Browse Articles
Discover research articles across all indexed journals
Development of an AI-based transcriptional surrogate for epigenetic aging in breast cancer.
e12575 Background: In our previous omics studies of breast cancer in Mexican patients, we observed an unexpected enrichment of mutational signature 1, a clock-like process positively correlated with age in both normal and malignant tissues, particularly in hormone receptor–positive tumors. This finding was independently replicated in a cohort of Mexican women residing in the United States. Given that breast cancer in Mexico is typically diagnosed at least ten years earlier than in other populations, we hypothesized that hormone receptor–positive tumors in Mexican women may exhibit features of accelerated biological aging relative to patients’ chronological age. Methods: To address this, we developed a machine learning–based transcriptional surrogate of Horvath’s epigenetic clock using gradient boosting (XGBoost), using large-scale public transcriptomic and DNA methylation datasets (N = 1,514 samples). The model was trained to predict DNA methylation age directly from gene expression profiles and demonstrated high predictive performance. Feature selection was based on both algorithmic importance and biological relevance, allowing prioritization of genes linked to aging biology and facilitating functional interpretation of molecular changes across the replicative and tumor evolutionary timeline. To further characterize biological processes associated with methylation age, we analyzed DNA methylation profiles from public breast cancer cohorts (N = 3,901). Samples were stratified based on chronological age and calculated epigenetic age using multiple clocks, including Horvath, PhenoAge, and GrimAge. Differential methylation and pathway enrichment analyses were performed. Results: This approach enabled the identification of transcriptional patterns consistent with age acceleration in Mexican patients, as well as higher estimated transcriptomic age. Integrative analyses identified methylation-associated pathways and molecular features that differed according to aging metrics, revealing biological processes associated with distinct measures of chronological and epigenetic aging in breast tumors. Conclusions: These findings support the value of combining artificial intelligence–based transcriptional surrogates of epigenetic clocks with large-scale methylation profiling and biological prioritization to provide a novel framework for studying accelerated tumor aging in breast cancer, with potential relevance for biomarker development and clinical risk stratification.
Efficacy and safety analysis of T-DXd versus anti-TROP2 ADCs therapy in HER2-low advanced breast cancer.
e15046 Background: Trastuzumab deruxtecan (T-DXd) and anti-Trophoblast cell-surface antigen(TROP)2 antibody–drug conjugates (ADCs) have emerged as key treatment options for patients with HER2-low advanced breast cancer, yet direct real-world comparisons are lacking. This study evaluated the relative efficacy and safety of T-DXd versus anti-TROP2 ADCs in this population. Methods: This real-world study included patients with HER2-low advanced breast cancer treated with either T-DXd or an anti-TROP2 ADC (sacituzumab govitecan, datopotamab deruxtecan, or SKB264) between May 2021 and June 2025. The primary endpoint was progression-free survival (PFS); secondary endpoints included objective response rate (ORR), clinical benefit rate (CBR), and safety. Results: A total of 220 patients were included, with138 in the T-DXd group and 82 in the anti-TROP2 ADC group. The median follow-up time was 14 months. The median age was 52 years (range, 31-82) in the T-DXd group and 53 years (range, 27-74) in the anti-TROP2 ADC group. The median treat line was 4 (range, 1-14) in the T-DXd group and 3 (range, 1-13) in the anti-TROP2 ADC group ( P < 0.001). Median PFS was significantly longer in T-DXd group than in the anti–TROP2 ADC group (5.5 vs 4.0 months; HR 0.60; P = 0.001). The ORR was 53.6% vs 32.9%, and CBR was 71.0% vs 56.1%. In subgroup analyses, among patients with HER2 IHC 2+, median PFS was 7.0 months (95%CI, 6.0 to 7.5) with T-DXd and 3.5 months (95%CI, 3.0 to 8.0) with anti-TROP2 ADC (HR = 0.54, 95% CI: 0.34-0.88, P = 0.012). In patients with HER2 IHC 2+ based on advanced-stage biopsy, T-DXd achieved a median PFS of 6.5 months (95%CI, 5.0 to 8.5) compared with 3.0 months (95%CI, 2.0 to 12.0) for anti-TROP2 ADC (HR = 0.47, 95% CI: 0.25-0.88, P = 0.018).No new safety signals were observed. The most common adverse events in the T-DXd group were elevated aspartate aminotransferase(61.6%), nausea(58.0%), and fatigue(53.6%), while in the anti-TROP2 ADC group were nausea(52.4%), leukopenia(50.0%), and anemia(43.9%). Grade 3–4 neutropenia occurred in 9.4% of patients receiving T-DXd and 14.6% receiving anti-TROP2 ADCs. Conclusions: In real-world treatment scenarios, T-DXd demonstrated superior efficacy and manageable safety compared with anti-TROP2 ADCs in patients with HER2-low advanced breast cancer. This study also shows that patients with HER2 IHC 2+, especially at the advanced stage derive greater benefit from T-DXd treatment.
DREAMM-9 final analysis: Belantamab mafodotin (belamaf), bortezomib, lenalidomide, and dexamethasone (BVRd) for transplant-ineligible (TI) newly diagnosed multiple myeloma (NDMM).
7503 Background: DREAMM-9 (NCT04091126) was a phase 1 dose and schedule evaluation study of BVRd in adults with TI NDMM. Methods: Patients (pts) received BVRd in 1 of 8 belamaf dosing cohorts at an induction/maintenance schedule of Q3/4W (SHORT: 1.9, 1.4, 1.0 mg/kg), Q6/8W (STRETCH: 1.9, 1.4 mg/kg), Q9/12W (step-down [S/D]: 1.9 for 1 dose then 1.4 mg/kg; 1.4 for 1 dose then 1.0 mg/kg), or Q12W (1.0 mg/kg). The primary endpoint was safety. Responses, minimal residual disease negativity (MRD[-]) in complete response or better (CR+), and pt-reported outcomes (PROs) were assessed. Median dose intensities (mDIs) were calculated as mg/kg/21 days in induction and mg/kg/28 days in maintenance. Results: As of June 2, 2025, 118 pts were enrolled (intention-to-treat population). Median duration of follow-up was 15.9–47.1 months. Induction mDI generally decreased as planned dosing intervals increased and was lowest in S/D Q9/12W or Q12W cohorts. Maintenance mDI was similar across cohorts due to dose modifications. Overall response rates were ≥83% across cohorts. Cohorts with higher induction mDI had the deepest responses (CR+/MRD[-] 75/55% for pooled SHORT+STRETCH and 59/43% for pooled Q9/12W+Q12W). Grade (Gr) ≥2 ophthalmic exam findings (OEFs; best-corrected visual acuity changes/slit lamp findings [Keratopathy and Visual Acuity scale]) were 90% for pooled SHORT+STRETCH/69% for pooled Q9/12W+Q12W, while Gr 3/4 OEF rates were higher with higher induction mDI (79%/35%). Most cohorts (5/8) had no belamaf discontinuations due to Gr ≥3 OEFs, which occurred only in SHORT cohorts (n = 1 each). First Gr ≥2 OEFs resolved prior to end of treatment in 83–100% across cohorts. The long-interval group (Q9/12W and Q12W dosing schedules) remained largely below the vision-related function (VRF) deterioration threshold (12.5 points per Ocular Surface Disease Index) across most timepoints, suggesting a favorable VRF profile vs the SHORT interval group. Conclusions: BVRd had high response rates across all cohorts (≥83%), highlighting the promising efficacy of the regimen. Regimens with higher belamaf induction mDI had the deepest responses, while those with lower induction mDI had milder OEFs. Belamaf dosing with a higher induction DI is optimal to achieve deeper responses, and a lower maintenance DI (i.e., using longer schedules) improves tolerability, PROs, and maintains responses. Clinical trial information: NCT04091126 .
