Browse Articles

Malnutrition as a hidden driver of inpatient mortality in gastrointestinal cancer: A national analysis of the U.S. National Inpatient Sample, 2016–2023.

e23235 Background: Inpatient mortality among patients with gastrointestinal (GI) malignancies remains high, yet risk stratification during acute hospitalization is limited. Malnutrition is frequently underrecognized in administrative data but may represent a marker of physiologic vulnerability. We evaluated the association between malnutrition and inpatient outcomes among hospitalized adults with GI cancer in the United States. Methods: We conducted a retrospective, survey-weighted analysis of the National Inpatient Sample (NIS), 2016–2023. Adult hospitalizations (≥18 years) with any-diagnosis GI malignancy (ICD-10-CM C15–C20, C22–C25) were included. Admissions with palliative care coding (Z51.5) were excluded. Malnutrition was defined using a narrow definition (E43, E44, E46) and a broad definition (E43, E44, E46, R64). The primary outcome was in-hospital mortality. Secondary outcomes included length of stay (LOS), hospitalization cost, acute kidney injury (AKI), shock, mechanical ventilation, dialysis, and discharge disposition. Survey-weighted regression models adjusted for demographics, payer, socioeconomic status, admission characteristics, hospital factors, cancer site, calendar year, and teaching status. Results: The analysis included 748,734 unweighted GI cancer hospitalizations after palliative exclusion. Malnutrition was present in 21.5% of admissions using the narrow definition and 22.8% using the broad definition. Crude inpatient mortality was higher among malnourished admissions (narrow: 4.96% vs 2.41%; broad: 5.21% vs 2.29%; both p < 0.001). Malnutrition was associated with longer LOS (9.27 vs 5.59 days), higher mean hospitalization costs ($31,904 vs $22,409), and increased rates of AKI (26.4% vs 17.3%) and shock (3.13% vs 1.61%). After adjustment, malnutrition remained independently associated with increased inpatient mortality (narrow: aOR 1.86, 95% CI 1.80–1.93; broad: aOR 2.05, 95% CI 1.98–2.12). In sensitivity analyses restricted to hospitalizations with a principal diagnosis of GI cancer, the association was stronger (aOR 2.15, 95% CI 2.01–2.31). Malnutrition was also independently associated with longer LOS (β +3.74 days, 95% CI 3.67–3.81) and higher costs (β 0.402, 95% CI 0.386–0.418), corresponding to an approximate 49.5% relative cost increase. Conclusions: Malnutrition is common among hospitalized patients with GI cancer and is independently associated with higher inpatient mortality, longer LOS, and increased resource utilization. These findings support malnutrition as an underrecognized marker of acute vulnerability and a potential target for inpatient risk stratification and quality improvement in GI oncology care.

Proactive response initiative for symptom management (PRISM): A multi-institutional symptom monitoring program for veterans receiving systemic cancer therapy.

11115 Background: Symptom assessment and management in cancer care are challenging. Within the Veterans Health Administration (VHA), gaps remain in widescale implementation of standardized symptom assessment, technology support, and adaptable workflows. PRISM was developed as a multi-institutional, technology-enabled clinical program to improve early identification and treatment of cancer-related symptoms for Veterans with the ultimate goal of system-wide implementation. Methods: We adapted an evidence-based proactive symptom intervention for the Long Beach VA with a working group of clinicians, pharmacists, social workers, case managers, and ancillary staff. The intervention involved a peer specialist conducting weekly telephone-based assessments using a modified Veteran Symptom Assessment System (VSAS) for Veterans initiating systemic anti-cancer treatment. Symptom scores ≥4 on a scale of 0-10 or a score increase by 2 points automatically triggered Nurse Practitioner (NP) review and follow up. We conducted an evaluation of our first PRISM site and assessed Veteran retention, symptom burden, interventions, and process metrics to inform scale and spread for two future VHA sites this quarter. Results: From May 22, 2025 to December 1, 2025, 30 patients (pts) enrolled in PRISM; 4 (13%) withdrew < 1 month, 2 (6.7%) withdrew > 1 month, 2 (6.7%) completed the program, and 1 (3.3%) died. Frequently reported symptoms of any score were fatigue (n = 14; 61%), dyspnea on exertion (n = 6; 26%), and pain (n = 6; 26%). Of the 412 times symptoms were reported as ≥4 on a scale of 0-10, the most frequent were fatigue (22%), dyspnea with exertion (13%), and drowsiness (13%). For the 23 pts who remained in the program for > 1 month there were 308 peer specialist encounters and 205 NP encounters. NPs intervened in 91% (186/205) of encounters with a total of 961 interventions. The mean number of interventions per NP encounter was 3.67. The most frequent intervention was pt education 86% (827/961), medication prescriptions 13% (123/961), and alerting the primary oncology team 4.2% (41/961). The mean time interval between a VSAS survey and indicated NP follow up was 1.06 days. The call duration was a mean 9.25 minutes per encounter with the peer specialist and 17.3 minutes per encounter with the NP. Documentation and chart review took a mean of 10.5 minutes per peer specialist encounter and a mean of 14.4 minutes per NP encounter. Conclusions: Early results from our first real world PRISM implementation in the VHA demonstrate feasibility, retention, and frequent early identification of and intervention for symptoms using this proactive symptom management program embedded within existing pt oncology care. The process metrics regarding capacity for peer support and NP effort will inform our expansion to two implementation sites by June 2026 and an additional 10 sites by 2027.

