-free survival. Safety was characterized by higher rates of treatment-related serious adverse events, with no meaningful increase in treatment-related deaths. These findings support immune checkpoint inhibition as an effective adjunct to curative-intent multimodality treatment strategies in LA-HNSCC. Data summary table with event counts for each outcome. Trial EFS/DFS (events) (ICI vs Control) Serious treatment-related AEs(ICI vs Control) Treatment-related deaths(ICI vs Control) KEYNOTE-689 136/363 vs 159/351 69/361 vs 33/315 4/361 vs 1/315 NIVOPOST-OP (GORTEC 2018-01) 112/332 vs 140/334 104/312 vs 59/306 2/312 vs 2/306 IMvoke010 89/203 vs 92/203 7/202 vs 1/203 1/202 vs 0/203 Total 337/898 vs 391/888 180/875 vs 93/824 7/875 vs 3/824

Real-world evidence of pembrolizumab efficacy and safety in patients with NSCLC with high PD-L1 expression.

e20620 Background: Immune check-point inhibitors improved survival in patients with NSCLC and no driver mutations. Pembrolizumab, PD-1 inhibitor, compared to standard chemotherapy resulted in higher response rate and significantly longer progression free survival and overall survival in patients with high PD-L1 expression. Methods: Study was conducted at the University Hospital Center Zagreb, Croatia and assessed the real-world efficacy of pembrolizumab in 246 patients with PD-L1-high (PD-L1 expression ≥ 50%) metastatic non-small cell lung cancer (NSCLC). Patients were treated between March 2018 and April 2020, with exclusion criteria of positive EGFR mutations, ALK rearangements and ROS1 fusions. Results: The mean age of patients was 65 years, 69.5% were male and most patients (92.9%) had a history of smoking (52.7% were current smokers, 40.2% were former smokers). The most common histology was adenocarcinoma (67.5%). Brain metastases were present at diagnosis in 21% patients. Efficacy was assesed with median progression-free survival (PFS) of 22 months and a median overall survival (OS) of 32 months. At 12 months, 60.6% of patients were alive without disease progression. The objective response rate (ORR) was 48.1%, with a disease control rate of 75.8%. There was no significant difference in PFS and OS based on tumor type, smoking status, or PD-L1 assessment method. The only statistical significance was observed in mOS between patients with CNS metastases and those without (p = 0.013; HR = 1.80 (1.04-3.15). In 44 patients (18.1%), adverse event led to treatment discontinuation, and in 7 patients (2.9%) led to death. Among these leading to discontinuation, 50% were grade 1 and 2. The most common treatment-related adverse events of any grade causing treatment discontinuation were pneumonitis in 21/44 patients (47.2%) and, dermatitis in 9/44 (20.4%). Among serious adverse events, grade 3-5, pneumonitis was the most common in 6/44 patients (13.6%). Conclusions: Our study showed that pembrolizumab is effective and safe in a real-world setting, with comparable outcomes from clinical trials.

Lineage-specific outcomes and resource utilization associated with antiphospholipid syndrome and thrombotic microangiopathy phenotypes during U.S. cancer hospitalizations: National Inpatient Sample, 2018–2022.

