Browse Articles

NS5 polymerase stimulates RNA duplex unwinding by the Orthoflavivirus NS3 helicase

Prevalence of non-treatment following pancreatic cancer diagnosis in the United States: A population-based analysis, 2016–2022.

e16357 Background: Pancreatic cancer is an aggressive malignancy for which cancer-directed therapy can improve outcomes when feasible. However, the proportion of patients who do not receive any cancer-directed treatment after diagnosis remains poorly characterized at the population level. We sought to quantify the prevalence of non-treatment following pancreatic cancer diagnosis and evaluate patient-level patterns and temporal trends over recent years. Methods: We conducted a retrospective population-based study using the Surveillance, Epidemiology, and End Results (SEER) database. Adults diagnosed with pancreatic cancer between 2016 and 2022 were included. Cancer-directed treatment was defined as receipt of chemotherapy, radiation therapy, or surgery of the primary tumor, as recorded in SEER. The primary outcome was receipt of no cancer-directed treatment. Non-treatment rates were evaluated overall and stratified by age, disease stage, race, and ethnicity. Temporal trends in non-treatment were assessed across the study period. Results: Among 95,315 patients diagnosed with pancreatic cancer, 39.7% received no cancer-directed treatment. Non-treatment increased markedly with age, affecting 26.7% of patients younger than 65 years, 34.2% of those aged 65–74 years, and 55.3% of patients aged 75 years or older. Notably, 41.2% of patients with localized disease did not receive cancer-directed therapy, while non-treatment was observed in 48.9% of patients with distant disease. Non-treatment rates varied modestly by race and were similar between Hispanic and non-Hispanic patients. Over time, non-treatment rates remained persistently high, ranging from 38.5% to 40.5% between 2016 and 2022, with no meaningful improvement observed. Among patients aged 75 years or older, non-treatment consistently exceeded 53% across all years. Conclusions: In this large population-based analysis, nearly two in five patients diagnosed with pancreatic cancer received no cancer-directed therapy, including a substantial proportion with potentially treatable disease. Non-treatment was particularly common among older adults and remained unchanged over time despite advances in systemic therapy and supportive care. These findings highlight a persistent care-delivery gap in pancreatic cancer and underscore the need for strategies to improve treatment access, referral, and decision-making, especially for older patients.

Spatial localization of epithelial-mesenchymal transition and immune niches in non-small cell lung cancer via interpretable gene set activity analysis.

e20556 Background: The tumor microenvironment (TME) in non-small cell lung cancer (NSCLC) is highly heterogeneous, with specific spatial niches driving metastasis and drug resistance. Epithelial-mesenchymal transition (EMT) is a critical process in progression, yet identifying exactly where EMT occurs within the tissue architecture remains challenging with standard tools. We utilized GESSO, a novel spatially informed analysis framework, to dissect the functional architecture of the NSCLC microenvironment, aiming to resolve spatially confined EMT programs and distinct immune cell niches that standard methods fail to localize. Methods: We analyzed high-resolution NanoString CosMx spatial transcriptomics data from human NSCLC tissue (55,365 single cells). To quantify gene set activity, we applied GESSO (code available at https://github.com/YMa-lab/GESSO), a spatially informed computational framework we developed that integrates spatial coordinates with gene expression to infer spatially coherent activity scores. We focused on characterizing the spatial distribution of TGF-beta-induced EMT programs and immune cell signatures. To validate the precision of these biological findings, we benchmarked the spatial fidelity of these signatures against standard activity scoring methods (AUCell, ssGSEA, GSDensity). Results: The analysis revealed that EMT is not a diffuse phenomenon but is strongly localized to a specific spatial niche at the interface between the tumor core and the desmoplastic stroma. Within this interface, the analysis identified a potent EMT signature driven by the coordinated upregulation of mesenchymal markers COL1A1 and FN1, and the invasion-associated metalloproteinase MMP2. Importantly, activity within this specific spatial niche was positively correlated with tumor pseudotime (Pearson r = 0.57), suggesting a direct link between this localized stromal interaction and tumor progression stages. Additionally, the framework resolved distinct, spatially separated immune microenvironments that standard methods blurred: specifically isolating macrophage niches enriched for CD68 and C1QB, and T-cell activation niches enriched for CD2 and CCL5. In comparative validation, standard methods failed to resolve these specific immune niches or the confined EMT interface, producing diffuse signals that obscured these critical microanatomical structures. Conclusions: This study demonstrates that key drivers of NSCLC progression, including EMT and immune activation, are organized into distinct spatial niches rather than occurring uniformly across the tumor. By precisely mapping these functional interfaces, we provide evidence of spatially coordinated stromal-tumor interactions that may serve as localized targets for disrupting metastatic progression.

Indocyanine green fluorescence–guided navigation during gastrectomy for resectable gastric cancer: A systematic review and meta-analysis.

e16147 Background: Radical gastrectomy with lymphadenectomy is the standard surgical approach for curative treatmentof gastric cancer. Adequate lymph node retrieval is critical for accurate pathologic staging andsurgical quality. Indocyanine green (ICG) fluorescence–guided navigation has been introduced as anadjunct to standard gastrectomy to enhance lymphatic mapping; however, randomized studies havereported variable effects on nodal outcomes. A meta-analysis was conducted to compare ICG-guidednavigation versus standard gastrectomy using head-to-head randomized evidence. Methods: A systematic search was conducted in PubMed, Embase, and the Cochrane Central Register ofControlled Trials from inception to January 2026 to identify randomized controlled trials comparingICG-guided navigation versus standard gastrectomy in patients undergoing curative resection forgastric cancer. Eligible studies reported lymph node retrieval outcomes. Pooled analyses wereperformed using inverse-variance methods; results were expressed as mean differences (MDs) with95% confidence intervals (CIs). Results: Three randomized controlled trials were included, enrolling 209 patients (ICG n = 105; No-ICGn = 104). Total retrieved lymph nodes were significantly higher with ICG guidance (MD 12.70; 95% CI8.55–16.86; P < 0.00001; I² = 0%). In contrast, the number of positive lymph nodes did not differsignificantly between groups (MD 1.39; 95% CI −0.89 to 3.66; P = 0.23; I² = 0%). Overallheterogeneity was minimal across pooled outcomes. Conclusions: ICG-guided navigation during gastrectomy was associated with a significantly higher total lymphnode yield compared with standard surgery, supporting improved nodal retrieval for staging quality.However, the number of positive lymph nodes was comparable between approaches. These findingssupport the intraoperative utility of ICG guidance to enhance lymph node harvest, while underscoringthe need for larger randomized trials with standardized oncologic and long-term outcomes to define itsoptimal role in gastric cancer surgery.

