Browse Articles

Comparative performance of large language model derived and clinician documented ECOG-PS in predicting clinical outcomes in head and neck squamous cell carcinoma (HNSCC) undergoing concurrent chemoradiation (CRT).

e23101 Background: Clinician-documented ECOG-PS (dECOG-PS) often overestimates true ECOG-PS. The ability of large language models (LLMs) to infer functional status from clinical descriptors remains limited in non-English texts and with open-weight models suitable for low-resource deployment. Methods: Clinical notes from the visit immediately preceding CRT for HNSCC pts treated between 2011 and 2021 were identified. We optimized a zero-shot structured reasoning prompt with examples and applied it to six locally deployed LLMs: Gemma3-27B (GEM), GPT-OSS-20B (GPT), Ministral-3-14B (MIN), Mistral-Small-24B (MIS), DeepSeek-R1-14B (DPS), and Qwen3-14B (QWE) to infer ECOG-PS (llmECOG-PS), without outcome information. We then assessed llmECOG performance against dECOG-PS for five outcomes: incomplete CRT (≤200 mg/m² cumulative cisplatin), emergency department visits within 7 days (ER7D) and 30 days (ER30D), and mortality at 30 days (M30D) and 180 days (M180D). Discrimination was assessed using AUC, and effect sizes estimated by logistic regression odds ratios (OR) for ECOG-PS ≥2 versus 0–1. Results: Of 693 pts, dECOG was 0 (17.5%), 1 (73.9%), 2 (8.2%), and 3 (0.3%). Overall agreement (llmECOG = dECOG), down-classification (llmECOG < dECOG), and up-classification (llmECOG > dECOG) rates were: GPT 53.8%, 22.9%, 23.2%; MIS 51.8%, 4.1%, 44.1%; MIN 57.4%, 3.4%, 39.3%; GEM 65.6%, 5.3%, 29.1%; DPS 58.1%, 5.3%, 36.6%; and QWE 65.4%, 8.8%, 25.7%. MIN and GPT llmECOG showed superior discrimination for M30D (AUC: MIN 0.775, GPT 0.768 vs dECOG 0.651); however, only MIN outperformed dECOG for M180D (AUC 0.609 vs 0.594). All models outperformed dECOG for ER7D (AUC range 0.553–0.602 vs dECOG 0.549) and ER30D (AUC range 0.547–0.574 vs dECOG 0.546), except MIS and GEM. No model improved discrimination for incomplete CRT. Effect sizes for grouped ECOG (≥2 vs 0–1) were often higher and more frequently significant (OR > 1) for llmECOG than for dECOG (Table 1). Conclusions: Smaller, open-weight, privacy-focused model-derived ECOG showed improved discrimination compared to dECOG for emergency visits and mortality after CRT. This indicator may be useful for quality-of-care monitoring and risk stratification, particularly in low-resource settings. GPT MIS GEM MIN DPS QWE dECOG Incomplete CRT 1.6 (1.0-2.5) 1.5 (1.0-2.1) 1.5 (0.9-2.3) 1.3 (0.9-1.9) 1.3 (0.8-1.9) 2.1 (1.3-3.3) 2.3 (1.3-4.0) ER7D 2.4 (1.5-3.9) 2.2 (1.5-3.4) 1.5 (0.9-2.4) 1.7 (1.1-2.6) 1.9 (1.2-2.9) 2.3 (1.4-3.7) 2.0 (1.1-3.7) ER30D 1.8 (1.2-2.7) 1.9 (1.4-2.6) 1.5 (1.0-2.3) 1.7 (1.2-2.4) 1.7 (1.2-2.4) 1.8 (1.2-2.8) 1.9 (1.1-3.2) M30D 7.5 (1.2-57.2) 2.7 (0.4-20.9) 3.4 (0.4-20.1) 9.1 (1.3-178.7) 0.7 (0.0-4.5) 3.9 (0.5-24.0) 3.6 (0.2-28.8) M180D 1.8 (0.9-3.4) 1.6 (0.9-2.9) 2.1 (1.1-4.0) 2.4 (1.3-4.3) 1.5 (0.8-2.8) 2.4 (1.2-4.6) 2.3 (1.0-5.0)

Patterns of recurrence after perioperative FLOT for resectable gastric cancer: A systematic review and meta-analysis.

