Linear Mixed-effects Models Research Articles

Short-term effects of a research-scale oyster cage aquaculture system on sediment transport, water quality, and seagrass meadow health in Copano Bay, TX, USA

Competing uses for nearshore coastal space, such as aquaculture, have complex environmental and ecological interactions with surrounding seagrass meadows. Oyster aquaculture is among the fastest-growing aquaculture segments in the United States, and it brings concerns such as increased sedimentation from farm maintenance operations and altered water quality through oyster filtration. Changes in seagrass coverage and growth are common indicators of ecological health used to determine anthropogenic impacts on nearshore environments. This study characterized the effects of a research-scale adjustable long-line oyster aquaculture system on the health of adjacent seagrass meadows in Copano Bay, TX. Four Halodule wrightii meadows were identified at various distances from the research site: 100 m upstream (upstream), directly adjacent (0 m), 30 m downstream (30 m), and 60 m downstream (60 m). Sites were monitored for 1) seagrass health, 2) water quality parameters, and 3) sediment deposition. Over the 18-week sampling period, no significant differences (p>0.05) were found for water quality variables or sediment grain-size analysis using a one-way ANOVA. A linear mixed-effects model was used for repeated measures of seagrass data, with no effect of the site found on mean seagrass length, coverage, or maximum length (p>0.05). These results suggest that seagrass health indicators were unaffected by proximity to the oyster system. It was concluded that an oyster research aquaculture system of the type and size at the project location had no major negative or positive impact on seagrass meadow health due to high background variability and the overall minor footprint of the oyster farm across a comparatively large spatial extent.

Beta-gamma phase-amplitude coupling of scalp electroencephalography during walking preparation in Parkinson’s disease differs depending on the freezing of gait

IntroductionDespite using beta oscillations within the subthalamic nucleus as a biomarker of akinesia or rigidity in Parkinson’s disease, a specific biomarker for freezing of gait (FOG) remains unclear. Recently, scalp phase-amplitude coupling (PAC) measured through scalp electroencephalography (EEG) has emerged as a promising tool for analyzing brain function. In this study, we examined whether PAC could be a biomarker for FOG.MethodsWe enrolled 11 patients with Parkinson’s disease and recorded scalp EEG in preparation for and during gait while simultaneously assessing motor function, including FOG. We investigated changes in cortical PAC during walking with and without FOG and examined its correlation with the postural instability and gait difficulty (PIGD) score.ResultsPatient characteristics were as follows: mean age 59.1 ± 6.9 years, disease duration 13.9 ± 4.1 years, and seven men. Four trials were excluded from the analysis owing to artifacts. In the trials without FOG (n = 18), beta-gamma PAC in the sensorimotor area decreased during gait preparation (p = 0.011; linear mixed-effects model), which was not the case in trials with FOG (n = 6) (p = 0.64; linear mixed-effects model). Using a support vector machine, machine learning of PAC during preparation for walking predicted the presence of FOG with an accuracy of 71.2%. Conversely, PAC increased during walking in trials with FOG (p = 0.0042; linear mixed-effects model), and PAC 20 s after the start of walking was positively correlated with the PIGD score (correlation coefficient = 0.406, p = 0.032; Pearson’s rank correlation).ConclusionBeta-gamma PAC in the sensorimotor area during preparation for walking differs depending on the emergence of FOG. As gait symptoms worsened, beta-gamma PAC in the sensorimotor area during walking gradually increased. Cortical PAC may be a biomarker for FOG in Parkinson’s disease and may lead to the development of strategies to prevent falls in the future.

Phrasal frequency and literacy as predictors of individual differences in on-line processing and comprehension of english complex NP subject-verb agreement