Enhanced predictive value of KELIM for surgical resectability in HRD-positive high-grade serous ovarian cancer.

5590 Background: Neoadjuvant chemotherapy followed by interval debulking surgery is commonly used in the management of advanced ovarian cancer. However, there is currently no reliable tool to predict the likelihood of achieving complete cytoreduction after 3–4 cycles of chemotherapy. Surgical outcomes are classified as R0 (no residual disease), R1 (residual lesions <1 cm), or R2 (residual disease ≥1 cm). The KELIM index, calculated from CA-125 kinetics during chemotherapy, has emerged as a potential biomarker of tumor regression and chemosensitivity in advanced ovarian cancer. We aim to study the performance of KELIM score as predictor of complete cytoreduction. Methods: This single-center, retrospective study included 69 patients with advanced high-grade serous ovarian cancer (Stage III/IV) who underwent interval cytoreductive surgery following neoadjuvant chemotherapy between January 2020 and December 2025. Kelim index values were calculated using Biomarker-Kinetics and categorized as favorable (≥1) or unfavorable (<1) as per literature. Further, ROC curve analysis was performed to identify the optimal KELIM cutoff for predicting complete cytoreduction. Correlations with degree of cytoreduction (R0, R1, R2) were analyzed using comparative statistics. Associations with molecular markers (BRCA, HRD) were also explored. Results: The median age of patients was 62 years [IQR 57–70]. The median KELIM score was 0.85 [IQR 0.69-1.00]. KELIM demonstrated statistically significant predictive performance for R0, with an area under the curve (AUC) of 0.69 (95% CI, 0.56–0.82; p = 0.004). The AUC optimal cutoff was KELIM = 0.995, yielding a sensitivity of 53.9% and specificity of 92.6%. Notably, this value closely aligns with the cutoff reported in the literature (KELIM = 1.0), which in our cohort corresponded to a sensitivity of 51.3% and specificity of 96.3% (Youden index = 0.47). Additional analyses were conducted using the literature-defined cutoff of KELIM = 1.0. Patients were categorized into two groups: Group A (22 patients) with favorable (≥1) and Group B (47 patients) with unfavorable (<1) KELIM scores. Patients with a favorable KELIM were significantly more likely to achieve complete cytoreduction compared with those with an unfavorable KELIM (R0: 95.2% vs 41.3%, p < 0.001). The discriminative performance of KELIM was numerically higher in HRD-positive patients AUC 0.706 (95% CI, 0.28-1) compared with HRP patients AUC 0.563 (95% CI, 0-1). KELIM was associated with degree of cytoreduction: Median 1 [0.71-1.20] for R0, 0.79 [0.70-0.88] for R1, and 0.64 [0.48-0.70] for R2, p=0.019. Conclusions: KELIM is a good predictor of complete resectability in high grade serous ovarian cancer. The differential predictive performance among the HRD status should be further explored.

EfficientNetV2-guided knowledge distillation for breast ultrasound diagnosis: Scalable deep learning with international clinical validation.

e12571 Background: Breast cancer remains a leading cause of cancer-related morbidity worldwide, with imaging central to early diagnosis and treatment planning. Ultrasonography is widely used due to its safety and effectiveness in dense breast tissue; however, interpretation is operator dependent and varies across clinical settings. Although deep learning has advanced ultrasound-based cancer detection, many high-performing architectures impose substantial training and inference costs, limiting scalability. EfficientNetV2 introduces a training-aware design optimizing convergence and computational efficiency through progressive scaling and fused convolutions. We evaluated whether an EfficientNetV2-guided knowledge distillation framework could enable accurate, scalable breast ultrasound classification with international validation. Methods: We analyzed 8,116 breast ultrasound images (4,074 benign; 4,042 malignant) with pathology-confirmed diagnoses, independently reviewed by two board-certified radiologists with consensus adjudication. An EfficientNetV2 model served as the reference architecture. A compact EfficientNetB1 student model was trained using structured knowledge distillation with soft probabilistic supervision and regularized optimization. Performance was assessed on a held-out test cohort and externally validated on independent ultrasound datasets. Metrics included accuracy, sensitivity, specificity, F1 score, and AUROC. The distilled model was deployed within a cross-platform application and independently evaluated by over 40 physicians across six continents. Results: The EfficientNetV2 reference model demonstrated strong diagnostic performance ( > 99%). The distilled EfficientNetB1 achieved approximately 95% accuracy with balanced sensitivity and specificity and AUROC near 0.95 across external validation datasets. Performance remained consistent across heterogeneous ultrasound systems and geographic regions. More than 95% of the physicians reported the system to be clinically informative and suitable for routine diagnostic triage. Conclusions: Training-aware architectural design coupled with structured knowledge distillation enables lightweight deep learning models to retain clinically meaningful diagnostic performance while improving scalability and deployability. EfficientNetV2-guided distillation into an EfficientNetB1 architecture supports global implementation of AI-assisted breast ultrasound interpretation. Prospective studies are planned to assess impact on diagnostic confidence, biopsy utilization, and time-to-decision.

Major adverse cardiovascular events among prostate cancer patients in the VA, stratified by HIV status and androgen deprivation therapy.