Single-cell multi-omics and TCR sequencing to identify early tumor cavity homing and phenotypic evolution of CAR-T cells in glioblastoma.

2554 Background: Chimeric antigen receptor T-cell (CAR-T) therapy has shown remarkable efficacy in hematological malignancies but remains largely ineffective in glioblastoma. A systematic understanding of early CAR-T cell trafficking and phenotypic adaptation within the tumor immune microenvironment may provide critical insights into this limitation. Methods: We performed single-cell multi-omics profiling using the 10x Chromium Single-Cell 5' Immune Profiling platform on infused CAR-T cells and matched tumor cavity cerebrospinal fluid (CSF) obtained via an Ommaya reservoir at day 3 post-infusion. Paired transcriptome/TCR data enabled analyses of composition, clonal tracking, and phenotypic transitions within CAR-T populations following early tumor cavity homing. Results: Single-cell profiling of the infused CAR-T product revealed a T cell–dominant population, primarily composed of CD8⁺ T cells (62.24%) and CD4⁺ T cells (34.63%). Analysis of tumor cavity CSF demonstrated a heterogeneous immune landscape, including CD4⁺ T cells (32.79%), CD8⁺ T cells (20.83%), dendritic cells, monocyte/macrophage populations, and other innate immune subsets. Gene set variation analysis across all immune cell populations showed enrichment of memory-associated programs in the CSF, while exhaustion-related signatures were not prominent, indicating a relatively non-exhausted immune state at this early time point. TCR-based clonal tracking identified 188 CAR-T cells in the CSF that originated from the infusion product, corresponding to an early homing rate of 2.19%. These homed CAR-T cells were predominantly derived from cytotoxic and proliferating CD8⁺ T cell subsets, as well as effector CD4⁺ T cells. Notably, cytotoxic CD8⁺ CAR-T cells exhibited pronounced clonal expansion in the CSF, whereas proliferating CD8⁺ CAR-T cells underwent phenotypic transitions toward cytotoxic states following tumor cavity infiltration. In parallel, activated CD4⁺ CAR-T cells preferentially adopted memory-like phenotypes. Consistently, TCR-tracked CAR-T cells in the CSF displayed significantly stronger memory-associated transcriptional signatures compared with their clonal counterparts in the infusion product (CSF vs. product, -0.187 vs -0.088，p = 2.48 × 10⁻⁹). Conclusions: This study provides a single-cell and clonal-level view of early CAR-T cell trafficking and phenotypic adaptation within the glioblastoma tumor cavity. Early tumor cavity infiltration is a selective process driven by cytotoxic and proliferative CAR-T subsets and occurs in the context of preserved memory-associated programs with limited exhaustion. These findings suggest that the limited efficacy of CAR-T therapy in glioblastoma is likely to stem from barriers that emerge at later stages following initial tumor entry instead of immediate CAR-T dysfunction.

From pixel to precision: A systematic review of CT radiomics in the diagnosis of the <i>EGFR</i> mutation in non–small cell lung carcinoma.

e20001 Background: Lung cancer is the leading cause of cancer mortality, with non-small cell lung cancer being more common. Radiomics is a non-invasive approach to detect actionable mutations for targeted therapy, such as EGFR in NSCLC, as current techniques like tissue biopsy and liquid biopsy are invasive and not always feasible. Radiomics enables the extraction of the various quantitative features from the CT images for the statistical or machine learning analysis of tumor genotype. This systematic review sheds light on the CT-based radiomics model for EGFR mutation prediction in NSCLC and shows its potential for the non-invasive EGFR assessment. Methods: We conducted a very detailed literature search on the various databases, including PubMed, PubMed Central (PMC), MEDLINE, Cochrane Library, and Wiley Online Library, using combinations of keywords such as lung cancer, EGFR mutation, and radiomics. Following that, duplicate articles were removed and screened based on title, abstract, and full-text screening. Following that, a total of 14 articles was finalized. All the studies were retrospective studies evaluating the CT radiomics approach for EGFR mutation in lung cancer. Quality analysis of the studies was assessed using the Methodological Radiomics Score (METRICS), and studies were classified as moderate to excellent quality. Results: This systematic review included 14 studies that evaluated CT-based radiomics and prediction AI models for the non-invasive prediction of EGFR mutation status in NSCLC. All the studies were retrospective, and most of them were single-institution, with sample sizes ranging from approximately 80 to 700 patients. Radiomics models were consistent with the discriminatory performance, with AUC values typically between 0.70 and 0.90. Models that incorporated peritumoral features or habitat-based features outperformed those using intratumoral features alone. Integration of the radiomics with clinical variables or deep learning-derived features yielded improvement in the performance, often reporting higher AUC (~ 0.92). The above performance findings were primarily based on the internal validation. External validation was not common, and if performed, it was associated with reduced performance. Conclusions: The findings indicated that CT radiomics detects CT imaging features that correlate with EGFR-driven behavior of the tumor in NSCLC and serve as a non-invasive test. However, heterogeneity in imaging protocols, feature extraction, and modeling strategies limits generalizability. While CT radiomics cannot replace tissue biopsy or liquid biopsy in current times, CT radiomics can be utilized as a complementary tool when molecular testing is not available in resource-limited settings. Prospective, multicenter studies with standardized workflows and external validation are essential before routine clinical implementation.

Real-world challenges in the adoption of emerging therapies for extensive-stage small cell lung cancer in community settings.