Spatial remodeling of the tumor-immune microenvironment (TIME) on neoadjuvant chemo-immunotherapy (NCIT) in patients (pts) with hormone receptor positive (HR+) early breast cancer (eBC) by multiplex immunofluorescence (mIF).
e12645 Background: NCIT efficacy remains limited in HR+ eBC. Given the role of cell interactions in antitumor immunity, we evaluated TIME spatial dynamics in this subtype. Methods: This is an exploratory analysis from trial NCT02999477 randomizing pts to a monotherapy run-in (A: nab-paclitaxel; B: pembrolizumab) followed by both in combination (NCIT). Biopsies were obtained at baseline (BL), after run-in (week 3, W3), and on NCIT (W7). At each timepoint, 4 tissue sections were stained with 4 mIF panels: T-cell (CD4, CD8a, FOXP3), activated T-cell (CD8a, granzyme B, Ki67), antigen presenting cells (APCs) (HLA-ABC, HLA-DR, CD8a), and myeloid (CD68, CD206, HLA-DR), all with pan-CK and nuclear staining. Objectives were to characterize immune cell densities (cell/mm2), changes from BL, tumor–immune nearest-neighbor distance (μm), and explore associations with clinicopathological features, residual cancer burden (RCB) (0/I vs II/III, Kruskal–Wallis) and event-free survival (EFS) (Cox model) (p < .05; FDR q < 0.2). Results: 27 pts were included. Interpatient heterogeneity in immune cell density was observed across timepoints. No significant BL density differences were seen by trial arm, age (≤50 vs > 50), race (White vs non-White), stage (II vs III). At BL, RCB-0/I pts had numerically higher density of helper T cells (Th) (CD4+; median 2154.1 vs 890.0; p = .06), cytotoxic T cells (Tc) (CD8+; 728.1 vs 367.9; p = .08), activated Tc (6.1 vs 1.3; p = .06), APCs (HLA-ABC+, HLA-DR+: 124.6 vs 51.3; p = .36), and macrophages (CD68+CD206+; 63.4 vs 26.7; p = .25) vs RCB-II/III. On treatment, RCB-0/I pts had numerically greater % increase in cell density at W3 from BL than RCB-II/III for Tc (median 52.7 vs −62.0; p = .64), activated Tc (269.5 vs −83.6; p = .13), APCs (417.6 vs 24.9; p = .60), and macrophages (780.2 vs −61.4; p = .24). At W7, APCs continued to show greater % increase in RCB-0/I vs RCB-II/III (1583.3 vs 121.3; p = .38). At BL, distance from tumor area was numerically shorter in RCB-0/I vs RCB-II/III for Tc (median 55.2 vs 79.3; p = .20), activated Tc (1015.4 vs 2274.1; p = .39), APCs (216.4 vs 370.4; p = .39), and macrophages (276.4 vs 517.8; p = .50). On treatment (W3, W7), % changes from BL in tumor-immune distance were heterogeneous in RCB-0/I vs RCB-II/III across most immune cells. At W3, APCs (−52.2 vs 27.6; p = .24) and macrophages (−37.7 vs 20.2; p = .18) got numerically closer to tumor vs BL in RCB-0/I. At BL, pts with RCB-0/I had a numerically shorter distance between HLA-ABC+ cells and CD8+ cells vs RCB-II/III (median 20.5 vs 30.8; p = .36). A continued numerical % decrease from BL was seen at W3 (−57.1 vs 5.3; p = .12) and W7 (−65.2 vs −16.5; p = .38). No density or distance thresholds predicted EFS. Conclusions: TIME remodeling was heterogeneous, trending toward greater Tc and APC density and proximity to the tumor area in RCB 0/I. Clinical trial information: NCT02999477 .
Beyond cardiotoxicity: The hidden burden of functional disability in cardio-oncology survivorship.
e24014 Background: Cardio-oncology traditionally focuses on cardiotoxicity prevention and acute cardiovascular-events. However, the functional, vocational, and mental health challenges that define daily survivorship remain largely hidden, representing a critical gap in holistic care. Methods: Utilizing the 2024 National Health Interview Survey (NHIS), we analyzed a nationally representative cohort of U.S. adults (≥18 years) across four groups: (1) Controls (no cancer or CVD), (2) CVD only, (3) cancer only, and (4) cardio-oncology. We evaluated five patient-centered metrics reflecting quality of life: functional and work limitations, assistive equipment use, personal assistance needs, and overall disability. Multivariable regression models, using Taylor series linearization for complex survey weighting, were adjusted for age, sex, race, hypertension, diabetes, hyperlipidemia, smoking, psychological distress (K6 score), and insurance status. Results: Among 32,599 adults (Weighted N = 259.7M), 2.9% lived with cardio-oncology multi-morbidity (weighted N = 5.34M; mean age 73y). After multivariable adjustment, functional limitations were highest in the cardio-oncology (63.6%, OR 2.34 [95%CI 1.89-2.88]), exceeding both CVD only (OR 1.99) and cancer only (OR = 1.26) groups. This population was also at the highest risk of disability (OR 2.74) and work limitations (31.7%, OR = 2.53). One in five cardio-oncology survivors reported severe mobility impairment (difficulty walking/steps), with a need for assistive equipment (OR = 2.86) and personal assistance (OR = 1.38). Conclusions: Cardio-oncology survivors face a functional and vocational burden significantly exceeding that of cancer or CVD alone. Care models must expand beyond event prevention to integrate routine functional screening and early rehabilitation to preserve quality of life in this high-risk population. Adjusted odds ratios by disease phenotype. Outcome CVD Only (vs. Control) Cancer Only (vs. Control) Cardio-Oncology (vs. Control) OR (95% CI) OR (95% CI) OR (95% CI) Functional limitation 1.99 (1.75–2.28) 1.26 (1.14–1.39) 2.34 (1.89–2.88) Uses special equipment 2.56 (2.20–2.98) 1.26 (1.08–1.46) 2.86 (2.32–3.53) Needs assistance 1.51 (1.17-1.94) 0.84 (0.64-1.12) 1.38 (0.99-1.92) Work limitation 2.12 (1.88-2.39) 1.19 (1.07-1.33) 2.53 (2.11-3.03) Disability 2.20 (1.90–2.56) 1.11 (0.96–1.29) 2.74 (2.26–3.32) Note: Models adjusted for age, sex, race, hypertension, diabetes, hyperlipidemia, smoking status, psychological distress, and insurance type. Reference group = Adults with no history of cancer or CVD. CVD = Cardiovascular disease; OR = Odds Ratio; CI = Confidence Interval.
