Citations

Abstract Objective In acupuncture therapy, the accurate location of acupoints is essential for its effectiveness. The advanced language understanding capabilities of large language models (LLMs) like Generative Pre-trained Transformers (GPTs) and Llama present a significant opportunity for extracting relations related to acupoint locations from textual knowledge sources. This study aims to explore the performance of LLMs in extracting acupoint-related location relations and assess the impact of fine-tuning on GPT’s performance. Materials and Methods We utilized the World Health Organization Standard Acupuncture Point Locations in the Western Pacific Region (WHO Standard) as our corpus, which consists of descriptions of 361 acupoints. Five types of relations (“direction_of”, “distance_of”, “part_of”, “near_acupoint”, and “located_near”) (n = 3174) between acupoints were annotated. Four models were compared: pre-trained GPT-3.5, fine-tuned GPT-3.5, pre-trained GPT-4, as well as pretrained Llama 3. Performance metrics included micro-average exact match precision, recall, and F1 scores. Results Our results demonstrate that fine-tuned GPT-3.5 consistently outperformed other models in F1 scores across all relation types. Overall, it achieved the highest micro-average F1 score of 0.92. Discussion The superior performance of the fine-tuned GPT-3.5 model, as shown by its F1 scores, underscores the importance of domain-specific fine-tuning in enhancing relation extraction capabilities for acupuncture-related tasks. In light of the findings from this study, it offers valuable insights into leveraging LLMs for developing clinical decision support and creating educational modules in acupuncture. Conclusion This study underscores the effectiveness of LLMs like GPT and Llama in extracting relations related to acupoint locations, with implications for accurately modeling acupuncture knowledge and promoting standard implementation in acupuncture training and practice. The findings also contribute to advancing informatics applications in traditional and complementary medicine, showcasing the potential of LLMs in natural language processing.

Abstract Background Delirium is a common complication of older people in hospitals, rehabilitation and long-term facilities. Objective To assess the worldwide use of validated delirium assessment tools and the presence of delirium management protocols. Design Secondary analysis of a worldwide one-day point prevalence study on World Delirium Awareness Day, 15 March 2023. Setting Cross-sectional online survey including hospitals, rehabilitation and long-term facilities. Methods Participating clinicians reported data on delirium, the presence of protocols, delirium assessments, delirium-awareness interventions, non-pharmacological and pharmacological interventions, and ward/unit-specific barriers. Results Data from 44 countries, 1664 wards/units and 36 048 patients were analysed. Validated delirium assessments were used in 66.7% (n = 1110) of wards/units, 18.6% (n = 310) used personal judgement or no assessment, and 10% (n = 166) used other assessment methods. A delirium management protocol was reported in 66.8% (n = 1094) of wards/units. The presence of protocols for delirium management varied across continents, ranging from 21.6% (on 21/97 wards/units) in Africa to 90.4% (235/260) in Australia, similar to the use of validated delirium assessments with 29.6% (29/98) in Africa to 93.5% (116/124) in North America. Wards/units with a delirium management protocol [n = 1094/1664, 66.8%] were more likely to use a validated delirium test than those without a protocol [odds ratio 6.97 (95% confidence interval 5.289–9.185)]. The presence of a delirium protocol increased the chances for valid delirium assessment and, likely, evidence-based interventions. Conclusion Wards/units that reported the presence of delirium management protocols had a higher probability of using validated delirium assessments tools to assess for delirium.

IntroductionAlzheimer's disease (AD) is a neurodegenerative disease with neurogenic fiber tangles caused by amyloid-β protein plaques and tau protein hyperphosphorylation as the pathological manifestations. This study was based on multi-omics to investigate the mechanisms and immune characterization of AD.Material and methodsBased on bulk RNA-seq (GSE122063 and GSE97760), we screened potential biomarkers for AD by differential expression analysis and machine learning algorithms. Then, we elaborated the expression characteristics and immune functions of the above biomarkers by scRNA-seq (single-cell RNA sequencing) data analysis (GSM4996463 and GSM4996462) and immune infiltration analysis.ResultsFive biomarkers (RBM3, GOLGA8A, ALS2, FSD2, and LOC100287628) were identified using machine learning algorithms. Single-cell analysis revealed distinct expression patterns of these biomarkers in astrocytes from AD samples compared to normal samples. Additionally, three key biomarkers were selected based on interaction networks, and the diagnostic models indicated high diagnostic efficacy for these biomarkers. Based on immune infiltration and correlation analyses, RBM3, GOLGA8A, and ALS2 were all highly correlated with CD8 T cell content in the immune microenvironment of AD.ConclusionsThe biomarkers identified in this study demonstrate significant diagnostic potential for AD. Notably, the downregulation of RBM3 in astrocytes and the decreased presence of CD8 T cells infiltrating brain tissue are potential risk factors for AD.

