Browse Articles

Impact of niraparib maintenance on progression-free survival and overall survival: An updated meta-analysis.

5591 Background: Niraparib is an oral poly(ADP-ribose) polymerase (PARP) inhibitor used as maintenance therapy in advanced ovarian cancer after platinum response. Mature OS data were lacking, so we performed an updated meta-analysis. Methods: PubMed, Embase, and CENTRAL were searched from January 2023 to January 2026 for randomized controlled trials (RCTs) of niraparib maintenance in advanced ovarian cancer. Hazard ratios with 95% confidence intervals were pooled using random-effects models for TFST, OS, overall PFS, and subgroup PFS by BRCA, HRD status, and newly diagnosed patients. The study is registered in PROSPERO with ID CRD420261294461. Results: Four RCTs (PRIMA, NOVA, NORA, and PRIME) including 1935 patients were analysed, of whom 1,291 received niraparib, and 644 received placebo. Niraparib significantly prolonged the TFST (HR 0.54, 95% CI 0.42-0.69; I²=77.6%; p<0.0001). No statistically significant improvement in OS was observed (HR 0.94, 95% CI 0.82-1.08; I²=10.0%; p=0.398). Niraparib significantly improved overall PFS (HR 0.42, 95% CI 0.31-0.58; I²=83.8%). PFS in BRCA-mutated patients (HR 0.33, 95% CI 0.24-0.45; I²=46.0%), HRD-positive patients (HR 0.41, 95% CI 0.30-0.56; I²=58.6%) (all p<0.0001). HRD-negative patients (HR 0.56, 95% CI 0.35-0.89; I²=52.4%), and newly diagnosed (HR 0.55 95% CI 0.39-0.80, I²=79.1%) (all p≤0.015). Conclusions: Niraparib provides durable disease-control benefits across molecular subgroups and significantly delays the need for subsequent therapy, however, mature data does not demonstrate a corresponding overall survival advantage. These findings refine patient selection and highlight the need for strategies that translate PFS gains into survival benefit. Updated meta-analysis of niraparib efficacy outcomes in advanced ovarian cancer. Study TFST (HR) Overall Survival (HR) Overall PFS (HR) PFS in BRCA mutation (HR) HRD+ PFS (HR) HRD− PFS (HR) Newly Diagnosed (HR) PRIMA 0.74 (0.62–0.89) * 1.01 (0.84–1.23) * 0.66 (0.55–0.78) * 0.43 (0.31–0.59) * 0.51 (0.40–0.66) * 0.67 (0.50–0.89) * 0.66 (0.55–0.78) * PRIME 0.45 (0.34–0.59) 0.63 (0.38–1.03) * 0.45 (0.34–0.60) 0.40 (0.23–0.68) 0.48 (0.34–0.68) 0.41 (0.22–0.75) 0.45 (0.34–0.60) NOVA (gBRCA) 0.57 (0.41–0.78) * 0.85 (0.61–1.20) * - 0.27 (0.17–0.41) 0.38 (0.24–0.60) - - NOVA (non-gBRCA) 0.58 (0.45–0.74) * 1.06 (0.81–1.37) * - - - - - NORA 0.39 (0.29–0.52) * 0.86 (0.60–1.23) * 0.32 (0.23–0.45) 0.22 (0.12–0.40) 0.22 (0.12–0.40) - - Total Result (95% CI) 0.54 (0.42 – 0.69) 0.94 (0.82–1.08) 0.42 (0.31–0.58) 0.33 (0.24–0.45) 0.41 (0.30–0.56) 0.56 (0.35–0.89) 0.55 (0.39–0.80) I² (%) 77.6 10 83.8 46 58.6 51.4 79.1 p value <0.0001 0.3979 <0.0001 <0.0001 <0.0001 0.0146 0.0014 *Indicate updated values. TFST: time to first subsequent therapy; PFS: progression-free survival; HR: Hazard ratio; CI: confidence intervals.

Spatiotemporal characteristics and geophysical drivers of the fractal dimension and b-value in North China seismicity

The North China Craton Destruction Zone, one of continental China’s most seismically active regions, exhibits complex spatiotemporal seismicity patterns due to its intricate geology. Using the North China Block earthquake catalog (1970–2025), we applied a stochastic declustering method based on the spatiotemporal Epidemic-Type Aftershock Sequence (ETAS) model to process the data. We then calculated the spatial fractal dimension and the b -value of seismicity using the correlation dimension and the maximum likelihood method, respectively. These analyses were conducted to explore their relationships with regional tectonic setting, stress state, and geophysical field characteristics. Declustering increased the spatial fractal dimension of background seismicity from 1.20 to 1.45, effectively reducing spatial clustering without significantly altering the statistical characteristics of the b-value. The fractal dimension is heterogeneous, with high values concentrated in tectonically fragmented zones like the Fen-Wei seismic belt and the central-northern North China Plain, while low values correspond to stable areas such as the Ordos Block. The fractal dimension exhibited systematic temporal changes, characterized by a decreasing before the mainshock and a recovery afterward. This pattern potentially reflects a transition from strain localization, akin to rupture nucleation, to post-seismic stress readjustment, suggesting it may reflect pre-seismic processes. The spatial relationship between fractal dimension and the b -value is complex, showing negative correlation in extreme value zones and positive correlation in moderate value zones. Different parameter combinations ( b -value and fractal dimension) can indicate distinct seismic hazard states. Quantitative analysis reveals the strongest positive correlation between fractal dimension and maximum shear strain rate, and a negative correlation with terrestrial heat flow, confirming that seismic spatial complexity is primarily driven by tectonic forces and modulated by deep thermal state. This study provides refined fractal-based criteria for characterizing seismicity and informing medium- to long-term seismic hazard assessment in the region.

Rapid and deep diesel desulfurization using bimetallic Cu-Fe/TiO2 nano-photocatalysts optimized by response surface methodology

Ideological and political education behavior recognition by artificial intelligence transformers

Genome-wide chromatin profiling reveals a nonlimiting role for RXR in macrophage-like cells stimulated with multiple nuclear receptor agonists

LIF expression as a predictive biomarker in metastatic renal cell carcinoma.

