Browse Articles

Persistence of aggressive biology: Outcomes of ¹⁷⁷ Lu-PSMA-617 in <i>BRCA</i> -altered mCRPC—Systematic review and meta-analysis.

e17053 Background: BRCA1/2 alterations define an aggressive subtype of metastatic castration-resistant prostate cancer (mCRPC) associated with adverse clinical outcomes. Although ¹⁷⁷Lu-PSMA-617 improves survival in unselected mCRPC, its efficacy in BRCA-altered disease remains incompletely characterized. We conducted a systematic review and meta-analysis to evaluate outcomes of ¹⁷⁷Lu-PSMA-617 in BRCA-altered mCRPC. Methods: A comprehensive PRISMA-guided literature search was conducted on PubMed, Cochrane, and Embase to identify seven eligible studies reporting clinical outcomes of ¹⁷⁷Lu-PSMA-617 in patients with BRCA-altered mCRPC. Overall survival (OS) hazard ratios (HRs) comparing BRCA-altered vs. non-altered patients were pooled using a random-effects model with Hartung-Knapp adjustment. A complementary single-arm proportional meta-analysis was performed to estimate pooled PSA50 response rates in patients with BRCA alterations. Inter-study heterogeneity was quantified using the I² statistic. Results: Seven studies (predominantly retrospective cohort analyses; one randomized controlled trial) met inclusion criteria, contributing aggregate data on approximately 728 patients treated with ¹⁷⁷Lu-PSMA-617. Three studies provided comparative OS data (total n = 292 patients: 50 with BRCA1/2 alterations and 242 without), while two studies reported BRCA-specific PSA response rates (n = 39 BRCA-altered patients). In the comparative studies, patient characteristics included median ages ranging from 62 to 72 years and frequent prior exposure to androgen receptor pathway inhibitors (ARPI), taxanes (docetaxel and/or cabazitaxel), and, in select cohorts, prior PARP inhibitors (PARPi). BRCA-altered patients exhibited inferior OS compared with non-BRCA counterparts (pooled HR 2.01, 95% CI 0.53–7.64; I² = 11.8%). The single-arm meta-analysis yielded a pooled PSA50 response rate of 51.3% (95% CI 35.9–66.4; I² = 0%) among patients with BRCA alterations. Limited data on other endpoints, including radiographic progression-free survival (rPFS medians 4.5-17.3 months in BRCA subgroups across studies) and deeper PSA responses (PSA90 rates 9-40% in small BRCA1/2 subsets), indicated preserved biochemical and radiographic activity despite the observed survival detriment. Toxicities and discontinuation rates due to adverse events were inconsistently reported and sparsely detailed across studies. Conclusions: In this meta-analysis, BRCA-altered mCRPC treated with ¹⁷⁷Lu-PSMA-617 was associated with worse overall survival compared with non-BRCA disease, despite preserved PSA responses. These results highlight the aggressive biology of BRCA-altered mCRPC and support further biomarker-stratified studies to optimize patient selection and treatment sequencing.

Rethinking infection prevention after modern CLL and NHL therapy: A real-world study.

e19029 Background: Infections are a major source of morbidity in patients with hematologic malignancies, particularly those who develop secondary hypogammaglobulinemia during systemic therapy. Evidence supporting intravenous immunoglobulin (IVIG) use for infection prevention is derived from trials predating modern targeted therapies. Current guidelines allow for IVIG or prophylactic antibiotics in high-risk patients, yet real-world practice varies and contemporary comparative effectiveness data remain limited. Methods: We conducted a retrospective cohort study using the TriNetX Network. Patients with chronic lymphocytic leukemia or non-Hodgkin lymphoma initiating systemic therapy were identified, with therapy start defining the index date. Systemic therapies included Bruton tyrosine kinase inhibitors, venetoclax-based regimens and anti-CD20 monoclonal antibodies. Exposure groups were defined by receipt of IVIG, prophylactic antibiotics, or neither within 0–90 days after index. Outcomes were assessed beginning at a 90-day landmark to reduce immortal time bias. The primary outcome was a composite serious infection endpoint. Relative risk was assessed from days 90–270, and time-to-event analyses from days 90–455. Propensity score matching included demographics, systemic steroid exposure, respiratory comorbidities, baseline infection-related diagnoses, medication classes, and albumin as a proxy for baseline nutritional and functional status. Results: After matching, the antibiotics versus IVIG comparison included 257 patients per cohort, with 67 versus 79 events (RR 0.83, 95% CI 0.60–1.14; HR 0.80, 95% CI 0.58–1.10). Antibiotics versus neither included 263 patients per cohort, with 69 versus 53 events (RR 1.41, 95% CI 0.97–2.05; HR 1.39, 95% CI 0.97–1.98). IVIG versus neither included 538 patients per cohort, with 153 versus 104 events (RR 1.54, 95% CI 1.19–1.98; HR 1.72, 95% CI 1.34–2.21). Conclusions: In this large contemporary real-world analysis, neither IVIG nor prophylactic antibiotics demonstrated superiority over the other for prevention of serious infections in patients with hypogammaglobulinemia receiving modern systemic therapy. Neither strategy was associated with improved outcomes compared with no prophylaxis. The higher incidence of infections observed among patients receiving IVIG likely reflects preferential use of IVIG in patients perceived by clinicians to be at higher baseline risk, rather than a causal adverse effect of IVIG itself. These findings highlight the absence of clear comparative benefit among commonly used prophylactic approaches and underscore the need for prospective studies to better define patient selection and optimize infection prevention strategies. Residual confounding remains possible due to the inability to directly match baseline immunoglobulin levels, prior infection burden, and other granular measures of disease severity.

Pharmacokinetics and pharmacodynamics evaluation in phase I trial of APX-343A, selective NOX inhibitor targeting CAF-mediated immunosuppression in patients with advanced solid tumors.