TPS622 Background: Approximately 50% of newly diagnosed invasive breast cancers are stage 1, with the majority being ER/PR-positive, HER2-negative. Genomic assays such as the Oncotype DX have identified patients (pts) with reduced risk of distant metastasis and without benefit from chemotherapy added to endocrine therapy, freeing them from excess toxicity. Genomic assays are also recognized as prognostic for in-breast recurrence (IBR) after BCS and could similarly allow de-escalation of adjuvant radiotherapy (RT). Reducing overtreatment is of interest to pts, providers, and payers. Methods: We hypothesize that breast-conserving surgery (BCS) alone is non-inferior to BCS plus RT for IBR and breast preservation in women intending endocrine therapy (ET) for stage 1 invasive breast cancer (ER and/or PR-positive, HER2-negative with an Oncotype DX Recurrence Score [RS] of ≤18). Stratification is by age (<60; ≥60), tumor size (≤1 cm; >1-2cm), and RS (≤11, >11-18/MammaPrint Low). Pts are randomized post-BCS to Arm 1 with breast RT using standard methods (moderate or ultra hypo- or conventional-fractionated whole breast RT with/without boost, or APBI) with ≥5 yrs of ET (tamoxifen or AI) or Arm 2 with ≥5 yrs of ET (tamoxifen or AI) alone. The specific regimen of ET in both arms is at the treating physician’s discretion. Eligible pts are stage 1: pT1 (≤2 cm), pN0, age ≥50 to <70 yrs, s/p BCS with negative margins (no ink on tumor), s/p axillary nodal staging (SNB or ALND), ER and/or PR-positive (ASCO/CAP), HER2-negative (ASCO/CAP), and Oncotype DX RS ≤18 (diagnostic core biopsy or resected specimen); (a “low risk” MammaPrint is permissible if completed as part of usual care prior to screening). Primary endpoint is IBR (invasive breast cancer or DCIS). Secondary endpoints are breast conservation rate, invasive in-breast recurrence, relapse-free interval, distant disease-free survival, overall survival, patient-reported breast pain, patient-reported worry about recurrence, and adherence to ET. We assume a clinically acceptable difference in IBR of 4% at 10 yrs to judge omission of RT as non-inferior (10-yr event-free survival for RT group is 95.6% v 91.6% for the omission-of-RT group). BR007 is powered to detect non-inferiority with 80% power and a one-sided α=0.025, assuming that there would be a ramp-up in accrual in the first two years (leveling off in yrs 3-5); 1,670 pts (835 per arm) are required for randomization. Conservative loss to follow-up is 1%/yr. Some T1a pts screened may have Oncotype DX scores >18, making them ineligible for the study. In the accrual process, 1,714 pts will be required to register to ensure that our final randomized cohort is 1,670 pts. Current accrual (02-05-2024) is 805 screened and 716 randomly assigned. Clinical trial information: NCT04852887 .

: Cerebral Edema (CE) is the final common pathway of brain death. In severe neurological disease, neuronal cell damage first contributes to tissue edema, and then Increased Intracranial Pressure (ICP) occurs, which results in diminishing cerebral perfusion pressure. In turn, anoxic brain injury brought on by decreased cerebral perfusion pressure eventually results in neuronal cell impairment, creating a vicious cycle. Traditionally, CE is understood to be tightly linked to elevated ICP, which ultimately generates cerebral hernia and is therefore regarded as a risk factor for mortality. Intracranial hypertension and brain edema are two serious neurological disorders that are commonly treated with mannitol. However, mannitol usage should be monitored since inappropriate utilization of the substance could conversely have negative effects on CE patients. CE is thought to be related to bloodbrain barrier dysfunction. Nonetheless, a fluid clearance mechanism called the glial-lymphatic or glymphatic system was updated. This pathway facilitates the transport of cerebrospinal fluid (CSF) into the brain along arterial perivascular spaces and later into the brain interstitium. After removing solutes from the neuropil into meningeal and cervical lymphatic drainage arteries, the route then directs flows into the venous perivascular and perineuronal regions. Remarkably, the dual function of the glymphatic system was observed to protect the brain from further exacerbated damage. From our point of view, future studies ought to concentrate on the management of CE based on numerous targets of the updated glymphatic system. Further clinical trials are encouraged to apply these agents to the clinic as soon as possible.