4548 Background: Clear cell renal cell carcinoma (ccRCC) is biologically heterogeneous, and optimal selection between VEGF-targeted and ICI–based therapies remain an unmet clinical need. Leukemia inhibitory factor (LIF), an IL-6 family cytokine implicated in angiogenesis and immune regulation, is associated with poor prognosis in ccRCC; however, its predictive value for treatment selection has not been defined. Methods: Baseline tumor LIF expression was evaluated in 2,026 patients with metastatic ccRCC across 12 clinical and molecular cohorts, including CheckMate 009/010/025/214, JAVELIN Renal 101, IMmotion150, TCGA-KIRC, CPTAC-3, and institutional datasets. RNA-seq data were batch-corrected, normalized, and immune-deconvoluted (CIBERSORTx). Among 1,306 patients with treatment data, 1,058 with complete clinical covariates were included in treatment-by-biomarker interaction analyses. Results: LIF-high tumors (median cutoff) were associated with younger age (60 vs 63 years, p<0.001) and increased sarcomatoid features (17% vs 8%, p<0.001). A significant treatment-by-LIF interaction was observed for progression-free survival (PFS; interaction p=0.006). High LIF expression predicted inferior PFS on TKI monotherapy (HR 1.26, p<0.001), while this adverse association was absent with ICI-based regimens (monotherapy or combination; HR 0.70, p<0.001). No significant interaction was observed for overall survival. LIF expression discriminated progressive disease from disease control (p<0.001). Immune deconvolution revealed that LIF-high tumors exhibited an inflamed but immunosuppressive microenvironment, characterized by enrichment of macrophages (M0 and M2), neutrophils, and regulatory T cells, with preserved CD8⁺ T-cell infiltration. LIF expression correlated strongly with CXCL8 (R=0.45), and more modestly with FGF2 and VEGFA, consistent with VEGF-independent angiogenic signaling. Conclusions: LIF identifies a biologically distinct subset of metastatic ccRCC characterized by angiogenic and immunosuppressive features and poor outcomes with TKI monotherapy. These data support LIF as a predictive biomarker to inform treatment selection and highlight the LIF/CXCL8 axis as a potential therapeutic target in advanced RCC. Treatment-by-LIF interaction analysis for progression-free survival in metastatic clear cell renal cell carcinoma. Groups N HR 95% CI P-value Full Cohort* 1792 1.09 1.01 – 1.14 0.017 Treatment interaction Model § Main effect (LIF in TKI) 1058 1.26 1.108 - 1.443 <0.001 LIF × IObased (IO and IO+IO) 1058 0.70 0.567 - 0.854 <0.001 LIF × IO+VEGF 1058 0.89 0.73 - 1.079 0.230 LIF × MTORi 1058 0.84 0.689 - 1.035 0.104 Stratified Analyses § : TKI Monotherapy 377 1.27 1.114 - 1.455 <0.001 IO/IO+IO 194 0.87 0.741 - 1.03 0.107 IO+VEGF 364 1.15 0.996 - 1.337 0.057 *Adjusted for IMDC, Age, Sex, Sarcomatoid Features, and Treatment modality. § Adjusted for IMDC, Age, Sex, Sarcomatoid Features.

FF-10832 liposomal gemcitabine monotherapy or combination with pembrolizumab in patients with advanced non–small cell lung cancer.

8560 Background: FF-10832, a stable liposomal formulation of gemcitabine with a prolonged half-life (mean t 1/2 of 26 hours compared to ~3 hours for gemcitabine HCl), has demonstrated single agent activity in patients (pts) with solid tumors. In pre-clinical models, FF-10832 has yielded superior anti-tumor activity compared to gemcitabine alone and in combination with PD-1 inhibitors. Following a safety run-in, we investigated the clinical activity of FF-10832 in an expansion study, administered alone or in combination with pembrolizumab (PEM), in pts with advanced/metastatic (m) NSCLC (NCT05318573). Methods: In the open-label single arm safety run-in, 40 mg/m 2 FF-10832 and 200 mg PEM every 3 weeks was established as safe and tolerable in 12 pts including 6 mNSCLC patients with unlimited prior therapies. The trial was expanded using this regimen. Pts were randomized to receive FF-10832 alone (mono) or FF-10832 +PEM (combo). Eligibility stipulated up to 3 prior therapies in the advanced/metastatic setting with prior disease progression (PD) on platinum and PD-(L)1-directed therapy. Pts receiving FF-10832 monotherapy could add PEM to FF-10832 upon PD determined by RECIST 1.1. To be evaluable for anti-tumor activity, pts must have had at least one RECIST assessment ≥6 weeks after baseline. All pts were evaluable for safety. Results: A total of 41 NSCLC pts were treated with FF-10832 (N=21) or FF-1032+PEM (N=20); median # of prior therapies was 2 for both arms. RECIST evaluability was achieved in 30 pts (14/16; mono/combo); 11 pts (7/4; mono/combo) came off study before evaluation for anti-tumor activity. No objective responses were observed. Stable disease (SD) was recorded in 9/14 pts in the mono arm and 13/16 in the combo arm; 4 pts (1 mono/3 combo) with >20% tumor shrinkage. Median time on study for the mono and combo arms was 9 (1 – 48.1) and 12.1 (1.1 – 44) weeks, respectively. Median (95% CI) PFS in months: mono arm, 2.7 (1.4 – 4.3); combo arm, 2.8 (1.9 – 5.5). Five mono pts began FF-10832+PEM treatment upon PD, two that remain on combination therapy with SD for 23 and 33 weeks after progressing on FF-10832 monotherapy. Related AEs were similar in mono vs combo arms; those in ≥15% of pts in each arm included nausea, vomiting, fatigue, rash, infusion-related reaction, and anemia. Related grade ≥3 AEs occurred in 33% (mono) and 45% (combo) of pts. The only attributable grade ≥3 events in >1 pt were in the combo arm (n): anemia (3), fatigue (2), dyspnea (2), and lymphopenia (2). FF-10832 PK showed a profile consistent with that previously reported (terminal t 1/2, ~30 hours). Overall survival data are pending. Conclusions: FF-10832 alone or in combination with PEM in platinum-CPI refractory pts proved tolerable with a safety profile matching previous experience. There were no objective responses, but the majority of evaluable pts had SD at 1 st assessment, some with tumor shrinkage. Aggregate mPFS was 2.7 mos. Clinical trial information: NCT05318573 .