e16094 Background: Despite curative-intent resection and perioperative chemotherapy, gastric and gastroesophageal cancer (GC/GEJ) remains an aggressive malignancy with high rates of disease recurrence. Perioperative fluorouracil, leucovorin, oxaliplatin, and docetaxel (FLOT) is the current standard of care and improves survival outcomes; however, detailed anatomic patterns of recurrence in the FLOT era remain incompletely characterized. Improved understanding of site-specific recurrence may inform surveillance strategies and guide the development of rational systemic and locoregional therapies. Methods: We performed a systematic review and meta-analysis of three studies (one prospective randomized trial [ESOPEC] and three retrospective cohorts) reporting anatomic patterns of recurrence in patients with GC/GEJ treated with perioperative FLOT. Site-specific recurrence risks were pooled using random-effects meta-analysis of proportions with 95% confidence intervals (CI). Outcomes included peritoneal/pleural, liver, lung, lymph node, bone, brain, other distant, and isolated local recurrence. Between-study heterogeneity was assessed using the I² statistic. Results: Four studies encompassing 329 patients with documented recurrence were included. Distant recurrence was the predominant pattern, with the highest pooled recurrence risk was observed in the peritoneal/pleural cavity (0.13; 95% CI, 0.09–0.18; I² = 81%), followed by lymph node recurrence (0.07; 95% CI, 0.03–0.16; I² = 90%) and liver metastases (0.08; 95% CI, 0.05–0.12; I² = 69%). Lung metastases were less frequent (0.04; 95% CI, 0.03–0.06; I² = 0%). Brain (0.02; 95% CI, 0.01–0.04; I² = 4%), bone (0.03; 95% CI, 0.01–0.05), and other distant sites (0.03; 95% CI, 0.01–0.05; I² = 62%) were uncommon. Isolated local-only recurrence remained infrequent (0.05; 95% CI, 0.03–0.10; I² = 78%). Conclusions: In the contemporary FLOT era, recurrence after curative-intent resection of GC/GEJ cancer is predominantly systemic, with the highest absolute risk observed in the peritoneal/pleural compartment. Liver and nodal metastases represent the next most frequent failure sites, whereas isolated local recurrence remains less common. These findings indicate that most recurrences after treatment arise from occult systemic or peritoneal disease rather than from inadequate local control, and support prioritizing improved biomarker-driven systemic and peritoneal-directed strategies in future clinical trials over intensification of local therapy alone. Peritoneal-directed approaches remain investigational, and optimized systemic control represents the principal unmet need in resectable GC/GEJ cancer.

Spatial transcriptomics–guided computational pathology model to stratify docetaxel benefit in metastatic hormone-sensitive prostate cancer: CHAARTED trial (ECOG-ACRIN E3805).

5002 Background: The CHAARTED trial demonstrated improved overall survival (OS) with the addition of docetaxel (DTX) to androgen deprivation therapy (ADT) in metastatic hormone-sensitive prostate cancer (mHSPC), but many patients do not benefit and experience significant toxicity. Clinical factors (e.g., metastatic burden) are suboptimal for predicting DTX benefit, indicating a need for better biomarkers. We developed ST-DoxPCa (Spatial Transcriptomics–Guided Docetaxel Therapy Stratification in Prostate Cancer), an artificial intelligence (AI)-driven biomarker that infers intratumoral gene expression from routine H&E slides (via virtual spatial transcriptomics) to stratify patients by likelihood of benefiting from DTX. Methods: ST-DoxPCa was developed using paired H&E and spatial transcriptomics from the HEST1K prostate cohort to train a Vision Transformer–based regressor to predict a 208-gene panel at spot/patch level; predictions were distilled into a biologically informed 26-gene signature and summarized into patient-level features capturing expression level and heterogeneity. In CHAARTED, a Cox model for overall survival (OS) was trained and internally validated in the ADT-only arm (n = 154; 50/50 split: n = 77 train, n = 77 validation) to derive an ST-DoxPCa risk score and fixed median threshold. This locked model was then applied to the ADT validation cohort (n = 77) and the held-out ADT+DTX arm (n = 129), and OS was compared between treatment arms within biomarker-defined risk groups using Kaplan–Meier estimates and Cox hazard ratios. Results: ST-DoxPCa stratified patients into biomarker-defined risk groups using aggregated spatial-expression features from a biologically informed gene panel; key genes included AR, PTEN, FASN, FOXA1, and ACPP. In ST-DoxPCa-positive (high-risk) patients, adding DTX to ADT significantly improved OS versus ADT alone (HR = 0.53, 95% CI 0.31–0.90; p = 0.018), with no benefit in ST-DoxPCa-negative (low-risk) patients (HR = 1.32, 95% CI 0.74–2.34; p = 0.34). Conclusions: ST-DoxPCa is a novel spatial transcriptomics guided histology biomarker that predicts differential benefit from docetaxel in mHSPC. ST-DoxPCa identified a subset of patients who achieve substantial survival gains from upfront DTX and another subset with no benefit, enabling personalized treatment intensification. ST-DoxPCa has the potential to spare low-risk patients unnecessary chemotherapy while directing therapy to those most likely to benefit. Clinical trial information: NCT00309985 .