Efficacy and safety analysis of anlotinib combined with etoposide plus ifosfamide in the treatment of children and adolescents with pulmonary metastatic osteosarcoma.
11525 Background: Osteosarcoma is a highly malignant and aggressive tumor, predominantly occurring in children and adolescents under 20 years of age. It exhibits early metastatic potential, with 15%–20% of patients presenting with distant metastasis at diagnosis. Among these, pulmonary metastases account for 85% of cases. The 5-year survival rate after pulmonary metastasis is <20%. This study retrospectively analyzes the efficacy and safety of anlotinib combined with etoposide and ifosfamide in treating pediatric and adolescent patients with pulmonary metastatic osteosarcoma. Methods: This study retrospectively screened children and adolescent patients with pulmonary metastatic osteosarcoma who received etoposide + ifosfamide ± anlotinib at Northwest Women's and Children's Hospital from May 2018 to December 2024. The study was divided into a combination group (anlotinib + etoposide + ifosfamide) and a control group (etoposide + ifosfamide). The dosage of anlotinib: 8 mg for patients < 12 years old, qd, d1-d14, q3w, 12mg for patients ≥12 years old, qd, d1-d14, q3w. Treatment continued until disease progression or intolerable toxicity, with a maximum of 4 chemotherapy cycles. The primary endpoint of the study was objective response rate (ORR), and secondary endpoints included disease control rate (DCR), progression-free survival rate (PFSR), and safety. Results: The data cut-off date was December 2024, The study enrolled 38 patients (combination group: n=18, control group: n=20). Efficacy analysis showed the combination group achieved higher ORR (22.2% vs 15%, p =0.057) and significantly better DCR (66.7% vs 45%, p =0.015), and 3-month progression-free survival rates (PFSR) of 44.0% versus 30%. Subgroup analysis indicated superior outcomes with 12 mg versus 8mg anlotinib ( p <0.05). Biomarker evaluation revealed patients with high BRCA1/VEGFR2 expression had lower 3-year survival than those with intermediate/low expression, while low PDGFR expression was associated with poorer survival compared to intermediate/high levels. The combination group exhibited higher incidence rates of treatment-related adverse events (TRAEs) compared to the control group, including epistaxis (44.4%), hand-foot syndrome (27.8%), abnormal thyroid function (22.2%), and proteinuria (16.7%). All reported adverse events were grade 1-2 in severity, with no grade 3-4 severe adverse reactions observed. Conclusions: The combination therapy of anlotinib with ifosfamide and etoposide demonstrated superior efficacy compared to chemotherapy alone in pediatric patients with pulmonary metastatic osteosarcoma, showing higher ORR, DCR, and 3-month PFSR. These results indicate that anlotinib combined with neoadjuvant chemotherapy can more effectively delay tumor progression and prolong progression-free survival in these patients.
Efficacy of nimotuzumab combined with concurrent chemoradiotherapy for patients with postoperative loco-regional recurrent esophageal squamous cell carcinoma: A phase II study.
4067 Background: The clinical therapeutic effect for patients with postoperative loco-regional recurrent esophageal squamous cell carcinoma (ESCC) is not satisfactory. Nimotuzumab, an EGFR-targeted humanized antibody, showed efficacy in the treatment of ESCC. We explored the efficacy of nimotuzumab combined with concurrent chemoradiotherapy (CCRT) for patients with postoperative loco-regional recurrent ESCC. Methods: Eligible patients (18-75 years) with postoperative loco-regional recurrent ESCC were enrolled in this single-arm phase II study, received nimotuzumab (400mg on D1, 200mg/w for 6 weeks) combined with CCRT (chemotherapy:TP regimen, radiotherapy: 45-50.4Gy/25-28f/5-6w), sequentially nimotuzumab 200mg every 3 weeks plus chemotherapy (patients with partial response or stable disease). The primary endpoint was objective response rate (ORR), the second endpoints were disease control rate (DCR), overall survival (OS), progression-free survival (PFS), and local-regional control (LRC) and adverse events. Results: From Sep 2021 to Apr 2024, 44 patients were screened, ultimately 42 patients were included in the Full Analysis set. The median age was 69 years (range: 48-75). 12 (28.57%) patients achieved complete response, 24 (57.14%) patients with partial response, 5 (11.90%) patients with stable disease, and 1 (2.38%) patient with progressive disease. ORR was 85.71%, and DCR was 97.62%. The failure patterns showed that 10 (23.81%) patients had loco-regional recurrence, and 9 (21.43%) had distant metastases after CCRT. The median follow-up was 29.3 months (95%CI: 25.7-38.0). 3-year OS was 51.8% (95%CI: 36.9-72.8%), with the median OS was not reached. 3-year PFS was 33.4% (95%CI: 20.4-54.7%), and the median PFS was 15.1 months (95%CI: 11.8-NE), 3-year LRC was 46.1% (95%CI: 29.9-71.0%), the median LRC was 30.8 (95%CI: 16.6-NE). The most common Graded 3-5 adverse events (AEs) included leukopenia, thrombocytopenia, hemoglobin decreased, radiation esophagitis, radiation-induced lung injury, fatigue, and radiodermatitis. 4 patients experienced Grade 3-5 radiation esophagitis, in which 2 patients developed esophageal fistulas (1 patient died). No nimotuzumab-related serious AEs occurred. Conclusions: Nimotuzumab combined with CCRT showed promising ORR and survival with tolerable toxicity in patients with postoperative loco-regional recurrent ESCC. Clinical trial information: ChiCTR2200063875.
Predictors of long-term survival in esophageal carcinoma: Is it the same for all stages?
