Browse Articles

Treatment patterns and survival outcomes in colon cancer patients with autism spectrum disorder: A propensity score–matched retrospective multicenter cohort study.

e15729 Background: Patients with autism spectrum disorder (ASD) face documented barriers to cancer care, yet oncologic outcomes in this population remain understudied. We investigated treatment patterns and survival outcomes comparing patients with ASD versus those without ASD with colon cancer. Methods: Using the TriNetX network, we identified adults with colon cancer diagnosed from 2015 to 2025. Patients with ASD (n = 368) were propensity score–matched 1:1 to non-ASD controls on demographics, comorbidities, and tumor characteristics (stage, mutations, etc.). Primary outcomes included receipt of any therapy and NCCN-concordant systemic therapy, hospitalization resource utilization, adverse events, and all-cause mortality across multiple time windows (3 months, 6 months, 1 year, 3 years, and any time). A prespecified subgroup analysis examined stage IV patients (n = 94 per arm) and stage I–III patients (n = 274 per arm). Measures of association (RR) and time-to-event hazard ratios (HRs) were evaluated. Results: Cohorts were well-balanced (SMD < 0.1). In the full cohort (N = 368/368), ASD patients received significantly less systemic therapy (chemotherapy/biologics/immunotherapy) at 3 months (13.3% vs 19.3%; RR 0.69, p = 0.028; HR 0.65, p = 0.018), 3 years (28.5% vs 35.9%; RR 0.80, p = 0.033; HR 0.74, p = 0.023), and any time (30.7% vs 38.6%; RR 0.80, p = 0.025; HR 0.73, p = 0.014); receipt of any therapy also trended lower at 3 months (19.6% vs 25.0%; RR 0.78, p = 0.075; HR 0.74, p = 0.051). ASD was associated with fewer ED visits at 6 months (16.0% vs 21.2%; HR 0.71, p = 0.046) and lower sepsis at 3 months (2.7% vs 5.7%; RR 0.48, p = 0.043; HR 0.46, p = 0.039). Despite these treatment differences, there were no significant differences in palliative care, bowel complications, or all-cause mortality in the full cohort (any-time mortality: 22.6% vs 23.4%, p = 0.79). In prespecified subgroup analyses, stage I–III patients had comparable colectomy rates. In contrast, excess mortality was concentrated in stage IV ASD patients: mortality was higher at 1 year (28.7% vs 14.9%; RR 1.93, p = 0.022; HR 1.98, p = 0.034), 3 years (45.7% vs 27.7%; RR 1.65, p = 0.010; HR 1.58, p = 0.064), and any time (47.9% vs 33.0%; RR 1.45, p = 0.038; HR 1.41, p = 0.037), with a concomitant trend toward lower systemic therapy at 3 months (12.8% vs 22.3%; RR 0.57, p = 0.091). Conclusions: In a propensity-matched real-world cohort, adults with ASD and colon cancer were less likely to receive NCCN-concordant systemic therapy. While overall outcomes were similar in the full cohort, stage IV ASD patients experienced markedly higher mortality - highlighting ASD as a high-risk equity population and an urgent target for autism-informed pathways to ensure timely, guideline-based treatment.

A first-in-human (FiH) phase I study of GFS202A, a GDF15/IL-6 bispecific antibody, in advanced cancer patients (pts) with pre-cachexia or cachexia.

12055 Background: Cachexia is a common syndrome in cancer pts, associated with weight loss, impaired quality of life, and poor treatment outcomes. GDF15, a pleiotropic factor of TGF-β super family, has been verified to be correlated with cancer cachexia and anorexia. IL-6 contributes to cancer cachexia progression via both central and peripheral pathways. GFS202A, a first-in-class bispecific antibody dual-targeting GDF15 and IL-6, improves symptoms in preclinical cachexia models with lean body mass and adipose tissue weight increased. Here, we report the preliminary data from the ongoing FiH study of GFS202A. Methods: This is the FiH phase I study (NCT06898255) of GFS202A in advanced cancer pts with pre-cachexia or cachexia. Eligible pts are required to have weight loss in the six months prior to the first dose and decreased appetite, with or without anticancer therapy. Enrolled pts will receive GFS202A Q3W at a starting dose of 5 mg for a 12-week treatment period, followed by a 12-week follow-up period. An accelerated titration design, combined with the Bayesian Optimal Interval (BOIN) and backfilling design, is employed. The primary objectives are to assess safety and determine the recommended dose ranges for future clinical studies. The secondary and exploratory objectives include characterization of PK profiles, PD biomarkers, preliminary efficacy and exploratory biomarkers. Results: Data cutoff at Dec 26, 2025, 14 pts had received at least one dose of GFS202A ranging from 5 to 400 mg Q3W (5mg: N=1; 25mg: N=4; 100mg: N=4; 200mg: N=4; 400mg: N=1). The overall safety profile was favorable. No DLT was observed. Adverse events (AEs) of any cause were reported in all pts. Six pts experienced at least one treatment-related AEs (TRAEs). Only one pt had G3 hypertension assessed related to GFS202A, which alleviated during the treatment. The pt did not experience any other cardiovascular events and had no clinically meaningful changes in lipid profiles. All other TRAEs were G1. No ≥ 2pt experienced an identical TRAE. GFS202A exposure was approximately dose-proportional from 5-200 mg, and the mean half-life was 82–138 hours. GDF15 was not detectable immediately at the end of infusion from the 5mg cohort, and durable inhibition of GDF15 was observed from the 100mg dosing cohorts during 21-day dosing interval. CRP decreased significantly post-infusion, with suppressions sustained through Week 6 in 12 out of 13 pts in 5-200 mg Q3W groups. Weight gain was observed across different dose groups, with least-squares mean increases of 1.8, 1.5 and 2.5 kg at Week 6 in the 25, 100, and 200 mg Q3W groups, respectively. Conclusions: GFS202A demonstrated favorable safety and encouraging efficacy with systemic inflammation alleviated and weight gain. The data support that dual-targeting GDF15 and IL-6 is a feasible way to mitigate cancer cachexia clinically. Future studies are warranted. Clinical trial information: NCT06898255 .

