Browse Articles

Second primary malignancies after CAR-T versus SCT in multiple myeloma.

e19556 Background: Multiple myeloma (MM) survival has improved with autologous stem cell transplantation (SCT) and immunotherapy, including chimeric antigen receptor T-cell (CAR-T) therapies, shifting attention towards late complications. As patients live longer, the incidence of second primary malignancies (SPMs) is an increasingly important survivorship issue. While SPMs have been documented in patients with MM, comparative data evaluating SPM risk among patients treated with CAR-T versus SCT alone remain limited. We compared the incidence of SPMs in MM patients treated with SCT and/or CAR-T using a large real-world database. Methods: We performed a retrospective cohort study using the TriNetX network. Two cohorts were constructed: Cohort (1) MM patients with SCT, without CAR-T therapy (MM-SCT-No CAR-T); and Cohort (2) MM patients with CAR-T with/without SCT (MM-CAR T+/- SCT ). Patients with non-MM malignancies prior to the index were excluded. The index date was the date of SCT or CAR-T administration. Propensity score matching was performed to balance baseline characteristics. Results: Mean follow-up was 949 and 314.5 days for cohorts 1 and 2 respectively. 263 patients in cohort 2 had received both SCT and CAR T. After propensity score matching, the analysis included 406 patients in each cohort, with balanced baseline characteristics. Secondary neoplasms occurred in 17 patients (4.25%) in cohort 1 and 83 patients (20.5%) in cohort 2 (risk difference -16.44% (-20.644,-11.843%); OR 0.172, 95% CI 0.1-0.296; HR 0.183 95% CI 0.109-0.309). Of these, 12 (2.97%) and 59 (14.568%) in cohorts 1 and 2 had hematologic malignancies (risk difference -11.598% (-15.411%,-7.784), OR 0.18, 0.095-0.34, HR 0.189 (0.101,0.351. Conclusions: In this real-world matched analysis, MM patients treated with CAR-T experienced a higher incidence of SPMs than those treated with SCT alone, particularly hematologic malignancies. These findings underscore the importance of incorporating secondary malignancy risk into long-term survivorship planning for CAR-T- treated patients and support the need for structured post-CAR-T surveillance and guideline-directed screening and treatment. Further studies are warranted to define therapy-specific and patient-level risk modifiers to guide the intensity and duration of long-term monitoring. Baseline characteristics of the two cohorts after matching. Characteristic MM-SCT-No CAR-T (N=406) MM with CAR-T +/- SCT (N=406) Standard difference Age at Index 66.2 +/- 8.8 66.7 +/- 8.9 0.049 Female 185 (45.5%) 175 (43.1%) 0.049 Male 221 (54.4%) 231 (56.8%) 0.049 Black or African American 90 (22.1%) 74 (18.2%) 0.098 White 273 (67.2%) 268 (66.0%) 0.026 Hispanic or Latino 10 (2.4%) 19 (4.6%) 0.119 Asian 10 (2.4%) 10 (2.4%) <0.001

Chromosomal instability and integrated tumor-margin immune ecosystem remodeling as drivers of progression of pulmonary ground-glass nodules.

8027 Background: Precise management of pulmonary ground-glass nodules (GGNs) is hindered by an incomplete understanding of the biological determinants driving the transition from indolence to aggressive progression. Methods: In a multicenter longitudinal study of 845 patients, we stratified GGNs into three trajectories via serial CT: Stable, ground glass opacity (GGO)-predominant progression, and solid-predominant progression. Integrative profiling was performed using WES, WGS, bulk RNA-seq, Visium HD, CosMx SMI, and imaging mass cytometry. We developed a deep learning model to predict solid-predominant progression using baseline radiographic features. Results: During a median follow-up of 55.8 months, our analysis indicated GGNs progression was driven predominantly by chromosomal instability rather than distinct point mutational profiles. Specifically, progressive GGNs harbored catastrophic structural variations, including chromothripsis, kataegis, and extrachromosomal DNA which were absent in stable lesions. Spatially, the immune microenvironments diverged sharply. While stable GGNs were quiescent, GGO-predominant growth lesions were significantly enriched with antigen-presenting macrophages (p = 0.00054), which facilitated the spatial aggregation of activated lymphocytes and promoted the formation of early tertiary lymphoid structure (TLS)-like niches through CCL19-CCR7 signaling. Conversely, in solid-predominant aggressive progression nodules, macrophages underwent a phenotypic shift from antigen-presenting toward SPP1+ state, accompanied by a marked reduction in CD74 and HLA-DR expression (p = 0.0019). Within the tumor core, SPP1+ macrophages suppressed T-cell effector function via SPP1-CD44 signaling. At the tumor margin, hypoxia-adapted endothelial cells exhibited significant upregulation of CD274 (PD-L1), contributing to T-cell inhibition, while activated fibroblasts deposited dense collagen matrices formed a physical barrier restricting B-cell infiltration. Together, these mechanisms established a coordinated immunosuppressive niche facilitating tumor growth. Leveraging these biological insights, we developed a deep learning model extracting radiomic features from the baseline CT nodule and an extended 2.5mm perinodular margin. This approach significantly outperformed models using the nodule alone or the nodule plus a 1mm margin (p < 0.05), achieving AUCs of 0.89 in the training cohort (n = 471), 0.85 in the internal test cohort (n = 202) and 0.83 in the external cohort (n = 117). Conclusions: Distinct radiographic progression trajectories of GGNs are underpinned by divergent chromosomal and spatial immune programs at the tumor core and margin, providing a biological foundation for margin-aware AI strategies to predict aggressive progression and guide optimal surgical intervention.