12157 Background: Despite the rapidly increasing trend in incidence of prostate cancer among persons with HIV (PWH), comprehensive data outlining outcomes and treatment toxicities of prostate cancer in this population remain sparse. Notably, both HIV infection and prostate cancer treatment independently increase the risk of cardiovascular disease. We sought to quantify the risk of major cardiovascular adverse events (MACE) among a large Veterans Affairs (VA) cohort of PWH and people without HIV (PWoH). Methods: We used patient data from the VA Corporate Data Warehouse and the VA Central Cancer Registry, including demographics, laboratory data, medications administered in inpatient/outpatient pharmacies, International Classification of Diseases 9 th Edition and 10 th edition diagnoses and procedure codes. Patients were included if they had diagnosis of prostate adenocarcinoma between 1/1/2008 and 12/31/2020. Patients were excluded if they had a diagnosis with any prior invasive cancer on or prior to prostate cancer diagnosis, diagnosis with prostate cancer prior to HIV positive diagnosis, or MACE within 6 months prior to prostate cancer diagnosis. MACE was defined as having a diagnosis code of myocardial infarction, stroke, or arrhythmia, within five years of start of treatment approach for prostate cancer (defined as androgen deprivation therapy (ADT) or active surveillance). We conducted a time-to-event analysis using the Cox proportional hazards model to examine the time to first occurrence of MACE, accounting for censoring and varying follow-up times. Results: A total of 87,144 prostate cancer patients were included (654 PWH and 86,490 PWoH). About 37.5% of PWH were < 60-years-old at time of prostate cancer diagnosis (compared with 19.6% of PWoH). Among 62,259 patients with low-risk or intermediate-risk localized prostate cancer, the risk of MACE was 24% greater among PWH compared with PWoH (HR 1.24, 95% CI 1.02-1.51, p = 0.03) after controlling for age, race, and cardiovascular risk factors at time of cancer diagnosis. Among PWH with prostate cancer, receiving ADT was associated with a 3.5 fold increased risk of developing MACE vs. those not receiving ADT (HR 3.51, 95% CI 2.25-5.49, p < 0.01). Among PWoH with prostate cancer, receiving ADT had a 2.7-fold increased risk of developing MACE vs. those not receiving ADT (HR 2.72, 95% CI 2.61-2.84). Conclusions: Both HIV infection and ADT increase the risk of MACE among prostate cancer patients. Future studies are warranted to explore the mechanistic links between HIV infection, ADT, and cardiovascular health.

Total neoadjuvant immunotherapy plus chemotherapy in Borrmann type IV gastric cancer (PERISCOPE-01 study): An open-label, single-arm, phase 2 trial.

4079 Background: Borrmann type IV (Borrmann-IV) gastric cancer (GC), or linitis plastica, is marked by diffuse infiltration, early metastasis, and poor prognosis, with limited benefit from conventional therapies. This study evaluated the efficacy and safety of a total neoadjuvant regimen combining immunotherapy and chemotherapy in Borrmann-IV GC. Methods: This trial (NCT06451211) enrolled patients with Borrmagnn-IV gastric cancer without distant metastasis. Participants received a total neoadjuvant regimen consisting of tislelizumab (an anti–PD-1 antibody) combined with platinum-based chemotherapy (oxaliplatin plus capecitabine or S-1) for 6 cycles at 3-week intervals, followed by radical surgical resection. The prespecified primary endpoint was the pathological response rate, defined as tumor regression grade (TRG) 0/1. Results: A total of 56 patients were enrolled, all patients had no distant metastasis confirmed by laparoscopic exploration and ascitic fluid cytology. The median age was 58 years; all patients were pMMR. In the efficacy analysis population (n=47), 41 patients completed 5–6 cycles and 6 patients completed 3–4 cycles of preoperative chemotherapy plus immunotherapy, no patients experienced disease progression leading to tumor metastasis during preoperative treatment. 47 patients underwent radical surgical resection, including 42 total gastrectomy, with an R0 resection rate of 98% (46/47). The prespecified primary endpoint of TRG 0/1 was achieved in 32% of patients (15/47; 95% CI, 21%–48%). Notably, 17% (n = 8) achieved pCR (ypT0N0), and 53% (n =25) were ypN0. Pathological response (TRG 0/1) was significantly higher in Lauren intestinal/mixed versus diffuse types (53% vs. 21%; p < 0.05), while efficacy was comparable between PD-L1 CPS ≥5 and <5 (44% vs. 30%). Grade 3/4 treatment-related adverse events occurred in 32% of patients (n=18), mainly thrombocytopenia and liver function impairment. Surgical morbidity (Clavien–Dindo grade II/III) occurred in 10.6% of surgical patients (5/47), including 1 patient with postoperative bleeding and 3 patients with anastomotic leakage; no perioperative mortality was observed. Conclusions: Total neoadjuvant tislelizumab plus chemotherapy demonstrated promising efficacy and an acceptable safety profile in patients with Borrmann-IV gastric cancer. Lauren classification may serve as a potential predictive biomarker and warrants further study. Clinical trial information: NCT06451211 .

Real-world efficacy and safety of tarlatamab in small cell lung cancer and extrapulmonary neuroendocrine carcinoma.

e20166 Background: Patients with extensive-stage small cell lung cancer (ES-SCLC) and extrapulmonary neuroendocrine carcinoma (EP-NEC) have limited treatment options after progression on platinum-based chemotherapy and immunotherapy. Tarlatamab, a DLL3xCD3 bispecific T-cell engager, is approved for previously treated ES-SCLC. While clinical trials have characterized its efficacy and toxicity, real-world data describing the timing and severity of cytokine release syndrome (CRS) and immune effector cell–associated neurotoxicity syndrome (ICANS) remain limited. Methods: We conducted a retrospective chart review of adults with ES-SCLC or EP-NEC who received ≥1 dose of tarlatamab between October 2024 and October 2025. Primary endpoints were the incidence and timing of CRS and ICANS. Progression-free survival (PFS) was estimated using Kaplan-Meier methods. Results: Of 26 patients screened, 22 met inclusion criteria. Median age was 64 years (range, 50–84), and 12 patients (55%) were female. Among the cohort, 18 patients (82%) were White, 4 patients (18%) were Black, and Hispanic ethnicity was reported in 7 patients (32%). Brain and liver metastases were present in 7 (32%) and 10 (46%) patients, respectively; one patient had EP-NEC. CRS occurred in 9 patients (41%), including grade 1 in 3 (14%), grade 2 in 5 (23%), and grade 4 in 1 (5%). ICANS occurred in 8 patients (36%), including grade 1 in 2 (9%), grade 2 in 6 (27%) with no grade ≥3 events. Median time to onset was 16.4 hours for CRS and 15.8 hours for ICANS, occurring predominantly during cycle 1. CRS and ICANS were managed with standard interventions, including tocilizumab and/or corticosteroids. One patient (5%) experienced grade 4 CRS and grade 2 ICANS, requiring vasopressor support and ICU admission. Median PFS was 4.1 months, with 11 patients (50%) remaining on therapy at the time of analysis (PFS to be updated at time of ASCO 2026 meeting). Conclusions: In this real-world cohort, tarlatamab was associated with early-onset CRS and ICANS, which were predominantly low-grade, though rare high-grade events occurred. Despite ICANS rates higher than those reported in clinical trials, toxicities were generally manageable with standard supportive measures. These findings support the feasibility of tarlatamab in diverse clinical populations when supported by appropriate clinical and oncology pharmacy-driven interventions.