AbstractWe present experimental evidence suggesting that frequency and literacy predict online processing and comprehension of subject-verb agreement constructions by adult native speakers of English. We measured participants’ eye fixations, reaction times, and response accuracy in a forced-choice task using audio-visual eye-tracking paradigm. Participants completed a battery of tasks, inc. the Literacy Rating Scale (Tarone et al., Literacy and Second Language Oracy-Oxford Applied Linguistics, Oxford University Press, 2013), Agreement Judgement Task (e.g., Veenstra et al., Frontiers in Psychology 5:783, 2014). The AJT involved matching an auditorily presented subject phrase to one of two images of easily distinguishable colours presented on a computer screen (e.g., stars, circles). Participants heard 42 test sentences, counterbalanced across the three types: Type 1 (e.g., ‘The stars with the circles are blue’), Type 2 (e.g., ‘The star with the circles is blue’) and Type 3 (e.g., ‘The star with the circles are blue’*. Type 1 and Type 2 constructions are considerably more frequent in writing than in speech (Miller et al., Spontaneous spoken language: Syntax and discourse, Oxford University Press on Demand, 1998) with Type 2 producing more attraction errors (Bock et al., Cognitive Psychology 43:83–128, 2001; Becker, L., & Dąbrowska, E. (2020). Does experience with written language influence grammaticality intuitions? UK Cognitive Linguistics Conference: University of Birmingham [conference presentation].). Data were analysed with linear mixed effects models and generalised additive models. Results show lower literacy participants took longer to process sentential cues and made more attraction errors. These findings support usage-based research showing frequency and experience effects on online comprehension of canonical and non-canonical constructions (Farmer, T. A., Misyak, J. B., & Christiansen, M. H. (2012). Individual differences in sentence processing. In Cambridge handbook of psycholinguistics (pp. 353-364)., Street, Language Sciences 59:192–203, 2017), detection and production of agreement attraction errors (Becker, L., & Dąbrowska, E. (2020). Does experience with written language influence grammaticality intuitions? UK Cognitive Linguistics Conference: University of Birmingham [conference presentation].) and demonstrate how linguistic and attentional processes interact (Tomlin and Myachykov, Attention and salience, Handbook of Cognitive Linguistics, 2015). They also complement corpus-based studies by providing evidence that native speakers are sensitive to observed distributions (Miller et al., Spontaneous spoken language: Syntax and discourse, Oxford University Press on Demand, 1998).

Assessment of measurement variability across automated biometry devices

Purpose: To evaluate measurement variability between different readily available automated biometric devices for the purposes of improving surgeon decision making for phakic intraocular lens (IOL) sizing. Setting: Private practice (Houston, TX, USA) Design: Prospective, single center, bilateral, non-randomized, open-label, observational study. Methods: This study included healthy eyes, with no prior eye surgery, and with a spherical equivalent of at least -1.00D. Orbscan II was compared to Argos, Atlas 9000, caliper, IOLMaster 500, IOLMaster 700, iTrace, Lenstar LS900, Pentacam HR, and Pentacam AXL Wave. Preoperative measurements included white to white (WTW), anterior chamber depth (ACD), and central corneal thickness (CCT). Linear mixed effect models were created to determine adjustment factor between the Orbscan II and other devices for WTW, ACD, and CCT measurements. Results: A total of 204 subjects (408 eyes) completed the study. All mean WTW measurements were significantly different compared to the Orbscan II (p < 0.01), except for Pentacam AXL Wave. All mean ACD and CCT measurements were significantly different compared to the Orbscan II (p < 0.01). Adjustment factors for WTW ranged from -0.65 mm to 0.24 mm. Adjustment factors for ACD ranged from -0.21 mm to -0.16 mm. Adjustment factors for CCT ranged from 19.9 µm to -36.0 µm. Conclusions: The biometers tested were not interchangeable with the Orbscan II. Adjustment factors for WTW and ACD may compensate for differences between devices and adapt ICL sizing recommendations.

Abstract PR018: Plasma proteomic biomarkers reveal biological insights about the tumor microenvironment in melanoma patients after PD1 blockade

