Citations

Abstract Background In the current clinical workflow of radiation oncology, therapists manually summarize the physician‐issued Computed Tomography (CT) simulation orders to prepare patients' CT simulations. This process increases the workload, introduces variability in documentation quality, and is prone to human errors. Purpose This study aims to address these challenges using a large language model (LLM) to automate the generation of summaries from the CT simulation orders and evaluate its performance. Methods A total of 607 CT simulation orders were collected from the Aria database at our institution. A locally hosted LLaMa 3.1 405B model, accessed via the Application Programming Interface (API) service, was used to extract keywords from the CT simulation orders and generate summaries. The downloaded CT simulation orders were categorized into seven groups based on treatment modalities and disease sites. For each group, a customized instruction prompt was developed collaboratively with therapists to guide the LLaMa 3.1 405B model in generating summaries. The ground truth for the corresponding summaries was manually derived by carefully reviewing each CT simulation order and subsequently verified by therapists. The accuracy of the LLM‐generated summaries was evaluated by therapists using the ground truth. Results Over 98% of the LLM‐generated summaries aligned with the ground truth in terms of accuracy. Our evaluations showed an improved consistency in format and enhanced readability from the LLM‐generated summaries compared to the corresponding therapist‐generated summaries. This automated approach demonstrated consistent performance across all groups, regardless of treatment modality or disease site. Conclusions This study demonstrated the high precision and consistency of the LLaMa 3.1 405B model in extracting keywords and summarizing CT simulation orders, suggesting that LLMs have great potential to assist in this task, reduce the workload of CT simulation therapists and improve radiation oncology workflow efficiency.

Background Immune checkpoint therapies targeting programmed cell death protein 1 (PD-1) and programmed death ligand 1 (PD-L1) have been met with limited clinical responses in the treatment of patients of varying cancer types. Current US Food and Drug Administration (FDA)-approved therapies function to prevent PD-1 binding to PD-L1, thereby preventing T-cell suppression. However, these therapies fail to prevent PD-L1 intrinsic signaling and recycling, which can render these treatments ineffective. Moreover, most research primarily focuses on how these antibodies affect tumor cells and their interactions with T cells. Myeloid cells, which also express PD-L1 and influence T-cell responses, are integral to the efficacy of these therapies. However, strategies that can target these cells to boost antitumor responses have not been effective. Methods Our group has developed a new anti-PD-L1 antibody (H1A), which disrupts PD-L1 recycling and redirects it towards degradation. Due to the loss of available PD-L1 protein, H1A effectively prevents PD-L1 intrinsic signaling. Using in vivo humanized PD-1/PD-L1 mouse tumor models, we evaluated the therapeutic efficacy of H1A against current FDA-approved PD-L1 blocking antibodies. Using human-derived immune cell in vitro assays, we pinpoint the biological consequences of H1A-induced PD-L1 degradation in myeloid cells. Results H1A demonstrated improved tumor control and established immunological memory responses in vivo in the treatment of tumors with moderate immunogenicity that are less responsive to current immunotherapies. This was supported by enhanced activation of myeloid cells (major histocompatibility complex (MHC)-II and CD80) and frequencies of effector T-cell populations found intratumorally. Human myeloid cells treated with H1A were also observed to have increased activation (MHC-II and CD80) and cytokine secretion. Similar to our in vivo findings, human peripheral blood lymphocyte cultures demonstrated increased frequencies of effector T-cell populations and greater tumor cell killing. Conclusions This work highlights the importance of targeting PD-L1 recycling and intrinsic signaling, specifically in myeloid cells, to boost antitumor responses. Our studies suggest that H1A can provide a solution to the lack of responses seen with current therapeutics, while also revealing previously unknown intrinsic functions of PD-L1 in myeloid cells.