Exploring the clinical significance of pathogenic variants in Lynch syndrome: A single-center retrospective review.

e22657 Background: Lynch syndrome (LS) is the most common hereditary cancer syndrome and is associated with increased malignant risk due to four pathogenic variants in mismatch repair genes. Although population-level genomic studies identify PMS2 and MSH6 as the most frequent variants, MLH1 and MSH2 account for most pathogenic variants in LS families. Understanding variant-specific cancer risks may improve early detection of LS-associated malignancies. We report mutation patterns, demographic characteristics, and family history trends in a LS patient cohort. Methods: We retrospectively identified patients with germline mutations consistent with Lynch syndrome at an urban US academic medical center (N=70). Demographic characteristics, germline mutations as well as personal and family histories of malignancy were extracted from the electronic health record, summarized descriptively, and stratified by mutation type and race. Results: Our cohort’s mutation profile for MLH1, MSH2, MSH6, and PMS2 mutations was 20, 20, 23, and 26%. This was particularly driven by black patients (N=14) in whom 42% of pathogenic variants were PMS2. The incidence of cancer type was similar across demographic groups. 76% of PMS2 carriers in our cohort had a positive family history of breast cancer, with breast cancer representing 37% of all malignancies seen in families carrying a PMS2 mutation, more than any other malignancy. This trend was not seen in other mutations. Conclusions: We present a single institution retrospective analysis of LS patients. Mutation profiles have been seen to vary by population, with most studies evaluating prevalence of LS mutations in the general population while ours evaluates pathogenic variants in confirmed LS families. We found a comparatively increased prevalence of PMS2 and MSH6 mutations, particularly in black LS families. Breast cancer prevalence was increased in LS families carrying PMS2 mutation (not other LS families), notable given breast cancer is not commonly viewed as LS-associated. Our findings highlight the need for investigation into mutation profiles for confirmed LS families and a possible correlation between PMS2 and breast cancer. Cancer type Family history (N, %) MLH1 Colorectal (4) Colorectal (16, 67%)Gastric (5, 21%)Endometrial, pancreatic, urothelial (1, 4%) MSH2 Colorectal (2)Ovarian (2) EndometrialSmall intestine Colorectal (11, 34%)Urothelial (5, 16%)Endometrial (4, 13%)Gastric, pancreatic, ovarian (3, 9%)Small intestine, biliary, brain (1, 3%) MSH6 ColorectalEndometrial (3)Sebaceous epitheliomaUrothelial Endometrial (10, 29%)Colorectal (9, 26%)Urothelial (6, 18%)Pancreas (5, 15%)Ovarian, brain (2, 6%) PMS2 Colorectal (2)EndometrialGastricBreast (2) Colorectal (13, 50%)Endometrial (6, 22%)Ovarian, pancreatic, urothelial (2, 8%)Gastric (1, 4%)Breast (15)* *Not included in percentage calculations.

Phase 2 study of PFS in third-line immune checkpoint inhibitor–resistant advanced NSCLC treated with the telomere-targeting agent ateganosine (THIO; 6-thio-2'-deoxyguanosine) sequenced with ICI.

e20610 Background: Despite advancements in third line treatments, long-term survival for advanced non-small cell lung cancer (NSCLC) remains suboptimal, with median survival follow-up of only 5.8 months 1 . Among patients treated with prior platinum chemotherapy and immune checkpoint inhibitors (ICIs), the median survival was reported to be 6.47 months 2 . Treatment options for ICI-resistant patients are limited. Ateganosine, a telomere-targeting agent that modifies telomeres in cancer cells, demonstrates improved progression-free survival (PFS) independent of PD-L1 expression. Methods: NCT05208944 is a phase 2, multicenter, open-label study that enrolled 79 patients with advanced NSCLC who relapsed after 1 - 4 prior treatments, including ICIs. At data cut off (Jan 15, 2026), in the third line therapy 22 patients treated with THIO (60, 180, or 360 mg) were evaluated for (PFS) and their prior PD-L1 expression at the time of study enrollment (C1D1). Results: In the third line therapy 17 out of 22 patients (77%) surpassed the benchmark value, and in the 180 mg dose group (n = 10) it reached 24.1 weeks compared to 2.5 months for the benchmark 3,4 . Exploratory biomarker analyses showed that baseline LDH levels were prognostic for outcomes. Across all patients, elevated LDH was associated with shorter PFS (log-rank p = 0.03, 2-sided) and OS (log-rank p = 0.01, 2-sided), with patients with no available LDH values excluded. In the third-line population, elevated LDH remained prognostic for OS (log-rank p = 0.04, 2-sided). THIO followed by cemiplimab was generally well tolerated in this difficult-to-treat population. The response to THIO and cemiplimab, demonstrated by partial response (PR) and stable disease (SD) was independent of baseline PD-L1 status. This indicates that THIO can be effective across patients regardless of their PD-L1 status. Conclusions: Ateganosine demonstrates clinically meaningful PFS improvement in third line patients with advanced NSCLC, independent of PD-L1 status. The improved PFS observed in patients treated with THIO in sequential combination with an immune checkpoint inhibitor (ICI), compared to standard chemotherapy, supports its potential to expand treatment options for ICI-resistant advanced NSCLC. Baseline LDH levels provide prognostic information across populations and may help identify patients most likely to benefit from therapy. Clinical trial information: NCT05208944 .

National trends and disparities in gallbladder cancer mortality in the United States, 1999–2023.

e16286 Background: Gallbladder cancer (GBC) is a rare but devastating gastrointestinal malignancy with an incidence of 122,000 new cases annually. GBC can have a varied presentation, and despite medical and surgical advancements, overall prognosis remails poor. Comprehensive, long-term national trends, and disparities in its mortality have not been recently characterized on a multivariable level. Methods: National mortality data for individuals aged ≥45 years in the United States were obtained from the Centers for Disease Control and Prevention (CDC) Wide-ranging Online Data for Epidemiologic Research (WONDER) Multiple Cause of Death database for the period 1999–2023. Deaths were identified using International Classification of Diseases (ICD-10 code) C23.0 as the underlying cause of death. Extracted variables included age, sex, race, Hispanic origin, U.S. census region, and level of urbanization. Temporal trends in mortality were evaluated using Joinpoint regression analysis to estimate the average annual percent change (AAPC) and corresponding 95% confidence intervals (CIs). Statistical significance was defined as p value < 0.05. Results: Between 1999 to 2023, a total of 54,454 deaths attributable to GBC were recorded in the United States.The national age-adjusted mortality rate was highest in 1999 (2.323 per 100,000 people) and declined steadily over the study period which reached 1.534 per 100,000 in 2023. This corresponded to a significant overall reduction in mortality, with an average annual percent change (AAPC) of −1.48% (95% CI: −1.64% to −1.31%; p < 0.001). There were significant downward trends in both females (AAPC: −1.65%) and males (AAPC: −1.31%), with mortality rates being consistently higher among females during study period. Racial and ethnic disparities were also very evident. The steepest decline occurred among Hispanic or Latino individuals (AAPC: −2.23%), followed by White individuals (AAPC: −1.87%) and Asian or Pacific Islander individuals (AAPC: −1.38%). In contrast, mortality rates among Black or African American individuals did not demonstrate a statistically significant change over the 25-year period (AAPC: 0.25%). In terms of region, the most significant decline was seen in Census region 1 (Northeast, AAPC: 2.01%, while Region 4 (West) had the least decline (AAPC: 1.02%). All urbanization categories experienced significant declines, with the greatest reduction in Medium Metro areas (AAPC: 2.50%). Conclusions: The significant national decline in GBC mortality since 1999 represents a meaningful progress. However, the absence of a corresponding reduction in mortality among some populations like Black or African American underscores a critical and persistent health inequity. Public Health focused interventions are needed to better understand the underlying drivers of these disparities and to promote equitable reductions in gallbladder cancer mortality.