8113 Background: Despite advances in the treatment of extensive-stage small cell lung cancer (ES-SCLC), outcomes remain poor, and uptake of novel therapies in community oncology settings is inconsistent. As the ES-SCLC treatment landscape rapidly evolves, understanding real-world knowledge gaps, operational barriers, and challenges in patient-centered care is critical to improving evidence-based practice. This study evaluated community oncologists’ confidence, practice patterns, and perceived barriers to integrating novel and emerging ES-SCLC therapies. Methods: From September–October 2025, 100 US community oncologists completed surveys assessing treatment selection, shared decision-making, adverse event (AE) management, clinical trial referral practices, and system-level barriers to care delivery for patients with SCLC. Results: Knowledge and training gaps were the most frequently cited challenges to incorporating novel therapies (55%), followed by electronic medical record workflow limitations (41%) and patient selection complexity (33%). While most oncologists reported a basic understanding of novel antibody-drug conjugates (72%), few actively followed emerging clinical data (19%). Delays in initiating first-line therapy were commonly attributed to unfamiliarity with available treatments or need for academic referral (56%), prolonged staging evaluations (53%), and limited infusion center capacity (39%). AE management represented a major unmet need: 90% reported difficulty educating patients and caregivers on early AE recognition, and 62% reported challenges distinguishing disease progression from treatment-related toxicity. These gaps directly affected patient-centered care, with barriers to shared decision-making including limited clinician knowledge of treatment options (80%) and difficulty tailoring communication to patient preferences (66%). Accordingly, improved AE management was identified as the single greatest opportunity to enhance quality of care (51%). Although most oncologists reported referring patients for clinical trial evaluation after progression most or all of the time (85%; 4-5 on 5-point Likert scale), enrollment was hindered by patient preferences (63%), limited infrastructure to identify and enroll eligible patients (51%), and poor performance status (40%). Conclusions: Community oncologists face significant knowledge and operational barriers to integrating novel therapies for ES-SCLC, particularly related to emerging mechanisms of action, AE recognition and management, patient communication, and clinical trial enrollment. Targeted educational and practice-based interventions are needed to support timely, patient-centered, and evidence-based care in the community setting.

Markers of kidney tubule dysfunction and injury and long-term risk of acute kidney injury following coronary artery bypass graft surgery

Background Approximately one in five patients undergoing coronary artery bypass graft (CABG) surgery will develop post-operative acute kidney injury (AKI). We sought to determine whether biomarkers of kidney tubule dysfunction and injury are associated with long-term risk of acute kidney injury (AKI) after coronary artery bypass graft (CABG) surgery. Methods We performed a cohort study using data from the REasons for Geographic And Racial Differences in Stroke (REGARDS), a national, population-based, longitudinal cohort study of 30,239 U.S. adults aged ≥45 years. Of 760 REGARDS participants who underwent CABG surgery after their REGARDS baseline visit, we excluded those with a history of dialysis or kidney transplant, missing laboratory data or with illegible or erroneous records, leaving 394 in the analysis. Exposures included urinary biomarkers of kidney tubule dysfunction (urine alpha-1 microglobulin [A1M], uromodulin [UMOD] and epidermal growth factor [EGF]) and injury (kidney injury molecule-1 [KIM-1]), measured at an average of 5.5 years before CABG surgery. The primary outcome was AKI development following CABG surgery, defined as an increase in serum creatinine ≥0.3 mg/dL from 48 hours prior to CABG surgery to end of hospitalization. Results Of 394 participants, 176 (45%) experienced post-CABG surgery AKI. Higher baseline urine A1M was associated with higher odds of AKI (adjusted OR 1.34 per 2-fold higher A1M, 95% Confidence Interval (CI): 1.00–1.80). Higher urine UMOD was associated with lower odds of AKI (adjusted OR 0.77 per 2-fold higher UMOD, 95% CI 0.62–0.95). Higher EGF showed a tendency towards lower odds of AKI (adjusted OR 0.79 per 2-fold higher EGF, 95% CI 0.59–1.05). KIM-1 was not associated with AKI (adjusted OR 0.92 per 2-fold higher KIM-1, 95% CI 0.77–1.10). Conclusions Select biomarkers of kidney tubule dysfunction, but not injury, are associated with future risk of AKI after CABG surgery. Biomarkers measured years before surgery may predict post-CABG surgery AKI.

Euphorbia hirta L. – mediated green synthesis of Fe, Ni and Fe-Ni alloy nanoparticles: Characterization and colloidal stability

4D‐Printed Spin Crossover Metamaterials with Giant Programmable Positive or Negative Thermal Expansion (Adv. Mater. 32//2026)

Influence of variation in physicochemical properties of pesticides on their adsorption isotherms and leaching behavior in egyptian clay soil

Abstract The physicochemical properties of pesticides specifically solubility, volatility, and molecular structure are the primary factors determining their mobility. These factors govern their movement within the plant and their transport through the soil. The present work evaluated how variation in physicochemical properties influences the adsorption and leaching behavior of nine commonly used pesticides in tomato fields. The study includes a representative selection of major insecticides, fungicides, and herbicides to provide a comprehensive overview of their behavior in the soil. The adsorption capacity of the bulk soil was evaluated using the batch equilibration technique following two models. According to the Freundlich model, tebuconazole and difenoconazole were exhibited the highest adsorption affinity. On the contrary, the Langmuir model identified pendimethalin as having the strongest binding affinity to the soil followed by indoxacarb. Regarding the leaching experiments conducted in packed Egyptian soil columns, metalaxyl demonstrated the highest mobility, with 15% of the applied amount recovered in the leachate followed by acetamiprid. In contrast, the remaining pesticides exhibited low mobility, with the highest concentrations retained in the upper soil profile. These findings are step forward to develop management practices that mitigate the risks of pesticides, ensuring the protection of agricultural water quality and prevent groundwater contamination.