e16152 Background: Survival outcomes in esophageal carcinoma depend on evolving prognostic factors across clinical stages. This study aimed to evaluate those stage-dependent predictors among treated patients. Methods: Patients with esophageal carcinoma were identified from the National Cancer Database (2004–2020). Survival was divided into following groups: 0–2, 2–5, or > 5 years based on months to last contact or death, without censoring. Ordinal logistic regression was performed to model the odds of belonging to a higher survival category. Adjusted odds ratios (aORs) were calculated for demographic, socioeconomic, tumor, and treatment-related factors. Results: Out of 37,429 patients, 16.9% had stage I, 43.5% stage II, 35.1% stage III, and 4.5% stage IV disease. In stage I, uninsured status (aOR 0.24, p < 0.05), Medicare status (aOR 0.71, p < 0.05) and reporting facilities in New England, South Atlantic, East North & South-central regions (aOR < 0.52, p < 0.05) were associated with worse survival, while tumor size had a minimal positive association (aOR 1.00, p = 0.05). In stage II, male sex (aOR 0.74, p < 0.001), higher comorbidity (aOR 0.87, p < 0.001), lower education (aOR 0.78, p < 0.05), Medicare status (aOR 0.76, p < 0.05) and adjuvant radiation (aOR 0.79, p < 0.05) were linked to poorer outcomes. Additionally, all the regions except Pacific (aOR < 0.79, p < 0.05) was associated with worse survival, while median income of $38K-48K (aOR 1.20, p < 0.05) had improved survival. In stage III, male sex (aOR 0.81, p = 0.001), uninsured (aOR 0.68, p < 0.05), Medicare status (aOR 0.71, p < 0.05) and longer time to systemic therapy (aOR 0.97, p < 0.05) were associated with lower survival, whereas well-differentiated histology (aOR 2.10, p < 0.05), absence of 30-day readmission (aOR1.47, p < 0.05) and care in early Medicaid expansion states (aOR 1.30, p < 0.05) improved outcomes. In stage IV, male sex (aOR 0.54, p < 0.05), and prolonged hospitalization (aOR 0.99, p < 0.05) were associated with worse survival, while combined surgery-radiation therapy (aOR 7.99, p < 0.05) was associated with improved survival. Conclusions: Survival drivers in esophageal carcinoma differ across AJCC stages. Sociodemographic factors predominate in early disease, whereas tumor biology and treatment quality drive outcomes in advanced stages. Targeted strategies addressing disparities in insurance coverage, comorbidity management, and access to multidisciplinary care may enhance stage-specific survival.
AI-enabled multimodal adjudication of revised IMWG progression after BCMA CAR-T.
7521 Background: Real-world ascertainment of clinically meaningful relapse after BCMA CAR-T is challenging because progression signals are distributed across laboratories, imaging, pathology, and clinician actions and are frequently embedded in unstructured external reports. The revised IMWG criteria (Kumar et al., IMS 2025) provide standardized definitions for imaging-based progression (PET/CT and WB-MRI) enabling automated recognition of radiologic and serologic progression thresholds. We developed an AI-enabled multimodal framework to automate real-time derivation of recognized IMWG progression across fragmented care settings. Methods: We analyzed 183 BCMA CAR-T treatment episodes with longitudinal routine laboratory data and independent dual-reviewer adjudication of progression dates. Automated detection evaluated every M-protein and FLC measurement longitudinally in a continuous IMWG rules engine, in addition to flagging new hypercalcemia. In addition, initiation of a new line of therapy (considered a progression event), together with large-language-model extraction of radiologic progression from PET/CT, WB-MRI, CT, X-ray, and MRI brain/spine reports. Performance was assessed using accuracy and specificity. Results: Serologic progression from real-world lab feeds was detected with high reproducibility, with FLC and M-protein achieving high accuracy (96.5% and 98.5%, respectively) and specificity (>97%), indicating minimal premature triggering across serial measurements. Hypercalcemia was rare but highly specific for progression. Radiology-based AI extraction achieved high accuracy (91%) and specificity (92.1%), enabling reliable identification of imaging-defined relapse. Initiation of a new line of therapy occasionally occurred before formal serologic IMWG thresholds were met, reflecting clinician-recognized relapse or that driven by non-serologic disease. Notably, 19% of adjudicated relapses were triggered by radiologic or marrow criteria when serologic IMWG thresholds were absent or lagging, including 12.5% radiology-only and 5.8% marrow-only events, aligned with our published data for post CART relapses (Abuhelwa et al Front. Oncol 2025). Conclusions: Automated lab-only approaches systematically underestimate true IMWG progression after BCMA CAR-T. An AI-enabled multimodal adjudication framework aligned with the revised IMWG criteria enables scalable, real-time, and reproducible progression capture for clinical trials and real-world datasets, supporting rapid endpoint determination, regulatory-grade retrospective analyses, and biologically faithful reconstruction of relapse patterns after CAR-T.
Impact of <i>STK11</i> / <i>KEAP1</i> / <i>SMARCA4</i> co-mutations in RAS-activated <i>KRAS</i> wild-type metastatic NSCLC treated with immune checkpoint inhibitors.
8565 Background: RAS-activated KRAS wild-type (RASa/KRASwt) non-small cell lung cancer (NSCLC) shares key biological features with KRAS-mutant (KRASm) tumors, including MAPK pathway activation. In KRASm NSCLC, co-mutations in STK11 , KEAP1 and SMARCA4 are associated with poor outcomes to immune checkpoint inhibitors (ICI). Whether these alterations similarly impact outcomes in RASa/KRASwt tumors is unknown. Methods: We analyzed patients with metastatic NSCLC treated with ICI at Dana-Farber Cancer Institute. RASa/KRASwt tumors were defined by pathogenic alterations in RAS–MAPK pathway genes ( NF1 , NF2 , BRAF , RAF1 , MAP2K1/2 , MAP3K1 , MAPK1 , PTPN11 , SOS1 , HRAS , NRAS , RIT1 , RASA1 , GNAS ). Co-mutations in STK11 , KEAP1 and SMARCA4 were evaluated using targeted exome sequencing. Survival analyses were performed using log-rank test and survival risks estimated with Cox regression models. Results: Among 1,215 patients included, 492 (40%) had KRASm tumors and 264 (22%) were RASa/KRASwt. Compared with RASa/KRASwt tumors, KRASm tumors were more frequent in female patients (65% vs 50%, p<0.001), ever-smokers (93% vs 86%, p=0.004) and non-squamous histology (98% vs 88%, p<0.001). The frequency of KEAP1 (21% vs 22%), STK11 (26% vs 20%) and SMARCA4 (10% vs 12%) mutations was similar between the groups. High tumor mutation burden (≥10 mut/Mb) was more frequent in RASa/KRASwt tumors (60% vs 46%, p<0.001), while PD-L1 expression, age and performance status were comparable. Within the RASa/KRASwt cohort, 103 patients (39%) had co-mutations in STK11 /KEAP1/ SMARCA4 and were more frequently ever-smokers compared with non co-mutated RASa/KRASwt (95% vs 80%, p<0.001), with otherwise similar clinical characteristics. As expected, co-mutations were associated with shorter median progression free survival (mPFS) and median overall survival (OS) in KRASm group (mPFS: 4.4 vs 6.2 mo., HR:1.33 (1.1-1.62), p=0.004; mOS: 11.4 vs 21.7 months(mo), HR:1.43 (1.16-1.78), p<0.001). However, among RASa/KRASwt tumors, concurrent mutations in STK11/KEAP1/SMARCA4 had no impact on survival outcomes (mPFS: 5.9 vs 4.1 mo., HR:0.79 (0.6-1.02), p=0.07; mOS: 15.8 vs 14.8 mo, HR: 1.02 (0.77-1.35), p=0.89). This differential effect was confirmed by a significant interaction between RAS status and co-mutational status (aHR 1.71, p=0.004). For OS, interaction analysis did not show a differential effect of co-mutations according to RAS status (aHR 1.43; p=0.073). Conclusions: Known co-mutations that confer worse outcomes to ICI in KRASm NSCLC do not appear to have the same impact in KRASwt tumors with other RAS activating mutations. RASa/KRASwt tumors with co-mutations behave more similarly to KRAS wild-type NSCLCs, highlighting biological heterogeneity within RAS-driven tumors and potential implications for patient selection for treatment intensification or de-escalation strategies.