Role of ATM as a prognostic biomarker in non small cell lung cancer: A systematic review and meta-analysis.

e20543 Background: Non-small cell lung cancer (NSCLC) constitutes the majority of lung cancer diagnoses and remains associated with poor survival despite advances in systemic therapies. DNA damage and repair (DDR) pathways are critical in tumor progression and therapeutic response. Ataxia telangiectasia mutated (ATM), a central DDR kinase, is frequently altered in NSCLC; however, the prognostic impact of ATM expression and ATM mutation remains incompletely defined. We conducted a systematic review and meta-analysis to evaluate the association of ATM alterations with survival outcomes in NSCLC. Methods: A systematic review and meta-analysis was conducted according to PRISMA guidelines, with protocol registration in PROSPERO (ID: 1076509). A comprehensive literature search was performed in PubMed, Embase, Scopus and Web of Science for studies published between January 2000 and May 1, 2024. The search strategy used Medical Subject Headings (MeSH) and keywords for “Non-Small Cell Lung Cancer” and “Ataxia Telangiectasia Mutated Proteins,” combined using Boolean operators. Studies were selected based on PICOS criteria: NSCLC patients; exposure defined as ATM expression or ATM mutation; comparison groups based on high vs low expression or mutant vs wild-type ATM; outcomes including overall survival (OS), progression-free survival (PFS), or disease-free survival (DFS); and cohort study design. Hazard ratios (HRs) with 95% confidence intervals (CIs) were pooled using a random-effects model. Study quality was assessed using the Newcastle-Ottawa Scale. Results: Four studies met inclusion criteria for quantitative synthesis. Two studies evaluating ATM expression showed no statistically significant association between ATM expression and OS (pooled HR 1.75, 95% CI 0.75-4.05; P=0.19), with substantial heterogeneity (I²=74%). In contrast, two studies assessing ATM mutation status demonstrated an association between ATM mutation and improved OS compared with wild-type ATM (pooled HR 0.69, 95% CI 0.47-1.00; P=0.05), with moderate heterogeneity (I²=62%). Conclusions: ATM mutation status, but not ATM expression, appears to be associated with overall survival in patients with NSCLC. These findings highlight the potential prognostic relevance of ATM alterations and support further investigation of DDR-related biomarkers in NSCLC. Prospective, well-powered studies are warranted to validate these observations and define their clinical implications.

Olanzapine for chemotherapy-related anorexia: Impact on body composition.

12065 Background: Chemotherapy-related anorexia and weight loss are associated with adverse outcomes in patients with advanced gastrointestinal and lung cancers. While olanzapine has demonstrated efficacy in improving appetite and weight gain, the impact of weight gain on body composition remains unclear. We evaluated changes in morphomic parameters among patients receiving olanzapine during chemotherapy. Methods: In this phase III, randomized, double-blind, placebo-controlled trial, adults (≥18 years) with untreated locally advanced or metastatic gastric, hepatopancreaticobiliary, or lung cancer were randomized to receive olanzapine 2.5 mg once daily or placebo for 12 weeks, along with standard chemotherapy and nutritional counseling. The primary endpoint was the proportion of patients achieving > 5% weight gain. Secondary endpoints included changes in skeletal muscle index (SMI), subcutaneous fat index (SFI), and visceral fat index (VFI) assessed at the third lumbar vertebra on CT scans. Morphomic analysis was performed in patients with paired baseline and interim CT scans. Results: A total of 103 patients (olanzapine, n = 53; placebo, n = 50) were included in the morphomic analysis. Baseline characteristics were comparable between groups. Weight gain > 5% was significantly more frequent in the olanzapine group compared with placebo (60% vs 9%, p < 0.0001). Both groups experienced a decline in mean SMI; however, among olanzapine-treated patients with > 5% weight gain, 63% demonstrated improvement or stabilization of SMI. Improvement in SFI was more frequent with olanzapine than placebo (62% vs 40%, p = 0.047). Among patients with > 5% weight gain, a higher proportion in the olanzapine group showed an increase in VFI compared with placebo (72% vs 60%, p = 0.006). Conclusions: Olanzapine improves weight gain with favorable changes in skeletal muscle and fat indices, indicating true nutritional gain rather than fluid retention in patients with chemotherapy-related anorexia. These findings suggest that olanzapine may favorably modulate cancer-related anorexia and treatment outcomes, warranting further validation in larger focused studies. Clinical trial information: CTRI/2020/08/027133.

Dynamics of tissue fatty acid profiles and association with radiotherapy toxicity in locally advanced nasopharyngeal carcinoma.