Rural-urban disparities in chemotherapy administration and quality certifications in Minnesota.

e13563 Background: Rural cancer patients experience disparities in access and outcomes compared to their urban counterparts. Quality certifications like ASCO's Quality Oncology Practice Initiative (QOPI) and the American College of Surgeons' Commission on Cancer (CoC) accreditation aim to standardize high-quality care. We examined rural-urban differences in chemotherapy administration sites and certifications in Minnesota. Methods: Using the Minnesota All-Payer Claims Database (2022), we identified billing providers with claims for carboplatin and paclitaxel (CPT: J9045, J9264, J9267). Providers were geolocated and assigned Rural-Urban Commuting Area (RUCA) codes, categorized as urban (RUCA 1-3) or rural (RUCA 4-10). QOPI certification status was obtained from ASCO, and CoC accreditation from ACS. Results: Among 104 providers, 55 (53%) were rural, accounting for 10,117 (27%) of 37,526 claim lines. No rural providers (0%) were QOPI-certified, compared to 16% in urban areas. Rural providers had lower CoC accreditation rates (26% vs 43% in urban areas). Conclusions: Rural Minnesota providers comprise over half of chemotherapy sites but deliver a minority of treatments and lack QOPI certification. Expanding QOPI and CoC programs in rural areas represents a key opportunity to improve equitable, high-quality cancer care delivery across America.

Restoring form, rethinking function: The dual-edged impact of maxillofacial prostheses on survival and complication outcomes in head and neck cancer—A propensity matched analysis.

e18125 Background: Maxillofacial prosthetic rehabilitation restores form and function in patients with head and neck cancer (HNC) following surgical resection. Beyond aesthetic and functional benefits, its impact on long-term survival and clinical outcomes remains understudied. This study evaluates survival and clinical outcomes, including mortality, severe sepsis, malnutrition, oral cavity diseases, respiratory diseases, and dental care utilization, in HNC patients with and without maxillofacial prostheses. Methods: This retrospective cohort study analyzed deidentified electronic health records from TriNetX research network, from 88 U.S healthcare organizations. Patients with malignant neoplasms of the lip, oral cavity, or pharynx were identified. Cohort 1 included patients without maxillofacial prosthesis (n = 103,153), and Cohort 2 included those with prosthesis (n = 703). Propensity score matching balanced cohorts on demographics, BMI, and ECOG, yielding 695 matched patients per cohort. The primary outcome was all-cause mortality. Secondary outcomes were severe sepsis, malnutrition, oral and salivary gland diseases, respiratory diseases, and dental care utilization. Analyses included risk ratios, odds ratios, Kaplan-Meier survival estimates, and event frequency analysis. Outcomes were assessed from one day after the index even t(first qualifying diagnostic or prosthetic procedure) with no specific end date. Results: After PSM, cohorts were well-balanced (standardized differences <0.10 ). Mortality risks were similar (25.1% vs 24.9%; risk ratio [RR] 0.993; P = .94), but prosthesis users had lower survival probability (43.60% vs 46.12%; hazard ratio [HR] 1.348; P = .001). Sepsis risk was higher in non-prosthesis users (5.0% vs 2.2%; RR 2.304 ; P = .005; HR, 2.900; P < .001). Malnutrition risks were similar (18.5% vs 15.7%; RR 0.849; P = .22). Oral /salivary gland diseases were more common in prosthesis users (64.5% vs 22.8%; RR 0.353; P < .001; HR, 0.379; P < .001), with higher mean events (7.775 vs 2.848; P < .001). Respiratory diseases showed similar risks (45.1% vs 39.2%; RR 0.868; P = .15), but non-prosthesis users had more mean events (6.035 vs 3.478; P = .008). Dental examinations were more frequent in prosthesis users (10.0% vs 2.6%;RR 0.263 [95% CI, 0.154-0.449]; P < .001; HR, 0.332; P < .001), with fewer events (1.386 vs 7; P = .002). Conclusions: Maxillofacial prosthetic rehabilitation in HNC patients has a multifaceted impact with lower risk of sepsis but higher oral cavity complications, higher dental care utilization, and lower overall survival. These findings highlight the need for multidisciplinary planning integrating prosthodontic, oncologic and supportive care to optimize functional and clinical outcomes.

OncoSphere AI: An agentic AI multidisciplinary clinical workflow support tool.