Correction: Acylation of Lysine 983 Is Sufficient for Toxin Activity of Bordetella pertussis Adenylate Cyclase

Effect of gut microbiota–derived testosterone on distant metastasis in non–small cell lung cancer.

8548 Background: Distant metastasis is the major cause of poor prognosis in non–small cell lung cancer (NSCLC). Emerging evidence suggests that gut microbiota and their metabolites influence tumor progression, but their role in NSCLC metastasis remains unclear. This study aimed to characterize gut microbiota and metabolic features associated with NSCLC metastasis and to elucidate the role and mechanism of the key metabolite testosterone. Methods: Fecal samples from 60 NSCLC patients (48 with distant metastasis, 12 without) were analyzed using metagenomic sequencing and untargeted metabolomics. Multi-omics integration with machine learning identified metastasis-associated metabolites. The effects of testosterone on NSCLC cell migration, invasion, and epithelial–mesenchymal transition (EMT) were evaluated in vitro (PC9, A549, H1299) and in vivo using a mouse metastasis model. Molecular mechanisms were investigated by Western blot, transcriptomics, Mendelian randomization analysis, and functional studies of FGF21. Finasteride was used as a pharmacologic antagonist. Results: Patients with metastatic NSCLC showed distinct gut microbiota profiles, with increased alpha diversity and altered community structure compared with non-metastatic patients. Opportunistic pathogens including Oscillospiraceae, Ruminococcus, and Actinobacteria were enriched in metastatic patients, while Fusobacteria was enriched in non-metastatic patients. Metabolomic analysis revealed significant enrichment of steroid-related metabolites in metastatic NSCLC, with markedly elevated testosterone levels. Testosterone demonstrated good predictive value for metastasis (AUC = 0.761). Testosterone significantly enhanced NSCLC cell migration and invasion and induced EMT, characterized by decreased E-cadherin and ZO-1 and increased N-cadherin and Snail expression. In vivo, testosterone promoted liver metastasis and EMT marker expression. Mechanistically, testosterone activated AKT/mTOR signaling, showed a causal association with AKT phosphorylation, and upregulated FGF21. Functional assays confirmed that FGF21 promoted EMT and metastasis via AKT/mTOR activation. Finasteride reversed testosterone-induced EMT, signaling activation, and metastasis by inhibiting the FGF21/AKT/mTOR axis. Conclusions: Metastatic NSCLC is associated with a distinct gut microbiota and metabolic profile. Gut microbiota–derived testosterone is a key metabolite that promotes NSCLC distant metastasis by activating the FGF21/AKT/mTOR pathway and inducing EMT. Finasteride effectively antagonizes this process, suggesting potential therapeutic value. The testosterone/FGF21/AKT/mTOR axis may serve as a biomarker and therapeutic target for metastatic NSCLC.

APS-TMA phenotypes as a failure-to-rescue signal during U.S. cancer hospitalizations: National Inpatient Sample, 2018–2022.

11041 Background: Failure-to-rescue (FTR), defined as mortality following major inpatient complications, reflects both patient disease pathology and hospital rescue capacity. Antiphospholipid syndrome (APS) and thrombotic microangiopathy (TMA) phenotypes are rare, high-acuity thrombotic syndromes observed during cancer hospitalizations, but whether excess mortality reflects independent risk factor versus complication-mediated risk is 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 diagnosis of malignancy were identified using ICD-10-CM codes (C00–C97; D45–D47). APS and TMA phenotypes were defined using any-diagnosis ICD-10-CM proxies for APS and TMA, including TTP/HUS, and combined as an APS/TMA phenotype. Because laboratory and medication data are unavailable in the NIS, exposures were analyzed as inpatient phenotypes rather than presumed diagnoses. Major complications included mechanical ventilation and acute kidney injury requiring dialysis. Failure-to-rescue was defined as in-hospital mortality among admissions with ≥1 major complication. National estimates incorporated discharge weights, hospital clustering, and stratification; multivariable models adjusted for All Patient Refined Diagnosis Related Group (APR-DRG) severity. Results: Among an estimated 4.81 million principal malignancy hospitalizations nationally, APS/TMA phenotypes occurred in 0.48% of admissions. Compared with non-phenotype admissions, APS/TMA phenotype hospitalizations had higher mortality (9.43% vs 4.35%) and higher rates of major complications, including mechanical ventilation (5.85% vs 2.57%) and dialysis-requiring acute kidney injury (2.11% vs 0.61%). In severity-adjusted modeling incorporating APR-DRG severity, APS/TMA phenotype was not independently associated with mortality (adjusted OR 0.99, 95% CI 0.70–1.40), indicating that excess mortality was largely mediated by complications and underlying illness severity. Conclusions: APS/TMA phenotypes identify a rare but high-acuity subgroup of cancer hospitalizations characterized by disproportionate ICU-level support and dialysis utilization, functioning as a national failure-to-rescue signal rather than an independent driver of mortality after severity adjustment. These findings highlight the importance of early recognition and severity-directed rescue strategies for high-risk thrombotic phenotypes during cancer admissions.