e18048 Background: Tissue levels of n-6 and n-3 fatty acids (FAs), reflected by erythrocyte membrane composition, modulate inflammatory processes and may influence tolerance to cancer therapy. Their dynamics during multimodality treatment for locally advanced nasopharyngeal carcinoma (LA-NPC) and predictive value for radiotherapy (RT) toxicity remain undefined. Methods: This prospective study enrolled 70 patients with LA-NPC (Stage III-IVA, AJCC 9th) receiving induction chemotherapy (toripalimab + cisplatin + nab-paclitaxel) for three cycles, followed by concurrent toripalimab and RT. Serial erythrocyte FA profiles, inflammatory cytokines, and nutritional indices were assessed at baseline (T0), each induction chemotherapy cycle(T1-T3), pre-RT (T4), mid-RT (T5; 40 Gy), and post-RT (T6-T7). Logistic regression and ROC analysis evaluated the n-6/n-3 ratio at T4 as a predictor of grade ≥3 radiation-induced oral mucositis (RIOM). Baseline profiles were also compared to 70 healthy volunteers, and differences between Stage III and IVA patients were examined. Results: Erythrocyte FA profiles in patients with LA-NPC demonstrated treatment-phase-specific dynamics. The n-6/n-3 ratio declined during induction chemotherapy, then increased sharply to a peak at mid-RT (T5: 10.10 ± 1.30 vs. T0: 7.38 ± 1.74; p<0.001), and declined post-RT. The peak n-6/n-3 ratio at T5 was driven by RT-induced depletion of omega-3 (EPA, DHA), while omega-6 levels remained stable, which also resulted in an increase in the AA/EPA ratio (T5:39.0±5.89 vs.T0:32.3±6.00; p<0.001). At the same time, a pro-inflammatory state (increased IL-6 and IFN-γ and decreased IL-10) was observed, accompanied by a decrease in serum albumin and body weight. The incidence and severity of RT-related adverse events peaked at T5 and were mainly ≥ grade 3 RIOM. At this time, the n-6/n-3 ratio was positively correlated with RIOM severity (r=0.60) and dermatitis risk (r=0.60), whereas EPA and DHA levels were negatively correlated (r=-0.4 to -0.9). Critically, the n-6/n-3 ratio at T4 was a strong independent predictor of subsequent severe RIOM (OR 6.15, 95% CI 2.43–15.55; p<0.001), with an AUC of 0.935 (cutoff 7.16; sensitivity 78.6%, specificity 93.0%). At baseline, compared with healthy volunteers, patients showed a pro-inflammatory metabolic profile, as indicated by significantly higher n-6/n-3 ratio, DTA, IL-6, and IFN-γ, and lower omega-3 levels (all p<0.05). There was no significant difference in FA profiles between stage III and IVA patients. Conclusions: Erythrocyte FA profiles undergo marked reprogramming during LA-NPC treatment, characterized by RT-driven omega-3 depletion. The pre-RT erythrocyte n-6/n-3 ratio is a potent predictive biomarker for severe RIOM, identifying a window for nutritional intervention.

Quality of life in patients with gastroenteropancreatic neuroendocrine tumors receiving ¹⁷⁷ Lu-edotreotide or everolimus: Results from the COMPETE study.

4171 Background: The Phase 3 COMPETE trial demonstrated a significant treatment effect in favor of 177 Lu-edotreotide over everolimus in patients with Grade 1/2, well-differentiated, gastroenteropancreatic neuroendocrine tumors; 177 Lu-edotreotide significantly improved progression-free survival (primary endpoint) and objective response rate (secondary endpoint) compared to everolimus. Nowadays, treatment choice also considers the impact of therapy on patients’ quality of life (QoL); here, we present QoL results from the COMPETE trial. Methods: The European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ)-C30 and EORTC QLQ-gastrointestinal neuroendocrine tumors (GI.NET) were completed by patients at baseline, each month during the first year, and every 3 months thereafter until disease progression or end of study. Mean change from baseline, duration of maximum health-related quality of life (HRQoL) improvement (until deterioration), and time to deterioration were calculated. Between-group comparisons were conducted using repeated measure analysis with factors for treatment, randomization stratification factors, visit, baseline value, and treatment-by-visit interaction; all HRQoL analyses were prespecified and exploratory. Results: Based on repeated measure analysis, mean change from baseline was in favor of 177 Lu-edotreotide: overall mean change (95% confidence interval [CI]) in global health score was 0.9 (-2.7, 4.4) in the 177 Lu-edotreotide arm vs -9.9 (-13.9, -6.0) in the everolimus arm (nominal p<0.0001) (positive change corresponds to improvement). Time to deterioration in global health status/QoL was longer in the 177 Lu-edotreotide arm vs the everolimus arm (Table 1). Similar trends were observed in the EORTC QLQ-C30 and EORTC QLQ-GI.NET subdomains. In the 177 Lu-edotreotide arm, 90/207 (43.5%) participants had a clinically meaningful improvement (≥10 points) in global HRQoL score, vs 31/102 (30.4%) participants in the everolimus arm. Among these participants, the median duration of improvement was 22.0 (95%CI: 10.1, not evaluable [NE]) vs 10.2 (95%CI: 3.3, NE) months in the 177 Lu-edotreotide and everolimus arm, respectively. Conclusions: Consistent with the overall COMPETE outcomes, this analysis demonstrated clinically meaningful and durable HRQoL benefits for 177 Lu-edotreotide compared with everolimus. Clinical trial information: NCT03049189 . Time to deterioration in global health status/QoL (≥10 points). Parameter / Statistics 177 Lu-edotreotide(N=207) Everolimus(N=102) Patients with ≥10 points deterioration, n (%) 114 (55.1) 78 (76.5) Median, months (95% CI) a 10.251 (7.359, 15.901) 2.267 (2.004, 3.253) Nominal stratified p-value b <0.0001 Stratified hazard ratio (95% CI) 0.392 (0.292, 0.527) a Estimated via Kaplan-Meier method. b Derived from a two-sided test between the two groups.

Integrated dual valorization of waste printed circuit boards for the synthesis of copper oxide nanoparticles and carbon-infused PVA membranes for dye removal

Comprehensive profiling and clinical utility of CSF-derived cell-free DNA/RNA for evaluation of solid and hematologic malignancies affecting the central nervous system.

3066 Background: Diagnosis and risk stratification of brain lesions is challenging when tissue biopsy is contraindicated due to anatomic or clinical factors. Cerebrospinal fluid (CSF)–based liquid biopsy using cell-free DNA (cfDNA) and RNA (cfRNA) offers a minimally invasive approach to detect neoplastic disease and inform management. We evaluated the clinical utility of a commercially available cfDNA/cfRNA next-generation sequencing assay applied to CSF specimens from patients with CNS lesions of uncertain etiology. Methods: Between 2023 and 2025, cfDNA, cfRNA, and CSF cell pellet RNA profiling using next generation sequencing (NGS) was performed on CSF samples from 1,559 patients, including 1,209 evaluated for suspected primary or secondary solid CNS malignancies and 350 evaluated for CNS involvement by hematologic malignancies. The assay interrogated single-nucleotide variants, insertions/deletions, copy number alterations, gene fusions, lymphocyte clonality, and tumor mutational burden. The depth of sequencing was between 25,000x and 30,000x for cfDNA. For quality control purposes, more than 80 million reads were required for accepting RNA results, while the required percentage of spliced RNA reads was above 20%. Clinical utility and actionability was further assessed in a single-institution cohort of 90 patients by correlating cfDNA/cfRNA results with cytology, flow cytometry, tissue biopsy, and clinical decision-making. Results: Among patients evaluated for solid malignancies in the entire cohort, the most common alterations involved TP53, KMT2C, PIK3CA, EGFR, and DNMT3A, while hematologic cases frequently harbored PIM1, TP53, MYD88, CD79B, TET2, and DNMT3A alterations. In the institutional cohort (median age 57 years), neoplastic cfDNA was detected in 40/90 samples, with recurrent alterations mirroring those of the larger panel. cfDNA profiling demonstrated 100% sensitivity for neoplastic DNA detection compared with current standard-of-care cytology and flow cytometry and identified neoplastic DNA in 37% of cytology-negative and 42% of flow-negative cases. Concordance with contemporaneous tissue biopsy yielded a positive predictive value of 89%, with a sensitivity of 68%, reflecting reduced detection in intraparenchymal lesions with limited CSF exposure. cfDNA results directly influenced treatment decisions in patients evaluated for new intracranial malignancy and for CNS metastasis or recurrence. Conclusions: CSF cfDNA/cfRNA profiling significantly enhanced diagnostic yield in CNS lesions of uncertain etiology, particularly when conventional cytology is negative, or biopsy is infeasible, while also providing actionable molecular information to guide therapy.