e19110 Background: Antiphospholipid syndrome (APS) and thrombotic microangiopathy (TMA) phenotypes are rare, high-acuity thrombotic syndromes observed during cancer hospitalizations; whether their inpatient impact differs between hematologic and solid malignancies at a national level remains unclear. Methods: A serial cross-sectional, hospitalization-level analysis was performed using the 2018–2022 Healthcare Cost and Utilization Project National Inpatient Sample (NIS). Adult hospitalizations with a principal malignancy were identified (ICD-10-CM C00–C97; D45–D47) and classified as hematologic malignancies versus solid tumors. APS/TMA phenotypes were defined using any-diagnosis ICD-10-CM proxies for APS/TMA (including TTP/HUS) and combined as an APS/TMA phenotype. Because the NIS lacks laboratory and medication data, exposures were analyzed as inpatient phenotypes. Outcomes included in-hospital mortality, mechanical ventilation, acute kidney injury requiring dialysis, length of stay (LOS), hospital cost (cost-to-charge ratio–adjusted), and APR-DRG severity. National estimates accounted for NIS discharge weights, hospital clustering, and stratification; outcomes were compared within lineage by APS/TMA phenotype status and across lineages among phenotype-positive admissions. Results: Among an estimated 4,808,274 cancer hospitalizations nationally, 11.8% involved hematologic malignancies and 88.2% solid tumors; the APS/TMA phenotype occurred in 0.48% of admissions. In solid tumors, APS/TMA phenotype admissions had higher mortality (8.74% vs 4.06%), mechanical ventilation (5.37% vs 2.42%), and dialysis-requiring acute kidney injury (1.19% vs 0.39%) versus non-APS/TMA admissions. In hematologic malignancies, APS/TMA phenotype admissions similarly had higher mortality (13.04% vs 6.55%), mechanical ventilation (8.33% vs 3.67%), and dialysis-requiring acute kidney injury (6.85% vs 2.26%). Among APS/TMA phenotype admissions, hematologic malignancies had longer LOS (17.69 vs 9.34 days) and higher costs ($91,950 vs $36,058) than solid tumors. Mean APR-DRG severity was higher in APS/TMA phenotype admissions for hematologic malignancies versus solid tumors (3.48 vs 3.20) and exceeded non-APS/TMA admissions in both lineages (hematologic: 2.94; solid: 2.51). Conclusions: APS/TMA phenotypes identify a rare but high-acuity subgroup of cancer hospitalizations, with greater inpatient severity and resource use in hematologic malignancies than solid tumors. These national estimates highlight lineage-specific differences in ICU-level care, dialysis use, LOS, and cost, supporting heightened surveillance and rapid escalation for phenotype-positive admissions.

Associations between systemic inflammation, nutritional status, and cardiometabolic diseases and risk factors among adults living in transitional rural communities in Ecuador

Background Chronic low-grade systemic inflammation, as indicated by elevated high-sensitivity C-reactive protein (hs-CRP), has been implicated in the pathogenesis of cardiometabolic diseases. Less is known about the role of total immunoglobulin E (IgE), a marker of type 2 T helper cell-driven inflammation, in such outcomes. Information on the relevance of these markers in rural populations in transitional rural communities remains scarce, despite their growing burden of non-communicable diseases. Objectives This study examined the associations between two inflammatory markers—CRP and total IgE—and cardiometabolic risk factors and diseases in adults from transitional rural communities in coastal Ecuador. Methods We conducted a cross-sectional analysis of 931 adults from ten rural agricultural communities. Standardized questionnaires, anthropometric measurements, and fasting blood samples were used to collect data on sociodemographics, body composition, biochemical risk factors, and inflammatory markers (CRP and total IgE). Cardiometabolic outcomes included hypertension, type 2 diabetes, metabolic syndrome, and history of vascular disease. Multivariable regression models accounting for clustering at household and community levels and adjusted for age, sex, and their interaction were used to examine associations. Results Elevated CRP (≥3 mg/L) was prevalent (50.9%) and significantly associated with hypertension (adjusted odds ratio [aOR] 1.69, 95% CI: 1.19–2.39), type 2 diabetes (aOR 2.53, 95% CI: 1.77–3.64), and metabolic syndrome (aOR 3.24, 95% CI: 2.30–4.58). Elevated CRP was also strongly linked to multimorbidity (aOR for ≥3–4 vs. no conditions: 5.96, 95% CI: 3.52–10.08, P  < 0.001), as well as insulin resistance, low high-density lipoprotein, high triglycerides, and multiple adiposity measures. CRP associations were attenuated after adjusting for body mass index, suggesting adiposity as a mediator. In contrast, elevated total IgE (≥140 IU/mL) did not seem to be associated with cardiometabolic diseases. Total IgE levels were higher in men and associated with short stature, illiteracy, and obesity. Conclusions Elevated CRP was strongly linked to cardiometabolic diseases and risk factors in this population, consistent with a model in which adiposity is a primary upstream driver of systemic inflammation. These findings highlight the importance of inflammation as a potential modifiable risk pathway and support the utility of CRP as a screening tool in low-resource transitional settings.