Hydroxyurea use and thrombotic outcomes in myelofibrosis patients with leukocytosis: A propensity-matched real-world analysis.

e18605 Background: Patients with myelofibrosis (MF), especially those with marked leukocytosis, face increased risk for thrombotic complications such as pulmonary embolism (PE) and venous thromboembolism (VTE). Hydroxyurea (HU) is often used to control blood counts. However, evidence for HU reducing thrombotic events in MF is limited, and contemporary real-world data on its impact in MF populations with significant leukocytosis remain sparse. Methods: We conducted a retrospective cohort study using the TriNetX research network. Adult patients with MF and leukocytosis (white blood cell count ≥25×10⁹/L) were identified. Patients with prior allogeneic hematopoietic stem cell transplantation were excluded. HU-exposed patients were compared with non-exposed patients using 1:1 propensity score matching. Matching criteria included age, sex, race, comorbidities, and baseline laboratory values when available. The study included 3,590 patients (1,795 per cohort). Primary outcomes were time to PE and VTE at 1, 3, and 5 years. Hazard ratios (HR) with 95% confidence intervals (CI) were estimated using Cox proportional hazards models. Results: In the propensity-matched cohort, HU exposure was not associated with a statistically significant reduction in thrombotic events. For PE, hazard ratios were numerically higher in the HU-exposed group at 1 year (1.5% vs 1.0%; HR 1.40, 95% CI 0.76–2.56; p=0.278) and 3 years (2.2% vs 1.8%; HR 1.29, 95% CI 0.80–2.08; p=0.289), with a modest increase observed at 5 years (2.8% vs 1.9%; HR 1.57, 95% CI 1.01–2.45; p=0.04). For VTE, no statistically significant differences were observed between HU-exposed and non-exposed patients at any time point (1 year: HR 0.95, 95% CI 0.63–1.44; p=0.821; 3 years: HR 0.97, 95% CI 0.69–1.36; p=0.85; 5 years: HR 1.14, 95% CI 0.84–1.57; p=0.40). Overall, PE and VTE event rates remained low and comparable between cohorts. Conclusions: In this large real-world cohort of patients with myelofibrosis and significant leukocytosis, hydroxyurea exposure was not associated with a statistically significant reduction in thrombotic events, including PE or VTE, over short- or long-term follow-up. Thrombotic event rates were low and generally comparable between hydroxyurea-exposed and non-exposed patients, with no consistent pattern suggesting a protective effect of cytoreductive therapy. Although a modest increase in PE risk was observed at 5 years, this finding was isolated and lacked a consistent temporal pattern, limiting its clinical interpretability. These findings suggest that the role of hydroxyurea in modifying thrombotic risk among MF patients with leukocytosis remains uncertain and highlight the need for individualized treatment decisions and further studies to better define its impact on thrombotic outcomes in this population.

Evaluation of a novel blood-based lung nodule triage test in elevated-risk subjects.

10547 Background: In the National Lung Screening Trial (NLST), 4,354 (6%) of LDCT scans were positive with Lung RADS scores of 3 or 4A, and 162 lung cancers (LCs) were diagnosed within one year of the associated scan. Despite the elevated risk of malignancy, referral for immediate diagnostic (Dx) workup is not recommended; as a result, diagnosis is delayed by an average of 106 days compared with subjects assigned a Lung-RADS 4B or 4X score, who are referred for immediate Dx workup. This delay highlights the need for methods to further risk stratify these subjects and enable more timely evaluation. Tumor-associated extracellular vesicles and particles (EVPs) are highly abundant in blood and carry surface biomarkers from their tumor cell of origin, providing a readily measurable analyte class. We report performance of an EVP-based Lung Nodule Triage Test in detecting LC as a function of Lung RADS score in plasma samples from NLST subjects. Methods: The Triage Test (biomarkers and classifier) was locked following a training study. Here we evaluated the Triage Test in a blinded case-control study nested within the ACRIN Biomarker Arm of the NLST. The study included LC cases diagnosed within 24 months of blood draw (n = 51 cases; 99 samples), including longitudinal samples from 48 subjects. Controls were age-matched NLST subjects who had no LC detected before or within the trial follow up period with Lung RADS v1.1 scores of 1 or 2 (n = 60). Triage Test results were determined blinded to case-control status, with unblinding done by ECOG-ACRIN. A static threshold set to 90% specificity was applied to Triage Test scores. Results: Among the 99 samples with a diagnosis of LC within 24 months of blood draw, the Triage Test exhibited 34% sensitivity for detection of LC. The Triage Test exhibited 43% sensitivity in samples associated with LDCT-positive screen results not recommended for immediate Dx workup. The projected positive predictive value (PPV) for the Triage Test applied as a reflex test for individuals with Lung RADS 3 and 4A is 13.8% (95% CI: 5.7% - 30.5%, bootstrapping test sensitivity and specificity). Conclusions: The blood-based Lung Nodule Triage Test exhibits clinically meaningful performance for detection of LC among subjects with LDCT Lung RADS 3 and 4A scores. The data supports further evaluation of the Triage Test for reflex testing of indeterminant lung nodules. Lung RADS Score 1 2 3 4A 4B/X 3+4A Total NLST LDCT Arm # LDCT Screens 39,840 28,674 2,282 2,072 1,898 4,354 74,766 NLST LDCT Arm # LC Dx Within 12M of Screen 26 100 32 130 428 162 716 NLST LDCT Arm LC Prevalence 0.07% 0.3% 1.4% 6.3% 22.6% 3.7% 1.0% NLST LDCT Arm Avg. Days to Dx vs Lung RADS 4B/X 93 101 128 84 n/a 106 n/a Present Study # Cases 23 37 5 9 25 14 99 Present Study Triage Test Sensitivity 30% (7/23) 32% (12/37) 60% (3/5) 33% (3/9) 36% (9/25) 43% (6/14) 34% (34/99) Modeled Implications of Reflex Testing # True Positive Workups 8 32 19 43 154 69 246 Modeled Implications of Reflex Testing # False Positive Workups 3984 2867 228 207 n/a 435 7287 Modeled Implications of Reflex Testing PPV 0.2% 1.1% 7.8% 17.3% n/a 13.8% 3.3%