Patterns of failure and efficacy for unresectable or recurrent gastric and gastroesophageal junction adenocarcinoma after immune checkpoint inhibitor therapy: A retrospective study.

e16491 Background: Immune checkpoint inhibitors (ICIs) combined with chemotherapy have been incorporated into clinical guidelines in multiple countries as a first-line treatment regimen for advanced gastric cancer. However, data regarding the patterns of failure associated with ICIs remain limited, and the optimal strategy for subsequent treatment remains controversial. Herein, we investigated the patterns of treatment failure following immunotherapy in patients with unresectable or recurrent gastric and gastroesophageal junction adenocarcinoma and further elucidated the impact of subsequent treatment strategies on patient outcomes. Methods: We retrospectively analyzed 193 patients with unresectable or recurrent gastric or gastroesophageal junction adenocarcinoma who received immune checkpoint inhibitor (ICI) combined with chemotherapy between January 2022 and April 2025. Kaplan–Meier method and log-rank test were used to perform survival analyses of different progression patterns and subsequent treatment strategies. Results: In the entire study cohort, a total of 152 patients (78.7%) experienced disease progression. Among the patients, the most common site of progression was the liver (n = 39, 25.7%), followed by the primary site or regional lymph nodes (n = 36, 23.7%) and peritoneal metastasis (including malignant ascites) (n = 26, 17.1%). Furthermore, oligoprogression occurred in 85 patients (55.9%), while 67 patients (44.1%) exhibited systemic progression. Patients with oligoprogression had both a longer median progression-free survival (PFS) (6.5 months vs. 4.7 months, P = 0.003) and a longer median overall survival (OS) (16.7 months vs. 9.9 months, P = 0.001) compared to those with systemic progression. Among patients who progressed after immunotherapy, continued application of chemotherapy (16.73 months vs. 8.33 months, P < 0.001) or immunotherapy (17.83 months vs. 9.63 months, P = 0.001) prolonged the median OS. Patients who received radiotherapy after progression showed a trend towards longer median OS compared to those who did not (15.90 months vs. 11.97 months, P = 0.06). Conclusions: Oligoprogression is the most common pattern of failure following immunotherapy for gastric cancer, with liver metastasis being the most frequent site of progression. Subsequent treatment with either chemotherapy or immunotherapy after disease progression is associated with improved survival. Receiving radiotherapy after progression shows a trend toward prolonged survival, suggesting its potential as an effective subsequent-line treatment.

Clinical outcomes of local treatments for oligoprogressive or oligorecurrent small cell lung cancer.

8090 Background: The role of local treatments in patients with oligoprogressive or oligorecurrent small cell lung cancer (SCLC) remains unclear. Our study aimed to describe the prognosis of patients with SCLC receiving a local treatment for oligoprogression or oligorecurrence in a real-life setting. Methods: The present study included all consecutive patients treated for SCLC at Lille University Hospital or Strasbourg University Hospital between January 2013 and March 2024, for whom local treatment was decided in a Multidisciplinary Tumor Board to control oligoprogression or oligorecurrence. Clinical data and characteristics of the local treatment were collected from medical records. Results: Of the 850 patients treated for a SCLC in both centers during the study period, 97 patients were eligible for inclusion. Of those, 84 received local treatment for oligoprogression (n = 47) or oligorecurrence (n = 37). At the time of oligoprogression or oligorecurrence, 82.1% of patients were being treated for extensive-stage disease. The brain was the predominant site of oligoprogression or oligorecurrence (71%) and the most commonly used local treatment was conformal radiotherapy (66.7%), followed by stereotactic radiotherapy (31%) and surgery (2.4%). Grade ≥3 adverse events were observed in only 2.4% of treated patients. After a median follow-up period of 11.9 months, the median overall survival (OS) and median progression-free survival (PFS), measured from the date of initiation of local treatment, were 12.4 months (95% CI: 10.7-18.9), and 3.4 months (95 %CI: 2.4-4.5), respectively. A good performance status (PS 0–1) was associated with better OS (HR 0.50; (95% CI: 0.26–0.97); p = 0.018). PFS was significantly longer in patients with limited-stage SCLC (HR 0.52; (95% CI: 0.31–0.89); p = 0.016) or with five or fewer metastases at the time of local treatment (HR 0.59; (95% CI: 0.37–0.94); p = 0.038). Most progressions observed after local treatment occurred distant from the treated sites (68.8%). Conclusions: This study demonstrates that the use of local treatments to control oligoprogression or oligorecurrence in SCLC is not uncommon in routine clinical practice. While this strategy is generally well tolerated, its efficacy remains limited. Further studies are needed to clarify its role in the era of tarlatamab development.

Intersecting risks: Myelodysplastic neoplasms and cardiovascular disease mortality in older adults across the United States, 1999–2020.