e13656 Background: Effective multidisciplinary cancer care is bottlenecked by the cognitive burden of synthesizing fragmented electronic health records across different specialties. Current AI tools often function as isolated chatbots rather than integrated workflow assistants. We developed and validated "OncoSphere AI," an Agentic AI multidisciplinary clinical workflow support tool designed to autonomously orchestrate the multidisciplinary review process by deconstructing complex cases into domain-specific insights. Methods: OncoSphere AI is a novel, multi-agent large language model–driven platform that simulates a multidisciplinary tumor board for cancer care. The architecture uses a parallel reasoning framework in which eight virtual specialty agents—Medical Oncology, Radiation Oncology, Surgical Oncology, Pathology, Radiology, Pharmacy, Clinical Research, and Nurse Navigation—independently analyze structured and unstructured patient data against specialty specific guidelines and contemporary peer-reviewed literature. System performance was assessed via built-in telemetry (task completion, latency, inter-agent agreement) and blinded review by eight expert oncologists using a structured survey to rate data extraction fidelity, guideline concordance, safety, and clinical utility compared with standard manual review. Results: The study cohort (n=20, median age 59.5 years; 55% female) had a median of 2.5 (range 0-7) prior lines of therapy and multiorgan metastatic disease (median: 2, range 1–4), including lung, breast, gastrointestinal (CRC, pancreatic, hepatobiliary), and genitourinary (prostate, renal) malignancies. OncoSphere AI demonstrated high workflow utility (mean rating: 4/5) and strong extraction fidelity, with more than 90% of biomarker extractions rated as a “perfect match” by expert reviewers. Operational efficiency gains were substantial with AI processing averaged 4.7 minutes per case versus 29.3 minutes for manual review, corresponding to an 84% reduction in time (p<0.001). System telemetry revealed adaptive compute behavior, with the platform autonomously allocating 1.73-fold more reasoning tokens to complex, refractory cases than to standard scenarios. Safety review identified only one critical contraindication issue in a heavily pretreated, multi-line setting. Conclusions: OncoSphere AI effectively automates the clinical “perception layer,” consistently transforming heterogeneous longitudinal data into decision-grade insights with high fidelity and marked time savings. By operating as a coordinated ensemble of domain-specialized AI agents rather than a single generic model, it offers a scalable approach to reduce physician cognitive burden and to standardize guideline-concordant care planning in diverse, heavily pretreated oncology populations.

The characterization and management of immune checkpoint inhibitor–induced pruritus: A retrospective review.

e24167 Background: Immune checkpoint inhibitors (ICIs) are monoclonal antibodies widely used in cancer treatment to block inhibitory pathways and activate antitumor immunity¹. ICIs can contribute to a vast array of dermatologic immune-related adverse events (d-irAEs) 2 . Pruritus is the second most common d-irAE, occurring in 13-20% of patients 3 , however significant gaps still remain in the understanding and management of ICI-induced pruritus. This study aimed to characterize ICI-induced pruritus, describe treatment patterns, and assess its impact on adherence to immunotherapy. Methods: A retrospective review of medical records for cancer patients treated with PD-1, PD-L1, or CTLA-4 inhibitors who had an ICD-10 pruritus code (L29.X) was performed. 971 individuals were manually verified to have ICI-induced, with a median age of 71 (range, 23-101). The cohort was 59.7% male, 62.9% White, 20.0% Asian, 9.3% Hispanic or Latino, and 3.0% Black or African American. Results: Pruritus onset occurred after a median of 2 infusion cycles (IQR, 1-4) and persisted for a median of 195 days (IQR, 70-440). Combined PD-1/CTLA-4 inhibitors had a significantly earlier symptom onset compared to PD-1 inhibitor monotherapy (t(415) = 6.73, p < 0.00001). Most patients received topical corticosteroids (96.0%) and/or antihistamines (91.5%). Systemic treatments included oral corticosteroids (14.8%), gabapentinoids (5.3%), and dupilumab (1.9%). Nearly 1 in 5 patients experienced QoL impairment, affecting sleeping, eating, and showering. 49.1% had a recorded CTCAE grade 4 of 1, 23.7% were Grade 2, and 27.2% were Grade 3. Overall, 12.2% of patients paused treatment due to pruritus, and 6.3% had treatment discontinued. Grade 3 patients were 12 times more likely to experience a treatment pause than Grade 1 patients (OR 12.40; 95% CI 5.91-29.18; p < 0.00001). Conclusions: ICI-induced pruritus has the potential to impact cancer treatment via both treatment pauses and treatment cessation. Dermatology involvement was low overall (36.5%), and 40% of patients with Grade 3 pruritus were not referred to a dermatologist, representing a large treatment gap. This data suggests an underutilization of dermatology care despite the high impact of severe pruritus on ICI interruption. Improved integration with dermatology has the potential to reduce unnecessary treatment pauses and support patient care.

Surgical outcomes of sleeve versus standard lobectomy in patients with non–small cell lung cancer: A systematic review and meta-analysis.