TPS2684 Background: Cancer-associated fibroblasts (CAFs) drive fibrotic remodeling and immune exclusion in solid tumors and contribute to resistance to immune checkpoint inhibitors (ICIs). CAF-mediated immune suppression is regulated by upstream oxidative signaling pathways. NADPH oxidases (NOXs) are upregulated in CAFs and promote fibrotic remodeling and myeloid cell–mediated inhibition of cytotoxic T-cell function, linking CAF activity to immune exclusion and immunotherapy resistance. APX-343A is a selective NOX1/2/4 inhibitor designed to modulate CAF-associated tumor biology. This ongoing Phase 1 study evaluates pharmacokinetic (PK) and pharmacodynamic (PD) components to assess translational biomarkers of NOX target engagement and CAF-driven immune modulation supporting dose selection and rational combination strategies with immunotherapy. Methods: This is a Phase 1, dose-escalation study (NCT07123415) evaluating the safety, tolerability, PK, PD, and preliminary antitumor activity of orally administered APX-343A as monotherapy (Part A; 100–600 mg BID) and in combination with pembrolizumab (Part B; 200–600 mg BID) in patients with advanced solid tumors. Dose escalation is conducted independently using dose-limiting toxicity–based escalation in Part A and a Bayesian Optimal Interval (BOIN) design in Part B to determine the maximum tolerated dose (MTD) and recommended Phase 2 dose (RP2D). PK assessments are conducted in Part A to evaluate dose proportionality and support RP2D determination, including Cmax, Tmax, area under the concentration–time curve (AUC), and accumulation. PD assessments use peripheral blood and tumor tissue samples to evaluate biomarkers of NOX pathway inhibition and CAF-driven immune modulation, including NOX isoforms (NOX1/2/4), CAF-associated markers, and immune-related cytokines and chemokines (e.g., interleukins, CCL2). Integrated PK/PD analyses explore exposure–biomarker relationships. This ongoing Phase 1 trial of APX-343A is actively enrolling, with Part A dose-escalation cohorts underway. Integrated PK and PD assessments are designed to establish translational evidence of NOX target engagement and CAF modulation, providing a mechanistic foundation for dose selection and rational combination strategies with ICI. Clinical trial information: NCT07123415 .

Temporal trends and disparities in liver cancer incidence in the United States, 1999–2022: A nationwide analysis.

e16216 Background: Liver cancer remains one of the most lethal malignancies in the United States, with incidence trends that have shifted over recent decades. Characterizing long-term temporal patterns and demographic disparities is critical to understanding the evolving epidemiology of this disease. Methods: We analyzed liver cancer incidence data from 1999 to 2022 using the CDC WONDER database. Incidence rates were stratified by sex, age group, and ethnicity (Hispanic vs non-Hispanic). Joinpoint regression was used to identify temporal phases, estimate annual percent change (APC), and calculate 95% confidence intervals (CIs). Results: Liver cancer incidence over the study period demonstrated three distinct temporal phases, with patterns differing by sex and ethnicity. Phase I (1999–2010) was characterized by a significant increase in incidence across most demographic groups, with the steepest rise among non-Hispanic males (APC +6.76%), followed by non-Hispanic females (APC +3.65%). Incidence trends among Hispanic males and females did not change significantly during this period. Phase II (2011–2015) showed a deceleration in incidence growth. Non-Hispanic males experienced slowing increases, while incidence among females and Hispanic populations largely stabilized without statistically significant changes. Phase III (2016–2022) demonstrated divergent trends. Hispanic males experienced a significant decline in incidence (APC −5.23%), while non-Hispanic males showed a downward trend that did not reach statistical significance (APC −2.86%). Incidence among females remained stable across both ethnic groups. Age-stratified analyses indicated that middle-aged adults, particularly males, exhibited the most pronounced phase-specific changes, whereas younger age groups often lacked sufficient case counts for stable trend estimation. Conclusions: Over more than two decades, liver cancer incidence in the United States has followed distinct temporal phases with marked demographic disparities. Early increases, driven largely by non-Hispanic males, have slowed or reversed, while Hispanic males have experienced a significant recent decline. These findings highlight persistent disparities and emphasize the need for targeted prevention, risk factor modification, and surveillance strategies as liver cancer epidemiology continues to evolve. These trends suggest that prevention and surveillance efforts should be dynamically tailored by sex and ethnicity to sustain recent gains and address populations in which liver cancer burden remains high.

Impact of human immunodeficiency virus on in-hospital outcomes among patients hospitalized with multiple myeloma: A National Inpatient Sample study, 2016–2022.

7552 Background: Outcomes among hospitalized patients with multiple myeloma (MM) and concomitant human immunodeficiency virus (HIV) infection remain incompletely characterized at a national level. We evaluated the association of HIV infection with in-hospital outcomes among MM hospitalizations in the United States. Methods: We analyzed 161,303 MM hospitalizations and identified HIV infection using ICD-10 diagnosis codes (B20–B24, Z21). To improve comparability, we restricted analyses to hospitalizations with length of stay (LOS) &lt; 30 days and complete covariate data, then performed 1:1 propensity score matching using demographics, hospital characteristics, Charlson comorbidity category, and key comorbidities. Primary outcome was in-hospital mortality. Secondary outcomes included LOS, total hospital charges, and inpatient complications. Results: Of 161,303 MM hospitalizations, 0.4% (n = 681) had HIV infection. After applying analytic restrictions, 458 HIV-positive hospitalizations were matched to 458 HIV-negative hospitalizations with good covariate balance; mean age was 60.0 years. HIV infection was not associated with higher in-hospital mortality (OR 0.68, 95% CI 0.37–1.21; p = 0.191). Median LOS was similar between groups (5 days in both); however, on log-transformed analysis HIV infection was associated with shorter LOS (LOS ratio 0.90; p = 0.033). HIV infection was associated with a nonsignificant trend toward lower total hospital charges (cost ratio 0.88, 95% CI 0.77–1.01; p = 0.061). HIV infection showed nonsignificant trends toward increased venous thromboembolism, including pulmonary embolism (OR 4.05; p = 0.078) and deep vein thrombosis (OR 1.88; p = 0.085). No significant differences were observed for mechanical ventilation or cerebrovascular accident. Conclusions: In matched analyses of MM hospitalizations, HIV infection was not associated with increased in-hospital mortality and was associated with shorter LOS despite similar median LOS. Total charges were not significantly different, though a trend toward lower charges was observed. Signals for increased venous thromboembolism warrant further study in larger cohorts.