Multidisciplinary AE management education for EGFR-mutated NSCLC: Mixed-methods QI outcomes.

e23368 Background: Anti-EGFR therapies for EGFR-mutated NSCLC have dermatologic and other adverse events (AEs) that benefit from standardized, multidisciplinary management in community practice. Methods: Three-phase mixed-methods quality-improvement initiative. Phase 1 used Premier Network encounter analytics from ~4,400 hospitals/health systems (Jan 2021–Apr 2024) and interviews with 5 community oncologists (Sep 2024) to define gaps. Phase 2 delivered an ACCME-accredited, on-demand 4-module video CME activity (1.25 credits) plus NPI-targeted microlearning (launched Feb 11, 2025) focused on AE monitoring, referral triggers, and patient counseling. Outcomes included participation, pre/post 3-item knowledge test, confidence ratings, and 3 case-based questions. Phase 3 included follow-up interviews (Nov 2025) and a separate ecological comparison of Premier encounters between baseline (Jan 2021–Apr 2024) and a later period (Jul 2024–Jun 2025). Results: 1,983 clinicians participated (385 enduring CME; 1,598 microlearning); 86% completed the enduring CME. Most challenging AEs were rash/pruritus (49%), ILD/pneumonitis (43%), and infusion-related reactions (37%). Mean test performance improved 31%→57%, and confidence managing dermatologic AEs improved 26%→44%. Case-based performance improved for selecting first-line therapy for EGFR exon 20 insertion NSCLC (+35%; p < 0.01), timing dermatology referral for rash during amivantamab treatment (+30%; p < 0.01), and counseling for nail changes/paronychia (+14%; p < 0.01). In Premier period comparisons, amivantamab encounters increased and patients were more racially/ethnically diverse. AE-related 90-day follow-up visits (overall and dermatologic) were less frequent for amivantamab in the later period (some comparisons p < 0.05; others NS). Dermatologic AE encounters for osimertinib were rare ( < 10 in each period). Post-activity interviews described more consistent documentation, earlier symptom reporting, and clearer referral workflows. Conclusions: ACCME-accredited multidisciplinary education improved clinician knowledge and scenario-based decision-making for AE management in EGFR-mutated NSCLC. Concurrently, Premier Network trends (including fewer AE-related 90-day follow-up visits for amivantamab in a later period) support testable hypotheses that standardized counseling, monitoring, and referral triggers may reduce downstream AE-related utilization; future work should evaluate these hypotheses using exposure-linked designs and post-launch analytic windows.

Effects of immune checkpoint inhibitor on the quality of life of patients with advanced/metastatic esophageal, gastro-esophageal junction, and gastric cancers: A systematic review of randomized controlled trials.

e16351 Background: Immune checkpoint inhibitors (ICIs) have demonstrated survival benefits in patients with advanced or metastatic upper gastrointestinal cancers; however, their impact on patient-reported quality of life (QoL) remains poorly characterized. This systematic review aims to evaluate the effect of ICIs on QoL outcomes in patients with advanced or metastatic gastric/gastro-esophageal junction cancer (GC/GEJC) or esophageal cancer (EC). Methods: PubMed, Scopus, and Cochrane databases were systematically searched for publications up to 18 January 2025 to identify randomized controlled trials (RCTs) reporting QoL outcomes in patients with GC/GEJC or EC receiving ICIs. QoL were assessed using the EORTC QLQ-C30, QLQ-OES18 and the EQ-5D visual analogue scale (VAS). Outcomes were pooled using random-effects meta-analyses, using mean differences (MDs) for change-from-baseline outcomes and hazard ratios (HRs) for time-to-deterioration outcomes. Results: Eleven RCTs including 6392 unique patients were analyzed (7 RCTs including 2060 patients with EC and 5 RCTs including 4332 patients with GC/GEJC). Pembrolizumab was evaluated in six RCTs, tislelizumab in 4, and nivolumab in 1. Compared with chemotherapy, ICIs did not worsen cancer-specific QoL (QLQ-C30/GHS: MD 0.55 [95%CI-1.65 to 2.75]; I2= 57%, p = 0.030) or general health status (EQ-5D VAS: 0.16 [-1.04 to 1.36]; I2= 0%, p = 0.690). Moreover, ICIs were not associated with earlier deterioration in global health status (HR 0.91 [0.80 to 1.03]; I2= 0%, p = 0.911) or physical functioning (0.88 [0.76 to 1.01]; I2= 1%, p = 0.363). No significant differences were observed for esophageal-specific symptoms, including dysphagia (1.03 [0.69 to 1.53]; I2= 74%, p = 0.021), reflux (1.01 [0.55 to 1.84]; I2= 87%, p < 0.001), appetite loss (0.99 [0.85 to 1.16]; I2= 0%, p = 0.642), nausea/vomiting (0.80 [0.59 to 1.08]; I2= 64%, p = 0.062), and pain (0.93 [0.81 to 1.08]; I2= 0%, p = 0.470). Conclusions: These findings support the tolerability of ICIs, demonstrating maintenance of health-related QoL in patients with advanced or metastatic upper gastrointestinal cancer without accelerated deterioration across functional or symptom domains.

Phase 1b dose-expansion study of IMC-002, a anti-CD47 monoclonal antibody, in patients with advanced triple negative breast cancer (TNBC).