e18581 Background: Myelodysplastic syndrome (MDS) predominantly affects older adults and carry substantial mortality. Cardiovascular disease (CVD) is a leading competing cause of death in MDS; however, long term national trends and demographic disparities in MDS-associated CVD mortality remain unexplored. We aimed to evaluate temporal, demographic, and geographic patterns of MDS and CVD-related mortality among older adults in the United States. Methods: We conducted a population-based descriptive study using the CDC WONDER database from 1999–2020. Deaths among individuals aged ≥75 years with both MDS (ICD 10: D46) and CVD (ICD 10: I00–I99) listed as underlying or contributing causes were included. Deaths were standardized to the U.S. population in 2000 in order to establish AAMRs. Temporal trends were assessed using Joinpoint regression to estimate annual percent change (APC) and average annual percent change (AAPC). Analyses were stratified by sex, race/ethnicity, age group, urbanicity, census region, place of death, and state. Results: From 1999 to 2020, 72,206 deaths occurred due to combined MDS and CVD among adults aged ≥75 years. Overall mortality increased from 1999–2010 (APC 2.17%, 95% CI 1.68–2.67), declined from 2010–2018 (APC −2.48%, 95% CI −3.33 to −1.63), and rose again from 2018–2020, though this increase was not statistically significant. Mortality rates were consistently higher in men than women (men: AAPC 0.19%; women: AAPC −0.38%). Non-Hispanic White individuals exhibited the highest mortality burden (AAPC 0.66%, 95% CI −0.06 to 1.39), whereas rates were lowest among Black adults (AAPC −0.16%, 95% CI −1.00 to 0.69). The highest mortality occurred in the 80–84-year age group (overall AAPC 0.47%, 95% CI −0.74 to 1.69). Geographically, AAMR was highest in Midwest (18; 95% CI 17.7 to 18.8) and was lowest in the South (14.2; 95% CI 14 to 14.4). State-level AAMRs varied greatly, ranging from 8.16 in Louisiana to 25.72 in Vermont. Non-metropolitan areas showed a higher AAPC (0.49%, 95% CI −0.30 to 1.30) compared with metropolitan areas (0.44%, 95% CI −0.10 to 0.99). Conclusions: Our findings demonstrate marked disparity in mortality rates of MDS associated CVD mortality across age, regions, race and geography. These highlight the need of care models and prioritizing targeted interventions for high-risk populations. Age-adjusted mortality rates (AAMR) according to demographic variables among the US population 1999-2020. Demography Age Adjusted Rates Age Adjusted Rate Lower 95% Confidence Interval Age Adjusted Rate Upper 95% Confidence Interval Overall Year 1999-2020 16.7 16.5 16.8 Sex Male 25.6 25.4 25.9 Female 11.4 11.2 11.5 Census Region Northeast 17.7 17.4 18 Midwest 18 17.7 18.3 South 14.2 14 14.4 West 18.4 18.1 18.7 Race American Indian or Alaska Native 8.4 7.2 9.7 Asian or Pacific Islander 11.5 11 12.1 Black or African American 9.8 9.4 10.1 White 17.5 17.4 17.7

Limitations in US cancer screening rate data and opportunities for improvement.

e22535 Background: Cancer screening is a cornerstone of cancer control, yet screening participation remains inconsistent across tumor types, populations, and geographies in the United States. Despite widespread reliance on screening rate data to inform clinical guidance, policy decisions, patient advocacy, and early detection innovation, the underlying data landscape is fragmented. Screening rates are derived from surveys, electronic health records (EHRs), and claims data; however, limitations across these sources may misrepresent screening participation and impede early detection. This study assessed the cancer screening data ecosystem to characterize critical limitations and identify prioritized solutions. Methods: Primary qualitative interviews were conducted with cancer screening data users (n = 8), including academic researchers and patient advocacy organizations, and data aggregators (n = 2) from national public health organizations. Secondary research included a structured review of publicly available survey-based, EHR-based, and claims-based screening data sources and relevant literature. Findings were synthesized and evaluated during a multistakeholder workshop with 20 participants representing researchers, clinicians, advocates, and industry experts to identify and prioritize solutions. Results: Across all major data sources, four pervasive categories of limitations were identified: data access, accuracy, consistency, and completeness. Screening data are dispersed across numerous sources with variable transparency, technical complexity, and cost, limiting accessibility and comparability. Accuracy is affected by recall and response bias in surveys, documentation and coding errors in EHRs and claims, and imperfect correction methods. Inconsistent tumor coverage, survey design changes, and irregular reporting cycles constrain longitudinal analysis. Data completeness gaps—including limited capture of underserved populations, risk factors, follow-up, and outcomes—restrict the ability to assess disparities or measure real-world impact. During the multistakeholder workshop, participants prioritized near-term solutions such as improved data education, standardized reporting practices, cross-source validation, and expanded data-sharing, alongside longer-term investments in registries and longitudinal infrastructure. Conclusions: Structural limitations in cancer screening data constrain the ability of stakeholders to optimize screening strategies, address inequities, and support innovation in early detection. Multistakeholder consensus suggests that targeted improvements, paired with longer-term system-level investments, can strengthen the screening data ecosystem. Improving data quality and usability is essential to ensuring screening data informs decisions that improve population-level cancer outcomes.

Predicting infiltrative hepatocellular carcinoma and efficacy of combination treatment for conversion therapy based on pretreatment CT/MR features.

e16290 Background: Infiltrative HCC (iHCC) has been demonstrated to exhibit the highest malignant potential and poorest prognosis. Unified imaging diagnostic criteria are lacking, and a practical predictive model may help to identify iHCC and guide clinical decision-making. This study aimed to compare clinicopathological features and outcomes between iHCC and non-iHCC after resection, develop and validate a imaging-based diagnostic model for iHCC, and assess its utility in predicting response to conversion therapy. Methods: This retrospective multicenter study analyzed 1158 patients with newly diagnosed large HCC (> 5cm) from three tertiary hospitals. Patients were allocated into training/test, external validation and application cohorts. Clinical, imaging (CT/MRI assessed per LI-RADS), and pathological data (classified per Japanese Society of Hepatology criteria), overall survival (OS), recurrence-free survival (RFS) and progression-free survival (PFS) were analyzed and compared. A pretreatment CT/MRI-based predictive model for iHCC was developed and validated. Diagnostic performance was calculated and the efficacy of hepatic arterial infusion chemotherapy (HAIC) combined with targeted and immune checkpoint inhibitors in predicted iHCC and non-iHCC patients was compared in the application cohort. Results: The iHCC exhibited higher malignant features and shorter RFS after hepatectomy compared to non-iHCC. The imaging predictive model (SWIRP score) incorporated ill-defined margin, wedged-shaped peritumoral arterial phase enhancement, irregular shape, pseudo-capsule enhancement at delayed phase and rim arterial enhancement. The AUC values in training, test and validation sets were 0.890 (95% CI: 0.857–0.923), 0.851 (95% CI: 0.781–0.920) and 0.835 (95% CI: 0.767–0.902) . According to the SWIRP score, the overall objective response rates (ORR, per mRECIST) of predicted iHCC and non-iHCC in the application cohort were 82.7% and 66.7% (p = 0.044). While the median PFS in the predicted iHCC group (16.6 months, 95%CI: 12.4-20.7) was shorter (10.4 months vs. 16.6 months, p = 0.003). Conclusions: Infiltrative HCC exhibites poorer malignancy and prognosis compared to non-infiltrative types. The SWIRP score demonstrated good performance for identifying iHCCs. While combined HAIC, targeted, and immunotherapy yielded higher ORR in iHCC, it correlated with shorter PFS, suggesting the need for optimized treatment strategies in this aggressive subtype. Baseline clinicopathological characteristics in the initial cohort. Characteristics Non-iHCC (n=722) iHCC (n=158) Bilobar tumor distribution 40 (5.5%) 19 (12.0%) AFP>200ng/ml 273 (37.8%) 90 (57.0%) MVI (+) 316 (43.8%) 86 (54.4%) Satellite Nodule (+) 100 (13.9%) 33 (20.9%) Hepatic capsule invasion (+) 484 (67.0%) 134 (84.8%) Poor differentiation tumor 64 (8.9%) 27 (17.1%)