Untreated advanced urothelial carcinoma (aUC) in the modern treatment era: Patterns of care from a decade-long national cohort.

4570 Background: Over the past decade, the therapeutic landscape for locally advanced/unresectable and de novo metastatic urothelial cancer (aUC) has transformed with the introduction of immune checkpoint inhibitors (ICIs) and antibody-drug conjugates (ADCs). These agents offer a more favorable benefit-risk and superior efficacy vs historical chemotherapy. We hypothesized that the proportion of untreated patients with aUC has declined over time, reflecting increased uptake of novel therapeutics. Methods: We analyzed the U.S. Flatiron Health EHR-derived de-identified database. We included patients diagnosed with aUC between Jan 1, 2011 and April 30, 2025. Patients were classified as treated if they received first-line (1L) systemic therapy within 6 months of advanced diagnosis. Calendar time was divided into three clinically relevant eras: pre-ICI approval (2011-2016), post-ICI but pre-ADC approval (2017-2019), and post-ADC approval (2020-2025). Outcomes included proportion untreated and time to treatment initiation (TTI) from advanced diagnosis. We estimated the cumulative incidence of treatment initiation using competing-risk methods, treating death before initiating as a competing event. Subgroup analyses evaluated patients with de novo stage IV disease at initial diagnosis. Results: We identified 12,849 patients with aUC (median age 73, 72.9% male, 78.7% white, 57.5% cis-ineligible, 78.3% bladder primary site). The proportion remaining untreated 6 months post advanced diagnosis declined across therapeutic eras: 36.8% untreated in the pre-ICI era, 27.4% in the post-ICI, pre-ADC era, and 22.7% untreated in the post-ADC era. Year-specific analysis showed steady improvement in treatment uptake: treatment within 6 months rose from 56% in 2011 to 79% in 2024 (Table). In the TTI analysis, overall, 25.1% had not initiated 1L therapy before death or data cutoff. Median TTI shortened across eras (63 days, 49 days, 45 days) and competing-risk cumulative incidence initiating therapy increased over time. For example, the probability of initiating 1L therapy within 90 days of aUC diagnosis across eras was 55.9% (pre-ICI), 66.1% (post-ICI, pre-ADC) and 70.7% (post ADC). Conclusions: In a large national cohort, untreated rates among patients with aUC declined markedly from 2011-2025 paralleling the introduction of ICIs and ADCs. Novel, better-tolerated therapies appear to have expanded real-world treatment eligibility and reduced early attrition. Temporal trends in treatment rates. Year 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 3 months 0.48 0.53 0.54 0.57 0.58 0.61 0.68 0.65 0.67 0.69 0.69 0.72 0.7 0.74 6 months 0.56 0.6 0.61 0.63 0.65 0.69 0.73 0.7 0.74 0.76 0.76 0.78 0.78 0.79 1 year 0.58 0.63 0.64 0.66 0.67 0.71 0.75 0.72 0.77 0.77 0.77 0.79 0.8 0.81

Interfacial Water Reorientation in Gadolinium‐Doped Ru/RuO <i> _x </i> Heterostructures Boosts Alkaline Hydrogen Oxidation Electrocatalysis

ABSTRACT The sluggish proton transfer in alkaline electrolytes severely limits hydrogen oxidation reaction (HOR) kinetics of fuel cells. Although the orientation of interfacial water strongly governs proton transport, precise control over its configuration remains challenging due to the inherently random distribution of water molecules. Herein, we report the synthesis of isolated Gadolinium (Gd) embedded into Ru/RuO x heterostructures and engineer the built‐in electric fields (BIEF) at the heterostructure for effectively tuning surface oxophilicity and directing the reorientation of interfacial water configuration to boost HOR catalysis of fuel cells. We find that isolated Gd atoms intensify the BIEF at the Ru/RuO x interface, driving strengthened asymmetrical charge redistribution and finely tune work function of catalyst. This electronic modulation in turn optimizes surface oxophilicity of active sites, enabling balanced hydroxyl species coverage and preferential stabilization of H 2 O ↓ ‐oriented water, thereby strengthening hydrogen‐bond network and constructing an efficient interfacial proton‐conduction channel. The resulting Ru/RuO x ‐Gd@C delivers an exceptional mass activity of 8.87 mA µg Ru − 1 and an exchange current density of 0.39 mA cm −2 , outperforming Pt/C by 6.6 and 2.0 times, respectively. An anion‐exchange‐membrane fuel‐cell assembled with Ru/RuO x ‐Gd@C achieves a PGM‐normalized peak power density of 16.3 W mg Ru −1 and operates stably for over 60 h at 0.2 A cm −2 .

The pancreatic cancer drug market

Surface Microstructural Reconstruction of CdS Buffers Enables Low‐Voltage‐Loss Kesterite Solar Cells With &gt;15% Efficiency

ABSTRACT The performance of kesterite Cu 2 ZnSn(S,Se) 4 (CZTSSe) solar cells is critically governed by the quality of the CZTSSe/CdS heterojunction; however, the morphology, crystallinity, and defect landscape of CdS buffer layers are intrinsically constrained by the inevitable competition between homogeneous and heterogeneous nucleation in widely used chemical bath deposition (CBD). Here, we report a simple yet effective surface microstructural reconstruction strategy based on chemical polishing that overcomes these challenges beyond the reach of conventional CBD process regulation. Specifically, polishing CZTSSe/CdS films with a Na 2 S/thiourea aqueous solution selectively removes low‐quality CdS particulates while inducing surface recrystallization and sulfur‐vacancy compensation. As a result, the CdS films exhibit markedly improved microstructure, enhanced crystallinity, and more homogeneous surface electrical properties. Benefiting from suppressed interfacial charge recombination and accelerated charge transport, kesterite solar cells achieve a champion efficiency of 15.3% with a high open‐circuit voltage ( V OC ) of 560 mV and a record‐low V OC deficit ( E g / e ‐ V OC ) of &lt; 0.5 V, significantly advancing kesterite photovoltaics toward low voltage loss. Moreover, this work establishes a broadly applicable post‐deposition paradigm for improving CBD‐CdS–based optoelectronic devices across a wide range of material systems and applications.