Weekend admission and inpatient outcomes in hospitalized patients with acute promyelocytic leukemia.

e18546 Background: Acute promyelocytic leukemia [APL] has historically been associated with significant early mortality, but it is now one of the most curable leukemias, with long-term survival rates >90%. Early recognition and prompt treatment initiation are critical to prevent early death. Although early mortality remains a significant challenge, the impact of weekend admission on hospitalization outcomes of patients with APL is unclear. We evaluated the association between weekend versus weekday admission and inpatient outcomes among patients with APL. Methods: Retrospective cohort analyses were conducted using data from the National Inpatient Sample (NIS) collected between 2017 and 2021. Using ICD-10 CM codes, adult (≥18 years) inpatient hospitalizations with a diagnosis of APL were identified. Multivariable logistic and linear regression models were used to examine the association between day of admission and inpatient outcomes among hospitalized patients with APL. Results: Among 4,959 hospitalizations with a diagnosis of APL, 19% of admissions occurred on a weekend. Patients admitted on weekends were more likely to be male (49% vs 43%, p=0.027), non-Hispanic Black (19% vs 11%, p<0.01) and insured by Medicare (72% vs 59%, p<0.001). There was no significant difference in weekend admissions by age. After multivariable adjustment, patients admitted over the weekend had higher odds of in-hospital mortality compared to those admitted during a weekday (Adjusted odds ratio [AOR] 1.80, 95% CI, 1.20 – 2.40). Weekend admission was associated with length of stay (LOS) >5 days (AOR 2.60, 95% CI, 1.80 – 4.50) relative to 5 days or less. Additionally, patients admitted over the weekend had $3,579 (95% CI, $43.6 – $78,964.5) higher total hospital charges compared to those admitted during a weekday. Conclusions: Our findings demonstrate a weekend effect on hospitalization outcomes among patients with APL, with weekend admissions associated with higher in-hospital mortality, longer hospital stays, and increased hospital costs. Further investigation is needed to identify the underlying causes of these differences and implement strategies to attenuate them. Impact of weekend admission on inpatient outcomes among hospitalized patients with acute promyelocytic leukemia. Weekend Admission Unadjusted Adjusted Mortality OR (95% CI) No Reference Reference Yes 1.23 (0.56–2.60) 1.80 (1.20–2.40) Length of stay OR (95% CI) 5 days or less Reference Reference More than 5 days 2.30 (1.20-3.60) 2.60 (1.80-4.50) Total Charges ß (95% CI) 49.05 (-654.35-178643.76) 3578.8 (43.64-78964.54) AOR=Adjusted odds ratio; CI=confidence intervals. Model adjusted for age, gender, race/ethnicity, median household income national quartiles, hospital region, hospital location, Charlson comorbidity index, insurance type, admission type and length of stay. Boldface indicates statistical significance.

PROGRESS: Implementation of a novel hybrid decentralized clinical trial to support precision oncology uptake at academic and community oncology practices in North Carolina.

e23150 Background: Rates of comprehensive molecular profiling (CMP), use of evidence-based genome-informed therapy (GIT), and clinical trial participation remain low in real-world data sets. We designed an interventional clinical trial (PROGRESS) using decentralized methodology to evaluate the impact of a centralized precision oncology navigator and pharmacist-led expert CMP review on GIT order rates and cancer outcomes in an academic practice and a network of community practice locations (CPLs) in North Carolina. We leveraged academic medical center (AMC) experience in hybrid decentralized clinical trials (hDCTs) to design a more accessible, community-friendly trial. Herein we describe the operational challenges and solutions. Methods: The PROGRESS study was operationalized as an hDCT in which local providers at CPLs were not engaged in research per the regulatory definition as they performed only assessments consistent with standard of care. Regulatory oversight was maintained by the AMC. A pre-launch meeting assessed CPL interest and feedback on study integration into clinical workflow. Launch meetings were then led by the AMC including PIs, study coordinator (SC), and regulatory staff. The AMC-based SC identified potential subjects via the shared electronic health record (EHR) and obtained electronic consent remotely; all follow up assessments were done via medical record abstraction. Descriptive qualitative and process-level metrics were used to identify challenges and guide workflow adaptations. Results: In addition to the AMC, three of four CPLs approached were onboarded with n = 40 subjects enrolled over 6 months. One CPL declined due to lack of provider time. Key onboarding facilitators included: a local study champion, integrating study requirements into clinical workflows, and addressing remote consent concerns. Post-launch enrollment barriers included: lack of multidisciplinary collaborators, participant language barriers, poor health literacy, research-fatigue (primarily at the AMC), and inefficient pre-screening tools. Adaptations to these barriers included a monthly enrollment newsletter, a protocol amendment to accommodate CMP order workflows, removal of burdensome surveys, translation of consent forms, and an efficient pre-screening report leveraging EHR-integrated CMP orders. Following these changes, enrollment increased by 50% in the second versus the first quarter. Conclusions: The implementation of PROGRESS using hDCT methodology generated practical insights to strengthen partnerships with CPLs, informing scalable approaches to expand access to precision oncology support and local clinical research opportunities. Clinical trial information: NCT06896162 .