Overall survival analysis of eSyM: A cluster randomized trial of an electronic patient-reported outcomes (ePRO)-based symptom management program.

11003 Background: ePRO monitoring can improve outcomes yet remains underutilized in routine cancer care. We conducted a pragmatic type II hybrid effectiveness–implementation trial using a cluster-randomized stepped-wedge design to assess an ePRO-based, EHR-integrated symptom management program (eSyM) across 6 health systems. Our prior analyses demonstrated that eSyM use, but not eSyM deployment, was associated with a significant reduction in emergency department visits and hospitalizations. Here, we report results of the overall survival (OS) analysis. Methods: Eligible patients were adults who received chemotherapy (CHEMO) or had surgery (SURG) for a gastrointestinal (GI), gynecologic (GYN), or thoracic (THOR) tumor from Jan 2018-Feb 2023. The outcome was OS 1 year after starting chemotherapy or postoperative discharge. The primary analysis compared patients treated pre- versus post-deployment of eSyM. A secondary analysis compared patients who were treated post-deployment and did versus did not use eSyM, where users were those who reported ePROs via eSyM at least once. Cox regression models accounted for sociodemographic, clinical, and health system factors. Analyses were stratified by treatment modality to account for heterogeneity of effect. Results: Health systems contributed data on 39,895 patients (selected characteristics in Table). After eSyM deployment, 45% of CHEMO and 53% of SURG patients used eSyM to report ePROs. Multivariable Cox regression analyses identified significant associations between eSyM deployment and OS for CHEMO (HR 1.11, 95%CI 1.03-1.21; P=0.007) and SURG (HR 0.86, 95% CI 0.76-0.98; P=0.026) patients. The adjusted absolute difference in 12-month OS probability for patients treated post-deployment vs. pre-deployment was -2.4% (P<.001) for CHEMO and 0.7% (P<0.031) for SURG. Comparing eSyM users vs. non-users from the post-deployment cohort only, OS was significantly better among CHEMO (HR 0.63, 95%CI 0.57-0.69; P<0.001; 12m OS diff 10.2%) and SURG (HR 0.58, 95%CI 0.49-0.69; P<0.001; 12m OS diff 2.8%) patients. Conclusions: Deployment of eSyM was associated with statistically significant but clinically marginal differences in OS. After eSyM was deployed, use of ePROs was associated with statistically significant and potentially meaningful improvements in OS. Considering that ePRO monitoring with eSyM decreases the need for acute care and may be associated with longer survival, strategies to improve adoption and implementation of ePRO-based symptom management programs may be warranted. Clinical trial information: NCT03850912 . CHEMO SURG Pre Post Pre Post # Patients 5149 7030 15917 11799 Age (median) 66 67 61 63 Female 54% 54% 72% 69% Charlson comorbidity ≥2 8% 7% 6% 7% GI, GYN, THOR 49%, 16%, 35% 45%, 17%, 38% 41%, 37%, 22% 43%, 33%, 24% # Deaths 1098 2037 546 614 12-mo OS (adjusted) 72.7% 70.2% 94.7% 95.4% 12 mo OS (users, non-users) -- 76.8%, 66.6% -- 96.0%, 93.3%

Age-related differences in patient burden in endometrial cancer: Findings from the international EXPRESSION XI/IMPROVE survey.

e17622 Background: Despite favourable survival in early-stage endometrial cancer (EC), patients experience a substantial disease- and treatment-related burden that negatively affect quality of life (QoL). The aim was to compare the most burdensome problems reported by patients with EC in different age groups. Methods: This international, anonymous survey was conducted within NOGGO, GCIG, ENGOT and ENGAGe. The 80-item questionnaire covered demographics, medical history, current symptoms, lifestyle, and self-perception. Eligible participants had a history of EC, regardless of stage or treatment. Four age-based subgroups were defined: (A) <40 years, (B) 40-70 years, (C) 71-80 years, and (D) >80 years. Analyses were descriptive and exploratory. Results: Overall, 2,761 women from 20 countries completed the survey. Of these, 36 were in subgroup A, 1,105 in subgroup B, 162 in subgroup C, and 162 in subgroup D. Treatment-related side effects were largely similar across age groups, though differences were observed in hair loss, exhaustion, pain, concentration, and lymphedema. Hair loss affected 34% of patients in group C, compared to only 18% in group D. In group A, almost 40% were affected by exhaustion, whereas only 24% of group D reported exhaustion. Pain was most prevalent in group A, at 25%, while in group D it was 13%. Concentration problems were reported by 22% of group A and only 15% of group D. Group A was also most affected by lymphedema, with 19% compared to 13% in group D. With regard to the problems of the past week, it was clear that patients in A reported significantly more worries (69% vs. 50 % (B) vs. 38 % (C) vs. 29% (D)), fears (56% vs. 44 % (B) vs. 25 % (C) vs. 20% (D)), sadness (53% vs. 33 % (B) vs. 23 % (C) vs. 18% (D)), and depression (22% vs. 14 % (B) vs. 9 % (C) vs. 7% (D)). Younger patients were also more concerned about their appearance (36% vs. 9 % (B) vs. 7 % (C) vs. 5% (D)). In contrast, more patients in group C and D complained of neuropathies (33% (C) and 30% (D) vs. 17 % (A) vs. 27 % (B)). Nausea, pain, sleep problems, nervousness, inflammation in the mouth area, obstipation, diarrhea, fever, and dry skin were equally distributed across all age groups. Conclusions: Treatment- and disease-related symptoms are common in endometrial cancer survivors, with younger patients reporting more exhaustion, concentration problems, and emotional distress, while elderly patients experience more neuropathies. Age-specific differences should guide follow-up care to improve quality of life. Clinical trial information: DRKS00025954.