Real-world 177Lu-PRRT experience in a lower-middle-income country (LMIC): Treatment delivery, tumor response, safety, and rechallenge patterns.

e16325 Background: 177Lu-peptide receptor radionuclide therapy (PRRT) is an established systemic option for metastatic neuroendocrine tumors (NETs). However, real-world evidence from LMIC settings remains limited. Affordability and infrastructure heavily affect access. We share our institutional outcomes, treatment delivery data, safety, and rechallenge utilisation. Methods: We analysed medical records of NET patients treated with 177Lu-PRRT at Sushrut Hospital, Mumbai, India. We summarized demographics, primary site, metastatic spread, PRRT cycles delivered, cumulative administered activity, treatment duration, toxicities, rechallenge use, and response at completion of 4 cycles were summarized descriptively Results: Twenty-six patients received 177Lu-PRRT. Median age was 47 years (range 18 to 70) and 65% were male (17/26). Primary sites were most commonly the pancreas (34.6%, 9/26) and the liver (23%, 6/26). Liver metastases were present in 69.2% (18/26), nodal disease in 57.7% (15/26), and bone metastases in 42.3% (11/26). Among 24 patients who received ≥2 cycles, the mean cycles delivered were 4.3 (median 4; range 2 to 7), with a mean cumulative activity of 596.8 mCi (median 592; range 282 to 1250). At completion of 4 cycles, best response was PR in 9, SD in 12, PD in 5, and CR in 0. This resulted in an ORR of 21.7% and a disease control rate of 69.6%. PRRT rechallenge was administered in 19% (5/26). In our center, a full 4-cycle PRRT course costs about USD 6,000, compared with the reported US Lutathera wholesale acquisition cost (WAC) of USD 55,896 per dose (USD 223,584 for 4 doses; drug cost only). Conclusions: In this real-world LMIC cohort, 177Lu-PRRT was deliverable in routine practice, achieving meaningful disease control (70%) without any Grade III/IV toxicities. Rechallenge was used in about 1 in 5 patients, showing that repeat treatment outside trials is feasible. The substantial affordability relative to the US highlights the potential for scalable PRRT access in resource-limited settings.

A boost for oncology care delivery: Improving hospital throughput with a redesign of multidisciplinary rounds.

e13555 Background: Hospital capacity constraints impact patient access, team workflows, and overall efficiency of care. Multidisciplinary Rounds (MDRs) are the cornerstone for patient flow activities. Capacity issues highlight the acute need for innovation and rapid change. MDRs are led by Registered Nurse (RN) Case Manager Navigators and had a rigid structure that hindered adaptability and participation for various oncology disciplines. Using stakeholder input, MDRs were redesigned to facilitate adaptability for oncology service-specific needs, coordination of activities prior to discharge, and increased collaboration. The purpose was to improve MDR efficiency and effectiveness to reduce discharge barriers and improve overall team satisfaction. Methods: In 2022, consultants were engaged to study our process and recommend nationally accepted MDR practices. With recommendation, we started full census reviews to establish comprehensive discharge plans. In 2024, stakeholder survey feedback highlighted inefficiencies and dissatisfaction with the revised MDR practices. Two strategies were used to address these concerns. 1: Survey data from hematology, solid tumors, and surgical cohorts were analyzed for themes. 2: A pilot was implemented, focusing discussions on patients within two days of discharge or with anticipated complex discharge plans. Results: Survey results revealed inconsistent practices and common themes across cohorts, such as the desired location of MDRs, timing, content, and key participants required by unit. The pilot led to significant improvements in participation among providers and other team members, effective communication, reduced total time, and greater verbalized satisfaction with the process. Lessons learned were leveraged to enhance MDR redesign for other services. The primary metric to gauge success of the revamped MDRs was accuracy of the estimated discharge date (EDD) by 9AM the day before discharge. The secondary metric was adoption of morning (AM) / afternoon (PM) discharge time designations, referred to as time of day (TOD). From baseline to pilot completion, EDD accuracy increased from 45% to 53% and TOD selection increased from 31% to 49%. Additionally, completion of discharge orders by 9AM increased by 37.5%, with 35% more patients leaving the hospital before noon. Our MDR change initiative led to quantifiable positive downstream outcomes. Conclusions: The latest MDR format led to highly effective processes showcasing teamwork and care quality. Customizing unit-specific MDRs improved participation, increased EDD accuracy and TOD selection. This allowed teams to better prepare patients for discharge and ensure discharge orders were entered by 9AM to improve patient discharge before noon. We were successful in incorporating varied perspectives in MDR redesign, which contributed to the overall goal of alleviating hospital capacity issues.

Updated clinical and translational results of olverembatinib (HQP1351) in patients with succinate dehydrogenase (SDH)–deficient tumors.