Abstract Most patients treated with immune checkpoint blockade (ICB) do not have durable treatment responses, stressing a critical need to identify early non-invasive biomarkers of response. Circulating biomarkers provide easy access for serial monitoring and can provide insight on the mechanisms of response to ICB. We performed plasma proteomics (2943 proteins) on 250 metastatic melanoma patients before and on ICB treatment and performed analysis to identify response- and time point-associated peripheral protein biomarkers. We next generated patient- matched peripheral blood lymphocyte (PBL) surface proteomic data to measure the proportions of immune cell subsets in circulation. We further obtained 13 tumor samples that were processed for single-cell RNA-sequencing (scRNA-seq). We fit linear mixed effects (LME) models to predict plasma protein abundances, using the sample time point, response status and interaction term between time point and response as covariates. We identified 397 proteins with a significant time point, response, or interaction effect. For proteins upregulated post-ICB, gene ontology analysis suggested these proteins were associated with inflammatory pathways involving the activation of effector immune functions. LME models fitted on the circulating immune cell proportions and parallel analysis of immune-related plasma proteins supported that post-ICB samples may be defined by increased immune activity, including T cell infiltration. Moreover, expression of genes corresponding to these post-ICB proteins in the tumor scRNA-seq data tended to be favored by immune subsets, namely those involved in immune cell recruitment and tumor reactivity, such as CXCL10+ macrophages and CXCL13+ T cells, respectively. On the other hand, expression of genes corresponding to plasma proteins upregulated in non-responders in the TME tended to be favored by suppressive myeloid subsets such as SPP1+ macrophages as well as malignant cells, suggesting the potential of immunosuppressive and pro-tumor biological processes to be involved in ICB resistance. Inference of cell-cell interactions in the TME further showed the involvement of these non-responder genes in potentially immunosuppressive and pro-tumor processes, including the PTPRC-MRC1 axis, involved in myeloid cell plasticity, the IGF1/2-IGF1R axis, involved in growth and survival of normal and neoplastic cells, and the GAS6-TYRO3 axis, associated with carcinogenetic mechanisms in multiple malignancies. We further identified a correlation between IL13 protein abundance in circulation and GAS6 expression in the tumor, suggesting that IL13 may reach the tumor from circulation to turn on the STAT6 pathway in receiving cells and upregulate GAS6 in turn to participate in this interaction with TYRO3. Together, these data represent one of the deepest peripheral biomarker studies using paired samples in melanoma patients treated with anti-PD1 therapy, and begin to elucidate the complex interplay between tumors and the systemic immune response within the host. Citation Format: Samuel J. Wright, Izabella Zamora, Milan Parikh, Deepika Yeramosu, Lynn Bi, Sarah Kang, Marijana Rucevic, Moshe Sade-Feldman, Baolin Liu, Ngan Nguyen, Jamey Guess, Alexis Schneider, David Lieb, Elliot Woods, William Michaud, Aleigha R. Lawless, Tatyana Sharova, Gyulnara Kasumova, Michelle S. Kim, Ryan J. Park, Russell W. Jenkins, Samuel J. Klempner, Ryan J. Sullivan, Keith T. Flaherty, Nir Hacohen, Genevieve M. Boland, Arnav Mehta. Plasma proteomic biomarkers reveal biological insights about the tumor microenvironment in melanoma patients after PD1 blockade [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr PR018.

Glucose measurements in sheep using a long-term continuous glucose monitor designed for humans

This study evaluated the feasibility of utilizing a continuous glucose monitors (CGM) designed for use in humans to measure glucose levels in sheep. Four Suffolk x Dorset sheep were fitted with jugular catheters and FreeStyle Libre 2 (Abbott®) glucose monitors. Glucose concentration from the CGM were compared with those from a glucometer and traditional assays during a hyperglycemic clamp, aiming to explore a broader range of physiological glucose concentrations in a controlled manner. Measurements were taken every five minutes during the infusion and every ten minutes post-infusion until baseline levels were restored. Relationships were analyzed using a linear mixed-effects model with glucometer readings as the response variables, GCM reading as fixed effects, and animal as random effect with significant level of p < 0.05. The results demonstrated a significant linear correlation between the CGM and glucometer readings both during the infusion (p = 0.0003) and afterward (p = 0.006). A glucose calorimetric test was used to confirm glucose concentrations on samples and used as gold standard. Although the glucometer and CGM did not differ from the assay method, they did differ from one another (p = 0.045). Upon more in-depth analysis, the random intercepts for animal were highly significant and one CGM showed values numerically much higher than other CGM and other glucose analysis methods. No difference among methods was observed (p = 0.715) when the outlier animal was omitted. While promising, CGM demands confirmation of initial readings and standardization against established methods before wider adoption in research or clinical applications.

One-year impact of behavioural interventions on schistosomiasis-related knowledge, attitude and practices of primary schoolchildren in Pemba, Tanzania

Abstract Background Elimination of schistosomiasis as a public health problem and interruption of transmission in selected areas are goals set by the World Health Organization for 2030. Behaviour change communication (BCC), coupled with other interventions, is considered an essential measure to reduce the transmission of Schistosoma infection. Focusing on elimination, we assessed the 1-year impact of BCC interventions on schistosomiasis-related knowledge, attitude and practices (KAP) of schoolchildren in hotspot schools versus low-prevalence schools that did not receive the interventions. Methods School-based cross-sectional surveys were implemented in 16 schools on Pemba Island, Tanzania, in 2020 and 2022, respectively. The schistosomiasis-related KAP were assessed in children attending grades 3–5, using pre-tested questionnaires. Between the surveys, in 2021, children from hotspot schools were exposed to BCC interventions. The difference in mean knowledge and attitude scores, respectively, between schoolchildren from hotspot and low-prevalence schools during the survey in 2022 was determined with a linear mixed-effect model. Results In the five hotspot schools that received BCC interventions, 315 children participated in the survey in 2020 and 349 in 2022. There was a 21.0% increase in children with moderate knowledge and a 13.8% decrease in no knowledge; a 8.3% increase in good attitude and a 19.2% decrease in poor attitude; 3.4% and 3.2% fewer children reported to use waterbodies for washing clothes or body, respectively. In the 11 low-prevalence schools without BCC interventions, 778 children participated in 2020 and 732 in 2022. The percentage of children with poor knowledge (56.4% and 63.1%) and poor attitude (55.3% and 53.1%) remained relatively stable from 2020 to 2022, but 4.9% and 3.0% less children reported to use waterbodies for washing clothes or their body, respectively. In 2022, the difference in mean knowledge scores was 0.8 [95% confidence interval (CI): 0.5−1.1] and the difference in mean attitude scores was 0.6 (95% CI: 0.4−0.7) between children in hotspot compared with low-prevalence schools. Conclusions After one year of implementation, the BCC interventions markedly improved the KAP of exposed children. Complemented by improved access to clean water and sanitation, BCC holds promise to contribute successfully to the achievement of schistosomiasis control and elimination targets. Trial registration ISRCTN, ISRCTN91431493. Registered 11 February. 2020, https://www.isrctn.com/ISRCTN91431493. Graphical Abstract