Magnitude and timing of clinical benefit associated with pathologic complete response in perioperative muscle-invasive bladder cancer: A reconstructed IPD analysis of NIAGARA trial.

4607 Background: Pathologic complete response (pCR) is widely used as an early efficacy endpoint in neoadjuvant or perioperarive bladder cancer trials. However, formal surrogacy analyses linking pCR to long-term clinical outcomes are lacking. In particular, the magnitude and timing of clinically meaningful benefit associated with pCR remain poorly defined. We aimed to quantify the absolute benefit and temporal emergence of event-free survival (EFS) and overall survival (OS) associated with pCR in NIAGARA trial. Methods: Individual patient data (IPD) were reconstructed from the most recent publicly available Kaplan-Meier curves of the NIAGARA trial, including stratification by pCR and non-pCR status for each treatment arm. The primary analysis compared outcomes between pCR and non-pCR patients within each treatment arm. Restricted mean survival time differences (ΔRMST) were estimated at 24 months for EFS and 36 months (mo) for OS. Time to benefit (TTB) was defined as the earliest time point at which ΔRMST reached ≥1 mo for EFS and ≥2 mo for OS. Exploratory analyses compared outcomes within pCR patients and within non-pCR patients between treatment arms, to explore whether the clinical benefit associated with pCR is purely prognostic or reflects treatment-related modulation within pathological response strata. Results: At 24 months, pCR was associated with a large and early EFS benefit across treatment arms. Among patients receiving durvalumab + cisplatin-gemcitabine (CG), pCR was associated with a ΔRMST of +5.87 mo (95% CI 4.78-6.96; p<0.0001) compared with non-pCR, with a TTB of 9 mo. Similar results were observed in the CG arm (ΔRMST of +5.63 mo, 95% CI 4.49-6.77; p<0.0001), with a TTB of 9 mo. At 36 months, pCR was also associated with a substantial OS benefit. In the durvalumab + CG arm, pCR was associated with a ΔRMST of +5.32 mo (95% CI 3.89-6.74; p<0.0001) versus non-pCR, with a TTB of 22 mo. Comparable findings were observed in the CG arm (ΔRMST +5.89 mo, 95% CI 4.30-7.47; p<0.0001), and a TTB of 20 mo. In exploratory analyses, no significant EFS or OS differences were observed between treatment arms among pCR patients (ΔRMST <1 month), with a TTB of 38 mo for EFS and not reached within 60 mo for OS. Among non-pCR patients, treatment-related differences were similarly modest (ΔRMST ~1 month), with delayed TTB (19 mo for EFS and 50 mo for OS). Conclusions: Using reconstructed IPD from NIAGARA, pCR identifies a large, early EFS benefit and a later OS benefit, informing, in the absence of formal surrogacy validation, a quantitative and temporal clinical interpretation of pCR. In contrast, no additional treatment-related survival benefit was observed within pCR or non-pCR subgroups, raising the hypothesis that survival differences may be largely attributable to pCR rather than to treatment-related modulation within response strata.

Racial and ethnic differences in melanoma survival in the era of immunotherapy.

e21588 Background: Since the introduction of immune checkpoint inhibitors (ICIs) in 2011, mortality in advanced melanoma has been decreasing steadily. Little is known about how sociodemographic factors affect survival in the post-ICI era. Methods: This was a retrospective cohort study of 15,676 patients with advanced melanoma (unresectable Stage III and Stage IV disease) included in the SEER database from 2000 to 2022. Descriptive analyses were conducted and stratified by racial/ethnic group - Non-Hispanic White (NHW), Hispanic (H), Non-Hispanic Black (NHB), and Other. Groups were compared across demographic variables using Fisher’s exact tests. Survival functions were compared using the log-rank test. Cox proportional hazards regression models were then fitted to estimate hazard ratios (HRs) for each group compared with NHWs, with and without adjustment for demographic variables (age, sex, year of diagnosis, marital status, income level, and urbanicity). Extent to which each variable contributed to survival prediction at different follow-up times was quantified by change in Brier score following permutation, with larger increases indicating greater variable importance. Survival analyses were repeated in subgroups diagnosed during 2000–2011 and 2012–2022, respectively. Results: There were 5,026 and 10,650 patients respectively in the pre-ICI (2000-2011) and post-ICI (2012-2022) groups. 14,135 (90.2%) were NHW, 982 (6.3%) H, 240 (1.5%) NHB and 319 (2%) Other. Younger age, female sex, higher income levels and being married consistently conferred a survival benefit (p < 0.05). There was no significant difference in the adjusted overall survival (OS) by racial/ethnic group in the overall time period (2000-2022) (p = 0.288), or in the pre-ICI group (p = 0.666). However, the post-ICI group showed a significant difference in adjusted OS (p = 0.008) with decreased survival in H and NHB patients compared to NHW. Variable importance analyses indicated that age and marital status contributed more to survival prediction than race in both periods. Conclusions: Disparities in survival by racial/ethnic groups after the advent of immunotherapy may indicate barriers in access to treatment. However, marital status played a more influential role in survival prediction than race in the variable importance analysis, highlighting the potential impact of social and demographic factors. Further research is needed to identify and address the underlying mechanisms behind these disparities. OS by racial/ethnic group (2012-2022). Unadjusted OS Adjusted OS HR P-value(compared with ref) Overall P-value HR P-value(compared with ref) Overall P-value NHW Ref Ref Ref Ref Ref Ref H 1.04 (0.94-1.14) 0.481 0.408 1.15 (1.04-1.27) *0.007 *0.08 NHB 1.17 (0.97-1.43) 0.107 0.408 1.23 (1.01-1.49) *0.043 *0.08 Other 1.01 (0.85-1.21) 0.886 0.408 1.12 (0.94-1.34) 0.194 *0.08 *represents statistical significance at p<0.05.