Fitness classification and induction intensity selection in AML practice: A contemporary real-world cohort study.

e18552 Background: Standard acute myeloid leukemia (AML) induction has historically relied on intensive cytarabine/anthracycline chemotherapy (IC). Many patients (pts), especially older adults or those with comorbidities or poor performance status are often not eligible for IC and face higher risk of treatment related toxicity and early mortality. Induction regimens have become increasingly individualized from improved supportive care and effective low intensity (LI) regimens, with recent trials supporting utility for LI regimens in broader patient populations. This highlights a need to explore current real-world treatment selection patterns. We retrospectively analyzed whether a predefined fitness stratification aligned with induction intensity selection and described early and 1-year mortality outcomes. Methods: We conducted a single-center retrospective cohort study of 70 treatment-naïve adults (age ≥ 18 years) with AML at a tertiary academic cancer center from 1/1/2019 to 12/31/2025. Fitness was classified using prespecified criteria: age ≥ 75 years, ECOG performance status ≥ 2, history of treated congestive heart failure or stable angina, chronic lung disease requiring baseline oxygen. IC included 7+3 or liposomal daunorubicin/cytarabine (CPX351). LI regimens included HMA ± venetoclax, or low dose cytarabine ± venetoclax. The primary endpoint was association between fitness and induction intensity selection. Secondary endpoints were 30- and 60-day mortality and 1-year overall survival (OS). Odds ratio (OR) and OS were reported descriptively. Results: Seventy total pts were included (median age 58 years); 36 pts were classified as fit and 34 as unfit. Induction intensity differed significantly by fitness. IC was administered in 78% (28/36) of fit and in 38% (13/34) of unfit pts (OR 5.65, 95% CI, 1.98–16.11, p = 0.0014), while LI regimen was selected in 22% (8/36) of fit and 62% (21/34) of unfit pts. Early mortality was low overall and did not differ by fitness at 30 days (2.8% [1/36] fit vs 0% [0/34] unfit, p = 1.0) or 60 days (2.8% [1/36] fit vs 8.8% [3/34] unfit, p = 0.35). Estimated 1-year OS was 82% (95% CI, 64%–92%) in fit and 60% (95% CI, 42%–75%) in unfit pts (log-rank p = 0.069). Conclusions: In a contemporary AML induction cohort, prespecified fitness classification strongly aligned with induction intensity selection, supporting fitness informed pathways in real-world decision making. Early mortality was low across fitness groups, potentially reflecting improved supportive care and risk adapted induction selection. 1-year OS numerically favored fit pts but did not reach significance, likely due to few events. These real-world data demonstrate current practice patterns and provide benchmarks for counseling and supportive care planning. Broader, large multicenter cohort studies with time-to-event analyses are needed to improve outcomes for adult AML patients.

Cancer patients' comfort with artificial intelligence in oncology care and first-round "AI chemo-teach" quality improvement initiatives.

e13652 Background: Artificial intelligence (AI)-powered tools are rapidly advancing oncology care, but patients’ acceptance of AI is crucial for patient-centered implementation. This study aims to analyze key factors that predict oncology patients’ comfort with various AI applications, including the effect of clinician oversight, and to apply these findings to an “AI chemo-teach" quality improvement (QI) initiative using AI-generated chemotherapy educational materials in adult patients with new gastrointestinal malignancies. Methods: This ongoing QI project initially used convenience surveying among adult cancer patients being seen at UCLA Health oncology clinics from October to December 2025. Data was collected regarding demographics, health literacy scores via BRIEF (a validated screening tool), familiarity with and attitudes towards AI, and comfort with AI-assistance in five theoretical use cases: administrative tasks, treatment decision support, and patient messaging as well as AI-generation of chemotherapy education materials (with and without clinician review of outputs). Comfort was assessed on a 10-point Likert scale. Statistical analyses included Kruskal-Wallis tests, Mann-Whitney U tests, and linear mixed-effects modeling. Results: Among 47 patients surveyed, most were male (51.1%), White (46.8%) or Asian (21.3%), and had a bachelor’s degree or higher (55.3%). Median age was 62 years, and 7 patients (14.9%) screened with low health literacy. Patient comfort was highest for AI-assistance with administrative tasks (6.53 ± 3.25), followed by treatment decision support (5.85 ± 3.22) and patient messaging (5.11 ± 3.07) (p = 0.067). Adding clinician review to AI generation of chemotherapy education materials significantly improved patient comfort by 3.19 points (from 4.19 ± 3.19 without clinician review to 7.38 ± 3.39 with clinician review; p < 0.001). Higher familiarity with AI (β = 0.40, p < 0.001) and positive attitude towards AI (β = 0.58, p < 0.001) were the strongest predictors of high comfort, and low health literacy significantly predicted poor comfort (p < 0.05). Age, sex, race/ethnicity, and education level were not significant predictors. Conclusions: Oncology patients’ comfort with AI in clinical care trended higher for lower-risk administrative tasks. Clinician review of AI-generated chemotherapy educational materials significantly improves comfort scores, and a similarly positive effect may be seen in other use cases. Familiarity and positivity with AI were dominant drivers of comfort, supporting a role for patient education about AI itself. Ultimately, AI integration in oncology care for low-risk clinical tasks for which providers remain “in-the-loop" are likely to be accepted by patients. Findings are informing the "AI chemo-teach" QI initiative for which the first plan-do-study-act cycle launched in December 2025.