ASPRIA: A single-arm phase 2 trial of atezolizumab with sacituzumab govitecan to prevent recurrence in triple-negative breast cancer.

TPS644 Background: Despite marked improvements in neoadjuvant chemoimmunotherapy (NACT) and adjuvant treatment for patients with early-stage triple negative breast cancer (TNBC), over 30% of patients with residual disease after NACT experience disease recurrence or death within 36 months. Understanding which patients are at greatest risk of relapse and designing effective interventions to prevent disease recurrence are major unmet needs. Circulating tumor DNA (ctDNA) can detect molecular residual disease (MRD) in TNBC and has demonstrated prognostic and predictive ability in the neoadjuvant setting; ctDNA thus represents a promising biomarker to select at-risk patients. Sacituzumab govitecan is a Trop-2 targeted antibody-drug conjugate leveraging a topoisomerase 1 payload that is approved for metastatic HR+ breast cancer and TNBC. Atezolizumab is an immune checkpoint inhibitor that interferes with the binding of PD-L1 to PD-1. The ASPRIA trial hypothesizes that ctDNA can identify patients with TNBC and residual disease after NACT at greatest risk for relapse, and that the combination of sacituzumab govitecan and atezolizumab may effectively treat MRD through cytotoxicity and anti-tumor immunity, leading to ctDNA clearance and improved clinical outcomes. Methods: The ASPRIA trial is an open-label, single-arm phase II trial evaluating sacituzumab govitecan in combination with atezolizumab in patients with early-stage TNBC. Eligible patients have residual invasive disease in the breast and/or lymph nodes following curative-intent therapy and detectable ctDNA, as assessed by a personalized, tumor-informed assay (Signatera, Natera, Inc.). The study includes a screening phase, with ctDNA presence assessed for the first 60 individuals to ensure feasibility, with a minimum prevalence threshold of approximately 15%. Patients with detectable ctDNA will undergo restaging imaging, and those without evidence of metastatic disease proceed to treatment. Up to 32 ctDNA-positive patients will receive atezolizumab 1200 mg IV on day 1 plus sacituzumab govitecan 10 mg/kg IV on days 1 and 8 of a 21-day cycle for 6 cycles. On-treatment ctDNA assessments will occur at weeks 3, 9, and at end of therapy. Participants will be followed for recurrence and survival every 6 months for 3 years. The primary objective is to determine the rate of ctDNA clearance after 6 cycles of therapy. Secondary objectives include estimating the ctDNA prevalence, assessing clearance after 1 and 3 cycles, evaluating recurrence risk and survival outcomes (iDFS, RFI, DMFS), and characterizing safety (per CTCAE v5). Correlative studies will examine predictors of ctDNA positivity, associations between treatment response and Trop-2 or PD-L1 expression in primary and residual tumors, temporal ctDNA dynamics in relation to relapse, and patient perspectives on ctDNA testing. Clinical trial information: NCT04434040 .

Palliative neck radiation therapy (RT) in combination with lenvatinib/pembrolizumab (L/P) in BRAF wild-type anaplastic thyroid cancer (BRAF-WT ATC).

6117 Background: BRAF-WT ATC carries a poor prognosis due to limited effective therapies and rapid, often fatal locoregional (LR) progression. Although L/P has shown some efficacy, response rates remain modest (36-52%). Prior studies have shown that complete surgical resection of the primary tumor significantly improves overall survival (OS). Methods: We retrospectively reviewed patients with BRAF-WT ATC treated with palliative neck RT (PNRT, ≤45 Gy) administered five days before to 21 days after first-line L/P, between January 2016 and August 2025. Primary objective was to evaluate the efficacy and safety of PNRT combined with L/P. Primary endpoints were progression-free survival (PFS), locoregional PFS (LPFS), and OS, defined from RT start, and estimated using Kaplan-Meier method. Secondary endpoints were overall response rate (ORR) and neck response per RECIST v1.1, changes in surgical morbidity assessed by the Thyroid Neck Morbidity and Complexity (TNMC) score, and proportion of patients proceeding to surgery. Results: Twenty-six patients met inclusion criteria. Median age was 66 years (range, 36-82) and 14 were male (54%). 23/26 (88%) had stage IVC disease. 21/26 (81%) received ‘Quadshot’ regimen (14Gy in 4 fractions), two 20 Gy in 5 fractions and three 30 Gy in 10 fractions. Median time from RT to L/P start was 2 days (range, –4 to 21). Median follow-up was 18.5 months (95% CI, 11.7-25.4). After PNRT and L/P, ORR was 72%, and 72% achieved a complete or partial response in the neck. Median OS and PFS were 10.0 (95% CI, 3.4-16.7) and 7.6 (95% CI, 4.4-10.8) months respectively. Median LPFS was not reached. Median TNMC score (0-4 scale, 4=unresectable) decreased from 4 (range, 2-4) at baseline to 3 (range, 0-4) at time of best response with a mean reduction of 1 point (range, 0-3). Ten patients (38%) proceeded to R0/R1 neck surgery after a median of 5.4 months (range, 1.6-9.5) from RT, with absence of residual ATC in the surgical specimens in 70% (7/10). OS and PFS were significantly longer in patients who underwent surgery compared with those who did not, with median OS 22.8 vs 6.1 months (p=0.002) and median PFS 12.7 vs 3.9 months (p=0.034). Seventeen patients (68%) had disease recurrence/progression: 1 LR, 10 distant, and 6 both LR and distant. All but one (6/7) LR recurrences were in patients who did not undergo surgery. PNRT combined with L/P was well tolerated, with most adverse events grade ≤ 2. One patient had a fistula 3 months after RT requiring emergency tracheostomy. Conclusions: Palliative neck RT combined with L/P for BRAF-WT ATC appears safe and may improve surgical resectability in patients with initially unresectable disease or high baseline surgical morbidity. Patients who proceeded to neck surgery had high pathologic complete response rate (70%) and significantly prolonged OS and PFS with low rate of locoregional relapse.