Immune profiling of tumor-infiltrating lymphocytes in patients with primary and metastatic soft tissue sarcomas.

e23544 Background: Soft tissue sarcomas (STSs) are rare and heterogeneous cancers with limited responsiveness to immune checkpoint inhibitors (ICIs), with most STSs exhibiting low expression of neoantigens and immune response genes. While tumor-infiltrating lymphocytes (TILs) can be readily isolated from many STSs, resistance of TILs to ICIs remains poorly understood. To investigate TIL phenotypes and function, we performed comprehensive profiling by flow cytometry of TILs obtained from STS patients undergoing routine surgical resections. Methods: Patients consented to an IRB-approved tumor banking protocol to collect clinical data and specimens. TILs were extracted/expanded from fragments of fresh STS surgical specimens by culture in AIMV media containing human IL-2 for 10-15 days. Cryopreserved TILs along with matched peripheral blood mononuclear cells (PBMCs) were analyzed using multiplex flow cytometry to assess TIL phenotypes, memory/activation states, and checkpoint protein expression. A subset of TIL samples were stimulated with PMA/ionomycin followed by flow cytometry/ELISA to assess cytokine production. Clinical and demographic data were correlated with TIL and PBMC immune variables using linear models with empirical Bayesian methods, adjusted for multiple comparisons, and followed by principal components analysis. Results: 23 patients were profiled, with 21 patients having evaluable paired samples, with most patients having liposarcoma, leiomyosarcoma or GIST. 43.5% of patients had no prior systemic therapy. TILs demonstrated a significantly higher CD8:CD4 ratio (median 1.05 [IQR 0.56–2.65] vs PBMCs 0.25 [0.14–0.43], p<0.0001). Tumor-resident memory cells (CD8⁺CD103⁺CD39⁺) were enriched in TILs (4.3% [1.6–10.5] vs PBMCs 0.09% [0.02–0.16], p<0.0001). T regulatory cells were overall rare but more prevalent in older patients and males. Most TILs were effector memory phenotype (CD8 61.2% [41.5-70.6] and CD4 81.6% [62.8-89.1]), whereas PBMCs showed higher central memory cells. Most CD3- TILs and PBMCs were CD56+ NK cells. Expression of PD1 was higher in TILs vs. PBMCs (mean MFI 7057 ± SEM 831.8 vs. PBMCs 1924 ± SEM 150.4, p=<0.0001), with trends towards higher expression of TIM3 and LAG-3 in TILs. No correlations were observed between STS types or prior treatments with quantity or phenotypes of TILs. Full analysis of clinical correlations and functional activity will be presented at the meeting. Conclusions: STS TILs display features of antigen experience with exhaustion, including higher CD8:CD4 ratio, greater EM populations, and higher checkpoint expression relative to matched PBMCs. While limited by small sample size and patient heterogeneity, our data will help inform future clinical trials of combination ICI therapies to increase TIL infiltration, lessen exhaustion, and overcome resistance, to improve clinical outcomes.

Clinical validation of a blood-based multimodal immune-response score in a real-world cohort of advanced non–small cell lung cancer.

e20581 Background: Immune checkpoint inhibitors (ICIs) have transformed care for advanced non-small cell lung cancer (aNSCLC), but predictive biomarkers remain imperfect. We previously developed a multimodal immunotherapy response score (MIRS), which integrated ctDNA epigenomic signatures with microsatellite instability (MSI) and tumor mutational burden (TMB) from a single blood draw, and validated its use as a predictive biomarker for ICI response. In this study, we assess the ability of MIRS to predict clinical outcomes in real-world patients treated with ICIs. Methods: MIRS was trained and validated using de-identified patient data from InfinityAI Data Library and expressed as percentile based on MIRS distribution with >20,000 aNSCLC samples. In this study, we validated the signature in an independent cohort of 32 aNSCLC patients by analyzing baseline plasma on Guardant360 Liquid (Guardant Health, Palo Alto, CA); 27/32 received PD-1/PD-L1 monotherapy and 7/32 had tumor PD-L1 TPS <1%. Patients with MIRS ≥50% were defined as MIRS-High. The primary endpoint was real-world progression-free survival (rwPFS). We fitted Cox proportional hazards models (adjusted for covariates such as sex, age, tissue PD-L1 expression, histologic subtype, ECOG, TNM stage and baseline methylation tumor fraction) and reported adjusted hazard ratios (aHR) with 95% CIs. Median rwPFS was estimated by Kaplan–Meier method. Model discrimination was summarized using concordance index (c-index). Results: MIRS-High patients (18/32) had significantly longer rwPFS (median 15.1 vs 7.2 months; aHR 0.24, 95% CI 0.07–0.82, p=0.02; c-index = 0.80). As a continuous variable, MIRS percentile was associated with improved rwPFS (p=0.04). Stratification by PD-L1 ≥50% (n=14) showed a consistent, but not statistically significant trend toward longer rwPFS (aHR 0.60, 95% CI 0.19–1.87, p=0.38; c-index = 0.75). When both biomarkers were combined, patients with PD-L1 ≥50% or MIRS-High (n=25) had significantly longer rwPFS (aHR 0.24, 95% CI 0.07–0.90, p=0.034). All complete responses (n=4) were MIRS-High (median MIRS score of 80th percentile in aNSCLC); patients with progressive disease (n=6) had a median MIRS of 27th percentile in aNSCLC (partial responders (n=15) had a median of 51th percentile in aNSCLC, and stable disease cases (n=7) had median of 55th percentile in aNSCLC). Conclusions: Patients with MIRS-High scores had a 76% lower adjusted hazard of progression or death even after accounting for key covariates. Additionally, all complete responses were MIRS-High and MIRS was associated with significantly improved PFS compared to known biomarkers such as tissue-based PDL1. These data indicate that the multimodal score taken via a single baseline blood sample strongly predicts ICI benefit and may be useful in ICI vs chemo combination decisions pending further ongoing validation.