Tunable ionic and photonic kinetics in InGaCdO broadband optoelectronic synapses for multimodal perception

Biological neural systems achieve multidimensional coupling of perception, memory, and processing through hierarchical information-integration mechanisms, enabling remarkably energy-efficient multimodal information computing. Inspired by this biological paradigm, we report an amorphous indium–gallium–cadmium–oxide (InGaCdO) optoelectronic synaptic transistor that exhibits tunable synaptic dynamics arising from the rich kinetics of ion-gated and photogenerated carriers in the oxide channel. Under optical and electrical stimuli, the device demonstrates typical synaptic plasticity behaviors, including excitatory postsynaptic current, paired-pulse facilitation, and spike-number-dependent plasticity. Leveraging this multimodal processing capability, we construct a reservoir computing system with integrated multisensory fusion. The multimodal system demonstrates a notable improvement in visual–audio recognition tasks, achieving a high accuracy of 91.3%. This study provides a device-level foundation for multimodal information-processing architectures and may facilitate the development of multisensory fusion intelligent systems and soft robots.

Ligands tuned Zn-MOFs with superior electrocatalytic performance for BPA sensor construction

ALP‐Responsive Luminescent Nanosheets Promote Precise Hepatitis Theragnostic by Co‐Targeting NLRP3 and Cholestasis

ABSTRACT Cholestatic hepatitis is a progressive inflammatory liver disease characterized by disrupted bile acid homeostasis, excessive reactive oxygen species (ROS) generation, and chronic inflammation, remains clinically challenging due to limited diagnostic precision and lack of effective, targeted therapies. Here, we developed a multifunctional theranostic nanoplatform, CyP‐CuGA‐UDCA nanosheets (NSs), that integrates therapeutic intervention with enzyme‐responsive disease monitoring. The platform features copper‐gallic acid (CuGA) nanozymes with superoxide dismutase‐, peroxidase‐, and catalase‐like activities for effective ROS scavenging and NLRP3 inflammasome inhibition. Co‐loaded with ursodeoxycholic acid (UDCA) to alleviate bile acid toxicity, and with an alkaline phosphatase (ALP)‐responsive near‐infrared fluorescent probe (CyP), the nanoplatform enables targeted treatment and real‐time imaging of cholestatic lesions. In a DDC‐induced mice model, CyP‐CuGA‐UDCA NSs significantly suppressed pro‐inflammatory cytokine production, reduced hepatic macrophage infiltration, attenuated oxidative stress and hepatocyte apoptosis, and alleviated fibrosis, outperforming monotherapies. Concurrently, the ALP‐activated fluorescence allowed precise visualization of cholestatic progression in vivo. This study presents a first‐in‐class nanostructured system that couples NLRP3 inflammasome modulation and bile acid regulation with enzyme‐specific diagnostics, offering a robust strategy for precision therapy and monitoring of cholestatic hepatitis.

Identification and validation of palmitoylation-associated biomarkers in major depressive disorder

Allelic variation alters expression and antigen presentation of MR1 allomorphs

Neoadjuvant targeted therapy versus chemoimmunotherapy in lung adenocarcinoma: A dual-center comparative study of surgical and survival outcomes.

e20071 Background: Although neoadjuvant chemoimmunotherapy (NCI) and targeted therapy (NTT) represent primary therapeutic strategies for locally advanced lung adenocarcinoma, their relative influence on surgical feasibility and long-term prognosis has not been fully elucidated. Methods: This retrospective, two-center investigation analyzed patients with lung adenocarcinoma who received radical resection following NCI (n=97) or NTT (n=105) between January 2020 and December 2022 at the Cancer Hospital CAMS and the First Affiliated Hospital of China Medical University. We compared surgical metrics (minimally invasive approach frequency, thoracotomy conversion rates, and morbidity), pathological response in the tumor bed and regional nodes, and survival outcomes (3-year OS and RFS) between the two cohorts. Results: Relative to the NCI group, the NTT group exhibited a significantly lower conversion rate (8.6% vs. 18.6%, P=0.04), decreased postoperative drainage (180 mL vs. 240 mL, P<0.01), and a reduced incidence of postoperative pneumonia (2.9% vs. 10.3%, P=0.03). While NCI achieved a superior major pathological response rate (9.3% vs. 1.9%, P=0.02), NTT was associated with improved 3rd-year RFS (73.4% vs. 58.6%, P=0.03). This survival advantage for NTT was particularly substantial in the gene mutation-positive subgroup (3rd-year OS: 89.5% vs. 78.9%, P=0.01; 3rd-year RFS: 72.0% vs. 53.7%, P=0.02). Conclusions: In lung adenocarcinoma, NTT facilitates reduced surgical complexity and fewer postoperative complications compared with NCI. Furthermore, NTT offers enhanced long-term survival benefits, especially for patients harboring specific gene mutations.