Conditional survival in advanced-stage epithelial ovarian cancer: Implications for survivorship planning and patient counseling.

e17500 Background: Advanced-stage epithelial ovarian cancer (EOC) is characterized by biological heterogeneity and poor outcomes. Traditional survival statistics fail to capture the dynamic, time-dependent nature of prognosis. Conditional survival provides a framework for updating prognosis over time and offers clinically meaningful information for patient counseling and survivorship planning. In this two-decade real-world study, we applied conditional survival analysis to evaluate how prognosis evolves as patients live longer and remain progression-free. Methods: We retrospectively analyzed patients with FIGO stage III–IV epithelial ovarian cancer who underwent primary or interval cytoreductive surgery between 2004 and 2024. Overall survival (OS) and progression-free survival (PFS) were estimated using the Kaplan–Meier method. Conditional overall survival (CS-OS) and conditional progression-free survival (CS-PFS) were calculated for additional 1- and 5-year intervals among patients who had survived or remained progression-free for 6 months, 1 year, 3 years, or 5 years after surgery. Results: A total of 808 patients with FIGO stage III–IV epithelial ovarian cancer were included. Median OS and PFS were 4.37 and 1.70 years, respectively, and approximately 11% survived beyond 10 years. Peritoneal dissemination and platinum resistance were independent predictors of poor survival. Conditional survival improved with increasing time survived. The 1-year CS-OS increased from 87% at 6 months to 95% at 5 years, while the 5-year CS-OS rose from 49% to 66%. Similarly, the 1-year CS-PFS increased from 89% to 95%, and the 5-year CS-PFS improved from 44% to 62%. In stage-stratified analyses, 5-year CS-OS increased from 43% to 58% in FIGO stage III disease and from 26% to 44% in FIGO stage IV disease. Conclusions: Conditional survival analysis provides time-updated prognostic information that complements traditional survival estimates and more accurately reflects long-term outcomes. Incorporating conditional survival into clinical practice may support more individualized patient counseling and inform survivorship planning, particularly for patients who remain alive and progression-free beyond the initial high-risk period. Conditional survival rates based on previous time survived (CS-OS) or on time without progression (CS-PFS) in patients with advanced epithelial ovarian cancer. Time already survived/progression-free CS-OS at 1 year (95% CI) CS-OS at 5 years (95% CI) CS-PFS at 1 year (95% CI) CS-PFS at 5 years (95% CI) 6 months 87 (84–90) 49 (45–53) 89 (86–92) 44 (40–48) 1 year 88 (85–91) 49 (45–53) 90 (87–93) 46 (42–50) 3 years 91 (88–94) 57 (52–62) 93 (89–96) 54 (50–58) 5 years 95 (92–98) 66 (61–71) 95 (91–98) 62 (57–67)

Cytokine profiles in HPV-associated cervical intraepithelial neoplasia.

e22517 Background: Persistent high-risk human papillomavirus (HPV) infection is a well-established driver of cervical intraepithelial neoplasia (CIN), yet progression risk varies among patients. Cytokines shape antiviral immunity and tissue remodeling, and cytokine profiling is emerging as a cost-effective tool to stratify CIN risk and refine prognostic assessment in this population. Objective: Our study evaluated serum and cervical mucus cytokine profiles in HPV-negative normal cytology and in low-grade squamous intraepithelial lesion (LSIL) and high-grade squamous intraepithelial lesion (HSIL), stratified by HPV genotype and viral load. Methods: Ninety women aged 18–80 years were enrolled, including 30 healthy controls and 60 HPV-positive patients with histologically confirmed CIN (LSIL, n = 30; HSIL, n = 30). Levels of 27 cytokines (pg/mL), including FGF, GM-CSF, IFN-γ, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-17A, TNF-α, VEGF, and selected chemokines, were measured in serum and cervical mucus using the Bio-Plex Pro Human Cytokine 27-plex assay (Bio-Rad, USA) on a Luminex 200 platform. The IL-10/IL-17A ratio was calculated. Cytokine profiles were stratified by HPV genotype [HPV(−), HPV(+) 16/18, HPV(+) 31–73] and viral load [HPV(−), low ( < 5), high (≥5)]. Data were reported as median (Q1–Q3). Statistical analyses included one-way ANOVA with pairwise t-tests and Bonferroni correction; p < 0.05 was considered significant. Results: Serum: LSIL was characterized by increased GM-CSF, IL-4, IL-7, and IL-8 with decreased IL-2, IL-5, IL-17A, and angiogenic factors (FGF/VEGF); the IL-10/IL-17A ratio was reduced (p < 0.05). HSIL showed increased IL-7, IL-10, TNF-α, IL-8, MCP-1, and FGF with decreased IL-2, IL-5, and IL-17A; the IL-10/IL-17A ratio was increased (p < 0.05). Cervical mucus: LSIL demonstrated increased IL-10, IL-1β/IL-1ra, IL-2, IL-4, TNF-α, and FGF with reduced IL-8 and IP-10. HSIL showed broader immune activation with increased IL-10, IL-12p70, IL-17A, IL-1β, IL-6, TNF-α, GM-CSF, FGF, and VEGF, alongside decreased IL-9, IL-8, and IP-10 (p < 0.05). Conclusions: CIN grade was associated with distinct systemic and local cytokine signatures, with HSIL showing broader proinflammatory activation in cervical mucus and a grade-dependent shift in the serum IL-10/IL-17A ratio, consistent with compartmentalized immune dysregulation in HPV-associated disease.