A novel approach to improving clinical trial referral and enrollment for patients with cancer: A cross-sectional analysis.

e23310 Background: Participation in clinical trials is fundamental to advancements in cancer treatment. Recent data estimates that only 7.1% of patients with cancer are enrolled into treatment trials. At present, ClinicalTrials.gov is one of the largest, publicly available online repositories of research trials. A Clinical Trials Navigator (CTN) program was launched in 2019 and analyzed the usability of this tool alongside four other repositories, finding outdated, insufficient, or inaccurate information. The CTN program developed a novel methodology for creating cancer-specific ‘Master Lists’ that house a more accurate repository of clinical trial information. The objective of this abstract is to review the successes and challenges with the novel Master List in facilitating the referral and enrollment of patients with cancer in clinical trials. Methods: Master Lists were developed by skilled clinical trial navigators and maintained monthly by performing searches on five different research trial repositories (ClinicalTrials.gov, Canadian Cancer Trials, Clinical Trials Ontario, Canadian Cancer Clinical Trials Network, and Q-CROC). Patients interested in participating in the CTN program are able to self-refer or be referred by a health care professional. The Master List is used to search for eligible studies. Weekly reviews identify and prioritize the available potential trials. The patient, sponsors, and principal investigators are contacted by a CTN MD as part of the MD FollowUp program to review interest and availability of the preferred trials. The Master List is updated in response to information gained from direct contact with study coordinators throughout follow up. Results: Throughout this study period (March 3, 2025 – January 25, 2026),133 people participated in the CTN program and a total of 114 clinical trial sites were contacted pertaining to 78 unique clinical trials in Canada and the United States. The Master List was updated 61 times for these 78 trials with information not available on ClinicaTrials.gov. We have compared accrual onto clinical trials pre and post addition of the MD FollowUp program. Prior to the implementation of the MD FollowUp program, 18% of those referred to trials were accrued to interventional trials. Post implementation of the MD FollowUp program, this improved to a 42% accrual rate. Conclusions: Overall, this study highlights a solution to improve accessibility of clinical trial information. This dynamic Master List with consistently updated information allowed an improvement in successful referral to accrual ratio, thereby improving the efficacy of the clinical trials unit. The Master List serves as a highly comprehensive and accurate resource that is successfully used to enroll patients with cancer into clinical trials. This is expected to translate into improved clinician engagement in the clinical trials process.

Clinical impact of pulmonary embolism in hospitalized patients with breast cancer: A National Inpatient Sample analysis.

e22623 Background: Pulmonary embolism (PE) is a serious thromboembolic complication in hospitalized patients with malignancy, including breast cancer, yet nationwide data on its inpatient burden and outcomes remain limited. This study aims to determine PE prevalence among hospitalized breast cancer patients and evaluate its associations with comorbidities, inpatient complications, procedures, and clinical outcomes using a nationally representative database. Methods: A retrospective cohort study was conducted using the National Inpatient Sample (NIS), including 982,110 hospitalizations with a diagnosis of breast cancer. Patients were stratified into those with PE and without PE. Baseline demographics, comorbidities, inpatient complications, and procedures were compared. Multivariable logistic regression was used to estimate adjusted odds ratios (aORs) with 95% confidence intervals (CIs) for clinical outcomes, adjusting for demographic and clinical confounders. Results: Pulmonary embolism was identified in 1.5% (n = 14,740) of hospitalized breast cancer patients, while 98.5% (n = 967,370) did not have PE. The cohort was predominantly female (99%) in both groups. Patients with PE had a significantly higher prevalence of medical comorbidities, including hypertension, diabetes, chronic pulmonary disease, obesity, anemia, coagulopathy, and congestive heart failure (all p < 0.001). PE was associated with markedly higher rates of acute inpatient complications, including acute respiratory failure, shock, sepsis, acute kidney injury, and need for mechanical ventilation (all p < 0.001). Patients with PE also demonstrated greater utilization of invasive and critical care–related procedures, including central venous catheterization and thrombolytic or vascular interventions. On multivariable analysis, pulmonary embolism remained independently associated with worse in-hospital outcomes, including increased mortality, prolonged length of stay, higher total hospital charges, and non-routine discharge disposition compared with breast cancer patients without PE. Conclusions: In this large nationwide NIS analysis, pulmonary embolism occurred in a clinically meaningful proportion of hospitalized breast cancer patients and was independently associated with increased morbidity, healthcare utilization, and adverse inpatient outcomes. These findings underscore the importance of heightened risk stratification, early diagnosis, and optimized inpatient management of PE in breast cancer patients.

Machine learning–based transcriptomic signatures to predict treatment outcomes across targeted and immunotherapy regimens in renal cell carcinoma.

4526 Background: Despite available tyrosine kinase inhibitors (TKIs) and immune checkpoint inhibitors (ICIs), reliable biomarkers guiding frontline advanced RCC treatment remain limited. Existing signatures lack generalizability across therapeutic regimens. We developed a data-driven machine learning (ML) framework to predict survival outcomes and therapeutic response. Methods: Transcriptomic and clinical data were analyzed from 733 patients across two frontline treatment cohorts, sunitinib (n = 376) and avelumab plus axitinib (n = 357), derived from JAVELIN Renal 101. A multi-algorithm feature selection framework was applied to identify transcriptomic signatures associated with progression-free survival (PFS) and overall survival (OS). Prognostic performance was evaluated using the concordance index (C-index). Predictive models for therapeutic response, including disease control, were developed using PFS-derived gene signatures and assessed by area under the curve (AUC). External validation was performed in an independent cohort from The Ohio State University Total Cancer Care (OSU TCC) (n = 114). Results: ML-derived transcriptomic models consistently stratified patients into distinct risk groups with improved prognostic discrimination compared with standard clinical classifiers. In the sunitinib cohort, the best-performing models achieved C-indices of 0.72 for PFS and 0.81 for OS, outperforming IMDC (0.59 and 0.66). In the validation set of the sunitinib cohort, high-risk patients exhibited worse outcomes, with hazard ratios of 3.00 for PFS (P < 0.001, 95% CI, 2.06–4.39) and 13.42 for OS (P < 0.001, 95% CI, 7.78–23.13). In the avelumab plus axitinib cohort, C-indices reached 0.70 for PFS and 0.79 for OS. Consistent risk stratification was observed in the validation set, with hazard ratios of 3.16 for PFS (P < 0.001, 95% CI, 2.07–4.83) and 4.69 for OS (P < 0.001, 95% CI, 2.65–8.30). For response prediction, the models demonstrated predictive performance, with the Naive Bayes model achieving a validation AUC of 0.83 for disease control in both sunitinib and avelumab plus axitinib cohorts. The model showed significant risk stratification in an external validation cohort (OSU TCC). Conclusions: This study presents a multi-cohort transcriptomic framework with prognostic and predictive utility in advanced RCC. By outperforming established clinical risk classifiers and enabling prediction of regimen-specific therapeutic responses, this ML-based approach supports biomarker-informed frontline treatment selection.