Browse Articles

Decoding spatial tumor heterogeneity via multi-view radiomics for long-term survival prediction in advanced NSCLC immunotherapy.

e20643 Background: Current radiomics models for NSCLC immunotherapy primarily focus on short-term outcomes (e.g., 6–12 months' durable clinical response), neglecting the dynamic evolution of tumor ecosystems. We address this gap by innovatively integrating subregional heterogeneity parameters with multi-view subspace learning, enabling quantitative characterization of spatial intratumoral heterogeneity and extended survival prediction (12–30 months). Methods: A multicenter cohort of 414 stage IIIB-IV NSCLC patients receiving chemoimmunotherapy was analyzed. Baseline CT-derived tumor subregions (10 partitions via SLIC) provided 110 radiomic features/subregion. Four feature views—tumor core, peritumoral zone, subregional textures, and intratumoral heterogeneity (ITH)—were fused using subspace learning with adaptive constraints. Mutual information (MI) and random survival forests (RSF) optimized feature selection for PFS and OS prediction, respectively. An accelerated failure time model validated performance across 12-/18-/24-/30-month intervals. Results: The model demonstrated superior long-term predictive capacity, achieving 24-month OS AUC = 0.742(95%CI:0.698–0.786) and 30-month PFS C-index = 0.607 in external cohorts. Subregional heterogeneity parameters (original_gldm_GrayLevelVariance) outperformed conventional whole-tumor features, reducing cross-center variability by 15.3% (ITH vs. tumor-view models). Multi-view fusion captured synergistic spatial patterns: elevated tumor core heterogeneity (original_glcm_SumSquares) with disordered peritumoral textures (original_glszm_ZoneEntropy) identified high-risk patients (median OS = 14.2 vs. 28.6 months, p < 0.001). Conclusions: This study establishes the first radiomics framework leveraging subregional spatial heterogeneity for dynamic survival prediction beyond 2 years. Its clinical applicability is enhanced by stable performance across heterogeneous treatment protocols and scanner platforms, offering a practical tool to address immunotherapy resistance.

Real-world survival benefit of glucagon-like peptide-1 receptor agonists (GLP-1 Ras) concomitant with cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) and endocrine therapy in hormone receptor-positive (HR+)/HER2− metastatic breast cancer: A large propensity-matched analysis.

1070 Background: Obesity increases hormone receptor-positive (HR+) breast cancer risk through adipose-derived estrogens and inflammation. In metastatic HR+/HER2- disease, endocrine therapy (ET) combined with cyclin-dependent kinase 4/6 inhibitors (CDK4/6i) is standard first-line therapy, improving survival. The widespread use of glucagon-like peptide-1 receptor agonists (GLP-1 RAs) for diabetes and obesity management induces meaningful weight loss and may enhance ET efficacy by reducing body fat mass, which modulates estrogen levels and cancer-related inflammation, albeit their impact on mBC survival in patients receiving ET plus CDK4/6i remains unclear. Methods: Using the TriNetX Global Collaborative Network, we retrospectively identified two cohorts of patients with mBC receiving endocrine therapy (ET) plus a CDK4/6i: 26,689 patients treated with ET+CDK4/6i alone and 604 patients who also received a GLP-1 RA initiated within 3 months of CDK4/6i start. Propensity score matching (PSM) was applied to balance cohorts for age, race, body mass index, heart failure, hypertension, diabetes mellitus, fulvestrant use, and type of CDK4/6i. Overall survival (OS) was estimated using the Kaplan–Meier method, and hazard ratios (HRs) were calculated to compare OS between cohorts. Results: After PSM, 604 matched pairs of patients were identified in the ET plus CDK4/6i and GLP-1 RA groups, respectively (mean age +/- standard deviation:61.4 ±11.3 and 61.8 ±11.8 years), with well-balanced baseline characteristics. After a median follow-up of 18.8 months (interquartile range, IQR 27.9) in the non-exposed cohort and 15.8 months (IQR 24.5) in the non-exposure and in the exposure cohorts, median OS was 67,9 months among patients receiving GLP-1 RA compared with 49 months in those not receiving GLP-1 RA (HR 0.70; 95% CI,0.56–0.89; P=0.003), corresponding to a 30% reduction in the risk of death mortality risk. Conclusions: This is the largest real-world, propensity-matched analysis demonstrating a significant OS benefit with GLP-1 RAs added to ET+CDK4/6i in HR+/HER2− mBC. These provocative findings warrant prospective validation to elucidate underlying mechanisms—such as metabolic reprogramming or immune modulation—and explore GLP-1 RAs as a novel therapeutic strategy in this setting.

Atezolizumab (A) plus pertuzumab/trastuzumab/hyaluronidase (PHESGO) in patients (pts) with solid tumors with <i>ERBB2</i> alterations (alt): Results from the Targeted Agent and Profiling Utilization Registry (TAPUR) study.

2644 Background: TAPUR is a phase II basket study evaluating the antitumor activity of commercially available targeted agents in pts with advanced cancers with specific genomic alts. Results of a cohort of pts with solid tumors with ERBB2 alts treated with A+PHESGO are reported. Methods: Eligible pts had measurable disease, ECOG performance status (PS) 0-2, adequate organ function, and no remaining standard treatment (tx) options. Genomic testing was performed in CLIA-certified, CAP-accredited labs. Dosing for A was 1200 mg IV delivered every 3 weeks (wks). PHESGO was dosed every 3 wks, with a loading dose of 1200 mg/600 mg/30,000 units, then 600 mg/600 mg/20,000 units, until progression. Primary endpoint was disease control (DC) per investigator defined as objective response (OR) or stable disease (SD) of at least 16 wks duration (SD16+) per RECIST v.1.1. Simon 2-stage design tested null DC rate of 15% vs. 35% (power = 0.85; α = 0.10). If ≥2/10 pts in stage 1 had DC, cohort expanded to stage 2; otherwise, the cohort was closed. Cohorts closed prior to reaching the protocol-specified sample size of 28 used alternative thresholds set forth in the protocol to maintain the α level. For n=20, 6 pts had to have DC to reject the null (power = 0.74). Secondary endpoints were OR, progression-free survival (PFS), overall survival (OS), duration of response and SD, and safety. Results: The cohort expanded to stage 2 but closed before reaching the planned sample size. 23 pts with 6 tumor types (colorectal [CRC; 14], gallbladder [GB; 3], stomach [3], breast [1], pancreas [1], small intestine [1]) with ERBB2 amplification (amp; n=16), ERBB2 overexpression (n=2), ERBB2 mutation (mut; n=2), and ERBB2 amp and mut (n=3) were enrolled. 3 pts were not evaluable. 2 partial responses (both GB, ERBB2 amp) and 2 SD16+ (both CRC, ERBB2 amp) were observed for a DC rate of 25% (1-sided 90% CI, 10 to 100) and an OR rate of 10% (95% CI, 1 to 32). The null hypothesis was not rejected (p=0.26). 6 pts had tx-related grade 3 AE/SAEs: acute kidney injury, ALP increase, dehydration, diarrhea, infusion related reaction, lymphopenia, maculo-papular rash, pneumonitis and sepsis. Conclusions: A+PHESGO did not demonstrate sufficient antitumor activity in pts with ERBB2 -altered solid tumors to warrant further study. However, the cohort did not reach its planned accrual, limiting statistical power for demonstrating efficacy. Other tx should be considered for these pts, including tx offered in clinical trials. Clinical trial information: NCT02693535 . Demographics (N=23) and efficacy outcomes (n=20). Median (Med) age, years (range) 61 (43, 90) ECOG PS, No. (%) 0 10 (44) 1 11 (48) 2 2 (9) Prior systemic regimens, No. (%) 0-2 ≥3 16 (70)7 (30) DC (OR plus SD16+) rate, % (1-sided 90% CI), p-value 25 (10, 100), p=0.26 OR rate, % (95% CI) 10 (1, 32) Med PFS, wks (95% CI) 9 (8, 16) Med OS, wks (95% CI) 35 (16, 52)