e20084 Background: Sleeve lobectomy is a parenchyma-preserving alternative to standard lobectomy for centrally located non–small cell lung cancer (NSCLC), yet its perioperative superiority compared with standard lobectomy remains unclear. This meta-analysis aimed to compare surgical and postoperative outcomes between these two procedures. Methods: A systematic search of PubMed, Embase, the Cochrane Library, and Web of Science was performed from inception to August 2025. Observational studies published in English comparing sleeve and standard lobectomy in NSCLC were included. Pooled estimates were calculated using a random-effects model, and results were expressed as mean differences (MDs) or odds ratios (ORs) with 95% confidence intervals (CIs). All statistical analyses were performed at R software (version 4.4). This study is registered in PROSPERO under the protocol number CRD420251128001. Results: From 1058 articles, we included 4 retrospective cohort studies encompassing 1,235 patients. Sleeve lobectomy was associated with a significantly longer operative time (MD = 25.93 minutes; 95% CI 2.49–49.36; p = 0.041; I² = 53.5%) but showed no significant difference in length of hospital stay (MD = –0.30 days; 95% CI –2.92–2.32; p = 0.671) or overall mortality (OR = 3.88; 95% CI 0.00–3951.25; p = 0.243; I² = 47.5%) compared with standard lobectomy. Similarly, there were no significant differences between the two technique regarding overall surgical complications (OR = 1.11; 95 % CI 0.57-2.17; p = 0.470; I2 = 0%) atrial fibrillation (OR = 1.17; 95% CI 0.36–3.86; p = 0.701; I² = 70.3%), pneumonia (OR = 0.84; 95% CI 0.45–1.59; p = 0.454; I² = 5.4%), prolonged air leak (OR = 0.88; 95% CI 0.47–1.67; p = 0.491; I² = 0%), or overall complications (OR = 1.11; 95% CI 0.57–2.17; p = 0.470; I² = 0%). Conclusions: Sleeve lobectomy offers comparable perioperative safety and postoperative outcomes to standard lobectomy in patients with NSCLC, with a modest increase in operative time. These findings support sleeve lobectomy as a safe, function-preserving, and potentially promising surgery option for NSCLC.

Defining when and how to support palliative care communication: International consensus on timing and clinician training priorities for adolescents and young adults with cancer.

12103 Background: Early, developmentally appropriate palliative care communication is an international standard for adolescents and young adults (AYAs) with cancer, yet remains inconsistently implemented. Clinicians report uncertainty regarding the optimal timing of palliative care discussions and around what training they need. This study aimed to establish international consensus on when key palliative care topics should be introduced with AYAs across the cancer trajectory and to identify clinician training priorities. Methods: A three-round international Delphi study was conducted with multidisciplinary health professionals who had worked with at least five AYAs who had died from cancer or its complications. Participants rated the appropriateness of introducing four palliative care communication topics (prognosis/goals of care, emotional/existential issues, quality of life, and end-of-life–related medical care) across prognosis levels and treatment phases. Participants also rated the importance of clinician training topics and likelihood of attending different training modalities. Consensus was defined a priori as ≥80% agreement. Results: Seventy-seven experts from 14 countries completed the final round. Consensus indicated that emotional and existential issues were appropriate to discuss early, including during active cancer treatment. In contrast, end-of-life–related medical decisions were considered appropriate later, most commonly following relapse, disease progression or poor prognosis. Discussions about prognosis, goals of care and quality of life were generally considered appropriate once cure was uncertain. Conversation timing was the only training topic to reach consensus as a priority (85.8%). Although no training modality reached formal consensus, learning from bereaved family members was the most highly endorsed approach (72.8%), followed by experiential and interdisciplinary learning formats. The presence of complex family or culturally and linguistically diverse dynamics did not fundamentally alter clinicians’ views on timing, but was associated with a need for greater team support and time. Conclusions: This international consensus provides clear guidance on when different palliative care topics should be discussed with AYAs with cancer and identifies timing as the primary area in which clinicians require additional support. Strong clinician endorsement of experiential and family-informed training approaches highlights opportunities to strengthen workforce capacity for earlier, developmentally appropriate palliative care communication.

Molecular circulating tumor DNA (ctDNA) profiling from patients (pts) treated with zanidatamab + chemotherapy (CT) in first-line (1L) HER2-positive (HER2+) advanced or metastatic gastroesophageal adenocarcinoma (mGEA).