Elisrasib (D3S-001), a next-generation GDP-bound KRAS G12C inhibitor, as first-line therapy for <i>KRAS</i> G12C mutation–positive non–small cell lung cancer (NSCLC).

8511 Background: Elisrasib (D3S-001) is a next-generation KRAS G12C inhibitor (G12Ci) designed to improve target engagement (TE) efficiency and overcome growth factor-induced nucleotide exchange. This unique MoA distinguishes elisrasib from prior G12Ci and has demonstrated robust and durable anti-tumor activity in PDX models, and in combination with immunotherapy (IO) in immunocompetent models. Current ongoing phase 1/2 trial (NCT05410145) evaluates elisrasib as monotherapy (mono) or combination (combo) with pembrolizumab as first line (1L) therapy for advanced NSCLC harboring G12C mutation. Methods: Treatment naïve patients (pts) with G12C mutant advanced stage NSCLC were eligible. Elisrasib is administered 600mg QD orally in a 21-day cycle as mono or combo with pembrolizumab i.v. 200mg Q3W. The key objectives included safety and efficacy. ctDNA dynamic was analyzed by Guardant360 CDx or OncoCompass Target panels. Results: As of 06 Jan 2026, 43 pts and 52 pts received mono and combo. Median study follow up was 8.5m and 5.7m in mono and combo, respectively. In mono, 41 (PD-L1 TPS [22C3]: 21 &lt;1% and 20 ≥1%) out of 43 pts were efficacy evaluable. Overall ORR was 78.0% (32/41). Subgroup ORRs in TPS &lt;1% and ≥1% were 76.2% and 80.0%, respectively. Median PFS and DOR were immature. 6m PFS rate was 68.9% and 6m DOR rate was 77.2%. Above results provide first time evidence of G12Ci monotherapy in 1L NSCLC. In combo, 48 (PD-L1 TPS [22C3]: 17 &lt;1%, 11 1-49%, and 20 ≥50%) out of 52 pts were efficacy evaluable. Overall ORR was 81.2% (39/48). Subgroup ORRs in TPS &lt;1%, 1-49%, and ≥50% were 70.6%, 72.7% and 95.0%, respectively. Median PFS and DOR in the overall combo population were immature. 6m PFS rate was 74.6% and 6m DOR rate was 80.5%. Toxicity profile is summarized in Table. Baseline ctDNA G12C+ was detected in 90% (37/41) of mono and 80% (37/46) of combo. 35 mono and 25 combo pts completed on-treatment ctDNA analysis, with 83% and 100% achieved molecular response (≥90% G12C MAF reduction), respectively. PK at 600mg QD achieved C trough exposure of ~5nM and ~3nM with mono and combo, respectively, with overlapping variabilities at steady state, both well above the required exposure (1nM) for complete TE. Conclusions: Both elisrasib monotherapy and in combination with pembrolizumab show strong efficacy and good tolerability as 1L treatments for G12C-mutant NSCLC, warrant for randomized study to evaluate elisrasib as a potential new standard of care. Clinical trial information: NCT05410145 . TRAEs* 1L NSCLC Mono (N=43) 1L NSCLC Combo (N=52) Any Grade 41 (95.3%) 48 (92.3%) ≥G3 3 (7.0%) 17 (32.7%) LFT TRAEs* by PT (≥G3) ALT increased 0 4 (7.7%) AST increased 0 3 (5.8%) *TRAEs for combo cohort is related to elisrasib and/or pembrolizumab.

Liquid biopsy at the finish line: End-of-treatment ctDNA predicts relapse and survival in diffuse large B-cell lymphoma.