Background Immune-related adverse events (irAEs) remain largely unpredictable, potentially affecting multiple organ systems and occurring at almost any point during and even occasionally after immune checkpoint inhibitor (ICI) treatment. To identify populations at risk for these immune-mediated toxicities, we analyzed genetic characteristics and immune markers associated with clinically significant irAEs. Methods We carried out a genome-wide association study on 373 white patients receiving ICI treatment. We identified single nucleotide polymorphisms associated with irAEs. Blood cytokine profiling and peripheral blood mononuclear cell RNA sequencing were performed at pretreatment baseline and 6–8 weeks after ICI initiation. Findings were validated in two external cohorts. Results We identified genetic haplotypes in VWA8 (Von Willebrand Factor A Domain Containing 8) , OSBPL6 (Oxysterol Binding Protein Like 6), and ADAMTS9-AS2 (ADAM Metallopeptidase With Thrombospondin Type 1 Motif 9 Antisense RNA 2) associated with grade ≥2 irAEs. Patients carrying risk haplotypes for one or more genes exhibited significantly greater rates of grade ≥2 (OR 3.02; 95% CI 1.83 to 5.02; p<0.001), grade ≥3 (OR 3.59; 95% CI 1.93 to 6.64; p<0.001), and multiple type irAE (OR 2.60; 95% CI 1.53 to 4.39; p<0.001). Serum CCL3 levels were significantly elevated in individuals carrying risk haplotypes (p=0.03). Gene expression analysis demonstrated activated autoimmune and inflammatory pathways in the genetic risk group. Conclusions Novel polymorphisms in VWA8, OSBPL6, and ADAMTS9-AS2 may impact immune pathways, promote inflammation, potentiate autoimmune phenotypes, and convey risk of irAE in ICI-treated patients.

ABSTRACT Objectives This portion of the Geriatric Emergency Department (GED) Guidelines 2.0 focuses on delirium in the emergency department (ED). Methods A multidisciplinary group applied the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach to assess the certainty of evidence and develop recommendations related to older ED patients with possible delirium. Results The GED Guidelines 2.0 Delirium Work Group derived six evidence‐based recommendations for risk stratification, diagnosis, and brain imaging. To reduce universal screening, the Delirium Risk Score may be used to identify older adults at low risk for delirium, though the evidence certainty is very low. In adults over 65 admitted to ED observation units, Zucchelli's risk assessment tool (threshold ≥ 4) may stratify delirium risk, also with very low certainty. For adults over 75, the REDEEM Score may be used to identify low‐ or high‐risk individuals, again with very low certainty. For diagnosis, 4AT, bCAM, CAM‐ICU, mCAM, AMT‐4, or RASS may be used to rule delirium in or out, based on very low certainty. The Delirium Triage Screen (DTS) may be used to rule out, but not to rule in, delirium, also with very low certainty. For diagnostic imaging, there is very low certainty of evidence to recommend for or against obtaining a head CT as part of the evaluation for older ED patients with delirium. All recommendations are conditional, reflecting very low certainty of evidence due to the lack of high‐quality ED‐based studies and comparative effectiveness research. Conclusion Rigorous ED‐based research is needed to strengthen evidence and guide delirium care for older adults in geriatric emergency medicine.

ABSTRACT Background The impact of psychosocial health and socioeconomic status (SES) on long‐term lung transplant survival remains unclear. Methods We conducted a single‐center, retrospective 2:1 matched case‐control study of adult lung transplant recipients (1999–2019). Cases survived ≥6 years post‐transplant; controls survived <6 years. Matching was based on transplant year (±5 years) and indication. Psychosocial health and SES were assessed using the Psychosocial Assessment of Candidates for Transplant (PACT) score, Mayo Clinic Housing‐Based Socioeconomic Status (HOUSES) index, and Area Deprivation Index (ADI). Secondary covariates included insurance, marital status, employment, caregiver relationship, and proximity to the transplant center. Conditional logistic regression adjusted for age and gender was used. Results Among 189 patients (63 cases, 126 controls), PACT score, HOUSES index, and ADI were not significantly associated with long‐term survival. Patients with non‐family caregivers had lower odds of survival compared to those with a spouse or partner (OR = 0.16, p = 0.022). Regional residents had lower odds of long‐term survival compared to local recipients (OR = 0.55, p = 0.001). Conclusions Pre‐transplant psychosocial and SES measures were not significantly associated with long‐term survival. Caregiver relationships and geographic proximity may have a greater impact on outcomes, highlighting the importance of support systems and access to care.