Real-world outcomes of lisocabtagene maraleucel (liso-cel) in patients (pt) with relapsed or refractory (R/R) chronic lymphocytic leukemia (CLL): First results from CIBMTR.

7024 Background: Liso-cel is an autologous, CD19-directed CAR T cell product with established efficacy and safety across B-cell malignancies. In TRANSCEND CLL 004 (NCT03331198), liso-cel demonstrated promising efficacy and manageable safety in pts with R/R CLL, including those previously exposed to Bruton tyrosine kinase inhibitor (BTKi) and B-cell lymphoma 2 inhibitor (BCL2i). Here, we report real-world effectiveness and safety of liso-cel in pts with R/R CLL. Methods: This observational study evaluated US pts with R/R CLL who received commercial liso-cel (05/2024–09/2025) and had ≥ 1 postinfusion assessment reported to CIBMTR. Effectiveness outcomes were ORR, CR rate, duration of response (DOR), PFS and OS; safety outcomes included AEs of special interest, nonrelapse mortality (NRM) and death. Results are descriptive. Results: Of 45 eligible pts, 64% were male; median age was 62 y (range, 44–83; ≥ 65 y, 44%). In pts with available data, all had ECOG PS 0–1 (43/43) and 57% (21/37) had ≥ 1 comorbidity, most commonly (> 10%) cardiac (24%) or pulmonary (16%). Elevated LDH was observed in 35% (15/43) and unmutated IGHV in 91%. Pts had a median of 6 (range, 1–10) prior therapies and 71% received liso-cel as fifth-line or later treatment (tx). Most pts (84%) were previously exposed to covalent BTKi and BCL2i, and nearly half (47%) had prior pirtobrutinib tx. Bridging therapy was used in 48% (21/44) of pts. At a median (95% CI) follow-up of 6.0 mo (5.4–7.1), ORR was 84% (70–93), and CR rate was 55% (39–70); 81% (13/16) of pts in CR with available data were MRD negative. Median (95% CI) DOR, PFS, and OS were not reached. Among responders (n = 36), 6-mo DOR was 89% (70–96.5); 6-mo PFS and OS were 77% (60–87) and 87% (72–95), respectively. Among pts with any-grade (gr) cytokine release syndrome (80%; gr ≥ 3, 4%) or immune effector cell–associated neurotoxicity syndrome (36%; gr ≥ 3, 13%), no gr 5 events occurred. Clinically significant infections were reported in 40% of pts, and at 30 d after infusion, 18% had persistent gr 4 thrombocytopenia and/or neutropenia. Rates were low of HLH/MAS (7%), second primary malignancies (2%; 1 case of basal cell carcinoma 34 d postinfusion), tumor lysis syndrome (7%), and gr 3/4 organ toxicity (9%). Of the 5 deaths, 2 each were due to relapse/progression or infection, and 1 pt had multiple causes reported; 6-mo NRM was 5% (95% CI, 0.8–14). Among the 49% of pts intended for outpatient infusion, 55% (12/22) were hospitalized postinfusion, with a median (range) time to hospitalization of 5.5 d (1–20) and a stay of 5.5 d (1–26). Conclusions: In routine clinical practice, liso-cel demonstrated response rates higher than in clinical trials, with a consistent safety profile. These real-world data support liso-cel as an effective option for pts with R/R CLL, including heavily pretreated or high-risk pts, and the potential for outpatient administration.

A radiomics-based model for the prediction of WHO/ISUP in clear cell renal cell carcinoma using contrast-enhanced CT indicating response to TKI therapy.

e16522 Background: WHO/ISUP grade is a significant risk factor for the prognosis of patients with clear cell renal cell carcinoma (ccRCC) and effects the response to tyrosine kinase inhibitors (TKIs) for advanced-stage patients. The purpose of this study was to develop a fully-automated model that can predict WHO/ISUP grade based on three-phase CT images, and may implicate the TKIs response. Methods: A total of 373 patients with ccRCC from three medical centers were retrospectively included in the study, with 261 in the training set and 112 in the testing set. CT images of 166 TCGA-KIRC cohort were used to explore the different expressed genes and enriched biological pathways related to the radiomics model. All CT phases were aligned to the venous phase and used to evaluate a presenting deep learning model (Kidney and kidney tumor segmentation 2023, KiTS23) for kidney tumor segmentation. Radiomics features were extracted from the tumor of original CT phases. Linear discriminant analysis was used to develop three models based on transcriptomic features, radiomics features, and both features combined. Models were evaluated by area under curve, sensitivity, and specificity. Results: The average dice coefficients of kidney tumor segmentation were 0.87 in the training set and 0.83 in the testing set. For WHO/ISUP grade prediction, in the testing set, the model based on radiomics (AUC = 0.801) outperformed the model based on transcriptomic features (AUC = 0.783). The hybrid model based on transcriptome and radiomics features achieved the best performance in both the training set (AUC = 0.911) and testing set (AUC = 0.859). Moreover, the hybrid model also provided the highest accuracy (0.930), sensitivity (0.714), specificity (0.972), positive predictive value (0.833), and negative predictive value (0.946). The TCGA-KIRC cohort were divided into high- and low-risk group based on the radiomics model prediction, and the differential expressed gene in high-risk group were significantly enriched on the pathway of EGFR tyrosine kinase inhibitor resistance. Conclusions: The fully-automated model based on transcriptome and radiomics features can accurately predict the WHO/ISUP grade of patients with ccRCC, and implicate the TKIs response for advanced-stage patients.