4050 Background: Zanidatamab is a dual HER2-targeted bispecific antibody that binds to HER2 extracellular domains 2 and 4. In the phase 3 HERIZON-GEA-01 trial, 1L zanidatamab ± tislelizumab + CT significantly prolonged progression-free survival (PFS) and, with tislelizumab, yielded significant overall survival benefits versus trastuzumab + CT in HER2+ mGEA. Consistent results were previously reported in a phase 2 trial of zanidatamab + CT. ctDNA profiling provides a noninvasive approach to identify pts who could benefit from HER2-targeted therapies, and on-treatment ctDNA reductions may be associated with response. Here we report ctDNA profiling from the phase 2 trial. Methods: The phase 2 trial (NCT03929666) evaluated zanidatamab + CT (mFOLFOX6, CAPOX, or 5-FU/cisplatin [FP]) in 1L mGEA. Pts in part 1 had HER2-expressing (IHC 3+ or 2+) mGEA. Pts in part 2 had HER2+ (IHC 3+ or IHC 2+/FISH+) mGEA by central assessment (ccHER2+). Primary endpoint was confirmed objective response rate (cORR). PFS was a secondary endpoint. Plasma ctDNA samples were collected prior to cycle 1 of zanidatamab (baseline [BL]) and on treatment (cycle 2 day 15 for zanidatamab + CAPOX or FP; cycle 3 day 1 for zanidatamab + mFOLFOX6) for NGS by Guardant360 and/or Guardant Infinity. Results: 46 pts were enrolled (zanidatamab + mFOLFOX6 [n = 24], CAPOX [n = 20], FP [n = 2]). Most (41 [89%]) pts had ccHER2+ mGEA. cORR was 76.2% (95% CI 60.5, 87.9) and median PFS was 12.5 (95% CI 8.2, 21.8) mo. BL ctDNA was available from 39 pts; 29 had on-treatment samples, and 16 had matched BL and on-treatment samples. The concordance was 94% (34/36) between centrally assessed ERBB2 amplification by FISH in tissue and ERBB2 amplification in plasma ctDNA by NGS. Among 34 pts with detectable ERBB2 amplification at BL, median PFS was 18.6 (95% CI 12.0, NE) mo. All pts with available samples and confirmed complete (CR) or partial (PR) response or stable disease (SD) had undetectable on-treatment ERBB2 amplification in ctDNA. Most pts with ERBB2 indel, fusion, or SNV (79% [15/19]) or with PIK3CA amplification or SNV (86% [12/14]) at BL had CR or PR. Nearly all (94% [15/16]) pts with available matched samples had a > 80% decrease in ctDNA from BL to on-treatment assessment on the Guardant Infinity panel. On-treatment ctDNA samples from pts with progressive disease or SD showed recurrence/persistence of ERBB2 alterations and co-alterations across PI3K/AKT and RAS/MAPK pathways. Additional data will be presented. Conclusions: Molecular ctDNA profiling showed that 1L zanidatamab + CT in HER2+ mGEA could markedly reduce total ctDNA levels and ERBB2 amplification detection early in treatment. Antitumor activity was observed among pts whose tumors harbored ERRB2 and PIK3CA mutations, both of which are associated with resistance to standard HER2-targeted therapies. Clinical trial information: NCT03929666 .

From awareness to action: A mixed-method community study on uptake of population-based breast cancer screening in an urban slum of Delhi.

e13574 Background: Delayed diagnosis contributes to poor breast cancer outcomes in low-resource settings. We evaluated the yield and feasibility of a population-based screening pathway from community to clinic in an underserved area of Delhi, India. We explored factors influencing screening uptake in the community. Methods: A community-based cross-sectional study was conducted among women aged > 30 years in an underserved urban resettlement colony of Delhi (n = 310). Screening protocols for the referral were tailored for the geography, and liaison was done by the team at various levels. ANM were trained for performing CBE and women with abnormal CBE were referred as per protocol. Necessary quantitative analysis and qualitative thematic analysis for screen-positive and refusal participants was done. Results: A total of 310 women were screened; 11.3% were classified as high risk, and abnormal CBE findings were detected in 8.4%. Participants were also taught about the Self-breast examination. A robust screening protocol and referral pathway for the urban slum was detailed. Mammography was completed by 28.6% of those eligible. Among those imaged, 40% had BI-RADS I, 30% BI-RADS II, and 30% had BI-RADS III findings. Qualitative analysis showed that women completing mammography (screen positive participants) demonstrated preventive health orientation and trust in early detection, whereas refusal was characterized by symptom-based and fatalistic health-seeking behaviour. Refusal was not attributed to dissatisfaction with healthcare providers. Conclusions: Community-based breast lump screening using CBE is feasible in urban resettlement settings but is challenged by poor community participation. Screening uptake is influenced primarily by health-seeking behaviour rather than healthcare dissatisfaction. Hence, the policy-level intervention advocacy should focus on knowledge and awareness, along with subsequent consequences, specifically in under-resourced settings.

Early burden reduction after two doses of cemiplimab in advanced non-melanoma skin cancer.