Introduction Social media platforms, particularly X (formerly Twitter), play an influential role in professional networking and information dissemination. This study aimed to examine X usage trends among faculty in a single department of surgery at a large academic institution. Methods Faculty X accounts were analyzed from January to June 2022. Metrics such as follower count, number of accounts followed, and post (formerly: tweets) types (original/quoted reposts) were studied. Themes of posts were categorized as personal, promotional, educational, or related to academic events/conferences. Associations between faculty rank, division, gender, and post metrics were explored using Mann-Whitney U and Kruskal-Wallis tests. Results Of 133 faculty members, 53 (39.85%) had X accounts, with 3 excluded due to privacy settings. A total of 3476 posts were analyzed, averaging 0-34 posts/month per faculty. Categories included professional (63%, n = 2174), educational (19%, n = 653), and personal (13%, n = 436). No significant associations were found between faculty gender/division and posting rates. However, faculty rank correlated significantly with primary post categories ( P < 0.01). Associate professors, and professors posted more about academic events (2% and 1%) than instructors or assistant professors (1%). The most followed faculty actively promoted colleagues’ work through reposting. Conclusion Less than half the faculty in a single department of surgery engaged with X. Those active predominantly used it for promotion, followed by education and personal use. Adoption remains limited, highlighting opportunities for growth in utilizing X for academic dissemination.

Abstract Context Preclinical studies demonstrate that treatment with calcitriol attenuates skeletal complications in mice with X-linked hypophosphatemia (XLH). Objective This work aimed to assess serum markers of mineral metabolism, nephrocalcinosis, skeletal microarchitecture, growth, and evidence of rickets following calcitriol monotherapy in children and adults with XLH. Methods A 1-year, prospective, single-arm, open-label study was conducted comparing baseline and 12-month outcomes. Participants were recruited from outpatient endocrinology clinics in the United States. Data were collected in a research unit. Eligible participants were aged 4 years or older, not pregnant, with baseline 25-hydroxyvitamin D of 20 ng/mL or greater and normal serum calcium. Participants were treated with calcitriol, with dose based on serum and urinary calcium concentrations, for 1 year. The primary end points were change in serum phosphate and nephrocalcinosis (all participants) and rickets severity score (children). Secondary end points included changes in additional markers of mineral metabolism, skeletal microarchitecture parameters via high-resolution peripheral quantitative computed tomography (all participants), and height z-scores (children). Results Serum phosphate did not change over 12 months of optimized calcitriol treatment. Rickets improved in 2 of 4 children with open epiphyses; height z-scores were unchanged. Nephrocalcinosis scores remained stable. Adults had increased cortical thickness at the radius diaphysis (P = .012). Serum alkaline phosphatase decreased in children (P = .021) and parathyroid hormone trended lower both in children (P = .065) and adults (P = .063). Four participants developed mild hypercalcemia and 2 developed hypercalciuria that resolved with calcitriol titration. Conclusion Calcitriol monotherapy is safe and well-tolerated in XLH, with modest benefits on laboratory indices of mineral metabolism and rickets.