11539 Background: SDH deficiency causes succinate accumulation and metabolic defects in gastrointestinal stromal tumor (GIST) and paraganglioma. Olverembatinib downregulates lipid uptake protein CD36, restricting fatty acid supply required for SDH-deficient GIST cell growth. We report updated clinical and translational results from study HQP1351SJ0003 (NCT03594422) of olverembatinib in SDH-deficient GIST and paraganglioma. Methods: Patients (pts) (≥ 12 years) with GIST and other solid tumors received olverembatinib 30 to 50 mg PO QOD. Primary cells from tumor tissues of 9 pts with SDH-deficient GISTs were used for ex vivo analyses. All samples were collected at Sun Yat-sen University Cancer Center, with informed consent from pts. Cell migration/invasion was evaluated by wound healing, Transwell migration, and Matrigel invasion assays. Lipid levels were quantitated by LC-MS. Knockdown of CD36 and p38 was done by siRNA. Protein levels were assessed by western blot. Results: As of November 13, 2025, a total of 32 pts, including 26 pts with IHC-confirmed SDH-deficient GIST, had received olverembatinib. The median age of pts with SDH-deficient GIST was 30 (range 13-56) years; 19 (73.1%) were female. Twenty-five pts (96.2%) had received ≥ 1 TKI and 13 (50.0%) ≥ 3. Six (23.1%) pts experienced PR as best response, and the median PFS was 25.7 (range 12.9-NR) months. An additional six pts with paraganglioma had received olverembatinib (all were SDH-deficient as determined by either negative SDHB IHC or SDHB gene mutation). The median age was 47.5 (range 31-61) years; all were male. Best responses were observed in four pts, with SD lasting ≥ 4 cycles (CBR, 66.7%); SD represents meaningful clinical benefit in this slow-growing tumor type. The median PFS was 8.25 (range 1.87-NR) months. SDH-deficiency was associated with a higher risk of tumor metastasis. In SDH-deficient (vs. SDH-competent) GIST cells, significantly higher linoleic acid levels were observed. Supplementation of free fatty acids promoted SDH-deficient cell migration, which could be inhibited by olverembatinib at 30 nM. In addition, siRNA knockdown of p38, a kinase target of olverembatinib, and treatment with two p38 inhibitors significantly downregulated CD36 protein levels, suggesting that p38 regulates cellular CD36. Further, western blot analysis showed that olverembatinib dose-dependently inhibited CD36 protein and phosphorylated p38 protein levels in primary cells from SDH-deficient pts. These findings suggest olverembatinib suppresses fatty acid-promoted migration via the p38-CD36 pathway. Conclusions: Olverembatinib showed sustained clinical efficacy in SDH-deficient GIST and paraganglioma. Mechanistically, olverembatinib inhibits fatty acid–promoted invasion and migration of SDH-deficient GIST by targeting the p38-CD36 signaling pathway. Clinical trial information: NCT03594422 .

Investigating reasons patients do not join a cancer clinical trial.

11092 Background: The perennial challenge to cancer clinical trials is accrual. To date, the focus to improve accruals rested on addressing patient knowledge, attitudes and beliefs rooted in patient reluctance to be experimented on, financial and practical barriers and issues of mistrust. We undertook this study to describe the impact of an educational intervention to democratize recruitment and improve understanding of and information about cancer clinical trials. Methods: We implemented a time series intervention (INT) identifying cancer patients at treatment decision points at 3 participating Cancer Centers, educating consented breast, liver and lung cancer patients about clinical trials in general and providing clinical trial match lists to oncologists and patients prior to treatment decision-making. We defined a treatment decision point as a new, recurrent or progressing cancer. We surveyed patients and their oncologists pre- and post-intervention about their knowledge, attitudes and beliefs (KABs) of clinical trials. We compared pre &amp; post KABs using Chi Sq and Fisher’s Exact T tests. Results: Of 523 consented patients at treatment decision points, 514 completed pre &amp; 368, post-INT surveys; 60% were non-white with no racial ethnic difference in response rates (70% BIPOC v 79% White; p = 0.07). Following an educational intervention, there was no significant change in positive attitude towards clinical trials (CT): 71% pre-intervention vs 77% post-intervention; p = 0.55, regardless of race-ethnicity (RE) (p = 0.5025). Black patients as compared to white patients, reported greater barriers to CT including: transportation (60% vs 37%; p &lt; 0.001) and family responsibilities (46% vs 25%; p = 0.001). There was no RE difference in other barriers: patient report of physicians talking with them about clinical trials (89% vs 65%; p = .36), insurance (77% vs 74%; p = 0.6051); out of pocket expenses (64% vs 53%; p = 0.12); time off work (41% vs 37%; p = 0.5457). 25/28 (89%) treating oncologists of consented patients responded. Oncologists reported being less likely to offer CTs to patients who they perceived were reluctant to: get placebo (48%), be randomized (48%), have increased family responsibilities (52%), would have difficulty getting time off work (45%) or had comorbidities (40%). Conclusions: More than ¾ of patients have a positive attitude towards clinical trials with no racial-ethnic difference. Black patients reported more barriers to trial enrollment than white patients. Despite Black patients having more barriers that would reduce oncologists’ likelihood of offering CTs to patients, there was no RE in patients’ report of oncologists talking with them about CTs. Clinical trial information: NCT05146297 .

Time from diagnosis to CAR T as a predictor of outcomes in transformed follicular lymphoma: A CIBMTR real-world analysis.

e19016 Background: CD19-directed CAR T-cell therapy is effective in relapsed/refractory transformed follicular lymphoma (tFL), but the prognostic impact of the interval from diagnosis to CAR T infusion remains unclear. We evaluated the association between time from initial diagnosis to CAR T therapy and clinical outcomes in a large real-world cohort. Methods: This retrospective multicenter study utilized the Center for International Blood and Marrow Transplant Research (CIBMTR) registry, based on the P-6131 dataset by Thiruvengadam et al. Adult tFL patients who received CAR T-cell therapy were included. Patients stratified based on the interval from initial lymphoma diagnosis to CAR T infusion as &lt;24 months versus ≥24 months. Survival outcomes were estimated using the Kaplan–Meier method. Cox proportional hazards regression models were used to calculate crude and adjusted hazard ratios (HR) with 95% confidence intervals (CI). Logistic regression was used to assess associations with binary outcomes. Statistical analyses were performed using SPSS, with p-values &lt;0.05 considered statistically significant. Results: Among 923 tFL patients treated with CAR T therapy, the mean interval from initial diagnosis to CAR T infusion was 22.8 months (SD 34.6). A total of 706 patients (76.5%) underwent CAR T within 24 months of diagnosis, while 217 (23.5%) received CAR T at ≥24 months. Patients with a ≥24-month interval demonstrated significantly improved outcomes. Median overall survival (OS) was 42.45 months in the ≥24-month group versus 29.14 months in the &lt;24-month group (p = 0.004). While the crude OS benefit was significant (HR 0.691, 95% CI 0.537–0.887; p = 0.004), this association was insignificant after adjustment for disease and treatment characteristics (HR 0.765, 95% CI 0.538–1.089; p = 0.137).Progression-free survival (PFS) was markedly longer in the ≥24-month group, with median PFS of 29.44 months compared with 8.85 months in the &lt;24-month group (p &lt;0.001). This difference remained significant after multivariable adjustment (adjusted HR 0.696, 95% CI 0.514–0.943; p = 0.019).At 2 years, relapse incidence was significantly lower in patients with a ≥24-month interval (33.0% vs 49.9%; OR 0.503, 95% CI 0.363–0.697; p &lt;0.001). Two-year treatment-related mortality was similar between groups (11.1% vs 8.65%; OR 1.301, 95% CI 0.79–2.143; p = 0.301). Rates of cytokine release syndrome (OR 1.026, 95% CI 0.712–1.476; p = 0.892), and ICANS (OR 1.385, 95% CI 0.991–1.936; p = 0.056) were comparable. Conclusions: An interval ≥24 months from diagnosis to CAR T was associated with superior PFS and reduced relapse in tFL. Early progression likely reflects aggressive, CAR T–resistant biology, underscoring the need for risk-adapted and novel strategies for early-transforming disease.

