Mahadevia, Himil

7548 Background: Glucagon-like peptide-1 (GLP-1) receptor agonists are widely used for the treatment of type 2 diabetes and obesity, and emerging evidence suggests their benefits beyond glycemic control, including cardiovascular effects and potential anti-inflammatory and anti-proliferative properties. Therapeutic advances in multiple myeloma (MM) have improved the overall life-expectancy in our patients. However, metabolic comorbidities remain common, likely driven by disease burden, functional decline, ongoing inflammation, and treatment related effects. Our study evaluates the impact of GLP-1 receptor agonists on outcomes in patients with MM, exploring potential interactions between metabolic modulation and disease outcomes. Methods: We conducted a real-world retrospective cohort study using the TriNetX Global Collaborative Network, covering January 2000 to December 2025, encompassing data from 168 global healthcare organizations. Patients aged 18 and above with multiple myeloma were identified and then stratified into two groups based on treatment with GLP-1 agonists or not. The two groups were then propensity-matched based on age, sex, race, and various comorbidities. We assessed outcomes including overall mortality, myocardial infarction (MI), ischemic stroke, sepsis, anemia, pancreatitis, venous thromboembolism (VTE), and ED visits. Time-to-event analyses were performed using Kaplan-Meier methods and Cox proportional hazards models. Results: We identified 1716 patients with MM and obesity who received GLP-1 receptor agonists, and 13,491 patients with MM and obesity who did not receive GLP-1 therapy. After propensity matching, each cohort consisted of 1712 patients with well-balanced baseline characteristics. We followed these patients for 5 years from treatment initiation. Patients with MM who received GLP-1 had a significantly lower risk of overall mortality (Hazard Ratio [HR]: 0.423, 95% CI: 0.351-0.509, p-value <0.001), MI (Risk Difference [RD]: -3.6 %, 95% CI: -5.4 % to -1.7 %, p-value <0.001), ischemic stroke (RD: -1.7 %, 95% CI: -3.3 % to -0.1 %, p-value = 0.04), sepsis with and without shock (RD: -3.0 %, 95% CI: -4.9 % to -1.1 %, p-value = 0.002), and anemia (RD: -9.6 %, 95% CI: -12.9 % to -6.4 %, p-value <0.001). There was no statistically significant difference in the rates of VTE, ED visits, and pancreatitis between the two cohorts. Conclusions: Our study revealed that patients with MM with obesity who received GLP-1 therapy had significantly lower rates of overall mortality, MI, ischemic stroke, sepsis with and without shock, and anemia. Additional longitudinal cohort studies are required to explore the associations between these agents and inform clinical practice guidelines in these patients.

7064 Background: Chimeric antigen receptor (CAR) T cell therapy has revolutionized care for relapsed/refractory diffuse large B cell lymphoma (DLBCL), yet granular delineation of real-world immune effector toxicities, organ-specific sequelae and economic burden across demographic and socioeconomic strata requires further exploration. Methods: Using the National Inpatient Sample (2017–2022), we identified adult DLBCL hospitalizations receiving CAR T through ICD-10-PCS codes. Cytokine release syndrome (CRS) [all grades - 2021–2022], immune effector cell–associated neurotoxicity syndrome (ICANS) [all grades – 2022], infections, cardiac events and acute kidney injury (AKI) were identified through ICD-10 codes. Age was stratified into 21-40, 41-60, 61-80 and >81 years. Payer groups were Medicare, Medicaid, Private and Uninsured. Races were White, African American, Hispanic and others [Pacific Islanders, Native Indians, Asians]. Chi-square analysis was performed, and multivariate regression adjusted for demographics, comorbidities, and hospital characteristics. Results: Across >11,000 weighted hospitalizations, inpatient mortality, all-grade CRS and ICANS were comparable across age, race, and payer groups (p>0.05). Inpatient mortality was numerically higher in other minor races (6.6%) vs White (3.3%), African American (3.1%), and Hispanic (0.9%; p=0.230). Any grade CRS was numerically higher in other minor races (72.5%) than in White (62.3%) and African American (60.6%; p=0.614). Any grade ICANS dispersion included 29.5% (White), 39.1% (African American), 22.5% (Hispanic) and 11.1% (Other; p=0.204). A race-specific signal for ICANS grade 5 (p=0.030) was seen, but numbers were small. Among cardiac events, arrhythmias rose from 7.6% (age 21–40 years) to 57.1% ( age ≥81 years) (p<0.001). By payer, arrhythmias were highest in Medicare (31.2%; p<0.001). Other cardiac events including myocardial infarction, stroke and cardiogenic shock were similar across groups. Pneumonia rates varied by payer, peaking in the Uninsured (23.8%; p=0.010). Sepsis and septic shock rates showed no differences. AKI (p=0.007) rates showed differences but attenuated after adjustment. Hispanic and other minor races had lower adjusted costs than White (Odds ratio [OR] 0.9, p=0.046; OR 0.8, p=0.006), while ages 61–80 had higher adjusted costs vs 21–40 (OR 1.2, p=0.023). Conclusions: In this national cohort, CRS, ICANS and mortality were similar across demographic and payer strata, indicating consistent immune effector safety. Cardiac arrhythmia rates were higher among the elderly and Medicare beneficiaries. There were lower expenditures among minor race groups and higher costs in elderly. These findings highlight discrete domains of variability and suggest intensified cardiac surveillance in elderly and equitable resource allocation to optimize outcomes with CAR-T.