Exosome whisperers: Decoding SPHK1's role in cisplatin-induced immune escape of NSCLC.

e20622 Background: Cisplatin, a key NSCLC chemotherapy, exhibits both cytotoxic and immunomodulatory effects. It influences tumor-derived exosomes (TDEs), critical carriers of bioactive molecules that reshape the tumor immune microenvironment (TIME) and promote immune evasion. While cisplatin is known to alter TDE secretion, contributing to an immunosuppressive TIME, the specific molecular mechanisms and functional consequences remain unclear. This study aims to elucidate how cisplatin reprograms exosomal cargo and its subsequent impact on tumor progression and immune escape in NSCLC. Methods: Transcriptome sequencing analyzed cisplatin's impact on RNA profiles in serum exosomes from NSCLC patients. Exosome biogenesis and secretion were assessed via immunofluorescence and automated fluorescence detection. RNA pull-down with mass spectrometry identified key mechanisms for SPHK1 mRNA loading into exosomes. Cisplatin's effect on the SPHK1/S1P pathway was evaluated by Western blot, qPCR, and ELISA, while ChIP-qPCR and dual-luciferase assays verified E2F1-mediated transcriptional regulation. SPHK1's role in proliferation and PD-L1 expression was tested via proliferation assays and Western blot. Exosome uptake was observed using FISH and immunofluorescence; flow cytometry and immunofluorescence assessed recipient cell changes. An orthotopic lung cancer mouse model treated with exosomes underwent single-cell sequencing to evaluate tumor immune microenvironment alterations. Results: Cisplatin induces dual pro-tumorigenic effects in NSCLC: it enhances exosome biogenesis and promotes STAU1-mediated sorting of SPHK1 mRNA into these vesicles. Concurrently, cisplatin transcriptionally upregulates SPHK1 via the transcription factor E2F1. Exosomes enriched with SPHK1 mRNA are internalized by recipient cancer cells, driving proliferation and upregulating PD-L1 expression. In a lung cancer mouse model, treatment with exosomes from SPHK1-knockdown cells (sh exo) inhibited tumor growth and remodeled the tumor immune microenvironment. Single-cell analysis revealed sh exos increased cytotoxic CD8⁺ T and NK cell infiltration while reducing Tregs and exhausted T cells, accompanied by enhanced expression of T-cell recruitment (CXCR6), cytokine signaling (Il2rg), and cytotoxic effector molecules (IFN-γ, GzmB, Pfn1). Conclusions: Cisplatin simultaneously triggers an exosome-mediated signaling axis that may counteract its efficacy by promoting tumor growth and fostering an immunosuppressive microenvironment through PD-L1 induction.

Multi-modal AI modeling to predict clinical trial outcomes and benchmarks pan-RAS versus KRAS G12D inhibition in KRAS G12D–mutant PDAC.

4020 Background: In pancreatic ductal adenocarcinoma (PDAC), uncertainty persists regarding the optimal positioning of pan-RAS inhibitors, allele-specific KRAS inhibitors, and chemotherapy backbones. While patient-derived organoids (PDOs) provide functional resolution of drug response, their clinical utility has been limited by cohort size and insufficient validation against trial outcomes. Methods: A biobank of 135 colorectal cancer (CRC) and PDAC PDOs was assembled with matched clinical annotation, omics characterization, and functional drug profiling. Multi-modal AI models integrating molecular features, PDO drug response, and clinical covariates were trained to predict patient-level progression-free survival (PFS) and longitudinal response, and to aggregate these predictions into cohort-level clinical trial outcomes. Model performance was evaluated by benchmarking predicted outcomes against published results across 22 treatment arms from multiple clinical trials. A pan-RAS inhibitor (daraxonrasib), a KRAS G12D-directed inhibitor (zoldonrasib), and FOLFIRINOX (FFX) were compared as single agents or in combination in a KRAS G12D-mutant PDAC subset (n = 23), with external benchmarking to published clinical data. Results: The AI models demonstrated robust predictive performance at both the patient and cohort levels. Patient-level predictions were driven by functional PDO drug response and established clinical factors, including performance status and prior treatment exposure. At the cohort level, predicted and published outcomes showed good agreement (R² = 0.63), with correct ranking of treatment arms by efficacy in 78% of cases. In pairwise arm-to-arm comparisons, the model identified the superior regimen in 80% of comparisons. When applied to KRAS G12D-mutant PDAC, virtual clinical trial simulations preserved the relative efficacy hierarchy observed in external datasets. Zoldonrasib showed heterogeneous predicted activity (simulated ORR 20% vs reported 30%), whereas daraxonrasib demonstrated more favorable predicted outcomes (simulated ORR 35% vs reported 36%). Combination with FFX further improved predicted efficacy for zoldonrasib, particularly in first-line simulations (simulated ORR 38% vs reported 63%). Conclusions: Multi-modal AI modeling trained on clinically annotated PDO collections enables calibrated simulation of clinical trial outcomes and comparative assessment of therapeutic strategies in KRAS-mutant PDAC. By capturing patient-level heterogeneity and cohort-level efficacy trends, this framework provides a scalable decision-support tool for arm selection, therapeutic positioning, and trial design in KRAS-directed drug development.