9583 Background: Advanced non-melanoma skin cancers (NMSC) frequently impose a high disease-related burden—severe pain, ulceration, intensive wound care—particularly in older and frail patients (pts). While cemiplimab has established antitumor activity, early pts-centered clinical benefit prior to first radiologic evaluation is under-described in routine practice. We quantified early changes in key burden domains after 2 infusions. Methods: A retrospective analysis was performed in a real-world cohort of pts with locally advanced/metastatic NMSC treated with cemiplimab. Assessments at T0 (≤7 days pre-treatment) and T1 (day 35–49; after 2 infusions). Pain intensity (0–10 Numeric Rating Scale, NRS), wound-care frequency (dressing changes/week), and analgesic exposure were extracted from structured clinical/nursing documentation; opioid exposure was quantified as daily morphine milligram equivalents (MME) using standard conversion. Primary endpoint: meaningful early pain benefit at T1 (≥2-point NRS decrease without analgesic escalation). Other endpoints: opioid de-escalation, wound-care reduction (≥2 fewer dressing changes/week), and a pragmatic early-burden composite (improvement in ≥2 of: primary pain benefit; opioid de-escalation; wound-care reduction). Paired Wilcoxon signed-rank and McNemar tests were used. Infection status summarized. Confounding-restricted sensitivity analyses excluded pts receiving RT/surgery within 8 weeks of treatment initiation. Results: 53 pts included; median age 80yr (IQR 71–88); 81% locally advanced disease; ECOG 0–1 56%. Baseline burden: median pain NRS 7 (IQR 6–8); baseline opioid use 60%; wound-care frequency median 4/week; ulceration G≥2 64.2%. At T1 median pain NRS improved to 3 (IQR 2–4) (median Δ −3; p<0.001). The primary endpoint was achieved in 83% (44/53; 95% CI 70.8-90.8). The proportion of pts receiving opioids decreased from 60% (32/53) at T0 to 22.6% (12/53) at T1 (p=0.00024). Median daily MME decreased from 20 mg (IQR 0–33) to 0 mg (IQR 0–14) (p<0.001). Among baseline opioid users (n=32), 100% reduced dose and/or discontinued opioids by T1 (32/32; 95% CI 88.6–100). Wound-care frequency decreased from median 4/week to 1/week (p<0.001), with wound-care reduction ≥2/week in 67.9% (36/53). The pragmatic early-burden composite was met in 79.2% (42/53). At first radiologic evaluation (median 91 days), RECIST evaluable in 51 pts; ORR was 76.4% (39/51) and disease control was 96.2% (50/52). Any immune-related AE occurred in 34% (18/53), with G≥3 in 1.9% (1/53). Infection decreased from 43% to 3.8% at T1. Conclusions: In an elderly, frail real-world population with NMSC and high supportive-care needs, cemiplimab was associated with rapid, meaningful early clinical benefit after two doses. Early burden endpoints may complement radiologic response to better capture patient-centered benefit and inform supportive-care planning in routine pathways.

A phase I, open-label, dose-escalation study of CAR001 (mRNA-engineered Nb-CAR.BiTE-γδ T cells) in patients with relapsed or refractory solid tumors.

TPS2674 Background: Solid tumors are characterized by a highly immunosuppressive tumor microenvironment (TME). Key immune checkpoint molecules, such as HLA-G and PD-L1, are frequently expressed and act as barriers that limit the infiltration and efficacy of conventional CAR-T cell therapies. CAR001 is an investigational, allogeneic, mRNA-engineered Nb-CAR.BiTE-γδ T cell therapy. By utilizing the innate homing capabilities of γδ T cells and a dual-targeting mechanism, direct binding of HLA-G by T cells and secreting PD-L1 bispecific T-cell engagers from T cells, CAR001 is designed to overcome TME-mediated resistance. This study evaluates the safety and potential activity of repeat intravenous dosing of CAR001 in patients with advanced solid tumors. Methods: Study Design: This is an ongoing, multicenter, open-label, phase I dose-escalation study utilizing a standard 3+3 design or similar escalation schema. Patient Population: Adults with histologically confirmed refractory or relapsed solid tumors (including, but not limited to, colorectal cancer, glioblastoma, and triple-negative breast cancer) who have exhausted standard of care. Intervention: Patients receive escalating doses of CAR001 via intravenous infusion. The protocol allows for repeat dosing to enhance therapeutic persistence. Objectives: The primary objective is to evaluate the safety and tolerability of CAR001 and to determine the maximum tolerated dose (MTD) or recommended phase 2 dose (RP2D). Secondary objectives include assessing preliminary antitumor activity per iRECIST or RANO criteria, as well as characterizing the pharmacokinetic profile of the Nb-CAR.BiTE-γδ T cells. Safety Monitoring: Adverse events (AEs) are monitored using CTCAE v5.0, with specific focus on dose-limiting toxicities (DLTs), cytokine release syndrome (CRS), and immune effector cell-associated neurotoxicity syndrome (ICANS). Trial Status: This trial is currently active and enrolling patients. As of the submission deadline, dose-escalation cohorts are ongoing. To maintain the integrity of the study and comply with ASCO Trials in Progress guidelines, no formal analysis of clinical endpoints or treatment outcomes is provided. Clinical trial information: NCT06150885 .