Abstract Background: Approximately 50% of newly diagnosed invasive breast cancers are stage 1, with the majority being ER/PR-positive, HER2-negative. Genomic assays such as the Oncotype DX® have identified patients (pts) with reduced risk of distant metastasis and without benefit from chemotherapy added to endocrine therapy, freeing them from excess toxicity. Genomic assays are also recognized as prognostic for in-breast recurrence (IBR) after BCS and could similarly allow de-escalation of adjuvant radiotherapy (RT). Reducing overtreatment is of interest to pts, providers, and payers. Methods: We hypothesize that BCS alone is non-inferior to BCS plus RT for in-breast recurrence and breast preservation in women intending endocrine therapy (ET) for stage 1 invasive breast cancer (ER &/or PR positive, HER2-negative with an Oncotype DX Recurrence Score [RS] of ≤18). Stratification is by age (&lt; 60; ≥60), tumor size (≤1 cm; &gt;1-2cm), and RS (&lt; 11, 11-18). Pts are randomized post-BCS to Arm 1 with breast RT using standard methods (hypo- or conventional-fractionated whole breast RT with/without boost, or APBI) with ≥5 yrs of ET (tamoxifen or AI) or Arm 2 with ≥5 yrs of ET (tamoxifen or AI) alone. The specific regimen of ET in both arms is at the treating physician’s discretion. Eligible pts are stage 1: pT1 (≤2 cm), pN0, age ≥50 to &lt; 70 yrs, s/p BCS with negative margins (no ink on tumor), s/p axillary nodal staging (SNB or ALND), ER &/or PR positive (ASCO/CAP), HER2-negative (ASCO/CAP), and Oncotype DX RS of ≤18 (diagnostic core biopsy or resected specimen). Primary endpoint is in-breast recurrence (invasive breast cancer or DCIS). Secondary endpoints are breast conservation rate, invasive in-breast recurrence, relapse-free interval, distant disease-free survival, overall survival, patient-reported breast pain, patient-reported worry about recurrence, and adherence to ET. We assume a clinically acceptable difference in IBR of 4% at 10 yrs to judge omission of RT as non-inferior (10-yr event-free survival for RT group is 95.6% vs 91.6% for the omission of RT group). BR007 is powered to detect non-inferiority with 80% power and a one-sided α=0.025, assuming that there would be a ramp-up in accrual in the first two years (leveling off in Yrs 3-5); 1,670 pts (835 per arm) are required for randomization. Conservative loss to follow-up is 1% per yr. Some of the T1a pts screened may have Oncotype DX scores &gt;18, making them ineligible for the study. In the accrual process, 1,714 pts will be required to register to ensure that our final randomized cohort is 1,670 pts. Current accrual (07-07-2023) is 555 screened and 488 randomized (~96% of predicted accrual). Support: U10CA180868, -180822, UG1CA189867, UG1CA189867; Susan G. Komen Foundation (JRW). NCT: 04852887. Citation Format: Julia White, Reena Cecchini, Eleanor Harris, Eleftherios Mamounas, Daniel Stover, Patricia Ganz, Reshma Jagsi, Stewart Anderson, Carmen Bergom, Valérie Théberge, Mahmoud El-Tamer, Richard Zellars, Dean Shumway, Guang-Pei Chen, Thomas Julian, Norman Wolmark. A phase III trial evaluating De-escalation of Breast Radiation (DEBRA) following breast-conserving surgery (BCS) of stage 1, HR+, HER2-, RS ≤18 breast cancer: NRG-BR007 [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO1-19-02.

Evidence from medicine and other fields has shown that gender diversity results in better decision making and outcomes. The incoming workforce of congenital heart specialists (especially in pediatric cardiology) appears to be more gender balanced, but past studies have shown many inequities. Gender‐associated differences in leadership positions, opportunities presented for academic advancement, and recognition for academic contributions to the field persist. In addition, compensation packages remain disparate if evaluated based on gender with equivalent experience and expertise. This review explores these inequities and has suggested individual and institutional changes that could be made to recruit and retain women, monitor the climate of the institution, and identify and eliminate bias in areas like salary and promotions.

Abstract This is a single Institute, prospective cohort study. We collected twenty- two postmenopausal women with pelvic organ prolapse planning to undergo vaginal hysterectomy with transvaginal pelvic reconstructive surgery, with or without a concomitant anti-incontinence procedure. Vaginal swabs and urine samples were longitudinally collected at five time points: preoperative consult visit (T1), day of surgery prior to surgical scrub (T2), immediately postoperative (T3), day of hospital discharge (T4), and at the postoperative exam visit (T5). Women experiencing urinary tract infection symptoms provided a sample set prior to antibiotic administration (T6). Microbiome analysis on vaginal and urinary specimens at each time point. Region V3-V5 of the 16S ribosomal RNA gene was amplified and sequenced. Sample DNA was analyzed with visit T1, T2, T5 and T6. Six (27.3%) participants developed postoperative urinary tract infection whose vaginal sample at first clinical visit (T1) revealed beta-diversity analysis with significant differences in microbiome structure and composition. Women diagnosed with a postoperative urinary tract infection had a vaginal microbiome characterized by low abundance of Lactobacillus and high prevalence of Prevotella and Gardnerella species. In our cohort, preoperative vaginal swabs can predict who will develop a urinary tract infection following transvaginal surgery for pelvic organ prolapse.