Capsule networks for cancer imaging applications.

e13687 Background: Computer vision models are increasingly used in clinical oncology, with several FDA-approved algorithms now available for diagnostic imaging and pathologic analysis. Ensuring safe clinical deployment of these models requires evaluation not only of their predictive performance but also the stability of their predictions. While convolutional neural networks (CNNs) and vision transformers (ViTs) are the most widely used computer vision models, their behavior has been shown to be unstable, with small pixel-level changes leading to incorrect predictions. This instability raises concerns, as noise or imaging artifacts could render models unreliable. Capsule Networks (CapsNets) offer unique advantages compared to CNNs and ViTs including compact model size and lower training data requirements, resulting in increased model stability. We evaluated the utility of CapsNets for oncology imaging applications by studying both predictive performance and model stability. Methods: Three oncologic datasets were selected for analysis, focused on lung CT nodule classification (n = 1,633), breast ultrasound malignancy detection (n = 780), and hematopathology classification (n = 17,092). For each task, we trained two CNNs (ResNet-18 and ResNet-50), one ViT (MedViT), and two CapsNets (DR-CapsNet and BP-CapsNet). Predictive performance was evaluated using area under the receiver operating characteristic curve (AUC). To test model stability, AUC was measured on slightly perturbed images generated by projected gradient descent. Latent space embeddings were inspected to identify underlying reasons for model stability. Results: All models demonstrated comparable baseline performance across all datasets, with AUCs ranging from 0.860-0.996 for CNNs, 0.889-0.998 for the ViT, and 0.856-0.993 for the CapsNets. CapsNets, however, exhibited greater model stability than the CNNs and ViT. On slightly perturbed images (ε = 0.032), CapsNets achieved stable AUCs ranging from 0.856-0.987 (1.65% average decline) across all datasets. The CNNs and ViT had considerably worse AUCs ranging from 0.305-0.712 (37.5% average decline) and 0.410-0.678 (36.9% average decline), respectively. Latent space analysis showed greater feature stability for CapsNets. Notably, CapsNets also achieved these gains with substantially smaller model sizes, using over 35% fewer parameters than all other models. Conclusions: CapsNets have significant utility in oncologic image classification. Compared to CNNs and ViTs, CapsNets consistently exhibited similar performance and superior model stability with the added benefit of decreased model size. Model Model Size (x 10 6 parameters) Baseline AUC Range Perturbed AUC Range Average % AUC Decline ResNet-18 11.2 0.892-0.996 0.559-0.712 29.8 ResNet-50 23.5 0.860-0.995 0.305-0.652 45.1 MedViT 31.1 0.889-0.998 0.410-0.678 36.9 DR-CapsNet 7.0 0.880-0.993 0.857-0.898 3.2 BP-CapsNet 7.0 0.856-0.989 0.856-0.987 0.1

Duration to treatment interval and survival outcomes in aggressive lymphoma: A U.S. academic single-center cohort.

e19048 Background: The influence of duration-to-treatment interval (DTI) on outcomes in aggressive lymphoma remains uncertain. Prior studies, predominantly population-based and conducted outside the United States, have reported worse outcomes among patients with short DTI, with some demonstrating independent associations after multivariable adjustment. Other analyses emphasize heterogeneity by cohort, subtype, and health-system factors, finding no independent DTI effect. We evaluated whether DTI independently predicts overall survival (OS) in a U.S. academic single-center cohort. Methods: We analyzed 129 consecutive adults with newly diagnosed aggressive lymphoma treated at a tertiary academic center. DTI was modeled as quartiles and prespecified categories (&lt;7, 7–13, &gt;13 days). Univariable and multivariable Cox proportional hazards models estimated hazard ratios (HR) for OS, adjusting for stage, age group, ECOG performance status, and insurance. Proportional hazards assumptions were tested; sensitivity analyses included stage-stratified models and a DTI×stage analysis. Results: Median age was 61 years (46.5% male). The most common subtype was diffuse large B-cell lymphoma (82.8%), followed by double hit lymphoma (6.2%), high grade B-cell lymphoma (6.2%), and primary mediastinal B-cell lymphoma (4.8%). Median DTI for the entire cohort was 13 days. Unadjusted analyses demonstrated a non-monotonic association between DTI quartiles and OS: Q2 (HR 0.47, 95% CI 0.22–0.99) and Q3 (HR 0.52, 95% CI 0.26–1.03) showed lower hazards versus Q1, while Q4 did not differ significantly (HR 0.82, 95% CI 0.41–1.63). The per-quartile trend lost significance after stage adjustment (HR 0.89, p=0.34). In a stage-adjusted model, DTI 7–13 versus 7 days was associated with lower mortality (HR 0.47, 95% CI 0.22–1.00, p=0.049). However, in the fully adjusted model, DTI was not independently associated with OS (7–13 vs 7: HR 0.69, p=0.39; &gt;13 vs 7: HR 0.74, p=0.37). Stage remained the dominant predictor (HR 3.29, p=0.003). Shorter DTI groups were enriched for advanced-stage disease, while longer DTI included more early-stage cases. Insurance status was not associated with OS or PFS. Sensitivity analyses including stage-stratified models and a DTI×stage interaction did not materially change effect estimates. Conclusions: In this U.S. single-center cohort, DTI was not independently associated with OS after adjustment for disease severity and clinical covariates. Apparent adverse outcomes with shorter DTI reflected clinical triage of advanced disease rather than treatment delay. The absence of association between insurance status and outcomes is reassuring, suggesting insurance-related DTI variation did not translate into inferior survival. These findings highlight confounding in DTI analyses and support the need for larger multicenter studies with causal inference approaches.