Trastuzumab deruxtecan (T-DXd) in patients (pts) with <i>HER2</i> -mutant (HER2m) non-small cell lung cancer (NSCLC) and central nervous system (CNS) metastases (mets): Results from DESTINY-Lung05 (DL-05).

e20500 Background: T-DXd (5.4 mg/kg) is approved in multiple countries for pretreated unresectable/metastatic HER2m NSCLC; conditional approval in China was supported by DESTINY-Lung02 (DL-02) and DL-05 results. In DL-02, T-DXd showed encouraging CNS activity in this population. Here, we report a post-hoc analysis of DL-05 assessing T-DXd in pts from China with HER2m NSCLC with/without baseline (BL) CNS mets. Methods: DL-05 (NCT05246514), an open-label, single-arm, Phase 2 study, evaluated T-DXd (5.4 mg/kg Q3W IV) in pts from China with metastatic HER2m NSCLC with disease progression on/after ≥1 prior anticancer therapy. Pts with pretreated asymptomatic / neurologically stable CNS mets were eligible. Brain CT or MRI scans were performed in all pts at BL and study end. Pts with CNS mets had brain scans every 6 weeks from enrollment date to 48 weeks, then every 9 weeks until RECIST 1.1-defined progressive disease. Additional brain scans were carried out as clinically indicated. In pts with/without BL CNS mets, systemic efficacy (confirmed objective response rate [cORR], disease control rate [DCR], duration of response [DOR], and progression-free survival [PFS] by independent central review [ICR] per RECIST 1.1; and overall survival [OS]), CNS-PFS by ICR per CNS-modified RECIST 1.1, and safety were assessed. CNS-cORR, CNS-DCR, and CNS-DOR were evaluated in pts with measurable BL CNS mets. Results: At data cutoff (November 4, 2024), 72 pts had received T-DXd; 30 pts had BL CNS mets, including 10 with CNS-measurable disease. Median (range) duration of follow up for all pts was 20.2 (2–27) months. In pts with and without CNS mets, median (range) treatment duration was 7.6 (0.7–26.9) months and 10.5 (0.7–26.6) months, respectively. Systemic efficacy data and CNS-PFS are in the Table. In pts with measurable CNS mets (n=10), CNS-cORR was 40.0% (95% CI 12.2, 73.8; n=4, one complete response was reported), CNS-DCR was 100% (95% CI 69.2, 100), and median CNS-DOR was not evaluable (NE). Drug-related Grade ≥3 adverse events were reported in 60.0% (n=18) and 52.4% (n=22) of pts with and without CNS mets, respectively. Conclusions: T-DXd showed antitumor activity in pts from China with metastatic HER2m NSCLC with and without BL CNS mets; consistent with previous analyses, promising CNS activity was observed. No new safety signals were reported. Results support T-DXd use in pretreated pts with HER2m NSCLC, including those with CNS mets. Clinical trial information: NCT05246514 . BL CNS mets(n=30) No BL CNS mets(n=42) cORR, % (n)*95% CI 46.7 (14)28.3, 65.7 64.3 (27)48.0, 78.4 DCR, % (n)*95% CI 90.0 (27)73.5, 97.9 92.9 (39)80.5, 98.5 Median DOR, months (95% CI)* 7.1 (5.3, 15.3) 14.0 (5.8, NE) Median PFS, months (95% CI)* 8.0 (5.8, 16.5) 13.0 (7.2, NE) Median CNS-PFS, months (95% CI) 7.8 (5.6, 9.9) NE (NE, NE) Median OS, months (95% CI)* 16.6 (13.3, 21.7) 24.0 (19.9, NE) *Systemic assessment.

Patient preferences for duration of adjuvant trastuzumab for HER2+ breast cancer.

e12518 Background: A recent meta-analysis of reduced duration adjuvant trastuzumab trials demonstrated that 6 months was noninferior to 12 months for early HER2+ breast cancer. However, US oncologists usually prescribe 12 months despite permissive NCCN guidelines indicating “up to 12 months.” Given apparent clinical equipoise, this treatment option could be an exemplar for patient choice and shared decision-making (SDM) in the clinic. Thus, we elicited patient preferences on adjuvant trastuzumab duration when presented with data from the meta-analysis and the PERSEPHONE trial as infographics with either point or range estimates of efficacy and toxicity. Methods: We surveyed women with breast cancer from December 2024 to June 2025 who were diagnosed within 5 years or receiving active treatment at the University of Chicago. Surveys presented an infographic that was developed through patient feedback and included 5-year disease-free (DFS) and overall survival (OS) from the meta-analysis for 6 vs 12 months. Surveys also presented data on cardiotoxicity causing trastuzumab to be discontinued permanently and severe fatigue during treatment, both from PERSEPHONE data. Out-of-pocket (OOP) cost estimates were based on simulations (Table). Patients were randomized to see efficacy and toxicity attributes as either point or range estimates. Patients indicated their binary choice for 6 or 12 months and their percent chance of choosing 6 months. Multivariable regression models were used to adjust for clinical factors. Results: Of 208 respondents, 65% had Stage I / II disease, 22% were HER2+, and 6% were receiving trastuzumab at the time of the survey. Most patients (70%) preferred 6 months in the discrete choice, and the median probability of choosing 6 months was 80% (IQR 50-95%). While patients did not differ in their percent chance of choosing 6 months when presented with point vs range estimates, patients who saw the range estimate had higher odds of making the binary choice of 6 months (aOR 2.27, p=0.01). Conclusions: When reviewing a patient-friendly infographic of trial data, most patients chose 6 months of adjuvant trastuzumab rather than 12 months. This suggests that an absolute difference of 1% in DFS and OS was not large enough for patients to prefer 12 months. This preference was even stronger when the infographic included uncertainty around efficacy and toxicity through range estimates. Oncologists wishing to actively engage in SDM with their patients could use this preference elicitation methodology to ensure that adjuvant trastuzumab duration aligns with patients’ preferences and values. Infographic attributes. Point Estimates Range Estimates 6 Months 12 months 6 Months 12 Months 5-Year DFS 85% 86% 84%-86% 85%-87% 5-Year OS 92% 93% 91-93% 92-94% Heart problems-trastuzumab stopped 3% 8% 2-4% 7-9% Severe Fatigue (during treatment) 9% 12% 8-10% 11-13% OOP Cost (estimated) $4,000 $8,000 $4,000 $8,000 Number of Appointments 9 18 9 18

A phase II study of the DNA plasmid-based vaccine STEMVAC in patients with metastatic triple-negative breast cancer (mTNBC).