2524 Background: IMC-002 is a novel therapeutic agent with a distinct mechanism of action. In a phase 1a dose-escalation study, IMC-002 demonstrated favorable safety and tolerability, and clinical activity was observed in patients with hepatocellular carcinoma (HCC) in a subsequent phase 1b trial. Here, we present results from the phase 1b expansion cohort in patients with triple-negative breast cancer (TNBC), focusing on safety, efficacy, and analyses of soluble biomarkers. Methods: Eligible pts had advanced TNBC with progression after ≥1 prior systemic therapy and ECOG PS ≤1. IMC-002 (20 mg/kg Q3W) was administered in combination with gemcitabine plus carboplatin or paclitaxel and continued until disease progression. Tumor response was assessed every 6 weeks per RECIST 1.1 and iRECIST. Baseline ctDNA was analyzed using the AlphaLiquid 1000 platform (1,023 -gene panel). To exclude germline and clonal hematopoiesis of indeterminate potential (CHIP) variants, paired PBMC sequencing was performed. The assay detected single nucleotide variants (SNVs), small insertions and deletions (INDELs), fusions, copy number alterations (CNAs), microsatellite instability (MSI), and blood-based tumor mutational burden (bTMB). Results: A total of 12 pts with advanced TNBC received IMC-002 in combination with gemcitabine plus carboplatin (n=9) or paclitaxel (n=3). Most pts had received ≥2 prior systemic therapy (n=11), and ECOG PS 1 was observed in 10 pts. TRAEs reported in > 1 pt included grade 3 hematologic events, namely anemia (n=6) and hemolytic anemia (n=5). Non-hematologic TRAEs were limited to grade 1–2 and included skin rash (n=9), vitreous floaters (n=4), photopsia (n=3), and IRR (n=2). Among 12 efficacy-evaluable pts, the ORR was 25%, the DCR was 75%, and the CBR (disease control ≥ 6 months) was 42%. Exploratory ctDNA analyses identified no MSI-H tumors. One patient with a partial response was BRCA-positive. Patients achieving clinical benefit showed a higher median maximum somatic allele frequency (MSAF) than those without clinical benefit (32.65 vs 1.00; p=0.19). No clear association between bTMB and clinical activity was observed. Conclusions: IMC-002 in combination with gemcitabine plus carboplatin or paclitaxel demonstrated encouraging antitumor activity with a manageable safety profile in heavily pretreated patients with advanced TNBC. Exploratory baseline ctDNA analyses did not identify molecular features clearly associated with clinical benefit. Clinical trial information: NCT05276310 .

GDF-15 in oropharyngeal squamous cell carcinoma (OPSCC): Association of an HPV-dependent biomarker with malnutrition and low skeletal muscle.

6110 Background: GDF15 is a stress-response cytokine implicated in systemic inflammation and cancer-related wasting. We assessed whether baseline circulating GDF15 captures an HPV-stratified nutritional/body composition phenotype and whether longitudinal GDF15 changes relate to treatment outcome in OPSCC. Methods: We analyzed a prospective cohort of OPSCC patients (pts) undergoing curative intent treatment (CIT; surgery and radiotherapy [RT]-based) with available baseline plasma GDF15. HPV status was defined by dual detection (p16 INK4a and HPV DNA). Baseline nutritional status (Patient Generated Subjective Global Assessment), symptom burden (nutrition-impact composite score), body mass index (BMI), and CT-derived skeletal muscle index (SMI) were analyzed when available. Response assessment was performed 8-12 weeks after CIT, classifying cases as no evidence of disease (NED) vs evidence of disease (ED). Longitudinal GDF15 was analyzed using linear mixed-effects models (LMEN) (random intercept) including interactions of time with response status and induction chemotherapy. Results: We included 138 pts (HPVnegative [−] n=82; HPVpositive [+] n=56). Stage III–IV was more frequent in HPV- vs HPV+ (85.4% vs 30.4%; p<0.001) and ECOG 0 less common (9.8% vs 51.8%; p<0.0001). RT-based treatment predominated (85.5%) mainly chemoradiotherapy ± induction (82.6%). Baseline GDF15 was higher in HPV- vs HPV+ (median 1591 vs 803 pg/mL; p<0.001). HPV− pts showed worse nutritional phenotype: lower BMI (22.7 vs 27.2 kg /m²; p=0.001), more severe malnutrition (37.8% vs 7.1%; p<0.001), higher symptom burden (1.90±0.94 vs 1.39±0.71; p=0.002) and higher sarcopenia (34.1% vs 16.1%; p=0.029). In pts with CT body composition (n=89), SMI was lower in HPV− vs HPV+ (45.6 vs 53.6 cm²/m²; p=0.010). Baseline GDF15 differed by nutritional status (p=0.014) and was associated with symptom burden and SMI (both p<0.05). In LMEN, GDF15 increased significantly from baseline to end of treatment (β=2264 pg/mL; p<0.001). Baseline HPV positivity was independently associated with lower GDF15 (β=−980 pg/mL; p<0.001). At the end of treatment, NED was associated with lower GDF15 compared with ED (β=−1141 pg/mL; p=0.011). Induction chemotherapy showed no significant effect on GDF15 change (β=−448 pg/mL at end of treatment; p=0.22). No significant interaction between time and baseline HPV status was observed (p=0.26) indicating comparable longitudinal trajectories of GDF15 in HPV+ and HPV-pts. Conclusions: Baseline GDF15 is strongly HPV-dependent and reflects a multidimensional nutritional and muscle phenotype. GDF15 increases during CIT and lower end-of-treatment GDF15 is independently associated with NED, whereas induction chemotherapy does not significantly modify GDF15 change.

Impact of SGLT2 inhibitors on clinical outcomes in diabetic patients with CLL receiving BTK inhibitors: A global propensity-matched study.