e19116 Background: PET/CT is the primary end-of-treatment (EOT) evaluation for diffuse large B-cell lymphoma (DLBCL); yet, false positives and insufficient prognostic clarity continue to be significant problems. Circulating tumor DNA (ctDNA)-based minimum residual disease (MRD) provides a tumor-informed, noninvasive way to assess recurrence risk. Methods: We used the MEDLINE, Embase, CENTRAL, and Web of Science databases to perform a systematic review and meta-analysis in accordance with PRISMA 2020 guidelines. DLBCL cohorts with end-of-treatment ctDNA-MRD status and survival outcomes were considered eligible research. The key outcome measures were progression-free survival (PFS) and overall survival (OS). Random-effects models calculated hazard ratios (HRs) and measured heterogeneity using I². Results: Eleven trials with 1,247 participants were included. Individuals with ctDNA-MRD had significantly worse progression-free survival (PFS) (pooled HR 5.48, 95% CI 3.96-7.58; I² = 42%) and overall survival (OS) (pooled HR 4.21, 95% CI 2.71-6.54; I² = 36%) compared to those without MRD. In MRD-positive patients, the 24-month recurrence rate was 68% compared to 14% in MRD-negative patients (absolute difference 54%, p &lt; 0.001), with ctDNA indicating return 4.2 months before clinical advancement. In studies reporting combined PET/CT, ctDNA-MRD showed superior predictive performance, with a higher positive predictive value (79% vs 41%) and negative predictive value (92% vs 78%) compared to PET positivity. Conclusions: End-of-treatment ctDNA-MRD is a strong predictor of DLBCL recurrence and survival, outperforming PET/CT for risk classification. Integrating ctDNA-MRD into post-therapy assessments may allow for early intervention and reduce the number of unnecessary biopsies. Clinical Takeaway: ctDNA-MRD at EOT identifies high-risk DLBCL patients missed by imaging and supports precision-guided post-remission management.

A generative model–assisted clinical decision support system for the postoperative management of colon cancer.

e15508 Background: Postoperative management of colon cancer relies on the accurate interpretation of complex pathology reports and adherence to evolving clinical guidelines. However, manual assessment is time-consuming and prone to numerical errors, resulting in staging misclassification and guideline-discordant therapy. Utilizing Large Language Models (LLMs) to bridge the gap between raw clinical data and standardized decision-making, while enhancing patient communication, remains an unmet clinical need. Methods: We developed and verified an LLM-based Clinical Decision Support System (CDSS) utilizing a dataset of 4,608 pathology reports from two tertiary hospitals and 409 cases from TCGA. The system implements a standardized "pathology-to-decision" workflow comprising five key modules: (1) Automated Staging: Extraction of unstructured data to automate TNM staging (AJCC 8 th edition) with explicit reasoning and field tracing; (2) Risk Stratification: Classification of MMR status and assessment of high-risk factors (e.g., lymph nodes &lt; 12, perineural invasion) to distinguish high-risk from low-risk patients; (3) Decision Support: Generation of guideline-concordant treatment recommendations and follow-up schedules calculated from the surgery date (NCCN Guidelines 2025 v4); (4) Prognostic RAG: A PubMed-based Retrieval-Augmented Generation module that synthesizes patient demographics and pathology features to answer queries regarding 5-year survival, recurrence windows, and genetic screening necessity; and (5) Patient Education: Simplification of complex reports into patient-friendly summaries. Results: Comparative analysis with manual expert review demonstrated that the CDSS significantly reduced the average time required for case processing by 98.3% (approx. 9.8s vs. 578.9s per case). In performance validation, the system achieved an accuracy of 96.8% for TNM staging and 97.9% for risk stratification, and reduced guideline-discordant treatment decisions. The PubMed-RAG module provided evidence-based prognostic information, and the patient summaries showed improved readability scores compared to original reports (Table 1). Conclusions: The CDSS improves postoperative staging accuracy, supports guideline-concordant treatment and surveillance, and facilitates patient-centered risk counseling. It provides a scalable approach to standardizing management of colon cancer across diverse clinical settings. Comparison of postoperative decision support approaches. Feature CDSS Manual Practice Staging Logic Deterministic (Rule-based) Experience-dependent Guideline adherence Real-time (AJCC/NCCN) Manual updates Traceability Fully traceable (Linked to source) Manual cross-checking Multilingual pathology support Yes Limited Patient support Integrated, guideline-based Time-consuming Time per Case &lt; 10 seconds 5-10 minutes

Sepsis risk factors and associated mortality in elderly mesothelioma patients with malignant pleural effusion.