TPS2698 Background: Adding immune checkpoint inhibitors for the treatment of mTNBC improves the efficacy of traditional chemotherapy in select populations. In populations who are not predicted to benefit from PD-1 checkpoint inhibitors, a vaccine may sensitize Th1 cells against cancer-associated antigens to facilitate immune-mediated tumor killing. STEMVAC is a safe, immunogenic plasmid DNA-based vaccine encoding T-helper 1 (Th1) selective epitopes from five antigens (MDM2, YB1, SOX2, CDH3, CD105) associated with breast cancer stem cells and the epithelial-mesenchymal transition. In addition to direct cytotoxic effects directed against cancer cells, we have demonstrated that vaccine-induced Th1 cells secrete IFN-γ, which upregulates suppressor of cytokine signaling 1 (SOCS1), leading to slowed tumor growth and increased vulnerability to cytotoxic agents. Combining STEMVAC with chemotherapy could thus have synergistic effects leading to enhanced cancer sensitivity to treatment and possible eradication. Methods: The study (NCT07078604) includes adults with PD-L1 negative mTNBC receiving standard therapies inclusive of chemotherapy, antibody-drug conjugates, and oral PARP inhibitors in the first and second line. Eligible patients will have histologically-confirmed mTNBC, radiographically measurable disease by RECIST v1.1, and a lesion amenable to biopsy. Key exclusion criteria include concomitant B-cell malignancies, ongoing systemic steroid use, and known hypersensitivity reaction to GM-CSF. Participants receive intradermal STEMVAC 300mcg with 100mcg of GM-CSF given during the nadir period of standard of care chemotherapy. Participants receive three priming vaccines given every 21-28 days concurrent with chemotherapy, followed by booster doses at 6 and 9 months then every six months thereafter. The primary endpoints are 1) observance of immunogenicity to one of the five vaccine antigens, measured by interferon-gamma enzyme-linked immunospot assay, and 2) safety. Secondary endpoints include objective response rate, real-world PFS2, overall survival, and magnitude of immunogenicity. Exploratory endpoints include quantitative levels of CD8+ infiltrating lymphocytes and expression of genes associated with epithelial-mesenchymal transition before and after vaccine priming. Current status: This trial is recruiting patients, with up to 20 patients expected to be recruited across multiple sites. Safety data and preliminary outcomes will be reported as they become available. This trial aims to evaluate whether STEMVAC would achieve clinically relevant immunogenicity when administered concurrently with chemotherapy for the treatment of mTNBC. The trial will evaluate the safety and preliminary efficacy of this novel combination and explore tumor-immune microenvironment biomarkers to gain mechanistic insight. Clinical trial information: NCT07078604 .

Comparing line of therapy selection methods in advanced non–small cell lung cancer (aNSCLC): A simulation study.

e23357 Background: External control arms utilizing real-world patients are increasingly used with single-arm clinical trials to demonstrate drug effectiveness in oncology. These patients often have multiple eligible lines of therapy (LoTs), the choice of which can meaningfully impact conclusions about effectiveness. Current LoT selection approaches include using the first eligible LoT (FEL), a random eligible LoT (REL), all eligible LoTs (AEL), and a novel method, stratified random LoT (SRL). We compare SRL to current selection approaches to identify an optimal approach in aNSCLC. Methods: Method performance was evaluated in a simulation study emulating trial-like EGFR+ aNSCLC patients and external controls for overall survival. Primary analyses were based on synthetic cohorts with typical prior LoT distribution between arm (Δ≈2), moderate treatment effect (HR≈0.73), and robust sample size (n≈750 per arm). We investigated the impact of varying LoT overlap, sample size, and effect size. Methods were evaluated on their ability to estimate effectiveness in naïve and weighted analyses using root mean square error, bias, and coverage. Results: Across primary analyses, SRL outperformed FEL with lower error and higher coverage, and performed comparably to or better than AEL. SRL more often overestimated treatment effects, whereas AEL tended to underestimate treatment effects. REL showed high error and bias, and low coverage, insufficiently addressed by weighting. SRL remained robust with smaller samples, null effects, and limited LoT overlap (Δ≈3); FEL and AEL degraded as overlap decreased. Conclusions: SRL demonstrated robust performance and unique value in disease settings where LoT strongly confounds clinical outcomes. Optimal LoT selection depends on therapeutic area and study context. Evaluation of LoT selection methods should incorporate multiple performance metrics; low bias alone may obscure poor inferential performance. Performance metrics. Primary Smaller Sample Size No Treatment Effect Less LoT Overlap Weighting approach Method HR RMSE Bias Coverage HR RMSE Bias Coverage HR RMSE Bias Coverage HR RMSE Bias Coverage SMR FEL 0.75 0.14 0.01 47% 0.81 0.19 0.08 44% 1.01 0.19 0.01 48% 0.98 0.91 0.24 17% REL 0.63 0.11 -0.10 32% 0.69 0.08 -0.05 73% 0.86 0.16 -0.15 26% 0.72 0.09 -0.02 67% AEL 0.77 0.06 0.03 85% 0.79 0.08 0.05 78% 1.05 0.08 0.04 86% 0.79 0.08 0.05 76% SRL 0.71 0.06 -0.03 84% 0.68 0.08 -0.05 72% 0.98 0.07 -0.02 88% 0.71 0.06 -0.03 82% sIPTW FEL 0.72 0.09 -0.01 66% 0.72 0.10 -0.02 61% 0.99 0.12 -0.01 66% 0.83 0.46 0.09 27% REL 0.60 0.14 -0.13 11% 0.59 0.15 -0.14 10% 0.82 0.19 -0.18 10% 0.66 0.10 -0.08 50% AEL 0.77 0.06 0.04 77% 0.74 0.05 0.01 86% 1.05 0.08 0.05 77% 0.80 0.08 0.06 66% SRL 0.69 0.06 -0.04 83% 0.67 0.08 -0.06 72% 0.96 0.08 -0.04 89% 0.70 0.07 -0.04 85% HR: Hazard Ratio; RMSE: Root Mean Square Error; sIPTW: Stabilized Inverse Probability of Treatment Weighting; SMR: Standardized Mortality Ratio.