5543 Background: Glucagon-like peptide-1 (GLP-1) receptor agonists are widely used for the treatment of type 2 diabetes and obesity. Emerging evidence suggests that these agents may have potential anti-tumor activity as GLP-1 signaling may influence pathways involved in cell proliferation and inflammation. Endometrial cancer is the most common gynecologic malignancy, and obesity represents a well-established, modifiable risk factor that may adversely affect oncologic outcomes through chronic inflammation and metabolic dysregulation. Our study aims to evaluate the impact of GLP-1 receptor agonists on outcomes among patients with endometrial cancer and comorbid obesity. Methods: A retrospective cohort study was conducted using the TriNetX Global Collaborative Network, covering January 2000 to December 2025, encompassing data from 168 global healthcare organizations. Patients aged 18 and above with endometrial cancer and comorbid obesity were identified and then stratified into two groups based on treatment with GLP1 agonists or not. The two groups were then propensity-matched based on age, ethnicity, medical comorbidities, and stage of cancer. Outcomes were assessed for five years from treatment initiation and included overall mortality, myocardial infarction (MI), ischemic stroke; development of peritoneal, lymph node, and visceral metastases; sepsis, pancreatitis, and ED visits. Time-to-event analyses were performed using Kaplan-Meier methods and Cox proportional hazards models. Results: After propensity matching, 6352 patients were included in each cohort with balanced baseline characteristics. Our analysis found that patients with endometrial cancer and comorbid obesity who received GLP-1 agonists had a significantly lower risk of overall mortality (Hazard Ratio [HR]: 0.45, 95% CI: 0.39-0.51, p-value <0.001), development of peritoneal metastases (Risk Difference [RD]: -2.16%, 95% CI: -2.74% to -1.58%, p-value <0.001), nodal metastases (RD: -0.41%, 95% CI: -0.72% to -0.09%, p-value =0.01), visceral metastases (RD: -3.76%, 95% CI: -4.60% to -2.93%, p-value =0.01), and sepsis with and without shock (RD: -1.42%, 95% CI: -2.14% to -0.70%, p-value <0.001) compared to patients who did not receive GLP-1 agonists. There was no statistically significant difference in the rates of MI, ischemic stroke, ED visits, and pancreatitis between the two cohorts. Conclusions: In this large real-world analysis, patients with endometrial cancer and comorbid obesity who received GLP-1 agonists had lower rates of overall mortality; peritoneal, lymph node, and visceral metastases, and sepsis with or without shock. Additional longitudinal cohort studies are imperative to explore the associations between these agents and inform clinical practice guidelines in this patient population.