e20063 Background: Age is an independent prognostic factor worse outcomes in mesothelioma patients. A significant proportion of sepsis related hospitalizations involve cancer patients, with sepsis playing a major role in cancer mortality and morbidity. However, specific data in elderly mesothelioma patients are limited. Our study seeks to identify determinants, trends and clinical profile and mortality outcomes of mesothelioma patients admitted with malignant pleural effusion. Methods: We used the 2016-2022 National Inpatient Sample to estimate the events of sepsis among elderly patients (ages 60 and older) with mesothelioma who were admitted with malignant pleural effusion. Multivariable regression models were adopted to identify independent predictors, and the overall mortality risks estimated. The data accounted for the weight, strata, and cluster for national estimates. Results: Among 4,395 elderly patients with mesothelioma and malignant pleural effusion, 325 (7.39%) developed sepsis during hospitalization. In multivariable analysis, independent predictors of sepsis included palliative care (AOR 3.53, 95% CI 1.81-6.89, p&lt;0.001), metastatic cancer (AOR 2.05, 95% CI 1.07-3.90, p=0.030), Hispanic ethnicity (AOR 3.38, 95% CI 1.21-9.47, p=0.021), and self-pay insurance status (AOR 9.79, 95% CI 1.24-77.11, p=0.030). Temporal trends showed increasing sepsis risk in recent years, with significantly higher odds in 2020 (AOR 4.27, p=0.035) and 2022 (AOR 6.56, p=0.006) compared to 2016. Traditional comorbidities including diabetes, chronic kidney disease, hypertension, congestive heart failure, and COPD were not independently associated with sepsis risk. Age, sex, hospital teaching status, geographic region, and socioeconomic status (by ZIP code income quartile) were also not significant predictors. Patients with sepsis experienced dramatically higher in-hospital mortality compared to those without sepsis (30.77% vs. 4.79%, p&lt;0.001). After adjusting for confounders, sepsis remained independently associated with a 7.6-fold increased odds of in-hospital death (AOR 7.58, 95% CI 2.78-20.67, p&lt;0.001). Conclusions: Although only a minority of the elderly mesothelioma patients with malignant pleural effusion incurred sepsis, it was associated with more than 7 fold higher odds of inpatient mortality. Sepsis risk was more profoundly driven by factors suggestive of advanced stage of cancer. These findings underscore the need for equitable care especially for those with late stage illness, marginalized insurance or ethnic backgrounds.

Gamma knife radiosurgery for brain metastases in the era of modern systemic therapy: A prospective real-world analysis using disease-specific GPA and RANO-BM.

e14026 Background: Gamma Knife radiosurgery (GKRS) is widely used for local control of brain metastases (BM), yet prospective real-world data incorporating disease-specific prognostic stratification and standardized response assessment in the era of modern systemic therapy remain limited. This study prospectively evaluated intracranial response, survival outcomes, and prognostic factors following GKRS. Methods: This was a prospective, non-randomized, single-institution observational study. Thirty consecutive adult patients (18–70 years; KPS ≥70) with 41 brain metastases treated with GKRS between March 2022 and June 2024 were included. Serial contrast-enhanced MRI at 1, 3, 6, and 12 months assessed treatment response using RANO-BM criteria. Disease-specific Graded Prognostic Assessment (GPA) scores were calculated based on primary tumor histology. Primary endpoint was local intracranial control (CR/PR/SD). Secondary endpoints included progression-free survival (PFS), overall survival (OS), and cause-specific mortality. Descriptive statistics were used; associations between prognostic variables and outcomes were explored using univariate analysis. Results: Median age was 57 years, and 63.3% were female. Lung (46.7%) and breast (33.3%) cancers were the most common primary tumors, and 96.7% of patients had extracranial disease. Most patients had solitary BM (66.7%). The median treated lesion volume was 1.5 cm³ (range, 0.03–15.4 cm³). Single-fraction GKRS was delivered in 95.1% of lesions, with a median marginal dose of 21 Gy. At last follow-up, local control was achieved in 63.3% of patients, with complete or partial response in 46.7%. Median intracranial PFS was 7 months and differed by GPA (6 months for GPA ≤2 vs 12 months for GPA &gt;2). Overall mortality was 53.3%, predominantly due to extracranial disease (40.0%). Median overall survival differed by GPA (10 months for GPA ≤2 vs 24 months for GPA &gt;2). No cases of radiation necrosis were observed. On univariate analysis, age, KPS, dose, treatment volume, primary tumor type, and GPA were not statistically significant predictors of local control. Conclusions: GKRS provides effective intracranial control with low neurological mortality in patients with limited brain metastases. Survival outcomes appear primarily driven by extracranial disease burden rather than intracranial response alone. Disease-specific GPA and RANO-BM offer a robust framework for outcome assessment in real-world GKRS practice. Further evaluation in larger prospective cohorts and real-world datasets may help refine patient selection and risk stratification.