Effects of malnutrition on in-hospital outcomes among young adults hospitalized with gastrointestinal cancers: A multicentric retrospective cohort study.

e16490 Background: Protein–energy malnutrition (PEM) is common in gastrointestinal (GI) cancers and may worsen inpatient outcomes. Contemporary national data describing the impact of PEM among young adults with GI malignancies are limited. Methods: We conducted a retrospective cohort study using the all-payer database of HCUP-NIS between 2018–2021. Young adults aged 18–39 years hospitalized with GI cancers were identified using ICD-10-CM malignant neoplasm codes C15–C26. PEM was identified using diagnosis codes for protein–energy malnutrition. Outcomes included in-hospital mortality, length of stay (LOS), total hospital charges, and discharge disposition. Survey-weighted analyses incorporating NIS discharge weights, hospital clustering, and stratification were used to generate national estimates. Multivariable survey-weighted logistic regression used to evaluate mortality and discharge disposition; multivariable survey-weighted linear regression evaluated LOS and charges, adjusting for age, sex, race/ethnicity, primary payer, ZIP-income quartile, elective status, calendar year, and age-adjusted Charlson Comorbidity Index. Results: The study included 58,910 weighted hospitalizations of young adults with GI cancers, of whom 11,915 (20.2%) had PEM. Compared with patients without PEM, those with PEM had higher comorbidity burden and acuity, including higher prevalence of metastatic disease (71.8% vs 53.1%). PEM was also associated with higher odds of in-hospital death (aOR 2.11 p &lt; 0.001. Discharge disposition also differed by PEM status; admissions with PEM had lower odds of discharge to home/home health versus non-home discharge (aOR 0.60; 95% CI 0.53–0.68; p &lt; 0.001). Unadjusted outcomes were worse among PEM admissions, including higher in-hospital mortality (8.4% vs 3.2%), longer LOS (10.43 vs 5.63 days), and higher total charges ($133,790 vs $83,702). In adjusted analyses, absence of PEM (No PEM) was associated with substantially lower inpatient resource utilization: LOS was 4.49 days shorter (β −4.492; SE 0.248; p &lt; 0.001) and total charges were $53,513 lower (β −$53,512.6; SE $5,527.6; p &lt; 0.001) compared with PEM. Conclusions: Among young adults hospitalized with GI cancers, PEM was present in approximately one in five admissions and was independently associated with markedly higher in-hospital mortality, longer LOS, and substantially higher hospital charges. These findings support routine inpatient nutritional risk screening and early nutrition-directed interventions as potential targets to improve outcomes and reduce resource utilization in this high-risk population.

Longitudinal assessment of circulating tumor DNA as a disease monitoring biomarker in soft tissue sarcoma.

e23523 Background: Soft tissue sarcomas (STS) comprise a heterogeneous group of malignancies with limited biomarkers for real-time disease monitoring. Circulating tumor DNA (ctDNA) represents a minimally invasive approach to assess tumor burden and treatment response; however, its clinical utility in STS remains incompletely defined. We evaluated ctDNA dynamics across diverse STS subtypes in a real-world clinical cohort. Methods: We retrospectively analyzed plasma-derived ctDNA from 57 patients with STS enrolled in a prospective EHR-based Comprehensive Bone and Soft Tissue Tumor Registry (NCT02677961) at The Ohio State University Comprehensive Cancer Center. ctDNA was assessed using Tempus (n = 34) and Signatera (n = 23) platforms. ctDNA levels were compared with radiographic tumor volume, histologic grade, and disease stage. Tumor burden was quantified using volumetric imaging, and clinical response was assessed by RECIST criteria or physician evaluation. Serial ctDNA measurements were available for a subset of patients. Results: ctDNA was detectable in over 75% of patients. ctDNA levels strongly correlated with tumor volume in both the Signatera (Spearman r = 0.84, p = 2.04 × 10⁻⁹) and Tempus (Spearman r = 0.63, p = 0.0375) cohorts. ctDNA detection was significantly associated with higher histologic grade (p = 0.0217), with Grade 3 tumors demonstrating the highest detection rates. ctDNA presence was not associated with surgical stage (p = 0.666). Among 20 patients with serial sampling, ctDNA dynamics closely mirrored changes in tumor burden during surgery, chemotherapy, and radiation. The sensitivity of ctDNA detection in patients with radiographically evident disease was 79.2%, and in 82% of cases, increases in ctDNA levels preceded radiographic evidence of disease progression across histological subtypes. Conclusions: In this real-world STS cohort, ctDNA levels were associated with tumor burden, histologic grade, and longitudinal treatment response across multiple sarcoma subtypes and assay platforms, with ctDNA increases frequently preceding radiographic disease progression. These findings support ctDNA as a promising biomarker for real-time disease monitoring in STS and provide a rationale for incorporation into prospective sarcoma clinical trials. Larger, multi-institutional studies are warranted to validate these observations.

Survival outcomes following resistance to immune checkpoint inhibition in advanced cutaneous squamous cell carcinoma.

9588 Background: Immune checkpoint inhibitors (ICI) are standard first-line therapy for advanced cutaneous squamous cell carcinoma (cSCC); however, 35% of patients experience disease progression. Evidence guiding optimal second-line management remains limited. We evaluated real-world outcomes of second-line treatment strategies in patients with cSCC resistant to ICI. Methods: A retrospective analysis of patients with unresectable locally advanced or metastatic cSCC treated with ICI monotherapy at two tertiary centres in Brisbane, Australia, was conducted. Patients with progressive disease (PD) were identified and classified as having primary resistance (PD &lt; 4 months) or acquired resistance (PD ≥4 months). Second-line strategies included local therapy (LT - surgery and/or radiotherapy), systemic therapy +/- LT or best supportive care (BSC). Outcomes included progression-free survival from second-line initiation (PFS2), and overall survival (OS) using Kaplan–Meier estimates. Results: Seventy-four patients were identified. Median age was 74 years. Most patients had locoregionally advanced disease (77%), head and neck primary tumours (70%), ECOG performance status 0–1 (76%), and were immunocompetent (70%). Primary resistance occurred in 54% and acquired resistance in 46%, of which 21% progressed after ICI cessation. Locoregional progression was observed in 70% of patients, while 30% progressed with distant disease. Management of ICI resistance included LT (N = 32), systemic therapy +/- LT (N = 16), and BSC (N = 26). An objective response rate of 65% was observed in patients undergoing systemic therapy +/- LT with responses to both ICI rechallenge and/or EGFR targeted therapy (7 of 11 patients). With a median follow up of 46.6m the Median PFS2 was 10.7m. PFS2 was numerically higher in patients undergoing systemic therapy vs LT alone (11.9m vs 8.4m). Of those that underwent LT alone, further PD was seen in 16 patients (50%), with half requiring systemic therapy. Median OS for the overall cohort was 19.0m. OS was numerically reduced in immunocompromised patients (11.6m) and in those with primary ICI resistance (6.9m) vs those who had acquired resistance (22.5m) and progression after ICI cessation (47.1m). Conclusions: Selected patients with advanced cSCC, particularly those with acquired resistance, derive meaningful benefit from second-line therapy following ICI progression. Outcomes remain poor for immunosuppressed patients or those with primary resistance, highlighting a major unmet clinical need and the importance of prospective studies in the post-immunotherapy setting.