Phlebotomy, adherence, and hematocrit control in patients with polycythemia vera in the United States: Real-world analysis.

6568 Background: Polycythemia Vera (PV) is characterized by erythrocytosis. A key treatment goal in PV, for which phlebotomy (PHL) is recommended, is achieving and maintaining hematocrit (HCT) levels < 45% to reduce thrombotic risk. However, real-world evidence suggests PHL alone may not achieve consistent and durable HCT control. Repeated PHL may also increase treatment burden and present logistical challenges. Methods: Data were collected using the Adelphi PV Disease Specific Programme, a United States (US) cross-sectional survey (August 2025-January 2026) with retrospective chart review. Physicians reported demographics, clinical characteristics, and treatment history for the next five patients (pts) with PV that met the inclusion criteria. Interim analyses are presented descriptively. Of those receiving ≥ 1 PHL in the 12 months prior to survey, pts receiving frequent PHL were defined as ≥ 3 in 6 months or ≥ 5 in 12 months; all other pts were classified as receiving infrequent PHL. Uncontrolled HCT was defined as any HCT ≥ 45% within the prior 12 months. Results: Twenty-seven physicians provided data for 143 pts; 86 (60%) received ≥ 1 PHL in the prior 12 months; median (interquartile range (IQR)) age was 65.0 (52.0-70.0) years, 77% were male, median (IQR) time since PV diagnosis was 11.0 (4.0-33.5) months, and 62% were high risk (defined as ≥ 60 years old or history of thrombosis). Half experienced ≥ 1 barrier to receiving PHL, including care coordination challenges (22%), lack of resources (12%), travel time burden (12%) and vein access challenges (10%). In prior 12 months, physicians reported that 38% of pts were not fully adherent to PHL (attended PHL appointments as recommended <90% of the time). At the most recent consultation, 51% of pts were receiving cytoreductive therapy in addition to PHL. Frequent PHL occurred in 24% of pts and infrequent in 76% of pts. Overall, 90% (75/83) had uncontrolled HCT: 89% for frequent (17/19) vs 91% for infrequent PHL (58/64). Since their PV diagnosis, 13% (11/86) of pts had experienced a thromboembolism (TE), and of these 11 pts, 45% (5/11) had experienced a TE within the 12 months prior to survey. For all pts who had ever experienced a TE, the median (IQR) time since their most recent TE was 20.0 (2.5-32.0) months. Conclusions: Within this study, pts receiving PHL for PV commonly experienced access and logistical barriers, and incomplete PHL adherence. Nearly all pts had evidence of uncontrolled HCT in the 12 months prior to survey, including both frequent and infrequent PHL groups, and TEs were observed despite PHL treatment. These findings highlight the need for more effective and sustainable approaches to prevent consequences, such as TEs, commonly associated with uncontrolled HCT, while also reducing pt burden.

Clinical utility of NGS in patient stratification for first-line chemotherapy in locally advanced/advanced GBC: A pilot study demonstrating the hypothesis.

e16244 Background: Gallbladder cancer is highly prevalent in Asian countries and is associated with poor survival outcomes. Platinum-based chemotherapy remains the standard of care in advanced disease; however, treatment responses are heterogeneous, and predictive molecular determinants of chemotherapy sensitivity are poorly defined. Methods: This retrospective study included 13 patients with locally advanced gallbladder cancer treated with first-line cisplatin and gemcitabine. Patients were stratified as responders or non-responders based on radiological response assessed by RECIST criteria at three months. Somatic mutation profiling was performed using the TarGT IndieGene panel covering 1,212 genes. Mutations were mapped to established cancer hallmark pathways to evaluate biological enrichment patterns while accounting for tumor heterogeneity. Results: Responders showed a relatively restricted hallmark enrichment pattern, predominantly involving sustained proliferative signaling, genomic instability, tumor invasion and metastasis, and evasion of growth suppressors, with most hallmarks exhibiting low-to-moderate enrichment scores (1+ to 3+). Hallmarks related to immune evasion, tumor-associated inflammation, and reprogramming of energy metabolism were largely absent in responders. The cumulative hallmark scores in this group ranged from 8 to 14 ( < 15), indicating limited oncogenic complexity.In contrast, non-responders demonstrated pronounced and consistent enrichment across multiple hallmark categories. High-level enrichment (3+) was frequently observed for sustained proliferative signaling, genomic instability, evasion of growth suppressors, resistance to cell death, sustained angiogenesis, and replicative immortality. Additionally, immune evasion, tumor-associated inflammation, and metabolic reprogramming—absent in responders—were selectively enriched in non-responders. These patterns resulted in higher cumulative hallmark scores ranging from 17 to 22 ( > 15). Non-responders also exhibited significantly shorter median PFS compared to responders (7 vs. 15 months; p = 0.03). Conclusions: Non-responding gallbladder cancers are characterized by extensive activation of multiple cancer hallmarks, reflecting greater biological complexity, heterogeneity, and adaptive capacity that may underlie resistance to platinum-based chemotherapy. This is a high impact finding. Validation in larger cohorts is warranted to adopt in routine clinics. The cumulative hallmark scores for each of the patients based on the p-value of the enrichment. id status hallmarks (10) Score 2 responder 7 14 7 responder 7 13 9 responder 7 8 10 responder 7 9 12 responder 7 13 1 nonresponder 7 17 3 nonresponder 8 22 4 nonresponder 8 19 5 nonresponder 9 21 6 nonresponder 10 22 8 nonresponder 9 21 11 nonresponder 7 20 13 nonresponder 8 20