7051 Background: Diabetic patients with chronic lymphocytic leukemia (CLL) receiving Bruton’s tyrosine kinase inhibitors (BTKi) face intersecting challenges of metabolic dysfunction, immune dysregulation, and treatment-related toxicities, including cardiovascular complications and increased susceptibility to infections. Sodium–glucose cotransporter 2 inhibitors (SGLT2i), have emerged as a pleiotropic cardio-renal-metabolic therapeutic class with benefits extending beyond glycemic control. We hypothesize that early initiation of SGLT2i in diabetic patients with CLL may mitigate BTKi–associated toxicities and improve long-term treatment tolerability and clinical outcomes. Methods: We used the TriNetX Global Collaborative Network to identify patients with CLL (C91.1, C83.00) and diabetes (E08–E13) treated with BTK inhibitors (ibrutinib, acalabrutinib, or zanubrutinib), comparing those with and without SGLT2 inhibitor exposure. Propensity score matching adjusted for demographic, metabolic, cardiovascular, and renal confounders, as well as hypoglycemic and cardiovascular medications. After matching, 1,102 patients (551 per cohort) were included. Outcomes were assessed within 3 years of the index event, and hazard ratios (HRs) with 95% confidence intervals were estimated. Results: The 3-year all-cause mortality was significantly lower with concurrent BTKi and SGLT2i compared with BTKi alone (HR 0.50, 0.38–0.66), corresponding to a 3-year overall survival of 77% versus 62% (p < 0.0001). Hematologic toxicity was also significantly lower in the SGLT2i cohort, including grade 3 thrombocytopenia (HR 0.632, 0.461–0.866), grade 3 anemia (HR 0.532, 0.422–0.671), and neutropenia (HR 0.618, 0.419–0.911). Infectious complications were less frequent in the SGLT2i group, with a significant decrease in neutropenic fever (HR 0.393, 0.221–0.698), sepsis (HR 0.686, 0.504–0.934), and pneumonia (HR 0.65, 0.506–0.835). Likewise, a lower incidence of Deep vein thrombosis and pulmonary embolism was found in the SGLT2i cohort (0.555, 0.35–0.879). Surprisingly, cardiovascular outcomes were not significantly different between groups, including hypertension (HR 1.10, 0.961–1.278), atrial fibrillation (HR 1.073, 0.86–1.339), and heart failure (HR 1.128, 0.919–1.385). Conclusions: SGLT2i use was associated with significantly longer 3-year overall survival and lower hazards of severe cytopenias and serious infections in diabetic patients with CLL receiving BTK inhibitors, with no meaningful differences in major cardiovascular outcomes. The mechanisms underlying these associations is unclear; however, emerging evidence suggests that SGLT2i may exert immunomodulatory effects and have been shown to reduce tumor cell proliferation in breast cancer models. Further studies are warranted to confirm the role of SGLT2i in diabetic CLL patients.

Prospective analysis of cardiovascular complications in HER2-positive breast cancer patients treated with anthracycline-free regimens.

e12545 Background: The assessment of cardiovascular complications (CVC) frequency and the predictive role of the HFA-ICOS scale is needed for pts with HER2-positive breast cancer (BC) receiving anthracycline-free therapy. Methods: 2,592 consecutive BC pts were prescreened to enroll 188 newly diagnosed stage I-III HER2-positive BC women (mean age 56 yrs) before start of HER2-inhibitors. Routine examination was added with cardiac biomarkers and speckle-tracking echocardiography every 3 month during 1 year and 3 months after the end of treatment. Anthracyclines were used 12 pts (6%). Results: The median follow-up was 7.8 months (ongoing). Cardiovascular comorbidities included arterial hypertension (AH) - 77 (41%), dyslipidemia - 140 (74%), obesity - 56 (30%), diabetes - 20 (10%), chronic kidney disease - 29 (15%), coronary artery disease - 5 (3%), atrial fibrillation - 9 (5%), chronic heart failure (HF) - 7 (4%), past ischemic stroke - 6 (3%), past myocardial infarction - 4 (2%) pts. The overall incidence of at least one CVC during antiHER2-therapy was 28%, including destabilization of AH — 15 (8%); newly diagnosed AH — 9 (5%); decrease in left ventricular global longitudinal strain > 15% from baseline — 19 (10%); asymptomatic decline in left ventricular (LV) ejection fraction (EF) < 50% — 5 (3%); new elevation of NT-proBNP level — 19 (10%); asymptomatic elevation of troponin level — 6 (3%); moderate hydropericardium — 5 (3%); deep vein thrombosis — 6 (3%); superficial vein thrombosis — 5 (3%); pulmonary embolism — 1 (0.5%); ischemic stroke — 1 (0.5%); symptomatic HF with LVEF decline < 40% — 1 (0.5%). Therapy was interrupted due to CVC in only two pts (ischemic stroke and symptomatic HF). In the multivariate analysis increased LV mass index was an independent risk factor for CVC (aOR = 1.022 per 1 g/m²; p = 0.017). By the HFA-ICOS scale, 160 (85%) pts were categorized as low/moderate risk and 28 (15%) as high/very high risk. Risk stratification by the HFA-ICOS scale showed no statistically significant prognostic value for overall CVC: frequency was 39% in the high/very high-risk group vs 26% in the low/moderate-risk group (p = 0.169). However, for the specific endpoint of cancer therapy-related cardiac dysfunction (CTRTD), the HFA-ICOS scale showed a statistically significant but clinically modest association (OR = 2.39, 95% CI 1.002–5.70, p = 0.045). The weak effect size (Phi = 0.147) suggests limited discriminatory power for CTRTD prediction in this cohort. Conclusions: CVC occur in 28% of HER2-positive BC on anthracycline-free therapy, AH was most frequent CVC. Prevalence of severe adverse events was 1,5% highlighting the overall controllability of most complications and supporting the manageable safety profile of trastuzumab-based regimens under active monitoring. The HFA-ICOS scale showed endpoint-specific utility for CTRTD, but limited value for overall CVC prediction.

ILCness to stratify lobular-like breast cancer biology and as an outcome independent of <i>CDH1</i> mutation in early ER-positive/HER2-negative breast cancer.