Incidence, predictors, and recovery of acute chemotherapy-related left ventricular dysfunction in breast cancer patients: A prospective US cohort study.

e12528 Background: Advancements in breast cancer therapies, including anthracycline-based regimens and targeted agents like trastuzumab, have improved survival but heightened risks of cardiotoxicity, particularly chemotherapy-related cardiac dysfunction (CTRCD). This study assesses the incidence, predictors, and recovery patterns of acute left ventricular dysfunction (LVD) in US breast cancer patients receiving contemporary chemotherapy. Methods: In this prospective observational study, 97 adults with newly diagnosed breast cancer underwent serial echocardiographic assessment, including left ventricular ejection fraction (LVEF) and global longitudinal strain (GLS), before, during, and after chemotherapy. CTRCD was defined as a &gt; 10% decline in LVEF, LVEF &lt; 50%, or a &gt; 15% relative reduction in GLS. Clinical, oncologic, and treatment-related variables were analyzed. Results: Approximately one in eight patients developed CTRCD during chemotherapy. Nearly half of cases were detected by mid-treatment, highlighting the value of early surveillance. Left-sided breast involvement and pre-existing hypertension were more frequently observed among affected patients. Estrogen receptor–positive disease demonstrated a higher association with CTRCD. Trastuzumab-containing regimens showed a disproportionately higher occurrence of left ventricular dysfunction compared with other therapies. Importantly, following chemotherapy modification and guideline-directed heart failure therapy, over 75% of patients demonstrated improvement in left ventricular function within three months. Conclusions: CTRCD occurs early during breast cancer therapy but is frequently reversible with prompt detection and intervention. Focused cardiac monitoring, particularly in patients receiving trastuzumab or with left-sided disease, may enable early recovery and continuation of oncologic therapy.

AI-driven risk stratification for distant recurrence in node-positive HR+/HER2− early breast cancer: Independent validation in the NSABP B-28 trial.

558 Background: Accurate prediction of risk of distant recurrence (DR) in HR+, HER2-negative early-stage breast cancer (EBC) is important for optimizing adjuvant therapy decisions, including treatment escalation with CDK4/6 inhibitors (CDK4/6i). Current guidelines recommend consideration of CDK4/6i for node-positive, clinically high-risk EBC, including N1 patients. However, it is not known which patients will benefit, and guideline recommendations suggest that the risks may outweigh benefits for patients with low risk of distant recurrence. Identification of these patients will help prevent over-treatment. RlapsRisk BC (RR) is an AI pathology-based test that integrates features from H&amp;E-stained whole-slide images with clinicopathologic data (age, tumor size, nodal status) to assess risk of distant recurrence in HR+, HER2-negative EBC. The RR model was previously developed and validated using 7 retrospective cohorts totaling 6,039 patients. Here we report validation of RR in the NSABP B-28 cohort of HR+ HER2-negative N+ patients who received post-operative chemoendocrine therapy. Methods: This blinded independent validation study of the pre-specified RR included a subset of NSABP B-28 patients with HR+ and HER2-negative EBC (n=731) and digitized whole slide images from primary tumors. The primary endpoint was distant recurrence-free interval (DRFI). The objective was to validate the prognostic utility of RR score for DR. Univariable and multivariable Cox models were performed. Results: The evaluable cohort had a median follow-up of 11.1 years and included 504 (69%) patients with N1 disease, 87% grade 2/3 tumors, and a median tumor size of 21mm. Across the entire cohort, RR classified 58% of patients as low-risk and 42% as high-risk. RR (high vs low) was significantly associated with DRFI: HR=3.1 (95% CI: 2.3-4.3; p &lt; .001). Estimated 10-year DR-free was 87.5% (95% CI: 83.9-90.4%) for low-risk vs. 65.0% (95% CI: 59.1-70.2%) for high-risk patients. In a multivariable Cox model, the histology-only score remained significant after adjusting for clinicopathologic data: HR=1.6 (95% CI: 1.4-1.9; p &lt; .001). In the N1 subgroup, RR identified 65.7% of patients as low-risk with an estimated 10-year DR-free of 90.6% (95% CI: 86.9-93.4%), compared to 69.8% (95% CI: 62.2-76.2%) for high-risk. Conclusions: RR demonstrates robust prognostic performance in clinically high-risk N+, HR+, HER2-negative EBC patients from the NSABP-28 trial, in which the majority of patients had N1 disease. Our results demonstrate that approximately two-thirds of N1 patients, identified as low-risk by RR, exhibited favorable long-term outcomes with standard chemoendocrine therapies alone. As the clinical landscape shifts toward broader CDK4/6i use, RR could be used to identify N1 patients for whom the benefit of treatment intensification may be minimal.