Practice patterns and outcomes by genomic risk in octogenarians with high-risk localized prostate cancer: A national real-world data analysis.
1660 Background: Chronologic age and traditional clinical-pathologic prostate cancer (PCa) risk factors (e.g. PSA, Gleason Score) are useful but incomplete predictors of prognosis. While some octogenarians may benefit from watchful waiting for clinically high risk (HR) and very high risk (VHR) PCa due to competing risks, others may survive more than a decade with curative prostate-directed treatment. We examined real-world data (RWD) on practice patterns and outcomes for octogenarians with HR PCa stratified by genomic classifier (GC) scores. Methods: A retrospective cohort study was conducted using a national database of clinical–genomic linkage of Decipher GC with longitudinal RWD (Veracyte, South San Francisco, CA). The cohort included 1,519 patients, ≥80-years-old with HR/VHR PCa who underwent GC testing with >1 year of follow-up. Patients were stratified by low (GC-Low) compared to intermediate or high (GC-High/Int) GC scores. Observational management was determined by absence of treatment with evidence of prostate cancer–related services (e.g. GC-High/Int-Obs). Freedom from distant metastases (FFDM), metastasis-free survival (MFS), and overall survival (OS) were modeled using Kaplan-Meier analysis and log-rank tests. Results: Most patients (88%, n=1,341) had GC-High/Int. The rate of observational management was significantly lower in the GC-High/Int compared to GC-Low cohorts (26% vs 38%, p=0.001). Among treated patients, there were similar rates of ADT only use (52% GC-Low vs 54% GC-High/Int) but amongst those who were treated with radiotherapy, concurrent ADT use was lower in GC-Low compared to GC-High/Int (55% vs 69%, OR 0.83, 95% CI 0.73-0.95, p=0.007). There were significant differences between groups for FFDM (p=0.003), MFS (p<0.001), and OS (p<0.001) (Table). Among observed patients, GC-Low had significantly better FFDM than GC-High/Int (HR 0.11, 95% CI 0.02–0.80; p=0.03). Conclusions: Geriatric patients with aggressive, localized PCa decide between potentially morbid, curative treatments and heterogeneous risk of distant metastases in the setting of limited life expectancy. We present RWD that patients ≥80 years old with NCCN HR/VHR PCa with low GC scores have <5% rate of metastatic disease within 5 years with observation alone. This hypothesis generating data suggests GC score may be a useful prognostic tool in addition to standard clinical-pathologic factors for elderly patients. Patient-specific risk calculators for cancer and non-cancer specific mortality may help this underserved but growing population. NCCN HR (%) NCCN VHR (%) FFDM Hazard Ratio (95% CI) 5-yr FFDM 5-yr MFS 5-yr OS GC-Low-Obs 58 (85%) 10 (15%) 0.11 (0.02-0.80) 98% 70% 71% GC-Low-Treat 90 (82%) 20 (18%) 0.76 (0.36-1.59) 82% 75% 91% GC-High/Int-Obs 250 (71%) 104 (29%) Reference 83% 37% 43% GC-High/Int-Treat 633 (64%) 354 (36%) 1.35 (0.93-1.98) 81% 59% 64%
Predictive role of gut microbiota and their metabolites in the efficacy of immune checkpoint inhibitor therapy for lung squamous cell carcinoma.
e20017 Background: Lung squamous cell carcinoma (LUSC) is the second most common subtype of lung cancer. Due to its low driver gene mutation rate, immunotherapy has become the first-line treatment for most patients. However, comprehensive tools to predict immunotherapy efficacy remain inadequately established. The role of gut microbiota and their metabolites in predicting response to immune checkpoint inhibitor (ICI) therapy for LUSC is largely unclear. This study aimed to explore the predictive value of gut microbiota and their metabolites for ICI efficacy in LUSC patients, as well as their regulatory effects on the immune system. Methods: Twenty-seven LUSC patients were enrolled and stratified into two groups by progression-free survival (PFS): responders (PFS ≥ 6 months, n = 16) and non-responders (PFS < 6 months, n = 11). A total of 54 fecal samples were collected at three time points: baseline (before immunotherapy), response (partial response [PR]), and progression (progressive disease [PD]). Integrated metagenomic and untargeted metabolomic analyses were performed within and between the two groups. Machine learning models were used to identify potential biomarkers linked to treatment outcomes. Results: Baseline gut microbiome composition was comparable between the two cohorts. During immunotherapy, non-responders showed progressive gut microbiota dysregulation, along with increased microbial diversity. Genera including Cryptobacterium, Faecalibacterium, Alistipes, and Lactimicrobium were identified as potential key determinants of treatment response before immunotherapy initiation. Moreover, relative abundances of probiotic genera (Romboutsia, Lactobacillus, Akkermansia, Lactiplantibacillus, Ligilactobacillus) and "pathogenic" genera (Finegoldia, Mycoplasma, Desulfobulbus, Peptoniphilus, Mycobacterium, Streptococcus) emerged as promising indicators for real-time monitoring of ICI efficacy. Metabolomic profiling revealed that lipid metabolic pathways play a pivotal role in immunotherapy, with multiple lipid metabolites showing significant level alterations. Furthermore, machine learning models identified static biomarkers (PG(i-19:0/PGE2), SM(d18:2(4E,14Z)/6-keto-PGF1α)) and dynamic biomarkers (DL-2-hydroxystearic acid, isoleucyl-valine, 22-hydroxydocosanoic acid, (9Z)-octadecenoic acid), which are highly promising for early prediction and dynamic monitoring of ICI efficacy. Conclusions: Our findings demonstrate that gut microbiota and their metabolites may modulate the efficacy of immunotherapy in LUSC patients. The identified static and dynamic biomarkers possess substantial potential for the early prediction and real-time monitoring of treatment responses.