576 Background: Invasive lobular carcinoma (ILC) represents a distinct histological subtype of breast cancer; however, its underlying molecular basis remains incompletely defined. Although loss-of-function mutations in CDH1 are considered a hallmark genetic event, substantial biological heterogeneity persists that cannot be fully explained by histology or mutation status alone. Methods: To address this limitation, we developed a continuous transcriptional framework, termed the ILCness score, to quantitatively capture lobular-like tumor biology. A tightly constrained principal component–based model was trained in the METABRIC cohort and fully fixed without retraining, enabling reproducible projection across independent datasets and expression platforms. Results: Unsupervised transcriptomic analyses demonstrated that invasive ductal carcinoma (IDC) and ILC did not form discrete expression clusters but instead distributed along a continuous molecular spectrum. The ILCness score robustly captured this continuum, with mixed-histology tumors occupying intermediate positions. Importantly, lobular-like transcriptional programs quantified by ILCness showed limited concordance with CDH1 mutation status. Across tumors, higher ILCness was consistently associated with reduced DNA repair and E2F target activity (Spearman |ρ| &gt; 0.4, P &lt; 0.001), whereas epithelial–mesenchymal transition displayed a context-dependent pattern. Tumor microenvironment analyses revealed that stromal and immune cell infiltration correlated more strongly with ILCness than with CDH1 mutation status. These associations were reproducibly validated in independent cohorts without model recalibration. Clinically, ILCness stratified prognostically distinct subgroups within histologically defined tumors, including IDC, independent of CDH1 mutation (disease-specific survival, P &lt; 0.01). Across multiple neoadjuvant and systemic therapy datasets, higher ILCness showed consistent directional associations with reduced pathological complete response to chemotherapy and differential response patterns to immune checkpoint inhibition and endocrine therapy. Conclusions: These findings establish lobular breast cancer as a continuum of transcriptional states rather than a discrete entity defined by histology or single-gene alterations. The ILCness score provides a scalable RNA-based framework for integrating molecular heterogeneity with tumor biology and clinical behavior, with potential implications for biologically informed patient stratification.

Real-world utilization and outcomes of hematopoietic stem cell transplantation in testicular cancer.

e17016 Background: Approximately 10–30% of patients with testicular cancer experience relapse or develop platinum-refractory disease, necessitating consideration for high-dose chemotherapy with hematopoietic stem cell transplantation (HSCT). However, contemporary real-world utilization patterns and outcomes for these patients remain incompletely characterized at a population level. Using a large federated electronic health record network (TriNetX), we identified adult patients with testicular cancer undergoing HSCT to describe their baseline demographics, pre-transplant chemotherapy exposure, and overall survival (OS) in routine clinical practice. Methods: We conducted a multicenter retrospective cohort study using the TriNetX research platform, a federated network of de-identified electronic health records from participating healthcare organizations. Adult patients (≥18 years) diagnosed with testicular cancer who underwent HSCT between January 1, 2010, and January 1, 2026, were identified using ICD-10 and procedure codes; patients with competing hematologic malignancies were excluded. The index date was defined as the date of the first HSCT. Baseline covariates included age, sex, race, ethnicity, and geographic region. Follow-up extended from the index date until death or the last recorded encounter. OS was estimated using the Kaplan–Meier method, and the association of covariates with mortality was evaluated using Cox proportional hazards models. Results: We identified 333 patients with testicular cancer who underwent HSCT. The mean age at transplant was 41 years (range 19–77). The cohort was predominantly White (76%) and 21% Hispanic. Prior to HSCT, the majority of patients had significant chemotherapy exposure: 224 (67%) received etoposide, 183 (55%) carboplatin, and 132 (40%) cisplatin. Additionally, 122 (36%) received ifosfamide and 95 (29%) paclitaxel. The median follow-up was 604 days (IQR 1,735), with a mean follow-up of 1,234 days. At the end of the observation window, 94 patients had died, corresponding to a crude mortality risk of 28.2%. The estimated 5-year overall survival was 65.3% by Kaplan–Meier analysis. The survival curve demonstrated an early decline in the first 12–24 months post-transplant, followed by a prolonged plateau consistent with long-term survivorship. Conclusions: The observed 5-year overall survival of approximately 65% indicates that HSCT achieves durable long-term survival for a substantial proportion of patients with relapsed or refractory testicular cancer in routine practice. These findings support the ongoing referral of appropriate candidates to specialized transplant centers and underscore the need for prospective studies to refine patient selection, optimize peri-transplant management, and evaluate novel consolidation strategies in this high-risk population.

Survival outcomes with contemporary treatments including immunotherapy for poorly differentiated neuroendocrine carcinomas of the head and neck.

6115 Background: Poorly differentiated Neuroendocrine Carcinomas of the Head and Neck (HN-NECs) are rare aggressive malignancies with limited data guiding therapeutic approaches, particularly in the metastatic setting, with cytotoxic chemotherapy (chemo) comprising the backbone of management. Despite the widespread use of immune checkpoint inhibitors/immunotherapy (IO) in advanced lung NECs, evidence regarding their utility in the extrapulmonary setting, particularly for HN-NECs, is scarce. We explored real-world outcomes with various therapeutic strategies, including IO, in patients with metastatic HN-NECs, using the National Cancer Database (NCDB). Methods: Patients with metastatic HN-NECs diagnosed between 2018 to 2023 were identified. Overall survival (OS) was analyzed using Kaplan-Meier estimations and Cox proportional hazards regression. Results: A total of 275 patients with metastatic HN-NECs were identified. Primary sites included larynx (n = 75, 27.3%), pharynx (n = 72, 26.2%), nasal cavity + paranasal sinuses (n = 67, 24.4%), and oral cavity (n = 61, 22.2%). Most common metastatic sites were liver (n = 128, 46.5%) and bone (N = 272; n = 116, 42.6%). Of these 275 patients, 212 (77.1%) received chemo, 95 (34.5%) received IO, 122 (44.4%) received radiotherapy (RT), and 46 (16.7%) underwent primary site surgery (surg). Among patients with available survival data (N = 227), receipt of chemo (n = 174) and/or IO (n = 78) were associated with prolonged OS vs. no chemo (p &lt;0.001) and/or no IO (p = 0.005), respectively. Use of RT (n = 105; p = 0.17) and surg (n = 39; p = 0.18) were not significantly associated with OS (Table). Among 174 patients who received chemo, concurrent IO (chemoIO; n = 73) was associated with significantly prolonged OS vs. chemo alone (14.1 months [m] vs. 9.9 m; HR, 0.67; 95% CI, 0.47 – 0.96; p = 0.03). On multivariable analysis adjusting for bone metastatic disease (the only other significant prognostic covariate on univariable regression), chemoIO remained associated with reduced mortality risk vs. chemo alone (HR 0.57; 95% CI, 0.39 – 0.82; p = 0.003). Conclusions: Akin to lung NECs, IO may improve OS in HN-NECs. Despite NCDB limitations, chemoIO was associated with prolonged OS over chemo alone. Further prospective evaluations are warranted, and results from ongoing studies (e.g., NCT05058651) are awaited. Survival comparisons with univariable Cox proportional hazards regression. Treatment Group Median OS in months (95% CI) HR (95% CI) p Chemo (vs. No Chemo) 11.3 (9.5 – 14.2) vs. 3.4 (2.7 – 6.2) 0.53 (0.38 – 0.74) &lt;0.001 IO (vs. No IO) 13.0 (10.8 – 17.1) vs. 8.1 (6.2 – 10.3) 0.63 (0.46 – 0.87) 0.005 RT (vs. No RT) 10.8 (8.8 – 14.6) vs. 8.5 (6.4 – 11.7) 0.81 (0.60 – 1.09) 0.17 Primary Site Surgery (vs. No Surgery) 12.5 (8.7 – 28.5) vs. 9.0 (7.7 – 11.7) 0.76 (0.51 – 1.14) 0.18 ChemoIO (vs. Chemo alone) 14.1 (11.1 – 19.3) vs. 9.9 (8.3 – 13.7) 0.67 (0.47 – 0.96) 0.03