Yttrium-90 SIRT followed by lenvatinib plus sintilimab for unresectable intermediate-to-advanced hepatocellular carcinoma: Phase II SIRLENS-90 trial.

4149 Background: Selective internal radiation therapy (SIRT) with yttrium-90 microspheres is an established locoregional therapy for unresectable hepatocellular carcinoma (HCC) and may induce immunogenic modulation, providing a rationale for PD-(L)1 blockade combinations. However, outcomes with SIRT–PD-(L)1 doublets suggest room for improvement. Lenvatinib has biological plausibility to augment radiation- and immunotherapy-mediated effects via VEGFR/FGFR pathway inhibition. We evaluated the efficacy and safety of sequential SIRT followed by lenvatinib plus sintilimab and explored serum protein biomarkers for outcome stratification. Methods: SIRLENS-90 (NCT05992584) is a single-center, open-label, single-arm phase II trial in unresectable BCLC B/C HCC (ECOG 0–1, Child–Pugh 5–7) suitable for SIRT after mapping and 99m Tc-MAA simulation. Patients received yttrium-90 resin SIRT (partition-model dosimetry) followed 3–7 days later by lenvatinib (8/12 mg daily) plus sintilimab (200 mg IV Q3W). Primary endpoint: PFS by mRECIST. Secondary endpoints: PFS by RECIST v1.1, OS, ORR, DCR, and safety (CTCAE v5.0). Exploratory Olink proteomics and LASSO-penalized Cox modeling derived a baseline risk score. Results: Thirty patients were treated (mean age 57 years [(SD 9.9], 93% male, 87% HBV-related; 43% BCLC B, 57% BCLC C; 43% macrovascular invasion; 53% bilobar disease; 37% extrahepatic metastases; mean max tumor diameter 9.6 cm [SD 3.8; range 3.2–16.1]). Median follow-up was 23.4 months. Median PFS was 15.8 mo (95% CI 8.3–20.7) by mRECIST (6-, 12-, and 18-mo rates: 77%, 53%, 40%) and 17.0 mo (95% CI 8.3–20.7) by RECIST v1.1. Median OS was not reached (12-, 18-, and 24-mo OS rates: 83%, 73%, 68%). ORR was 83% (95% CI 65–94) by mRECIST and 60% (95% CI 41–77) by RECIST v1.1; DCR was 90% (95% CI 74–98) by both. Intrahepatic ORR was 93% (95% CI 78–99) by mRECIST. Grade 3–4 treatment-related AEs occurred in 47%, with treatment-related SAEs in 10% and no treatment-related deaths. A baseline risk score incorporating IL-5 (protective), CXCL11 (adverse), extrahepatic metastasis, and bilobar disease stratified PFS (log-rank P = 0.001) and OS ( P = 0.02), with optimism-corrected C-index 0.758 (for PFS). Conclusions: Yttrium-90 SIRT followed by lenvatinib plus sintilimab demonstrated encouraging activity with manageable toxicity in unresectable intermediate-to-advanced HCC. The cytokine-integrated baseline risk score is hypothesis-generating and warrants external validation. Clinical trial information: NCT05992584 .

Real-world outcomes of crizotinib treatment for ALK+ and ROS1+ metastatic non-small cell lung cancer in Nepal.

e20689 Background: Lung cancer remains the most common and most lethal cancer in Nepal. Although targeted therapies have changed the treatment and prognosis of mutation-driven non-small cell lung cancer (NSCLC), access to these agents remains limited in low- and middle-income countries (LMICs) due to high costs and under-resourced health systems. Multistakeholder access programs offer a pathway to expand availability of molecular therapies in these settings. Here we describe treatment outcomes for patients with ALK- or ROS1-positive metastatic NSCLC treated at Patan Academy of Health Sciences (PAHS) in Nepal and supported by the Max Access Solutions program. Methods: A retrospective analysis was conducted of patients with ALK or ROS1-positive metastatic NSCLC treated with Crizotinib at PAHS and enrolled in Max Access Solutions between January 1, 2018, and June 30, 2025. Demographic data, treatment duration, and survival status were collected using the Patient Access Tracking System (PATS) and summarized descriptively. Results: Of the 930 patients with metastatic NSCLC, 123 patients were enrolled: 101 with ALK-positive (82.1%) and 22 (17.9%) with ROS1-positive disease. Median age was 53 years, with a near-equal male (52.5%) and female (47.5%) distribution. All patients initiated treatment shortly after diagnosis (median 1 month). At the time of analysis, 52 patients remained on therapy (median duration: 8 months), and 13 (25%) had been on treatment for over 2 years. Among the 71 patients no longer receiving therapy, the median treatment duration was 13 months. Reasons for discontinuation included death (51.3%), progression of the disease (36.4%), and loss to follow-up (12.3%). Conclusions: NSCLC continues to exert a significant toll globally. While mutation-driven therapies have improved the prognosis of ALK and ROS1-positive lung cancer in high-income countries, access remains limited in LMICs. This evaluation demonstrates the feasibility and clinical benefit of providing Crizotinib to patients with ALK and ROS1-positive metastatic NSCLC in a resource-limited setting. Timely initiation enabled sustained treatment for many patients, with some achieving long-term benefit beyond two years, comparable to outcomes reported in high-income settings. These findings highlight the importance of multistakeholder partnerships in improving access to and cancer outcomes in LMICs.