Expression of Concern: Genomewide Association Studies for 50 Agronomic Traits in Peanut Using the ‘Reference Set’ Comprising 300 Genotypes from 48 Countries of the Semi-Arid Tropics of the World

Extended depth of focus for optical coherence tomography based on the versatile diffractive optical element

Optical coherence tomography (OCT) faces a fundamental trade-off between transverse resolution and depth of focus (DOF). While diffractive optical elements (DOEs) have been utilized to extend the DOF, existing designs are typically customized for specific objective lenses, limiting their versatility. We propose an extended-DOF OCT system utilizing a configuration-independent DOE. By modulating the incident wavefront to evenly distribute beam energy along the optical axis, the DOE extends the DOF without requiring redesign when applied to different optical setups. Simulations and experiments demonstrate that the integration of a single DOE across different sample arms more than doubles the effective DOF while maintaining diffraction-limited lateral resolution. The improved deep-tissue and large-field-of-view imaging capabilities are validated through resolution targets, cleared tissues, and in vivo OCT angiography of the murine cerebral cortex and human nailfold microcirculation. Quantitative evaluations reveal significant improvements in the contrast-to-noise ratio and vascular density (VD). The proposed system offers a highly adaptable solution for high-resolution, extended-DOF OCT imaging, demonstrating significant potential for preclinical and clinical applications.

Naringenin and quercetin nanoparticles as a dual strategy for topical management of diabetic retinopathy

Self‐Evolving Interfacial Kinetic Highways via Reactive Separator Engineering for Durable Potassium Metal Batteries

ABSTRACT The practical deployment of potassium metal batteries (KMBs) is impeded by unstable interfacial chemistry. Currently, most research on KMBs uses the thick, inert glass fiber separator, which compromises energy density and lacks sufficient regulation of dynamic interfaces. Here, we report a reactive separator engineering strategy that transforms the inert separator into an active regulator to construct a robust, self‐evolving anodic interphase. Inspired by mechanistic screening that intercalation compounds are superior to conversion counterparts as separator modulators, intercalative WO 3 is utilized as a prototype to modify polypropylene separator and actively trigger in‐situ formation of the self‐evolving K x WO 3 interphase. Based on density functional theory calculations, this phase features a minimal K + migration barrier of 0.12 eV, acting as a potassiophilic reservoir to build a kinetic highway. The evolving W species construct a chemically stable skeleton that templates the growth of a spatially graded interphase, creating a rigid‐flexible coupling architecture to buffer volume fluctuations. Consequently, the K||Cu half cell achieves a low nucleation overpotential of 17 mV at 0.2 mA cm −2 , while the K||K symmetric cell delivers an ultralong lifespan of 6000 h at 0.05 mA cm −2 . Impressively, exceptional stability exceeding 2500 h with cumulative capacities of 57 095 mAh g −1 at 50 mA g −1 is realized for KMBs.

Thermodynamic natural gradient descent (NGD-T) regulates natural-gradient steps by a geometric speed-cost bound

OTOP2 proton channel couples luminal pH sensing to intestinal immune homeostasis

Determinants of cervical cancer awareness, procedural knowledge, and information sources among Nepali women undergoing HPV screening.

11110 Background: Successful implementation of cervical cancer screening in low- and middle-income countries (LMICs) depends on community acceptance. Although general awareness of cancer is rising, limited understanding of screening procedures continues to hinder participation. This study examined the factors influencing cervical cancer awareness, procedural knowledge, and sources of screening information among Nepali women undergoing HPV-based screening. Methods: HPV screening was conducted through two community health fairs in Janakpur, Nepal, in 2025. Clinician-collected samples were tested using an isothermal amplification assay. Survey data collected included participants’ age, education level, cervical cancer awareness, procedural knowledge on vaginal sampling, sources of screening information, and health decision-making practices. Multivariate logistic regression was used to identify factors associated with cervical cancer awareness, procedural knowledge, and information sources. Results: Valid survey data from 274 Nepali women, including 32 HrHPV-positive cases, were analyzed. Cervical cancer awareness was significantly associated with education. Women with secondary education or higher were more likely to have heard of cervical cancer (aOR 4.45; 95% CI 1.81-10.95, p = 0.001), compared to those with less education. However, education was not associated with specific knowledge of vaginal sampling. Procedural knowledge was instead hindered by low health autonomy (husband or in-laws as decision-maker: aOR 0.42, 95% CI 0.25-0.72, p = 0.002, vs. self-decision making). Women who tested HrHPV-positive were significantly less likely to understand the sampling procedure prior to testing than those who tested negative (aOR 0.33, 95% CI 0.14-0.78, p = 0.01). Regarding information sources, women aged 30-44 years relied more on community networks, i.e., female community health volunteers (FCHVs) than medical providers, compared to women < 30 years (aOR 1.89, 95% CI 1.01-3.53, p = 0.048). Conclusions: Cervical cancer awareness in Nepali women is driven by education, but procedural knowledge is limited by low health autonomy. Middle-aged and older women rely on FCHVs or their community networks to acquire information about cervical screening. Future implementation strategies should address interventions that empower women and improve understanding of screening procedures, as well as tailored training on FCHVs to promote cervical screening among middle-aged women. Keywords: Cervical Cancer, Procedural Knowledge, Awareness, Implementation Science, Health Literacy, Global Health