Erythropoietin receptor (EPOR) expression to identify a biologically aggressive tumor phenotype in clear cell renal cell carcinoma.

e16521 Background: Clear cell renal cell carcinoma (ccRCC) demonstrates substantial biologic heterogeneity and variable clinical outcomes. Erythropoietin receptor (EPOR) signaling has been implicated in tumor biology, yet its relationship to canonical therapy-resistance and immune-associated programs in ccRCC remains incompletely characterized. Methods: RNA sequencing and clinical data from The Cancer Genome Atlas kidney renal clear cell carcinoma cohort (TCGA-KIRC; n = 533) were analyzed. EPOR expression was z-scaled. Pre-specified gene signatures representing angiogenesis, mTOR/PI3K/AKT signaling, immune activation, epithelial–mesenchymal transition (EMT), metabolic reprogramming, and DNA damage repair were quantified using mean gene-level z-scores. Associations between EPOR and pathway signatures were evaluated using Spearman correlation with false discovery rate (FDR) correction. Overall survival (OS) was assessed using multivariable Cox regression adjusted for age, stage, and grade. Results: Higher EPOR expression was independently associated with worse OS (hazard ratio [HR] 1.29, 95% confidence interval [CI] 1.14–1.47, p < 0.001). EPOR expression demonstrated strong positive correlations with mTOR/PI3K/AKT signaling (ρ = 0.31, FDR < 0.001) and angiogenesis (ρ = 0.26, FDR < 0.001), and was also significantly associated with cytolytic immune activity, EMT, MYC targets, DNA damage repair, and oxidative phosphorylation programs (all FDR < 0.05). EPOR-high tumors showed significant enrichment of angiogenic, mTOR, and cytolytic immune signatures compared with EPOR-low tumors (all FDR < 0.001). Importantly, inclusion of these pathway-level signatures in multivariable survival models did not attenuate the prognostic impact of EPOR expression. Conclusions: EPOR expression identifies a biologically aggressive ccRCC tumor phenotype characterized by coordinated activation of angiogenic, mTOR-driven, immune, and invasive programs, while retaining independent prognostic significance. These findings support EPOR as an integrative biomarker of high-risk disease biology in ccRCC and provide a rationale for further investigation of EPOR-associated pathways in therapeutic resistance.

Neoadjuvant PD-1/PD-L1 inhibitors combined with chemotherapy in early-stage triple-negative breast cancer: A systematic review and meta-analysis.

e12639 Background: Neoadjuvant immune checkpoint inhibition targeting the PD-1/PD-L1 pathway has become an integral component of treatment for early-stage triple-negative breast cancer (TNBC). While improvements in pathological complete response (pCR) have been consistently reported, the durability of benefit and the relevance of emerging survival data warrant integrated evaluation. We conducted a systematic review and meta-analysis to synthesize contemporary evidence, incorporating recent efficacy and survival updates. Methods: PubMed/MEDLINE and Cochrane CENTRAL were systematically searched to identify randomized and non-randomized studies evaluating neoadjuvant PD-1/PD-L1 inhibitors combined with chemotherapy in early-stage TNBC. Both published manuscripts and major conference presentations were reviewed to capture the most recent trial updates. Randomized controlled trials comparing chemo-immunotherapy with chemotherapy alone were eligible for quantitative synthesis of pCR, while event-free survival (EFS), overall survival (OS), and safety outcomes were summarized qualitatively using updated data. Study selection and synthesis followed PRISMA guidance. Results: Across randomized trials, neoadjuvant PD-1/PD-L1 inhibitors combined with chemotherapy consistently increased pCR rates compared with chemotherapy alone. Meta-analytic synthesis demonstrated a consistent and clinically meaningful increase in the likelihood of achieving pCR with the addition of immunotherapy. Updated analyses from landmark trials showed durable improvements in EFS, with emerging signals of OS benefit in high-risk early-stage TNBC. Real-world observational studies qualitatively corroborated trial findings, supporting the effectiveness of chemo-immunotherapy beyond controlled trial settings. Immune-related adverse events were more frequent with immunotherapy-containing regimens but were generally manageable and consistent with established safety profiles. Conclusions: Neoadjuvant PD-1/PD-L1 inhibitors combined with chemotherapy improve pathological response and are associated with durable survival signals in early-stage TNBC. This updated synthesis integrating contemporary trial updates reinforces neoadjuvant chemo-immunotherapy as a key component of modern treatment strategies and underscores the importance of continued follow-up to define long-term outcomes.