The risk of hyperglycemia with PI3K inhibitors in cancer patients: A meta-analysis.

e15128 Background: Phosphoinositide 3-kinase (PI3K) inhibitors are a promising therapeutic class in cancer treatment and their clinical utility has been limited by toxicity, particularly high-grade hyperglycemia. Currently the overall risk of hyperglycemia in patients treated with PI3K inhibitors has not been well understood. We performed a meta-analysis on the risk of hyperglycemia in patients treated with PI3K inhibitors based on currently available published randomized control trial (RCT) data. Methods: A systematic meta-analysis was conducted including phase II–III RCTs evaluating PI3K inhibitors in cancer patients. The primary endpoints were the incidence and relative risk of all-grade and high-grade hyperglycemia. Pooled effect sizes were calculated using random- or fixed-effects models based on the heterogeneity of included studies. Results: A total of 5199 patients across 14 eligible RCTs were included for analysis. The summary incidences of all-grade and high-grade hyperglycemia were 32% (1654/5199; 95% CI, 24%-37%) and 13% (637/4977; 95% CI, 5%-16%), respectively. In comparison with placebo controls, PI3K inhibition significantly increased the risk of all-grade hyperglycemia (RR: 1.39; 95% CI, 1.04–1.75, P&lt;0.01). Compared to controls, PI3K inhibition was associated with a significantly increased risk of high-grade hyperglycemia (RR: 2.27; 95% CI, 1.74–2.80; p &lt; 0.001), with moderate heterogeneity (I² = 38.9%). Subgroup analyses revealed that control type significantly moderated toxicity risk (p &lt; 0.001), with the greatest risk observed in trials with placebo controls (RR: 2.60; 95% CI, 2.05–3.15) and active controls (RR:1.50; 95% CI, 0.49–2.52). PI3K inhibitor subtype also moderated risk (p &lt; 0.001). PI3K-α selective inhibitors were associated with the highest risk (RR: 3.80; 95% CI, 2.50–5.11), followed by pan-PI3K (RR: 2.15; 95% CI, 1.79–2.52) and PI3K/mTOR dual inhibitors (RR:1.46; 95% CI, 0.61–2.31). Class I α/β/δ inhibitors showed no significant elevation (RR: 1.38; 95% CI, -0.782–3.53). Conclusions: PI3K inhibitors substantially increased the risk of all-grade and high-grade hyperglycemia, and may vary with inhibitor subtypes.

Distinct patient and product-specific phenotypes associated with toxicity and resource utilization after CAR T-cell therapy: A Nationwide Inpatient analysis.

11193 Background: Chimeric antigen receptor (CAR) T-cell therapy improves outcomes in relapsed or refractory hematologic malignancies but is associated with significant toxicity and inpatient resource utilization. Nationally representative real-world data remain limited. We evaluated the incidence, phenotype-specific predictors, and inpatient consequences of CAR T-cell–related toxicities in U.S. practice. Methods: We conducted a retrospective cohort study using the 2023 Nationwide Inpatient Sample, identifying adult elective hospitalizations involving CAR T-cell infusion using ICD-10-PCS New Technology Group 7 codes for FDA-approved CAR-T therapies. Interhospital transfers were excluded. Cytokine release syndrome (CRS), immune effector cell–associated neurotoxicity syndrome (ICANS), and cytopenias were identified using ICD-10-CM codes during the index hospitalization; toxicity grade was unavailable. Functional status was assessed using APR-DRG loss of function severity. Primary outcomes were coded in-hospital CRS, ICANS, and cytopenias. Secondary outcomes included transfusion requirement, length of stay (LOS), and total inpatient cost, adjusted to 2023 U.S. dollars. Survey-weighted multivariable regression models evaluated predictors of toxicity and resource utilization. Results: The weighted cohort represented 3,405 CAR-T hospitalizations nationwide. Indications included multiple myeloma (42.6%), diffuse large B-cell lymphoma (32.0%), follicular lymphoma (9.0%), mantle cell lymphoma (8.6%), B-cell acute lymphoblastic leukemia (7.4%), and chronic lymphocytic leukemia (0.4%). Cytopenias were coded in 45.2% of hospitalizations, CRS in 62.6%, and ICANS in 30.1%. In adjusted analyses, cytopenias were associated with older age, B-cell acute lymphoblastic leukemia, moderate-to-severe functional impairment, and CAR-T product type. CRS risk varied by CAR-T product type and was higher among transplant recipients and patients with severe functional impairment. ICANS was associated with transplant history, selected comorbidities (including metastatic solid tumors), follicular lymphoma, older age, and CAR-T product type. Median LOS was 12 days (IQR, 9–16) and was longer among patients with neutropenia. Transfusion support was required in 6.6% of hospitalizations. Median total inpatient hospitalization cost exceeded $1.3 million and was independently associated with diffuse large B-cell lymphoma, B-cell acute lymphoblastic leukemia, neutropenia, and autologous leukapheresis. Conclusions: CAR T-cell therapy is associated with substantial inpatient toxicities, prolonged hospitalization, and high cost, with risk varying by patient and product specific phenotypes. These findings support pre-infusion risk stratification and resource-aligned inpatient care models for CAR-T delivery.