Beyond detection: Five-year screening behaviors and patient-reported perceptions in DETECT-A participants.

10534 Background: Multi-cancer early detection (MCED) blood testing may improve early cancer detection. The impact of MCED testing on anxiety and standard of care (SoC) cancer screening adherence is not fully understood. Detecting cancers Earlier Through Elective mutation-based blood Collection and Testing (DETECT-A) enrolled 10,006 women aged 65 to 75 years from the Geisinger Health System (GHS) with no previous history of cancer and assessed a single instance of MCED (CancerSEEK) testing in 9,911 subjects. Herein , we report cancer survey results, and SoC screening adherence over a 5-year follow-up period (5YFU). Methods: Medical record review and participant surveys administered at prespecified intervals, including approximately 5 years after enrollment, were used to assess patient experience, anxiety, and commitment to SoC screening. Decision regret/participation was assessed using a 5-item, 5-point agreement scale; perceived impact on anxiety, depression, and intention to adhere to SoC screening were assessed using 3-option directional responses (more/less/no change), with a ‘prefer not to answer’ option. Observed SoC screening adherence was assessed through August 2024. SoC screening eligibility (U.S. Preventive Services Taskforce criteria) and adherence data were available for DETECT-A subjects. Changes in adherence over time were assessed using a mixed effects model. Results: 8,114 subjects completed all survey questions. 97.3% expressed no participation regret; 94.8% noted a willingness to participate again. Most participants reported no changes in anxiety (89.4%) or depression (96.6%) levels since enrollment. 98.7% reported an equal or greater likelihood of adhering to SoC screening. SoC screening adherence was assessed for a subset of DETECT-A subjects who had at least one month of active GHS patient status (Breast and CRC) and were also eligible for screening (Lung, Breast, CRC). DETECT-A subject SoC adherence was high at consent (Lung: 28.0%, Breast: 88.4%, CRC: 73.9%) and remained high at one year (Lung: 38.5%, CRC: 82.3%) or two-year (Breast: 87.7%) follow-up post consent. Adherence remained high throughout the follow-up period. Conclusions: DETECT-A subjects were satisfied with study participation and reported little change in anxiety and depression. Reported commitment to SoC screening adherence was high and observed adherence to recommended screening modalities was maintained over the follow-up period. In a controlled setting of a prospective clinical trial, MCED testing did not contribute to anxiety or depression and did not interfere with adherence to SoC screening.

A single-arm, phase II study of sequential therapy with curative intent in de novo HER2+ metastatic breast cancer: The SAPPHO study.

TPS1147 Background: Due to dramatic improvements in neoadjuvant and adjuvant HER2-directed therapy (tx), most patients (pts) with early HER2+ breast cancer are cured. As a result, over half of pts newly diagnosed with HER2+ metastatic breast cancer (MBC) now present with de novo stage IV disease. Anti-HER2 tx has also significantly extended survival for pts with HER2+ MBC, with a subgroup of exceptional responders alive many years (yrs) after diagnosis. However, the paradigm for HER2+ MBC remains non-curative, and pts receive tx indefinitely with significant toxicities and costs. The SAPPHO study is investigating whether an intensification approach of sequential, non-cross resistant anti-HER2 tx followed by tx discontinuation is associated with long-term disease control in pts with HER2+ MBC. Methods: SAPPHO is an open-label, phase II, single-arm trial testing a sequential regimen of non-cross resistant, HER2-targeted tx with curative intent in pts with de novo HER2+ MBC. Eligible pts must have biopsy-proven, de novo MBC with high HER2 expression (3+ by immunohistochemistry). Pts with brain metastases are eligible upon receipt of local tx. Treatment consists of an induction regimen (trastuzumab-pertuzumab-taxane [THP] x 4 cycles, followed by trastuzumab deruxtecan [TDXd] x 6 cycles, followed by trastuzumab emtansine [TDM1]-tucatinib x 4 cycles), followed by a maintenance regimen (HP-tucatinib for 1 yr). Given results from DESTINY-Breast09, the sequence of TDXd-P x 6 cycles followed by THP x 4 followed by TDM1-tucatinib x 4 can be chosen as alternative induction tx per investigator and patient choice. Pts who remain progression-free after completing maintenance will stop all anti-HER2 tx. Endocrine tx will be continued for pts with hormone receptor+/HER2+ tumors. Tumor specimens from breast and a metastatic site are collected at baseline and the end of induction. Serial plasma samples for ctDNA analysis are collected at baseline, during treatment, and during follow-up. The primary endpoint is the probability of being progression-free and off anti-HER2 tx 4 yrs from the start of induction. With a sample size of 72 pts, the study is designed to have 91% power to reject the null hypothesis that the probability of being off anti-HER2 tx and progression-free is less than 24%, with an alternative hypothesis of 40%. Key secondary endpoints are overall survival, overall response rate by modified RECIST 1.1 after induction, and safety. Correlative endpoints include the relationship between ctDNA dynamics and outcomes. Patient-reported outcomes will be analyzed, including quality of life, illness intrusiveness, financial toxicity, anxiety, distress about cancer progression, perception of benefit and risk of progression. SAPPHO began enrollment in Q3 2024, and the study is open at 4 US sites within the Translational Breast Cancer Research Consortium (NCT05721248). Clinical trial information: NCT05721248 .

Determinants of lung metastasis in pediatric and young adult differentiated thyroid cancer: A meta-analytic study.