Safety and efficacy results from a first-in-human, phase 1/2 study of ASKG915, an anti-PD-1/pro-IL-15 bifunctional fusion protein, for patients with advanced solid tumors.

2607 Background: ASKG915 is a bifunctional fusion protein of PD-1 antibody fused with an IL-15 prodrug. Preclinical studies demonstrated the efficacy of ASKG915 in neoplastic models. Here, we present updated safety and efficacy results from a first-in-human, dose escalation and dose expansion of ASKG915 monotherapy in patients (pts) with advanced solid tumors. Methods: This open-label, multicenter, Phase 1/2 study in adult pts with advanced unresectable or metastatic solid tumors (NCT05867420) comprised two parts: dose escalation (Part 1) and dose expansion (Part 2). Pts with advanced solid tumors that were resistant/refractory to current standard treatment, and with at least 1 measurable lesion per RECIST 1.1, were eligible. The primary objective was safety. Secondary objectives included efficacy, pharmacokinetics, pharmacodynamics, and immunogenicity. Results: As of December 30, 2025, we have treated 104 pts at dose-escalation phase (n=19) and dose expansion phase (n=85). Primary tumor types included NSCLC (n=48, 46.2%), CRC (n=35, 33.6%), ovarian cancer (n=11, 10.6%), cervical cancer (n=4, 3.8%) and others (n=6, 5.8%). Most pts (83/104, 79.8%) had received two and more lines of prior treatment, and 71 pts (68.3%) had undergone previous immunotherapy. No dose-limiting toxicities was observed in dose escalation phase, and the maximum tolerated dose has not been reached up to 3 mg/kg. In SS set, any-grade treatment-related adverse events (TRAEs) were reported in 98/104 (94.2%) pts, with the most common being rash (39/104, 37.5%), anemia (37/104, 35.6%), and elevated aspartate aminotransferase (19/104, 18.3%). ASKG915 monotherapy exhibited dose-dependent efficacy across dose levels. In the EFR cohort, encouraging antitumor activity was observed at the mid-dose level or above in both late-line NSCLC that had progressed on prior immunotherapy and MSS CRC: a 30% confirmed partial response (PR) rate (3/10) in non-liver metastatic MSS CRC at the mid-dose level or above, and a 30% PR rate (3/10) in NSCLC patients at the high-dose level. Further efficacy data will be presented at the time of the conference. CD8+ T and NK cells demonstrated proliferative expansion at mid-to-high doses. Drug plasma exposure increased dose-dependently with low activated ASKG915 levels and no ADA impact after repeat dosing. Conclusions: ASKG915 demonstrated a promising clinical efficacy in pts with MSS CRC and NSCLC, with a well-tolerated safety profile. These results support further evaluation of ASKG915 as a monotherapy and in combination therapies. Clinical trial information: NCT05867420 .

A phase 3 trial of adagloxad simolenin/OBI-821 in patients with high-risk early-stage triple-negative breast cancer.

2551 Background: Adagloxad simolenin (AS) is a therapeutic cancer vaccine composed of Globo H linked to the carrier protein keyhole limpet hemocyanin. OBI-821 is an adjuvant administered with AS to strengthen immune response. This phase 3 trial assessed safety/efficacy of AS/OBI-821 in patients with triple negative breast cancer (TNBC; NCT03562637). Methods: Eligible patients for this randomized, open-label trial were adults with high-risk (≥1cm residual primary, ≥1 residual axillary node after neoadjuvant chemotherapy, or stage IIB or III cancer treated with adjuvant chemotherapy alone), primary localized early-stage, Globo H-positive (H-score ≥15) TNBC. Patients were required to have received ≥4 cycles of standard taxane- and anthracycline-based chemotherapy. Patients were randomized 1:1 to receive standard of care (SOC; observation alone [29%], capecitabine [69%], or a checkpoint inhibitor ± capecitabine [2%]) or AS/OBI-821 and SOC (observation alone [26.6%], capecitabine [69.6%], or a checkpoint inhibitor ± capecitabine [3.8%]). Patients in the AS/OBI-821 arm received 4 weekly doses of study drug administered via subcutaneous injection, followed by 4 biweekly doses, 4 doses every 4 weeks, then doses every 8 weeks up to 100 weeks. The primary endpoint was invasive disease-free survival (IDFS); secondary endpoints included safety assessed via treatment-emergent adverse events (TEAEs) and overall survival (OS). Patients were followed for up to 5 years after randomization for OS. Results: The intent-to-treat (ITT) population was comprised of 575 patients, with 286 in the AS/OBI-821 arm, and 289 in the SOC arm. Patients were a median age of 51 years (range 24-85), all were female, and most were white (52.7%). A total of 5.2% of patients had stage I cancer, 50.3% had stage II, and 37.1% had stage III. IDFS in the AS/OBI-821 arm was not statistically different than in the SOC arm (IDFS hazard ratio [HR], 1.23 [95% CI: 0.88, 1.73]; P = 0.23). Three-year IDFS rate was 54.5% for AS/OBI-821 arm and 56.4% for the SOC arm. Similar pattern was observed for OS with HR = 1.36 [95% CI: 0.77, 2.38). TEAEs occurred in 79.6% of patients in the AS/OBI-821 arm and 66.7% in the SOC arm. The most common TEAEs in the AS/OBI-821arm were palmar-plantar erythrodysaesthesia syndrome (18.7%), fatigue (11.6%), and headache (11.3%). Serious TEAEs occurred in 5.3% of patients in the AS/OBI-821 arm vs 6.2% in the SOC arm. One patient (0.4%) in the AS/OBI-821 arm had a TEAE of interstitial lung disease, assessed as unlikely related to treatment and one patient (0.3%) in the SOC arm had acute coronary syndrome that led to death. Conclusions: In this phase 3, randomized, open-label trial, AS/OBI-821 was well-tolerated, with a safety profile similar to SOC. IDFS and OS were similar between the AS/OBI-821 and SOC arms. The trial was terminated per recommendation by the Data and Safety Monitoring Board as it met prespecified futility criteria. Clinical trial information: NCT03562637 .