Prognostic and predictive value of circulating tumor cells (CTC) phenotypes in CRS-HIPEC patients with colorectal and appendiceal peritoneal carcinomatosis: Insights from a phase II randomized trial.

e15065 Background: CTCs are a minimally invasive biomarkers in metastatic colorectal/appendiceal cancer, however, standard assays relying on epithelial CTC (CTC EpCAM) enumeration to predict disease recurrence often fail as CTC EpCAM do not fully represent dynamic heterogenous tumor environment and the epithelial-mesenchymal transition. We quantified CTC EpCAM and mesenchymal CTCs (CTC FAPα) using serial processing and calculated a mesenchymal-to-epithelial CTC ratio (FAPα /EpCAM, Φ) which was an early predictor of the disease progression in PDAC. We evaluated Φ as a predictive biomarker of cancer recurrence in colorectal/ appendiceal peritoneal carcinomatosis. Methods: A correlative study was performed within a Phase II RCT comparing mitomycin-C vs. melphalan CRS-HIPEC for colorectal/ appendiceal peritoneal carcinomatosis. We calculated the Φ ratio at pre-op (prior to CRS-HIPEC on Day 0) and post-op intervals. Longitudinal collections were analyzed to evaluate phenotypic trends prior to radiographic cancer recurrence. Results: 67 pts underwent CTC correlative study. For pre-op (Day 0) Φ, 39 were evaluated after excluding pts with no CTC. Recurrence occurred in 29/39 pts with a mean time to recurrence of 11.8 months, for these pts pre-op mean Φ was 1.58 (median 0.69) was not statistically different from Φ for patients without recurrence (mean 1.26; median 0.53). For post-op MRD analysis, we included 31 pts whose CTC were collected between Month 1 to 3. 23 pts recurred with a mean time to recurrence of 14.9 m, with post-op mean Φ was 1.33 (median 0.79) was not statistically different from Φ for patients without recurrence (mean 1.96; median 1.83). For 11 recurred patients for which longitudinal CTC collections was within 3 months of the later collection, Φ increased in 7/11 (63.6%) and decreased in 4/11 (36.4%). With measurable baseline Φ, the median % increase prior to recurrence was +192.4% (mean +259.2%, range +37.5% to +622.2%). Paired CEA was available in 9/11, with concordant directionality in 4/9 (44.4%). Conclusions: Single-timepoint CTC Φ ratio (pre-op or post-op MRD) did not significantly predict long-term recurrence likely due to the prolonged mean time to recurrence. However, frequent longitudinal evaluation of Φ as a part of MRD/surveillance testing did predict disease recurrence for 63.6% of pts at an earlier timepoint compared to standard imaging. The low concordance with CEA suggests that CTC phenotypic shifting offers superior surveillance value in peritoneal carcinomatosis undergoing CRS-HIPEC. Clinical trial information: 03073694 . Measure Pre-op Φ Recurrence Pre-op Φ No Recurrence Post-op Φ Recurrence Post-op Φ No Recurrence N 29 10 23 8 Age (Yrs) 58.62 61.7 57.4 61 FAP/EPCAM (Mean/Median) 1.58, 0.69 1.26, 0.53 1.33, 0.79 1.96, 1.83 Mean months to recurrence 11.8 14.9

Prediction of axillary response in 5,262 patients with node-positive breast cancer treated with neoadjuvant chemotherapy: Results from the prospective multicenter AXSANA/EUBREAST-03/AGO-B-053 study.