e18107 Background: Distant metastases from differentiated thyroid cancer (DTC) occur more frequently in children and young adults (CAYA) than in older adults, with the lung as the most common metastatic site. This meta-analysis provides a comprehensive assessment of factors associated with lung metastasis at diagnosis, during follow-up, and lung metastasis–free survival in CAYA with DTC. Methods: This study was registered in PROSPERO. A systematic search of PubMed, Scopus, OVID, Cochrane, and Web of Science was conducted in February 2024. Retrospective, prospective, and case-control studies evaluating lung metastasis in patients aged ≤21 years with DTC were included. Lung metastasis was confirmed by imaging and/or pathology. Two reviewers independently performed study selection and data extraction. Random-effects meta-analyses evaluated demographic, pathological, and molecular predictors. Results: Thirty-two studies were included. At diagnosis, younger age, advanced T stage, nodal involvement, lateral lymph node metastasis, extrathyroidal extension (ETE), and multifocality were significantly associated with lung metastasis. Wild-type BRAF status was associated with a higher risk of lung metastasis compared with BRAF V600E . During follow-up, male sex, nodal disease, lateral lymph node metastasis, and ETE were significant predictors. Age ≤15 years was associated with worse lung metastasis–free survival. Conclusions: Nodal disease, ETE, advanced T stage, younger age, and multifocality are key predictors of lung metastasis in CAYA with DTC, while BRAF V600E mutation status appears protective. Factors predicting lung metastasis at diagnosis. No. of studies No. of patients OR/mean difference 95% CI P- value I 2 Male (vs. female) 11 1177 1.28 [0.87, 1.87] 0.21 18% Average age (years) 8 490 -1.65 [-2.46, -0.84] 0.0001 0% Age ≤10 years (vs. &gt;10) 9 1558 2.35 [1.72, 3.22] 0.00001 0% Age ≤15 years (vs. &gt;15) 8 1394 2.50 [1.56, 4.02] 0.0002 12% Papillary thyroid cancer (vs. follicular) 7 716 1.52 [0.59, 3.89] 0.39 20% T3/T4 (vs. T1/T2) 7 345 2.86 [1.45, 5.64] 0.002 0% pN1(vs. pN0) 8 984 4.14 [2.68, 6.39] 0.00001 0% Lateral lymph node (vs. no) 3 256 10.70 [3.71, 30.90] 0.0001 0% Average tumor size (cm) 4 617 0.70 [-0.07, 1.47] 0.070 72% ETE (vs. no) 3 792 2.55 [1.80, 3.61] 0.00001 0% Multifocal (vs. unifocal) 3 518 2.42 [1.56, 3.76] 0.0001 0% Wild type BRAF (vs. mutation) 6 419 5.71 [2.05, 15.87] 0.0008 0% Lung Metastasis on follow-up Male (vs. female) 6 549 2.39 [1.30, 4.41] 0.005 0% Average age (years) 3 211 -1.45 [-3.57, 0.67] 0.180 67% Papillary thyroid cancer (vs. follicular) 5 487 2.07 [0.60, 7.13] 0.250 0% T3/T4 (vs. T1/T2) 4 313 2.62 [0.81, 8.52] 0.110 15% pN1(vs. pN0) 5 479 3.89 [1.72, 8.83] 0.001 0% cN+(vs cN0) 2 246 7.02 [3.42, 14.40] 0.00001 0% Lateral lymph node (vs. no) 3 287 3.48 [1.79, 6.76] 0.0002 0% ETE (vs. no) 3 284 12.6 [4.74, 33.85] 0.00001 0% Total thyroidectomy (vs no) 3 281 2.46 [0.31, 19.55] 0.400 67%

A Korea–USA multicenter study of duodenal pancreatic fluid–derived exosomal CES1 and MAN2A1 as pancreas-proximal biomarkers for early pancreatic cancer detection.

10552 Background: Small pancreatic lesions often yield limited tissue via endoscopic ultrasound (EUS)-guided biopsy, and serum CA19-9 provides poor discrimination in high-risk settings. Duodenal pancreatic fluid (DPF), obtained during routine endoscopy/EUS, is a pancreas-proximal, low-noise biofluid that may enrich tumor-derived exosomal signals. We evaluated whether DPF exosomal proteins complement serum CA19-9 to discriminate PDAC from controls. Methods: DPF was collected during upper endoscopy/EUS with or without secretin. DPF-derived exosomes (n = 26) were profiled by LC-MS/MS, and CES1/MAN2A1 were prioritized by stage-pattern clustering. Candidates were quantified by ELISA in a Korea–USA multicenter cohort (n = 123, prespecified to detect a target AUC of 0.94; 63 PDAC, 60 controls; 4 Korean sites and 1 US site, combined with serum CA19-9 using a locked stacked ensemble model with LightGBM meta-learner trained only on out-of-fold predictions). High-risk controls included IPMN and chronic pancreatitis. Orthogonal validation used pancreatic juice from surgery or ERCP (PDAC n = 20; chronic pancreatitis n = 20). The primary endpoint was AUC for PDAC versus controls, compared with CA19-9 alone and in the CA19-9–normal subgroup; performance was also assessed in early-stage PDAC versus high-risk controls. 95% confidence intervals (CIs) were estimated by cross-validated resampling. Results: LC-MS/MS identified 2,460 proteins; 130 were differentially expressed (&gt; 2-fold, p &lt; 0.05). ELISA confirmed higher exosomal CES1 and MAN2A1 in PDAC versus controls (p &lt; 0.001). The multi-marker model achieved an AUC of 0.995 (95% CI 0.982–1.000), a sensitivity of 98.33% (95% CI 0.931–1.000), and a specificity of 100% (95% CI 0.925–1.000), outperforming CA19-9 alone (AUC 0.63). In patients with CA19-9 in the normal range (&lt; 37 U/mL), the model maintained high discrimination (AUC 0.931, 95% CI 0.746-1.000) with 100% sensitivity. For early-stage PDAC (stage I–II, n = 23) versus high-risk controls (n = 20, including IPMN), AUC was 0.976 (95% CI 0.924-1.000). In orthogonal validation, CES1/MAN2A1 in pancreatic juice were elevated in PDAC patients (p &lt; 0.001) but not in matched serum, supporting pancreas-proximal enrichment. Decision curve analysis showed a higher net benefit than CA19-9 across clinically relevant thresholds. Conclusions: A DPF exosome–derived CES1/MAN2A1 signature integrated with CA19-9 using a locked ensemble model showed high diagnostic discrimination and supportive biological plausibility across a Korea–USA multicenter cohort. These results suggest a pancreas-proximal exosomal biomarker that addresses a critical gap in early detection of pancreatic cancer. Clinical trial information: NCT07030348 .