Mortality from pulmonary embolism among patients with lung and bronchus malignancies: National trends and forecasting from CDC WONDER (1999–2030).

e20116 Background: Pulmonary embolism (PE) represents a major cause of preventable mortality in patients with lung cancer due to hypercoagulability, treatment-related risk, and prolonged immobility. Contemporary national trends quantifying the population-level burden of PE-related mortality in this high-risk group remain limited. We evaluated long-term mortality patterns and projected future burden of PE among U.S. adults with lung and bronchus malignancies. Methods: A retrospective population-based analysis was done using the CDC WONDER Multiple Cause-of-Death database (1999–2020). Adults aged ≥55 years with PE (ICD-10 I26) listed as the underlying cause of death and malignant neoplasms of the lung and bronchus (ICD-10 C34) listed as contributing causes were included. Age-adjusted mortality rates (AAMRs) per 100,000 population were calculated and stratified by sex, race/ethnicity, geographic region, and urban–rural classification. Temporal trends were assessed using Joinpoint regression to estimate annual percent change (APC). Time-series forecasting was performed using autoregressive integrated moving average (ARIMA) models with Box-Cox transformation and stationarity testing (ADF/KPSS), with residual diagnostics for model validation. Results: From 1999-2020, 40,198 deaths were attributed to PE among patients with lung and bronchus malignancies. Mortality was higher among males compared with females (54.6% vs 45.3%, p &lt; 0.05) and among non-Hispanic White individuals (85.2%). National AAMRs increased significantly during 1999–2009 (APC = 3.65%; 95% CI: 2.85–4.71; p = 0.010) and again during 2015–2020 (APC = 2.18%; 95% CI: 0.65–5.59; p = 0.004). Forecast modeling projected continued escalation in PE mortality, with AAMR reaching 3.09 per 100,000 by 2030 (95% CI: 2.61–3.50), supported by favorable model performance (RMSE = 0.070; Ljung-Box p = 0.296). Regional heterogeneity was observed, with the Northeast demonstrating early acceleration (APC = 5.81%, 1999–2008; p &lt; 0.001) and the Midwest exhibiting sustained increases (APC = 6.98%, 2008–2020; p &lt; 0.001). Both large central metropolitan areas (APC = 4.46%, 1999–2008; p &lt; 0.001) and rural non-core counties (APC = 2.50%, 1999–2020; p &lt; 0.001) demonstrated persistent upward trends. Wisconsin (AAMR = 2.6) and Alaska (2.5) exceeded the national mean mortality rate (2.03 per 100,000). Conclusions: PE-related mortality among older adults with lung cancer has increased significantly over the past two decades and is projected to continue rising. Persistent demographic and geographic disparities highlight critical gaps in thromboprophylaxis utilization, early PE detection, and equitable access to cancer-associated thrombosis care. Findings underscore the urgent need for risk-stratified preventive strategies and targeted intervention frameworks in high-risk populations.

Attenuation of long-term survival benefit of biological therapy in metastatic colorectal cancer (mCRC): Real-world data validation and molecular characterization.

e15589 Background: Incorporation of biologic drugs to chemo has improved overall survival (OS) of patients with metastatic colorectal cancer (mCRC), as shown by randomized phase 3 clinical trials. Long term survival patterns remain uncharacterized. Analyses of SEER Medicare data showed addition of biologics confer early survival benefit followed by attenuation and lower OS, suggestive of late mortality. We sought to validate our findings in a large real-world cohort to explore molecular and treatment related factors associated with long term outcomes. Methods: A total of 11,048 mCRC tumors underwent molecular profiling at Caris Life Sciences (592 gene, NextSeq; WES, WTS NovaSeq). Genetic ancestry was inferred using ancestry informative variants derived from gnomAD reference. Immune cells were estimated using Quantiseq (WTS). Patients (Pts) were stratified as chemo alone (CT) or added biologics: anti VEGF (CTV) or anti EGFR (CTE). Significance was calculated using X 2 /Fisher’s exact or Mann-Whitney U (p adjusted for multiple comparisons). Clinical outcome was from insurance claims; OS was calculated from tissue collection and real-world OS (rwOS) from therapy start to last contact/death. Cox proportional hazards model was used for Hazard Ratio and log rank for p value, estimating average treatment effects, recognizing potential non proportional and time varying hazards. Results: 4442 patients (40%) received CT, 984 (9%) CTE and 5622 (51%) CTV. Median age in CT was 65 yrs vs CTV (62) and CTE (61), p &lt; 0.05. Left sided tumors received CTE at a higher rate (12.4%) vs right sided (3.6%), p &lt; 0.001. African ancestry pts were more frequently treated with CTV (p = 0.015). CTE had improved OS vs CT; (27.5 vs 23.8 m, HR 0.86; 95% CI 0.79-0.94, p = 0.001), and rwOS (44.4 vs 40.6 m, HR 0.91; 95% CI 0.83-1.0, p = 0.04). Notably, OS curves crossed at ~ 50 m, with observed early benefit attenuated and reversed beyond 50 m. CTV had improved OS vs CT; (25.7 vs 23.7 m, HR 0.94; 95% CI 0.90-0.99, p = 0.013) with OS curves crossing ~ 40 m. No rwOS improvement was observed, while OS curves crossed at 30 m (OS favored CTV before cross point but favored CT after cross point). Molecular comparison showed significant differences including enriched Tregs in CT vs CTE (FC 1.24, q = 0.03) and CTV (FC 1.12, q = 0.03). Conclusions: In a large database of mCRC; CTE and CTV pts had better OS vs CT with crossing of the survival curves matching late mortality observed in the SEER data. CTV was the predominant treatment of choice, with benefit of adding anti VEGF up to but not lasting past 30 ms. For pts with survival 30-50 m, long-term use of CT with a biologic presented a key point beyond which treatment may no longer be beneficial. This work provides a foundation for investigating risks vs. benefits of selective, monitored deployment of biologics with CT over time.