505 Background: Axillary lymph node dissection (ALND) for node-positive breast cancer (BC) converting to clinical node-negativity after neoadjuvant chemotherapy (NACT) is increasingly being replaced by less invasive procedures, such as sentinel lymph node biopsy (SLNB) or targeted axillary dissection (TAD). Concerns remain regarding the use of de-escalated procedures in patients with high initial nodal burden. We aimed to determine factors associated with nodal response, with a focus on tumor biology and nodal burden, to enable less extensive surgery without compromising oncological outcomes. Methods: AXSANA is an ongoing study investigating oncological and patient-reported outcomes after different axillary procedures in cN+ BC treated with NACT. In the present analysis, the impact of nodal involvement and tumor biology on axillary response was analyzed. The entire dataset is continuously and systematically monitored for data quality assurance. Results: 7,071 patients from 288 sites in 26 countries were included between June 2020 and January 7 th , 2026. Of these, 5,262 had completed surgery at the time of analysis. 2,341 patients (44.5%) had HR+ HER2- disease, followed by HR+ HER2+ (1,244; 23.6%), triple-negative (1,053; 20.1%) and HR- HER2+ (618; 11.7%). 91.3% of patients had an invasive ductal carcinoma of no special type (NST). 1,158 (22.1%) had ≥ 4 suspicious nodes at time of diagnosis. The highest nodal pCR rate (ypN0) was observed in patients with HR- HER2+ disease (86.1%), followed by HR+ HER2+ (70.7%), triple-negative (68.5%), and HR+ HER2- (30.5%; p &lt; 0.001). Nodal pCR rate was higher in patients with NST tumors (55.6%), compared to those with invasive lobular (35.2%) and mixed histology (43.8%; p &lt; 0.001). Patients with higher Ki67 (p &lt; 0.001), higher grading (p &lt; 0.001), multicentric tumors (p = 0.001), and without lymphangitis carcinomatosa (p = 0.046) were more likely to achieve nodal pCR in the univariate analysis. In contrast, the number of suspicious nodes at the time of diagnosis was not associated with axillary response (ypN0: 54.4% in pts. with 1-3 suspicious nodes vs. 53.6% in ≥ 4 suspicious nodes; p = 0.670). In the multivariable analysis, receptor status, Ki67, and grading, but not the number of suspicious lymph nodes at the time of diagnosis, were significantly associated with axillary response to treatment. Conclusions: This large prospective analysis shows that tumor biology, rather than the extent of nodal involvement is associated with axillary response to NACT. This challenges current guidelines and the common approach of restricting surgical de-escalation to patients with a low axillary tumor burden at presentation, and suggests that the selection of candidates for a potential de-escalation should be based on tumor biology rather than the extent of axillary disease at diagnosis. Clinical trial information: NCT04373655 .

Impact of ACE inhibitors (ACEi) and angiotensin receptor blockers (ARB) on outcomes in hepatocellular carcinoma treated with atezolizumab–bevacizumab.

4117 Background: Immune checkpoint inhibitor–based combinations (ICIs), including atezolizumab–bevacizumab (AB), have dramatically changed the therapeutic landscape of unresectable hepatocellular carcinoma (HCC), improving survival and response rates of patients, although outcomes remain heterogeneous. Concomitant medications may modulate immunotherapy efficacy through systemic and tumor microenvironment–mediated effects. Inhibition of the renin–angiotensin system (RAS) has been suggested to potentiate the effects of immunotherapy in various solid tumors by acting on angiogenesis, fibrosis, and immune modulation. However, the clinical impact of RAS inhibition in HCC patients receiving ICIs remains controversial, with retrospective studies suggesting possible improved outcomes. Methods: Patients with unresectable HCC treated from 2023 to 2025 with AB in Italy were included in this retrospective multicenter analysis from the ARTE database. Patients receiving concomitant ACEi/ARB at AB initiation were compared with non-exposed individuals. OS and PFS were estimated by Kaplan–Meier method and analyzed using Cox proportional hazards models. ORR and DCR were analyzed using logistic regression. To address baseline imbalances, a propensity score adjustment for potentially confounding characteristics (age, ECOG PS, viral etiology, alcoholic liver disease, cirrhosis, diabetes, obesity, cardiovascular events, decompensated cirrhosis, ALBI score, vascular/biliary invasion, previous locoregional treatments, metastatic disease and alpha-fetoprotein levels) was performed. Results: A total of 538 patients were included in our analysis, of whom 36% were receiving ACE-i/ARBs and 64% were not. ACEi/ARB use was associated with worse OS (HR 1.23; 95% CI: 0.97-1.57, p = 0.092), and a trend toward lower DCR (OR 0.63, 95% CI 0.39–1.02; p = 0.06). No significant associations were observed for PFS or ORR. After propensity score adjustment to balance baseline confounders, the magnitude of the OS difference was attenuated: median OS was 17.0 months in ACEi/ARB users versus 19.4 months in non-users (HR 1.14 (95% CI: 0.86-1.51; p = 0.358)), suggesting that the unadjusted signal was partly driven by baseline imbalance. Conclusions: In this large multicentric cohort of HCC patients treated with AB, concomitant ACEi/ARB therapy did not improve clinical outcomes and was possibly associated with a negative survival trend. Further studies will be needed to assess the potential effect of concomitant medications and RAS inhibitors on HCC, given their widespread use and the burden of comorbidities related to this specific tumor.