Association between frailty status and osteomyelitis: A nested case-control study

Background Although frailty has been recognized as a predictor of various adverse outcomes in older adults, the association between physical pre-frailty or frailty and the risk of osteomyelitis remains unclear. Methods In this nested case-control study, data from 466,918 eligible participants recruited into the UK Biobank between 2006 and 2010 were used. Incident osteomyelitis cases were identified through linked electronic health records up to 19 December 2022 and matched to controls at a 1:5 ratio by age, sex, and assessment center from the baseline population meeting the eligibility criteria. Frailty status was assessed at baseline using a five-item phenotype adapted from the Fried criteria and categorized as non-frail, pre-frail, or frail. Conditional logistic regression was used to evaluate the association between frailty status and osteomyelitis after adjustment for socioeconomic factors, lifestyle behaviors, medication use, and clinical risk factors. Results Compared with non-frail individuals, participants with physical pre-frailty had a significantly higher odds of osteomyelitis (OR = 1.38; 95% CI: 1.16–1.64), with the odds further increasing among those with physical frailty (OR = 2.79; 95% CI: 2.05–3.81), demonstrating a clear dose–response relationship (P-trend <0.001). These associations between physical frailty status and osteomyelitis remained robust across subgroups defined by potential risk factors. Conclusions Physical pre-frailty and frailty were associated with higher odds of osteomyelitis, with the odds increasing across frailty categories.

Heat flux-dependent partial discharge behavior in a two-phase fluorinated liquid: Role of the microstructured heating surface

Fluorinated liquids are essential dielectric coolants used in immersion cooling and near-junction thermal management of high-voltage, high-power-density semiconductors. The impact of thermally induced bubbles on the partial discharge behavior of fluorinated liquids has been widely noted; however, the differences in their partial discharge characteristics with varying heat flux on smooth and microstructured surfaces remain unclear. Here, partial discharge experiments on a two-phase fluorinated liquid are conducted, and the relationship between boiling behavior and partial discharge characteristics is revealed. On the smooth surface, intensified boiling leads to a concurrent rise in discharge frequency and charge magnitude, accompanied by a continuous reduction in the partial discharge inception voltage. In contrast, the microstructured surface regulates the bubble dynamics, mitigating the deterioration of insulating performance as the heat flux increases. Moreover, the concentrated electric field at the sharp edges of the microstructures triggers partial discharges that fragment newly formed bubbles, thereby inhibiting the occurrence of severe discharges. This study lays the foundation for the synergistic improvement of thermal and electrical performance of immersion cooling or near-junction cooling based on fluorinated liquids for future high-voltage, high-power-density chips.

Carbohydrate polymers in oncology: From molecular interactions to therapeutic innovations

Exploring the conformational landscape of adenylate kinase and beyond with protein folding models

P4HA3 is dispensable for prolyl 4-hydroxylation of type I collagen during mouse development

Influence of menopausal status on the efficacy of adjuvant CDK4/6 inhibitors in hormone receptor–positive early breast cancer: A pooled meta-analysis of phase III clinical trials.

e12506 Background: Adjuvant cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) improve outcomes in hormone receptor–positive (HR+), HER2-negative early breast cancer, though trial results have been heterogeneous. MONARCH-E and NATALEE demonstrated significant invasive disease-free survival (iDFS) benefit, while PALLAS and PENELOPE-B did not. Menopausal status and endocrine therapy backbone represent potential sources of variability; however, there is a paucity of data exploring the influence of menopausal status on CDK4/6i efficacy. Methods: Phase III randomized trials evaluating adjuvant CDK4/6i in HR+/HER2− early breast cancer (PALLAS, PENELOPE-B, MONARCH-E, NATALEE) were identified. Trial-level hazard ratios (HRs) and 95% confidence intervals (CIs) for iDFS and overall survival (OS) were extracted separately for subgroups of pre/perimenopausal and postmenopausal participants. HRs were pooled using inverse variance and random-effects modeling (DerSimonian–Laird). Heterogeneity was assessed using I², and Cochran Q statistics. Subgroups were compared using the Deeks method. A sensitivity analysis was performed limiting the analysis to approved CDK4/6i (Ribociclib and Abemaciclib). Results: Among 4 trials comprising 17,749 participants, 45% were pre/perimenopausal and 55% were postmenopausal. OS analyses excluded NATALEE as OS data were not reported based on menopause status. Results of the analysis are shown in the Table. Pooled iDFS HR was 0.80 (I²=64%) in pre/perimenopausal patients and 0.79 (I²=0%) in postmenopausal patients. Pooled OS HRs were 1.02 (I²=80%) in pre/perimenopausal and 0.89 (I²=0%) in postmenopausal patients. These differences were not statistically significant for either iDFS or OS (subgroup difference p = 0.87 and 0.57). Compared to post-menopausal participants, those who were premenopausal had 2.9% fewer absolute IDFS events at 5-6 years. Conclusions: Randomized trials do not suggest any influence of menopausal status on the efficacy of adjuvant CDK4/6 inhibition. Higher heterogeneity among pre/perimenopausal patients suggests variability related to endocrine therapy backbone and patient risk, warranting further investigation. Meta-analysis of iDFS and OS with adjuvant CDK4/6i by menopausal status. Pre/Peri-menopausal Post-menopausal Subgroup difference p n HR (95% CI) Absolute difference at 5-6 years HR (95% CI) Absolute difference at 5-6 years Full cohort IDFS 4 0.80(0.67-0.97) 3.3%(n=3)^ 0.79(0.72-0.86) 4.2%(n=3)^ 0.87 OS 3 1.02(0.67-1.54) 0.2%(n=2)* 0.89(0.78-1.03) 1.2%(n=2)* 0.57 Sensitivity analysis limited to approved CDK4/6i IDFS 2 0.69(0.60-0.79) 4.9% 0.73(0.68-0.85) 4.9% n/a OS 1 0.72(0.55-0.94) 2.5% 0.91(0.75-1.09) 1.6% n/a ^PENELOPE-B excluded due to absence of IDFS subgroup data. *NATALEE/PENELOPE-B excluded due to absence of OS subgroup data.