Comparison of DNA+RNA co-detection and DNA-only NGS for driver gene identification and co-mutation profiling in Chinese patients with non–small cell lung cancer.
e20750 Background: Comprehensive identification of oncogenic drivers is essential for precision treatment of NSCLC. Although DNA-based NGS is widely used, it has limited sensitivity for gene fusions and splice variants. DNA+RNA co-detection enables simultaneous detection of DNA- and RNA-level alterations and may improve identification of actionable drivers. However, real-world comparisons of these strategies in Chinese NSCLC populations are limited. This study compared driver detection rates and co-mutation landscapes between DR co-detection and DNA-NGS. Methods: Two independent cohorts of Chinese patients with NSCLC were analyzed. The DR co-detection cohort included 1,309 patients tested using a 57-gene DNA+RNA co-detection panel, whereas the DNA-NGS cohort included 2,362 patients tested using a 733-gene DNA-only NGS panel. Driver gene detection rates for major actionable genes (EGFR, ALK, RET, ROS1, MET, ERBB2, BRAF, KRAS) were compared. Co-mutation patterns involving multiple driver genes were further analyzed to assess differences in molecular complexity captured by each testing strategy. Results: DR co-detection demonstrated superior performance in identifying fusion and structural-variant–driven oncogenes. Compared with DNA-NGS, DR co-detection yielded higher detection rates for ALK (6.95% vs 4.70%), RET (4.81% vs 1.44%), MET (5.96% vs 4.11%), ERBB2 (5.35% vs 4.40%), BRAF (4.35% vs 2.41%), and ROS1 (2.14% vs 1.48%), highlighting the added value of RNA-based interrogation for fusion and splicing events. In contrast, DNA-NGS showed slightly higher detection rates for classic point-mutation–driven genes, including EGFR (48.05% vs 47.39%) and KRAS (12.40% vs 11.92%), consistent with its technical strengths in SNV/Indel detection. Both methods identified patients harboring multiple driver alterations (79 in the DR co-detection cohort and 83 in the DNA-NGS cohort); however, the composition of co-mutation profiles differed substantially. DR co-detection preferentially identified fusion-involved co-mutation patterns, such as ALK/EGFR, RET/EGFR, ROS1/KRAS, and MET/EGFR, whereas DNA-NGS primarily detected SNV-dominant co-mutations, including EGFR/KRAS, EGFR/MET, and EGFR/BRAF. Overall, DR co-detection captured greater driver gene diversity and structural-variant complexity. Conclusions: In Chinese patients with NSCLC, DNA+RNA co-detection significantly improves the detection of fusion and structural-variant driver alterations and expands the identification of complex driver co-mutation landscapes, whereas DNA-NGS remains advantageous for classic point mutations. These complementary strengths support the clinical value of integrated DNA+RNA testing strategies for comprehensive molecular profiling and precision treatment decision-making in NSCLC.
Molecular and immune profiling of <i>BRAF</i> -mutated ( <i>BRAF</i> <sup>MUT</sup> ) non-small cell lung cancer (NSCLC).
8653 Background: BRAF MUT occur in 6-7% of NSCLC and comprise of biologically distinct classes showing variable therapeutic responses. While Class 1 benefits from targeted therapy, molecular and immune determinants distinguishing BRAF classes remain incompletely characterized. We performed comprehensive profiling to identify class-specific biomarkers. Methods: 51,692 (BRAF wild-type [ BRAF WT ]: n=48,288; BRAF MUT : n=3404) NSCLC specimens underwent DNA (592-gene panel/whole-exome) and/or RNA (whole-transcriptome) sequencing at Caris Life Sciences. BRAF MUT tumors were classified into 3 Classes, other pathogenic variants (O PV ), and unknown/unclassified variants (VUS). The tumor microenvironment (TME) was estimated using the QuanTIseq method. Overall survival (OS) and IO-associated survival (IO-OS) were derived from insurance claims and calculated from the date of tumor biopsy (OS) or IO initiation (IO-OS) to last contact using Kaplan-Meier estimates. Statistical significance was determined by Fisher’s Exact, chi-square, and Mann-Whitney U test with adjustments for multiple comparisons ( P <0.05). Results: Class 1 was significantly enriched for SETD2 mutations (mut) and more frequently PDL1+ (22c3, TPS>=1) while showing a lower prevalence of TMB-high (>=10 mut/Mb), TP53, STK11 and KEAP1 mut (Table). Class 2/3 tumors were moderately PDL1+ and TMB-high and enriched for STK11 and KEAP1 muts (Table). While both OPV and VUS were enriched for TMB-high and not PDL1, OPV was enriched in STK11, and VUS was enriched in TP53 mut (Table). Among BRAFMUT, Class 1 exhibited the highest infiltration of M1 macrophages (1.1-1.3-fold), neutrophils (1.2-1.3-fold) and the lowest infiltration of dendritic cells (0.4-0.8-fold). Class 2 and 3 had largely overlapping immune profiles. In metastatic NSCLC, only Class 1 and VUS demonstrated improved OS and IO-OS (vs BRAFWT, both p<0.05, Table). OPV had the poorest outcomes, exhibiting the shortest OS and shorter IO-OS relative to Class 1 and Class 2 (all p<0.05, Table). Conclusions: BRAF MUT classes represent distinct molecular, immune and clinical phenotypes. Improved OS and IO-OS in Class 1 and VUS (vs WT) appears to arise from divergent molecular and immune interactions. These findings underscore the importance of evaluating of BRAF MUT NSCLC beyond class 1-3, suggesting the expanded classification may have therapeutic and prognostic consequences. Molecular features (odds ratio) and survival outcomes (median months [95% CI]) among BRAF MUT classes. Genes Class 1 (n=695) Class 2 (n=741) Class 3 (n=701) O PV (n=123) VUS (n=936) WT SETD2 21.5 s 1.2 1.0 0.6 1.4 0.2 s TP53 0.4 s 1.0 1.0 0.8 1.6 s 1.1 s STK11 0.1 s 1.7 s 2.3 s 2.1 s 0.8 0.9 s KEAP1 0.1 s 1.5 s 1.7 s 1.5 1.1 1 TMB-high 0.3 s 1.8 s 2.3 s 2.3 s 3.3 s 0.6 s PDL1+ 3.4 s 1.3 s 1.1 0.9 1.0 0.7 s med OS 18(14-21) 14(12-18) 13(11-16) 8(4-11) 16(12-20) 12(12-12) med IO-OS 27(19-40) 20(15-24) 18(12-24) 8(2-15) 24(19-32) 17(17-18) P <0.05 denoted by s .
Effects of vancomycin on the gut microbiota and metabolic profile to sensitize microsatellite-stable colorectal cancer to anti-PD-1 immunotherapy.