Real-world use and prognostic performance of a 14-gene molecular risk assay in stage IA–IIA non-small cell lung cancer.

e20031 Background: Despite early diagnosis, stage IA-IIA non-small cell lung cancer (NSCLC) carries a high risk of relapse and mortality. The AIM-HIGH randomized trial demonstrated that a validated 14-gene molecular expression assay identified molecularly high-risk patients who benefited from adjuvant cisplatin-based chemotherapy. We evaluated the real-world applicability of these findings. Methods: This retrospective cohort study included 233 patients with resected stage IA–IIA NSCLC at Mount Sinai Medical Center from 2021 to 2024 who were stratified by genetic risk (GR) using a 14-gene molecular assay and by clinical risk using National Comprehensive Cancer Network (NCCN) clinicopathologic risk (CPR) features. Disease-Free Survival (DFS) rates were analyzed using Kaplan–Meier estimates with log-rank testing, and multivariable Cox regression. Associations were evaluated using chi-square or Fisher’s exact tests. Analyses were conducted using R. Results: The cohort was 50.2% female, 91% white, 79.4% Hispanic. Adjuvant chemotherapy was given to 5/118, 12/64, and 9/51 patients in the low-, intermediate-, and high-GR groups, respectively. Relapse rates by GR were 9.3%, 20.35%, and 21.6% for low, intermediate, and high risk, respectively, while relapse rates by CPR were 14.47%, 18.5% for low and high risk, respectively. Median DFS was not reached. At 36 months, DFS rates by GR were 84.3%, 66.2%, and 71.4% for low, intermediate, and high risk, respectively (p = 0.045), and 36-month DFS by CPR was 77.6% and 76.1% for low and high risk, respectively (p = 0.43). Adjuvant chemotherapy was not associated with improved DFS. For postoperative circulating tumor DNA (ctDNA), 36-month DFS rates were 20.0% in positive and 86.1% in negative patients (p &lt; 0.0001). On multivariable Cox regression, ctDNA positivity (HR 4.20, 95% CI 1.27-13.85; p = 0.019), high genetic risk (HR 2.16, 95% CI 1.09-4.26; p = 0.027), positive surgical margins (HR 4.61, 95% CI 1.24-17.17; p = 0.023), and PD-L1 expression (HR 2.58, 95% CI 1.16-5.74; p = 0.020) were independently associated with DFS. Conclusions: The 14-gene assay demonstrated superior prognostic stratification for relapse in resected IA–IIA NSCLC compared with traditional CPR features, in our primarly Hispanic cohort. Intermediate and high GR groups showed lower 36-month DFS, with the high-risk group also showing lower 12-month DFS, suggesting higher early relapse risk. Despite evidence supporting molecularly guided adjuvant chemotherapy, fewer than 19% of patients in intermediate- and high-GR groups received adjuvant treatment in our cohort, reflecting a gap of care that may contribute to higher recurrence. High GR, postoperative ctDNA positivity, PD-L1 expression, and positive surgical margins were independently associated with worse DFS and should be considered in postoperative risk assessment to guide individualized adjuvant treatment decisions.

Cancer burden among young adults in Brazil: Incidence, stage at diagnosis, and sociodemographic shifts.

e22564 Background: Cancer incidence among young adults (YA, 20–49 years) is rising globally. Information on risk factors and epidemiological behavior in this population is scarce, particularly in Latin America. Understanding temporal trends in YA may help explain underlying drivers of cancer increases and guide targeted public health strategies. Methods: We conducted a retrospective study including all patients with confirmed cancer (excluding non-melanoma skin cancer) at a tertiary center in Brazil from 2005 to 2024. Patients were stratified into YA (20–49 years) and late-onset cancer (LOC, ≥50 years). Annual incidence trends for the 10 most frequent YA cancers were analyzed using linear regression, calculating Average Annual Percent Change (AAPC). Clinical and sociodemographic variables at diagnosis (stage, diabetes, smoking, alcohol use, BMI ≥30, education, race) were analyzed using logistic regression. Results: Among 28,630 cases, 7,045 (24.5%) occurred in young adults (YA). Overall cancer incidence increased significantly in YA (AAPC 2.5%, p = 0.003) and in late-onset cancer patients (AAPC 3.85%, p &lt; 0.001). The ten most frequent cancers in YA were breast, lymphoma, head and neck (CCP), cervix, colon, thyroid, brain, leukemia, melanoma, and ovary. Considering annual percent increases, the largest AAPCs were observed for cervical cancer (5.6%), colon (4.21%), lymphoma (3.23%), melanoma (2.73%), thyroid (2.71%), and breast (1.11%). Analysis of clinical and sociodemographic variables in YA revealed significant temporal increases in advanced-stage cases (stage III–IV) for breast (OR 1.06, 95% CI 1.03–1.09), lymphoma (OR 1.19, 95% CI 1.13–1.26), head and neck (CCP, OR 1.29, 95% CI 1.21–1.38), cervix (OR 1.22, 95% CI 1.15–1.30), colon (OR 1.20, 95% CI 1.13–1.28), brain (OR 1.32, 95% CI 1.09–1.60), and melanoma (OR 1.55, 95% CI 1.13–2.12). Pardo/Black race significantly increased in breast (OR 1.03, 95% CI 1.00–1.06), lymphoma (OR 1.04, 95% CI 1.01–1.07), cervix (OR 1.07, 95% CI 1.02–1.13), and colon (OR 1.09, 95% CI 1.04–1.15). Obesity (BMI ≥30) decreased in colon cancer (OR 0.77, 95% CI 0.66–0.90). Conclusions: Cancer incidence is rising among young adults in Brazil, with more advanced-stage diagnoses and higher representation among Pardo/Black individuals. These findings highlight the importance of tailored public health strategies, early detection, and further research to reduce disparities and improve outcomes.