Cycling of Fe(II)/Fe(III) in the euphotic zone mediated by mineral photocatalysis

Improving availability and quality of serious illness documentation using large language models.

e13693 Background: Timely access to prior documented serious illness conversations (SICs) is critical for inpatient clinicians caring for patients with advanced cancer whom they may be unfamiliar with. Despite widespread implementation of structured electronic health records (EHR) modules intended to capture SICs, structured documentation remains sparse, with most SIC content embedded within large volumes of free-text clinical notes. We evaluated whether large language models (LLMs) could synthesize longitudinal clinical notes into actionable SIC summaries. Methods: The BRIDGE-SIC pilot clinical trial enrolled 58 patients with solid tumor admitted to an academic hospital with a predicted 90-day mortality ≥40%. For each patient, HIPAA-compliant LLM prompts identified and summarized SIC content from the prior 6 months of EHR records for 5 predefined SIC domains (illness understanding, prognosis, hopes, worries, and social supports). In this post-hoc analysis, Epic SIC module content was reviewed, and their completeness, word count, and date of last update were compared to those of the LLM summaries. A qualitative comparison was also conducted. Results: The LLM summaries identified SIC documentation within free-text clinical notes for 4 of 5 SIC domains for all patients (n = 58, 100%) and 5 of 5 for 87.9% of patients (n = 51). LLM summaries captured a mean of 4.9 (Standard Deviation = 0.33) of 5 domains and a mean of 16.1 (SD = 2.6) bullet points per patient across 5 SIC domains. In contrast, only 8 patients (13.8%) had any Epic SIC module entries; of those, only a mean of 2.8 (SD = 1.3) of 5 domains were filled. These entries typically lacked longitudinal context or detailed goals and contained a mean of 129.1 (SD = 120.3) words compared to the mean of 582.2 (SD = 163.7) words for the LLM summaries. Compared to SIC modules, LLM summaries included data that was a mean of 66.6 (SD = 129.1) days more recent. Qualitatively, LLM summaries provided time-ordered syntheses integrating patient priorities, symptom burden, and family context, whereas SIC module entries, when present, contained minimal information that reflected a fixed timepoint. Conclusions: LLMs can securely synthesize longitudinal free-text notes into concise SIC summaries that substantially outperform existing SIC modules in completeness, timeliness, and utility. Application of LLM SIC summaries in clinical care may improve goal-concordant decision-making for hospitalized patients. Clinical trial information: NCT07147023 . Illustrative comparison of illness understanding documented in the Epic SIC module vs. LLM summary for a patient. SIC module [date] Aggressive progression on TNBC. LLM summary [date] Patient understands cancer progression with new metastases in liver, bone, and lungs. Aware of current treatment (capecitabine) ineffectiveness and new treatment (eribulin) plan. Cognizant of potential side effects and limited prognosis if new treatment is ineffective.

Cost-effectiveness of first-line osimertinib plus chemotherapy versus amivantamab plus lazertinib for advanced <i>EGFR</i> -mutant NSCLC in the US and China.

e23169 Background: The FLAURA2 and MARIPOSA trials established overall survival benefits of osimertinib plus chemotherapy and amivantamab plus lazertinib, respectively, over osimertinib monotherapy in advanced non-small cell lung cancer (NSCLC) patients with epidermal growth factor receptor (EGFR) sensitizing mutation. This study aimed to evaluate the cost-effectiveness of osimertinib plus chemotherapy versus amivantamab plus lazertinib from the US and Chinese healthcare perspectives. Methods: A partitioned survival model with 28-day cycles was developed to simulate total costs, life-years, quality-adjusted life-years (QALYs), incremental cost-effectiveness ratio (ICER) over a 10-year time horizon. Clinical efficacy and safety inputs were derived from FLAURA2 and MARIPOSA. An anchored bridging approach with osimertinib monotherapy as the common comparator was applied. Costs and utilities were sourced from public bid-winning databases, local charges and published literature. A scenario analysis incorporated the amivantamab patient assistance program (PAP) in China. One-way and probabilistic sensitivity analyses (OWSA, PSA) assessed model uncertainty. Results: Compared with osimertinib monotherapy, the ICERs for osimertinib plus chemotherapy were $551,944/QALY (US) and $37,965/QALY (China). For amivantamab plus lazertinib, ICERs were $2,186,499/QALY (US) and $291,478/QALY (China). OWSA identified drug prices and utility values as key drivers. PSA indicated that in the US, osimertinib was cost-effective in 100% of simulations at a $150,000/QALY threshold. In China, at a $40,334.05/QALY threshold, osimertinib plus chemotherapy had a 66.1% probability of being cost-effective. The PAP scenario did not alter the conclusion for amivantamab plus lazertinib in China. Conclusions: Osimertinib monotherapy is the cost-effective option for advanced EGFR-mutant NSCLC in the US. In China, osimertinib plus chemotherapy represents the optimal regimen given current pricing. Amivantamab combinations are not cost-effective in either country at commonly accepted willingness-to-pay thresholds.

Frequency of mutations in homologous recombination genes in patients with different histological types of gastric cancer.

e16031 Background: Gastric cancer (GC) is the leading cause of cancer-related death worldwide. Despite advances in treatment methods, poor prognosis and low 5-year survival rates persist, especially in recurrent and metastatic GC. Given this problem, there is a need to search for biomarkers to study additional treatment methods in patients with GC. The purpose of the study was to estimate the frequency of mutations in homologous recombination genes in tumor material of patients with gastric cancer. Methods: The study was conducted on a group of 48 patients with stage IB-IIIC gastric cancer, aged 40 to 83 years (median 64 years), who underwent planned treatment at the National Medical Research Centre for Oncology of the Ministry of Health of the Russian Federation in 2021-2023. According to the classification by P. Lauren, the intestinal type was verified in 24 patients, diffuse – 4, mixed – 3, and unspecified – 17. Gene mutation screening was performed using NGS on the MiSeq Dx genetic sequencer (Illumina, Inc.) using the MiSeq Reagent kit v2 300 cycles. DNA libraries were prepared using the Solo-test ABC Plus kit (Oncoatlas). Mutation analysis was performed using the “maftools” package of the R 4.5.1 computing environment. Results: Based on the results of the NGS study, 22 genetic variants were detected: 11 somatic and 11 germline mutations. The frequency of mutations in the ATM gene was 40.9%, BRCA1 – 22.7%, BRCA2 – 13.6%, CHEK2 – 9.1%, FANCL – 9.1%, RAD51C – 4.5%. The frequency of mutations in the genes of homologous recombination in patients with different histological types of gastric cancer was 35.4% (n = 17), of which 64.7% (n = 11) were patients with the intestinal type, 23.5% (n = 4) with an unspecified type, and 5.9% (n = 1) each were patients with mixed and diffuse gastric cancer. Mutations in the ATM, BRCA1 , BRCA2 , and FANCL genes were detected in patients with the intestinal type of gastric cancer (n = 16), while in patients with an unspecified type, mutations were detected in the ATM , FANCL , CHEK2 , and RAD51C genes (n = 4). In patients with the mixed type, mutations were identified only in the ATM gene (n = 1), and in patients with the diffuse type, mutations were identified in the CHEK2 gene (n = 1). Conclusions: The detected genetic variants in the ATM , BRCA1 , BRCA2 , CHEK2 , FANCL and RAD51C in various histological types of gastric cancer are of interest for further research. These results open up possibilities for the treatment of patients with gastric cancer using PARPi inhibitors, which act on the principle of synthetic lethality in combination with homologous recombination deficiency.