e15580 Background: Microsatellite-stable (MSS) colorectal cancer (CRC) remains largely refractory to immune checkpoint blockade (ICB). While the gut microbiota regulates the tumor immune microenvironment, mechanisms linking antibiotic-induced dysbiosis to ICB efficacy remain unclear. We investigated whether vancomycin sensitizes MSS CRC to anti-PD-1 therapy and elucidated underlying mechanisms involving microbial metabolites and dendritic cell (DC) function. Methods: MSS CRC subcutaneous syngeneic models were established and treated with vancomycin, anti-PD-1, or their combination. Tumor growth was monitored longitudinally. Integrated multi-omics analysis (16S rRNA sequencing, metagenomics, and metabolomics) profiled gut microbiota and metabolic alterations. Molecular mechanisms were validated in bone marrow-derived DCs (BMDCs) and tumor-draining lymph nodes (TDLNs) using RNA sequencing, molecular docking, co-immunoprecipitation, and ubiquitination assays. Results: Compared with monotherapy, the combination of vancomycin and anti-PD-1 significantly suppressed tumor growth and prolonged survival. Microbiome analysis revealed that vancomycin dramatically altered gut microbiota composition, enriching beneficial bacterial taxa. Metabolomics identified marked downregulation of systemic and intratumoral asparagine levels in the combination group. Mechanistically, high asparagine levels impaired DC function specifically within TDLNs. Asparagine supplementation upregulated the E3 ubiquitin ligase FBXO21 in DCs. FBXO21 directly interacted with p100, promoting K48-linked ubiquitination and processing to p52, thereby aberrantly activating the non-canonical NF-κB pathway. This activation caused defective DC maturation and impaired antigen cross-presentation in TDLNs, hindering T cell priming. Conversely, vancomycin-induced asparagine depletion downregulated FBXO21, inhibited p100-to-p52 processing, and suppressed non-canonical NF-κB signaling, effectively restoring DC antigen-presenting capacity in TDLNs and promoting effector CD8+ T cell infiltration into tumors. Conclusions: Vancomycin enhances anti-PD-1 efficacy in MSS CRC by remodeling the gut microbiota and depleting asparagine. We identified a novel axis wherein asparagine downregulation inhibits FBXO21-mediated p100 processing and non-canonical NF-κB activation, restoring DC function in TDLNs. Targeting the microbiota-asparagine-FBXO21 axis offers a promising therapeutic strategy to overcome ICB resistance in MSS CRC.
Comparative effectiveness of advanced radiofrequency ablation electrode systems for hepatocellular carcinoma: A systematic review and network meta-analysis.
e16237 Background: Percutaneous radiofrequency ablation (RFA) offers a curative option for early-stage hepatocellular carcinoma (HCC) in nonsurgical candidates, but conventional techniques have higher local tumor progression (LTP) than resection due to inadequate margins and heat sink effects. Advanced electrode systems address these issues, yet comparative evidence is fragmented. This network meta-analysis synthesizes data to identify the optimal RFA electrode configuration for better HCC ablation outcomes. Methods: A total of 693 studies from PubMed, Embase, Scopus, and Cochrane were screened. RCTs and cohort studies of adults with HCC undergoing percutaneous RFA were included, comparing six electrode types: conventional single straight internally cooled monopolar (MonoRFA), internally cooled wet monopolar (ICW_MonoRFA), twin internally cooled wet bipolar/multi-electrode (TICW_RFA), separable clustered internally cooled switching monopolar/no-touch (SC_MonoRFA), expandable needle monopolar (Exp_MonoRFA), and no-touch multi-electrode RFA (NT-mbpRFA). Primary outcomes: CR and OS. Primary outcomes were complete remission (CR) and overall survival (OS). Frequentist network meta-analysis was performed in R (v4.5.2) using netmeta. Results: A total of 19 publications (28 direct comparisons, N = 2,421) evaluated six RFA electrode systems: MonoRFA (n = 1,041), SC_MonoRFA (n = 414), NT_mbpRFA (n = 463), ICW_MonoRFA (n = 218), Exp_MonoRFA (n = 243), and TICW_RFA (n = 42). SC MonoRFA has a significantly better OS compared to MonoRFA (OR: 0.48, p = 0.02). The LTP is significantly lower for SC monoRFA (OR: 0.33, p < 0.0001) and NT mbpRFA (OR: 0.34, p < 0.0001) compared to MonoRFA. Treatment effectiveness was worse for NT_mbpRFA versus MonoRFA (OR: 5.59, p = 0.03); however, it was highly confounded. There was no statistically significant difference for major complications and complete ablation rates. CR was highest for SC_MonoRFA (OR: 3.96, p < 0.01) and NT_mbpRFA (OR: 2.36, p > 0.05), although rates did not differ significantly from conventional MonoRFA. Cumulative ranking analysis (SUCRA/P-score) consistently positioned SC_MonoRFA as the top-ranked treatment for both LTP and OS, followed by NT_mbpRFA for LTP control. The network for the primary LTP outcome showed low statistical heterogeneity (I² = 0%) and no significant global inconsistency (p = 0.69), supporting the robustness of the comparative estimates. Conclusions: This network meta-analysis shows that separable clustered electrode RFA systems provide statistically superior local tumor control and OS in HCC versus conventional monopolar RFA, without increased major complications. Future research should focus on RCTs comparing it against other advanced systems in standardized populations, plus cost-effectiveness and implementation studies for global adoption.
Strategic alignment of quality improvement interventions to address barriers to adult vaccination in oncology.
e23285 Background: To address suboptimal vaccination rates in adult patients with cancer, the American Society of Clinical Oncology utilized a multi-year learning collaborative to develop, implement, and sustain clinical workflows to support increased COVID-19, influenza, and other routine vaccinations in adult cancer care centers. Methods: Eight participating health systems identified local barriers to adult vaccination and developed quality improvement (QI) interventions. Barriers were mapped to the SURE (Supporting the Use of Research Evidence) Framework, classifying obstacles into System constraints, Provider limitations, or Patient factors. Proposed and implemented interventions were coded according to the ERIC (Expert Recommendations for Implementing Change) framework clusters. A relational database was constructed to link specific identified barriers to their corresponding QI interventions. A complex, many-to-many relationship between the domains of barrier inputs (SURE) and intervention outputs (ERIC) was subsequently visualized using a Sankey diagram to illustrate the predominant flows to actual strategies used to address identified barriers. Results: The analysis classified distinct barriers across all three SURE domains, with prominent challenges identified in patient knowledge, staff Electronic Medical Record (EMR) proficiency, and System-level workflows and resource availability. Health systems prioritized interventions leveraging technology and optimizing workflows, with six of the eight teams implementing strategies categorized under the "Change Infrastructure" ERIC cluster, often involving integrating vaccination processes directly into the EMR to reduce staff burden and improve data capture. Seven teams implemented interventions to "Train and Educate Stakeholders" to boost staff knowledge and confidence, and four teams implemented strategies to "Engage Consumers" by empowering patients and addressing their concerns. Teams focused on formalizing and improving clinical workflows and addressing physical and logistical barriers to make vaccination more convenient. Conclusions: The strategic alignment of barriers and interventions demonstrates that a tailored, multi-faceted approach supports improved adult vaccination rates in oncology. By addressing specific patient, provider, and system-level challenges with targeted strategies, health systems can effectively overcome barriers to care. The implementation of diverse interventions across multiple cancer centers highlights the importance of context-specific adaptation and a systematic framework for guiding quality improvement initiatives.
A deep learning model for intraoperative diagnosis of lung adenocarcinoma using frozen section whole-slide images (CryoPath): A multicentric diagnostic study.