From multi-omics to clinical practice: BC-BioMIXER to predict PCR and guide individualized neoadjuvant chemotherapy for breast cancer.

e12551 Background: Early and accurate prediction of pathological complete response (pCR) is crucial for personalizing neoadjuvant chemotherapy (NACT) in invasive breast cancer, yet many high-performing models depend on costly multi-modal data not routinely available. Purpose: To develop and validate BC-BioMIXER, a biologically informed model that transfers multi-omics–derived knowledge to routine clinical data for pre-treatment pCR prediction. Methods: BC-BioMIXER was developed in a multi-modality cohort of 648 patients with invasive breast cancer (T2–4, any N, M0) with transcriptomic, proteomic, MRI, and clinical data. External validation was performed in three independent cohorts (total N=830): one multi-modality cohort, one clinical trial cohort, and one contemporary real-world cohort. All patients received NACT followed by surgery. The framework uses a teacher–student paradigm: a multi-omics teacher learns biologically integrated representations, which are transferred to a student model trained only on routine clinical data. Performance was benchmarked against a multi-modality reference model and assessed across cohorts, receptor-defined subgroups (HER2, HR), and treatment groups (NACT ± immune checkpoint inhibitors [ICI]). Prognostic value was evaluated using distant recurrence-free survival (DRFS), and clinical utility for immunotherapy was explored by comparing DRFS between NACT+ICI and NACT-alone within model-predicted pCR/non-pCR strata. Results: BC-BioMIXER achieved pCR prediction comparable to the multi-modality benchmark (AUC 0.82 vs 0.85; p=0.271) and showed consistent discrimination across validation cohorts (AUC 0.82, 0.81, 0.80; all p<0.001). Model-predicted pCR was associated with improved 3-year DRFS (HR=0.36; 95% CI, 0.20–0.67; p<0.001). In patients treated with NACT+ICI, BC-BioMIXER was numerically superior to PD-L1 alone for pCR prediction (AUC 0.84 vs 0.72; p=0.08). Notably, within the predicted non-pCR subgroup, NACT+ICI was associated with inferior DRFS versus NACT alone (HR=2.70; p=0.032), while no significant difference was observed in the predicted pCR subgroup. Conclusions: BC-BioMIXER transfers multi-omics biological knowledge to a routine-data model for robust pCR prediction before NACT. Its consistent external validation and ability to stratify outcomes under NACT±ICI support scalable, accessible precision oncology.

A novel prognostic index for survival prediction in ovarian cancer: Beyond FIGO staging.

5577 Background: Prognostication in ovarian cancer is traditionally based on FIGO stage, although survival is influenced by additional patient- and treatment-related factors. We developed a new scoring system integrating age, comorbidities, and disease burden to improve survival prediction. Methods: Patients with ovarian malignancies treated with curative intent between June 2016 and November 2024 were identified from a prospectively maintained database. Cox proportional hazards regression was performed for 4-year overall survival (OS). Regression coefficients of significant variables were used to construct a new index. Survival outcomes were assessed using Kaplan–Meier analysis. Predictive performance of new index was compared with FIGO staging using receiver operating characteristic (ROC) analysis. Results: Of 221 consecutive patients, 167 treated with curative intent were analyzed. Median age of the cohort was 52 years (IQR 22–78), and FIGO stage IIIC was most common (45.7%). Optimal cytoreduction was achieved in 80.8%, and 71.3% received adjuvant therapy. At a median follow-up of 38.3 months, recurrence and mortality rates were 32.9% and 26.2%, respectively. The new score incorporated age, total surgical PCI, hypothyroidism, diabetes mellitus, and hypertension with regression-derived coefficients of 0.044, 0.114, 0.329, 0.146, and 0.313 respectively. For each patient, a prognostic index (PI) was calculated as the weighted sum of regression coefficients corresponding to that patient’s covariate profile, using formula, where PI (Prognostic index)= ∑ (β 1 x age) + (β 2 x PCI) + (β 3 x C 1 ) + (β 4 x C 2 )+....(β k x C k ), C 1... C k = individual comorbidities (eg., diabetes, hypertension, etc), if present = 1, absent = 0). Based on the prognostic index, patients were classified into low- (33.75%), intermediate- (31.3%), and high-risk (31.3%) groups. Compared with the low-risk group, OS was worse in the intermediate-risk group (HR 2.37; 95% CI, 0.95–5.91; p = 0.06) and significantly inferior in the high-risk group (HR 7.89; 95% CI, 3.37–18.49; p < 0.001). Disease free survival (DFS) also showed a similar trend. The prognostic index demonstrated superior discrimination compared with FIGO staging (C-index 0.73 vs 0.68; AUC 0.71 vs 0.65) and remained robust on bootstrap internal validation. A prespecified subgroup analysis restricted to FIGO stage III patients was performed and demonstrated superior discrimination for overall survival compared with FIGO staging alone (AUC 0.71 vs 0.61), highlighting prognostic heterogeneity beyond anatomic staging. Conclusions: The prognostic index demonstrated improved discrimination compared with FIGO staging by integrating patient age, comorbidities, and disease burden. This practical, clinically applicable index may facilitate improved risk stratification beyond anatomic staging and may support individualized management strategies in ovarian cancer.