Can robotic surgery better preserve sensory function?: The significance and strategy of nerve preservation in axillary lymph node dissection.

582 Background: Conventional axillary lymph node dissection (C-ALND) carries a high risk of damaging the intercostobrachial nerve (ICBN), which remains a critical clinical issue affecting both psychological recovery and quality of life in breast cancer patients. While advancements in minimally invasive surgery provide improved visualization and precision, evidence supporting its clinical benefit in preserving sensory function is lacking. Methods: A retrospective analysis was conducted on breast cancer patients who underwent robotic-assisted axillary lymph node dissection (R-ALND, n = 53) and endoscopic-assisted axillary lymph node dissection (E-ALND, n = 58) between 2021 and 2025 (total n = 111). The anatomical integrity of the ICBN (T1-3) was systematically identified and recorded during surgery. Additionally, a blinded evaluator performed standardized bedside sensory testing (BST) on four key ICBN areas (medial upper arm, axilla, upper chest, and lateral chest wall) three months postoperatively, using a scoring system adapted from the ISCIPS consensus group for the current study. A multivariate linear regression model was applied for analysis. Results: Compared with E-ALND, R-ALND demonstrated a significantly higher ICBN-trunk preservation rate (77.4% vs. 63.1%, p = 0.033) and a greater proportion of areas with normal sensation at three months postoperatively (89.2% vs. 74.6%, p = 0.025). Multivariate linear regression revealed that T2-ICBN trunk preservation was the strongest independent predictor of sensory recovery at three months (β = 0.571, 95% CI 0.477–0.664, P &lt; 0.001), with an effect size (standardized β = 0.720) far exceeding other factors. Patients with T2-ICBN preservation achieved a 77.4% rate of completely normal sensation across all four key ICBN areas at three months. After adjusting for variables including T2-ICBN preservation, tumor stage, age and so on, robotic surgery showed a tendency to be an independent protective factor for sensory recovery at 3 months postoperatively compared with endoscopic surgery. (β = 0.110, p = 0.016). Conclusions: Through three-dimensional visualization and tremor-filtering technology, R-ALND significantly reduces the risk of iatrogenic ICBN injury and outperforms E-ALND in preserving short-term postoperative sensory function recovery. In the early phase with a limited learning curve, preserving the ICBN trunk (particularly the T2-ICBN) is recommended to retain most of the patient's sensory function.

Driving efficiency in oncology care: Tactical solutions for inpatient flow challenges.

e13566 Background: Hospital compression limits access to timely care with a negative impact on patients. These pressures hinder ability to consistently deliver right care, emphasizing the need for inpatient capacity and flow. Key barriers were variable discharge planning, inefficient workflows, and inconsistent rounding patterns. Our goal, as the largest comprehensive cancer hospital, was to develop core tactical interventions for timely discharges, keep all access points open, and transform inpatient flow utilizing a multipronged system-based approach. Methods: Our strategy to streamline discharge workflows combined 3 core tactics: process redesign, technology integration, and role-based accountability. Our guiding principle was an accurate estimated discharge date (EDD) validated during multidisciplinary rounds, signaling discharge readiness. Our process redesign was a cohort approach (hematology, surgery and solid tumor) to an oncology-specific discharge checklist integrated into the electronic medical record that aligned incomplete discharge tasks with upcoming EDD. Role based accountability empowered advanced practice providers as discharge delegates with early rounding and discharge order. Clinical nurses monitored readiness and escalated barriers. Technology supported real-time performance dashboards for missed opportunities with feedback loop. Three key performance indicators were chosen to measure success: 1) accurate EDD one day prior, 2) provider placed discharge orders before 9AM, 3) team completed discharges before noon. Results: Over 12 months following the baseline period, EDD accuracy one day prior improved sustainably across all service lines. Early discharge order by 9AM improved in hematology from 8% to 52%, surgery from 24% to 54%, and solid tumor from 8% to 44%. Our discharges before noon in hematology improved from 6% to 46%, surgery from 12% to 46%, and solid tumor from 10% to 44%. In our opinion, this remarkable change in our KPIs marks a milestone not previously reached by any organization. Gains are durable beyond 12 months, sustained by dashboard monitoring and system level governance. Conclusions: A multipronged tactical approach proved successful in a coordinated, data-driven inpatient flow design that was both effective and sustainable. Interventions should be conducted synchronously, not in isolation. These tactics are replicable and scalable to other institutions. In conclusion, standardized workflows, role-based accountability, and technology-based performance reduced variation and sustained improvements. Results over time. Population Hematology Surgery Solid tumor Period % 9AM Orders % DC by Noon % 9AM Orders % DC by Noon % 9AM Orders % DC by Noon 12 Month Baseline Period 8% 6% 24% 12% 8% 10% 12 Month Intervention Period 18% 14% 29% 18% 17% 16% 12 Month Post Intervention Period 52% 46% 54% 46% 44% 44% 4 Month Sustainment Period 63% 57% 64% 55% 52% 55%