1615 Background: Immune checkpoint inhibitors (ICIs) have revolutionized lung cancer treatment but cause adverse events, including immune-related colitis (ir-colitis). Accurate identification of ir-colitis is critical for improving patient outcomes and clinical research. However, current methods relying on ICD codes have substantial limitations: non-specific categories that fail to capture the cases or distinguish cases from infectious or ischemic etiologies, as well as incomplete coding that leads to missed cases. Large language models (LLMs), with advanced natural language understanding capabilities, offer a scalable approach for automated phenotyping using unstructured clinical notes. We aimed to develop and evaluate an LLM-based approach for identifying ir-colitis in patients with lung cancer receiving ICIs and compare it with ICD code based method. Methods: We identified 5,278 lung cancer patients with ICI administration at Mayo Clinic (2012-2025). To investigate ir-colitis occurrence within two years after ICI initiation, patients were stratified into three ICD-based cohorts: (1) IR-Colitis, with clinician-assigned ir-colitis-specific diagnosis codes; (2) Non-IR Colitis, with colitis diagnosis codes excluding ir-colitis; and (3) No-Colitis, absent of any colitis-related codes. We used three open-source LLMs to process clinical notes : Llama4, MedGemma, and Qwen. Prompts were designed to extract both gold standard evidence (explicit ir-colitis terminology, such as "immune-mediated colitis" and "checkpoint inhibitor colitis") and supportive evidence (such as pathology and lab results). Results: Llama4 demonstrated superior accuracy among three LLMs in manual evaluation and was thus selected for all cohort analyses. Among ICD-based IR-Colitis patients (n = 339), 82.0% had confirmatory evidence while 18% lacked evidence, suggesting potential false-positive coding. In the Non-IR Colitis cohort (n = 297), LLM-based extraction identified 43.8% patients in fact had ir-colitis, demonstrating substantial misclassification. In the No-colitis cohort (n = 4,642, random sample of 996 for analysis), the LLM identified 10.7% with documented ir-colitis, revealing substantial missing by administrative coding. Conclusions: LLM-based phenotyping demonstrates superior sensitivity for identifying ir-colitis compared to ICD codes alone, successfully uncovering cases missed by administrative coding. This scalable approach enables more comprehensive case identification for immunotherapy safety research and clinical surveillance. ICD-Based Cohorts LLM-Identified ir-Colitis Gold Standard Evidence Supportive Evidence Any Evidence ir-Colitis (n = 339) 277/339 (81.7%) 275/339 (81.1%) 278/339 (82.0%) Non-ir-Colitis (n = 297) 129/297 (43.4%) 126/297 (42.4%) 130/297 (43.8%) No Colitis (n = 996) 106/996 (10.6%) 104/996 (10.4%) 107/996 (10.7%)

10576 Background: Type 2 diabetes mellitus (T2DM) and excess adiposity are independent risk factors for pancreatic cancer. Pancreatic cancer risk within T2DM across body-mass index (BMI) strata is heterogeneous in real-world cohorts and lean T2DM individuals may represent a distinct high risk phenotype. We compared pancreatic cancer incidence and post-diagnosis survival across BMI strata. Methods: We performed a retrospective comparative outcomes analysis using the TriNetX network. Adults with T2DM (HbA1c ≥6.5%) were categorized as obese (BMI ≥30 kg/m²) or lean (BMI <25 kg/m²). The lean cohort additionally excluded any patient with prior BMI of >25 kg/m² to control for cancer associated weight loss. The index event was first T2DM diagnosis, and outcomes were evaluated after a 365-day window. Patients with pancreatic cancer prior to the outcome window were excluded. Propensity score matching was done using the following parameters: age, sex, smoking status, pancreatitis, pancreatic cysts, chronic kidney disease, gallstones, family history of pancreatic cancer, and HbA1c. Outcomes assessed were incident pancreatic cancer over 10 years and all-cause mortality among patients with pancreatic cancer and T2DM. Results: Total number of patients in the obese group was 1,759,075 with mean age of 57 ± 13.4 years and in the lean cohort were 205,193 with a mean age of 61.8 ± 16.6 years. After propensity score matching, there were 204,086 patients in each group. Mean HbA1c levels were 7.4 ± 1.8% in the obese group and 7.9 ± 2.2% in the lean BMI group. The incidence of pancreatic cancer was 0.16% (n = 335) in obese patients as compared to 0.23% in lean (n = 478) patients with a risk difference of 0.07% (p <0.001). Obesity was associated with lower pancreatic cancer risk (RR 0.692, 95% CI 0.603–0.797; p < 0.001). In the subgroup analysis of patients diagnosed with pancreatic cancer, obese patients had longer median survival (1,109 vs 651 days) with HR 0.853 (95% CI 0.793 - 0.918; p <0.001). Conclusions: Lean adults with T2DM had a higher incidence of pancreatic cancer and worse post-diagnosis survival than obese patients. These results identify lean T2DM as a distinct high-risk phenotype and demonstrate that pancreatic cancer risk within T2DM is heterogeneous and not solely attributable to excess adiposity. Further investigation is needed to elucidate the biological mechanisms underlying this association and its implications for clinical risk assessment. Pancreatic cancer incidence and associated risk estimates among obese and lean adults with type 2 diabetes mellitus. Cohort Pancreatic cancer incidence RD RR Obese T2DM(n = 204,086) 0.16% (n = 335) 0.07% (95% CI: 0.04 - 0.10), p < 0.001 1.44 (95% CI: 1.25 - 1.66) Lean T2DM(n = 204,086) 0.23% (n = 478) Abbreviations: CI, confidence interval; RD, risk difference; RR, relative risk; T2DM, type 2 diabetes mellitus.