Abstract Disclosure: N.M. Singh Ospina: None. Y. Chen: None. F. Du: None. J.P. Brito: None. H. Shao: None. Q. Liu: None. R.G. McCoy: None. T. Jiao: None. Background: The levothyroxine dose for patients with hypothyroidism is highly correlated with weight. In patients with hypothyroidism, TSH measurements are used to assess the adequacy of levothyroxine dosing.Since type 2 diabetes (T2D) and obesity are two common comorbidities for patients with hypothyroidism, and the significant weight loss associated with Glucagon-Like Peptide-1 Receptor Agonists (GLP-1 RAs) therapy, the initiation of GLP-1 RA on levothyroxine users may significant increase the need for levothyroxine dose reduction potentially requiring additional TSH tests to assure control. Objectives: To evaluate the timing and frequency of TSH testing within one year of initiation of GLP-1, or sodium-glucose co-transporter 2 (SGLT-2) inhibitors, in patients with T2D on levothyroxine treatment. Methods: We identified a cohort of T2D patients using 15% of Medicare fee-for-service claims data from 2005 to 2020. Cohort entry was defined by the initiation of GLP-1 or SGLT2. Patients needed to continuously receive the same dose of levothyroxine for &gt;6 months and have at least 1 TSH measurement before the index date. We excluded patients with hyperthyroidism, thyroid cancer, end-stage renal disease, HIV/AIDs, organ transplant, or who received medications that affect thyroid levels before the index date. The primary outcome was the timing and frequency of TSH testing within one year after GLP-1 RA initiation. Demographics, comorbidities, and medication use were balanced using 1:1 propensity score matching based on the index date. Results: After matching, no significant demographic differences were observed among the 2,107 patients in GLP-1 RA and SGLT-2 groups, respectively. Overall, 32.1%, 32.3%, and 35.6% were aged 65-70, 70-74, and over 75, with 69.3% female and 87.9% white. In the year before the index date, the average number of TSH tests was higher in the GLP-1 RA (1.16 ± 1.27) compared to the SGLT2 (1.07 ± 1.18, p-value: 0.001). In the year after the index date, the average number of TSH tests declined to 0.93 ± 1.18 for GLP-1 and 0.89 ± 1.13 for SGLT2 (p-value: 0.03). Most patients (48.6 vs. 49.5%) had no TSH tests, 41.3 vs. 40.6% had 1-2 tests, and 10.1 vs. 9.9% had more than 2 tests in the GLP-1 RA and SGLT-2 group, respectively. Fewer patients (32.3 vs. 33.6%) were tested within 0-6 months while the majority (67.7 vs. 66.5%) waited longer than 6 months. The relative risk of TSH testing between the GLP-1 RA group and the SGLT2 group was 1.03 (95% CI: 0.98 to 1.08) 1-year before the index date and 1.00 (95% CI: 0.94 to 1.06) 1-year after the index date. Conclusions: Despite the potential impact of GLP-1 RA on levothyroxine dosing, on average, half of patients did not receive a TSH test after treatment initiation. There were no significant differences in TSH testing patterns between GLP-1 and SGLT2 groups, with similar test frequencies and timing observed. Presentation: Saturday, July 12, 2025

Abstract Glioblastoma (GBM) is a highly aggressive brain tumor characterized by invasive growth, genetic heterogeneity and resistance to therapy. Alterations in the epithelial growth factor receptor (EGFR) are frequently detected in GBM, including gene amplifications, mutations and exon deletions, exemplified by the expression of EGFR variant III (EGFRvIII). To target both wild-type and EGFRvIII, we generated the humanized antibody A10, which reacts with the 287-302 loop that is fully exposed on the extracellular domain of EGFRvIII and conditionally exposed on cells expressing gene-amplified EGFR. However, A10 showed little or no reactivity for cells expressing normal levels of wild type EGFR. We engrafted antibody fragments derived from A10 on to ARTEMIS® Cell Receptor Platform, which introduces two separate receptors into engineered T cells to provide activation and co-stimulatory signals. Specifically, the A10 antigen-binding fragment (Fab) was fused to the N-terminus of a γδTCR-signaling domains (A10-AbTCR) while the A10 single chain variable fragment (scFv) was fused to a costimulatory signaling receptor (A10-CSR). The resulting A10 ARTEMIS® T cells were cytotoxic for tumor cell lines displaying amplified or overexpressed EGFR and EGFRvIII. Further, glioblastoma PDXs were sensitive to A10 ARTEMIS® T cell killing. This prompted an in vivo evaluation in EGFR-driven cancers. GBM39 PDX tumors (expressing amplified EGFRvIII) were implanted into the brains of NSG mice. Mice received one or two infusions. Treatment with one infusion of A10 ARTEMIS® cells resulted in prolonged survival of tumor bearing mice with 40 % of mice were long-term survivors (74 days) compared to a median survival of 32 days (control) or 52 days (mock T-cells). For mice receiving two A10 ARTEMIS® T cell infusions, there was 100% survival of mice beyond 60 days, while all control and mock treated mice were dead by 28 days. Further, treatments of EGFR-amplified MDA-MB-468 tumors resulted in complete tumor regression with one treatment of A10 ARTEMIS® cells. Thus, this new format of A10 engineered T-cells appears to be a particularly promising for treating cancers with high levels of EGFR or EGFRvIII including GBM. Citation Format: Antonella Antignani, Danielle Burgenske, Guangyan Xiong, Hongbing Zhang, Cheng Liu, Jann Sarkaria, David FitzGerald. Engineered T-cells, targeting overexpressed EGFR and EGFRvIII, produce complete remissions in a PDX model of human glioblastoma [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference on Molecular Targets and Cancer Therapeutics; 2025 Oct 22-26; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2025;24(10 Suppl):Abstract nr B056.