A multicenter non-randomized phase III study of sentinel node navigation surgery for early gastric cancer.

4076 Background: Sentinel node navigation surgery (SNNS) has the potential to minimize the extent of gastrectomy, thereby improving quality of life (QOL) after surgery for patients with early gastric cancer (GC). This prospective, multicenter, non-randomized, phase III study aims to evaluate the long-term outcomes of SNNS for early GC. Methods: A total of 13 institutions enrolled clinically diagnosed primary T1N0M0 gastric cancer patients with a single lesion (≤40 mm) and without previous endoscopic treatment. Sentinel nodes (SNs) were identified using dye and radioisotope tracers, then subjected to intraoperative rapid pathology. Patients with negative SN metastasis underwent individualized surgery consisting of limited stomach resection and SN basin dissection, while those with positive SN metastasis underwent standard gastrectomy with D2 lymph node dissection. The primary endpoint was five-year relapse-free survival (RFS). The secondary endpoints were overall survival (OS), SN detection rate, diagnostic accuracy for SNs, SN distribution, and the distribution of metastatic and non-SNs. Postoperative QOL was also assessed. RFS was assessed in the full analysis set (FAS), excluding patients who did not undergo intraoperative SN biopsy and pathological assessment during surgery. The data cutoff date for the analysis was May 31, 2025. Clinical trial information: jRCTs031180432. Results: A total of 187 patients were enrolled from May 2014 to May 2020. Of the 180 patients in FAS, 62.2% were male. The median (range) age was 64 (27–85). cT1a/cT1b was 18.3%/81.7%, respectively. The median follow-up time was 60.2 months (range: 9.0–62.8). The 5-year RFS rate (95% CI) was 95.6% (91.43–98.06). Since the lower boundary of the 95% CI exceeded the predefined threshold of 90%, the study met its primary endpoint. Events for RFS included one recurrence and three non-cancer deaths. SNs were identified in all patients, and LN metastasis was detected in 17 patients (9.4%) during surgery. Finally, 143 patients (79.4%) underwent individualized surgery consisting of limited stomach resection and SN basin dissection. The 5-year RFS rate (95% CI) for this cohort was 96.5% (92.03 - 98.86). There was no mortality after surgery. Conclusions: SNNS for early GC has been shown to be non-inferior to standard gastrectomy, with an acceptable toxicity profile. SNNS represents a promising alternative as a potential new standard treatment for early GC. Clinical trial information: jRCTs031180432 .

Evolution of the therapeutic landscape in metastatic uveal melanoma: A single-institution real-world analysis of treatment efficacy and patient survival.

e21529 Background: Fifty percent of patients with uveal melanoma (UM) develop metastatic disease (90% to the liver) which is associated with a median overall survival of only one year. Clinical management relies on extrapolation of data from cutaneous melanoma with immune checkpoint inhibitors (ICI), recently approved bispecific agents such as tebentafusp for HLA A02:01 positive patients, and liver directed Hepzato Kit (whole liver perfusion of high dose melphalan). Methods: Upon IRB approval, database search using ICD codes identified 590 patients (pts) who were diagnosed with ocular melanoma since 2014 at Huntsman Cancer Institute (HCI). We further screened for those pts who developed metastatic melanoma and underwent systemic therapy. Clinical data were collected by chart review. Results: 42 unique pts were identified, some treated by multiple lines of therapy. The majority received ICIs (12 anti-PD1 alone and 30 with combination anti-CTLA4 + anti-PD1); ORR were 17% (2/12) and 17% (5/30), with median OS 12.8 and 12.3 mon, respectively. Tebentafusp was given to 3 pts; all had PD but median OS was 23.2 mon. One of 5 pts treated by anti-VEGF +/- anti-PD1 had PR (20%) with median OS of 4.04 mon. Seven patients received Hepzato Kit whole liver perfusion of melphalan, with 5 receiving bridging immunotherapy prior to Hepzato initiation (3 tebentafusp and 2 ipi/nivo); 4/5 had a PR (80%) and one remaining pt had SD, vs 0/2 pts who received Hepzato alone had a response with only one SD. All pts treated at HCI for Hepzato perfusion tolerated without severe AEs. OS analysis of Hepzato cohort requires longer follow up. Translational studies are ongoing to further understand the clinical benefit. Conclusions: ICI had limited efficacy in UM. Tebentafusp and anti-VEGF showed evidence of activity. Combination of immunotherapy with Hepzato whole liver perfusion shows promise that warrants further investigation. Treatment N Median Age Sex (M/F) Best Response (CR/PR/SD/PD/Other) Median PFS (Mon) Median OS (Mon) Median Duration on Tx (Mon) Anti-PD1 12 64 7/5 1/1/4/6/0 2.98 12.78 7.48 Anti-PD1 + Anti-CTLA4 30 54.5 18/12 1/4/5/20/0 2.41 12.3 1.4 Tebentafusp 3 62 2/1 0/0/0/3/0 2.03 23.17 2.6 Anti-VEGF +/- anti-PD1 5 55 3/2 0/1/0/3/1 2.68 4.04 2.76 Hepzato 2 48.5 1/1 0/0/1/0/1 3.96 3.96 3.91 Hepzato + Tebentafusp 3 69 2/1 0/2/1/0/0 7.6 9.1 6.57 Hepzato + ipi/nivo 2 59.5 1/1 0/2/0/0/0 5.09 11.07 8.76