8024 Background: Intraoperative frozen section (FS) analysis is critical for guiding surgical decision-making in lung cancer. However, its diagnostic accuracy can be limited by time constraints, sampling issues, and inter-pathologist variability. Deep learning-based analysis of whole-slide images (WSIs) offers a new paradigm for real-time, objective intraoperative diagnosis. Methods: We developed and validated an artificial intelligence system, CryoPath, for the intraoperative diagnosis of lung adenocarcinoma (LUAD) from FS-WSIs. The system was trained and tested using a large, multicentric, retrospectively collected cohort comprising 3000 LUAD FS-WSI cases from seven medical centers in China. The model architecture is based on attention-challenge multiple instance learning framework with vision transformer for feature extraction. CryoPath was designed for end-to-end prediction of diagnostic subtypes (pre-invasive lesions, minimally invasive adenocarcinoma and invasive adenocarcinoma) and key pathological features such as extent of invasion and lymphovascular invasion from FS-WSIs. Results: The study included 1,042 patients, yielding 1,139 WSI with early-stage lung adenocarcinoma from multiple centers, which comprised a total of 637 WSIs with IAC, 402 with MIA, and 100 with PIL. The CryoPath model achieved an AUC of 0.931 for discriminating IAC from PIL/MIA, which achieved a sensitivity of 0.87 and a specificity of 0.85. In clinical decision simulation, the model-assisted strategy increased surgical decision accuracy from 75% to 88%. 15 WSIs classified as PIL on FS but upgraded to IAC on permanent sections by pathologist were corrected by CryoPath and Attention heatmaps generated by the model highlighted histologically relevant regions. CryoPath model achieved an AUC of 0.92 (95% CI: 0.90–0.94) for discriminating IAC from PIL/MIA on the test set. . In clinical decision simulation, the model-assisted strategy increased surgical decision accuracy from 75% to 88. Attention heatmaps generated by the model highlighted histologically relevant regions. Conclusions: The CryoPath system demonstrates for the first time that large-scale deep learning-based analysis of FS-WSIs can achieve high-accuracy, real-time, and interpretable intraoperative diagnosis for LUAD. This system has the potential to serve as a powerful adjunct tool for pathologists and surgeons, potentially reducing diagnostic variability, shortening intraoperative wait times, and providing an objective foundation for precise surgical strategy. A prospective, multi-center clinical trial is being planned to further evaluate its clinical utility.
Peptide receptor radionuclide therapy with <sup>177</sup> Lu-DOTATATE in pediatric, adolescent, and young adult neuroendocrine tumors: A single-center experience from Bambino Gesù Children’s Hospital.
10027 Background: Neuroendocrine tumors (NETs) in pediatric, adolescent, and young adult (AYA) patients are rare and often associated with hereditary syndromes such as Von Hippel Lindau (VHL) and type 1 neurofibromatosis (NF1). Peptide receptor radionuclide therapy (PRRT) with ^177Lu-DOTATATE is well established in adults with somatostatin receptor–positive NETs, but data in younger populations remain limited. We report our single-center experience evaluating feasibility, safety, and efficacy of PRRT in this age group. Methods: We retrospectively analyzed patients <20 years treated with ^177Lu-DOTATATE at Bambino Gesù Children’s Hospital between 2019 and 2025. Eligibility required somatostatin receptor expression confirmed by PET-DOTA imaging (Krenning score ≥2). Data included demographics, age at diagnosis and PRRT, tumor characteristics, prior therapies, PRRT cycles, cumulative dose, toxicity, and best response per RECIST. Follow-up status was assessed at last visit. Results: Six patients (median age at diagnosis 11 years, range 7–16; median age at PRRT 15 years, range 10–19) received PRRT for metastatic or unresectable NETs (pheochromocytoma/paraganglioma, n=5; bronchial carcinoid, n=1). All completed 4 cycles of ^177Lu-DOTATATE (median cumulative dose ~30 GBq). Prior treatments included surgery, chemotherapy (temozolomide), and MIBG therapy. No grade ≥3 hematologic or renal toxicity occurred; treatment was well tolerated. Best responses: 4 partial responses (PR), and 2 stable disease (SD). At a median follow-up of 18 months, all patients were alive: one in complete remission after surgery, others with stable disease or ongoing therapy. Genetic predisposition was present in 3 patients (VHL, NF1). SUVmax on baseline PET ranged from 1.2 to 43.8, correlating with Krenning score ≥2 in all cases. Conclusions: PRRT with ^177Lu-DOTATATE is feasible and well tolerated in pediatric, adolescent, and young adult patients with somatostatin receptor–positive NETs, achieving meaningful disease control with minimal toxicity. This single-center experience supports PRRT as a valuable option in selected cases and highlights the need for prospective studies and age-specific guidelines.
The role of FOXM1 in regulating immune memory in triple-negative breast cancer and targeting FOXM1 to develop a therapeutic vaccine.
1119 Background: Triple-negative breast cancer (TNBC) is an aggressive malignancy with limited therapeutic options and poor long-term outcomes. Although immune checkpoint blockade has improved survival for a subset of patients, durable responses remain uncommon, underscoring the need for strategies that enhance tumor immunogenicity and immune memory. We recently identified a novel function of the transcription factor FOXM1 in promoting an immune-suppressive tumor microenvironment (TME) in TNBC by regulating stress ligands and the STING pathway. This study investigates the role of FOXM1 in shaping antitumor immune memory and leverages FOXM1 targeting to identify tumor antigens for therapeutic peptide vaccine development. Methods: FOXM1 was knocked out in TNBC cell lines using CRISPR/Cas9, and syngeneic mouse models were utilized to evaluate tumor-immune interactions. Single-cell RNA sequencing (scRNA-seq) evaluated immune landscape remodeling following FOXM1 loss. Immunopeptidomics was performed to identify tumor-associated antigens enriched following FOXM1 loss. Candidate peptides were functionally validated using dendritic cell–based antigen presentation assays and patient-derived peripheral blood mononuclear cells (PBMCs). Clinical relevance was evaluated using TNBC patient datasets and immunotherapy response cohorts. Results: scRNA-seq of syngeneic tumors revealed that FOXM1 deletion reshaped the immune microenvironment, with increased CD8+ T cell and NK cell infiltration and expansion of memory T cell populations. Tumor-intrinsic FOXM1 suppressed antitumor immunity by downregulating MHC-I and the stress ligand ULBP1, disrupting signaling pathways required for NK and CD8+ T cell–mediated cytotoxicity. FOXM1-deficient tumors conferred durable protection upon tumor rechallenge, consistent with long-term immune memory. Analysis of patient datasets demonstrated that elevated FOXM1 and DNMT1 expression, coupled with reduced STING and ULBP1 levels, correlated with poorer survival and reduced responsiveness to immunotherapy. Leveraging enhanced antigen presentation in FOXM1-depleted cells, immunopeptidomics identified 275 differentially presented tumor antigens. Dendritic cell pulsing of the top 15 enriched peptides induced robust T cell responses in patient-derived PBMCs, supporting their immunogenicity and its potential to serve as bona fide candidates for a therapeutic vaccine tailored for TNBC patients. Conclusions: These findings establish FOXM1 as a clinically relevant regulator of immune evasion and immune memory in TNBC. Importantly, the identification of immunogenic peptides provides a strong rationale for the development of a peptide-based therapeutic vaccine, either as a monotherapy or in combination with immune checkpoint blockade, to improve long-term disease control in patients with TNBC.