MRI-free processing of tau PET images for early detection

Abstract Tau positron emission tomography (PET) imaging in Alzheimer’s Disease (AD) is becoming increasingly common to assess in vivo tau burden. MR images are often acquired to assist with processing of PET data, including for region-of-interest definitions in native space and for normalization to template space. However, in the real-world setting, corresponding MRIs may not be available and PET processing may require MRI-free pipelines. This is particularly important and challenging as the field moves towards early detection among clinically unimpaired (CU) individuals where changes in tau PET signal are expected to be subtle. We used two independent [18F]Flortaucipir tau PET datasets to evaluate whether MRI-free PET processing can detect subtle tau PET uptake differences in Amyloid+ (A+) CU individuals (preclinical AD) versus A-. Standardized Uptake Value Ratios (SUVRs) from MRI-free compared to MRI-based methods were evaluated using linear regression and linear mixed-effects regression models. Effect size differences between A+/- CU groups in MRI-free processed cross-sectional and longitudinal tau PET SUVRs were compared to differences quantified through MRI-based processing. Regional MRI-free SUVRs were highly correlated with MRI-based SUVRs within CU individuals (average ICC = 0.90 for ADNI CU and 0.81 for A4 CU). MRI-free and MRI-based pipelines resulted in similar estimates of cross-sectional and longitudinal differences between A- and A+ CU, even in early focal regions within the medial temporal lobe.

Age differences in mental state inference of sarcasm: Contributions of facial emotion recognition and cognitive performance

Abstract Objectives Older age is associated with poorer ability to accurately infer mental states, but some mental states are more complex than others. Sarcasm is a complex mental state because the literal and intended meaning of a speaker's words are in opposition. Individuals must rely on additional cues (e.g., facial expressions, intonation) for accurate inference. We hypothesized that understanding of sarcastic versus sincere exchanges would be more sensitive to age-related difficulty in mental state understanding. Method We examined accuracy at identifying sarcasm among 263 adults (ages 18-90 years) using videos of social interactions in The Awareness of Social Inference Test. Hypotheses were tested using a logistic linear mixed effects model predicting correct/incorrect trial-level responses. To characterize why sarcasm differed with age, we measured two abilities commonly implicated in mental state understanding: facial emotion recognition and cognitive performance. Results Sarcasm understanding declined with age, whereas understanding of sincere exchanges did not. Both better emotion recognition and cognitive performance related to better understanding of sarcastic but not sincere exchanges. Only cognitive performance showed an age-related effect such that the cognitive performance among the oldest participants facilitated their understanding of both sarcastic and sincere exchanges. Discussion We showed that individual variation related to age and social and cognitive performance is more pronounced when the use of multiple mental state cues is more (sarcasm) or less (sincerity) necessary for accurate understanding of social interactions. Naturalistic paradigms involving multiple mental state cues can address important questions about how older adults make decisions in the real world.

The effects of electronic cigarette use patterns on health-related symptom burden and quality of life: analysis of US prospective longitudinal cohort study data

ObjectiveThis study aimed to examine the association between e-cigarette (EC) use patterns and health-related symptoms (fatigue, pain, and emotional problems) as well as general quality of life (QoL).MethodsData were analyzed from 7,225 adults across Waves 1–6 of the US Population Assessment of Tobacco and Health (PATH) study. Current combustible cigarette (CC) or EC use patterns included dual CC/EC use, exclusive EC use, non-current use of CC or EC, and exclusive CC smoking. Multivariate linear mixed-effects models were used to investigate longitudinal associations between EC use patterns, and symptom burdens/QoL scores.ResultsThose who were not currently smoking or vaping reported the lowest fatigue, pain, and emotional problems, and the best QoL, among the four groups (all p < 0.001). Compared to exclusive CC smoking, exclusive EC use was associated with a significant decrease of 0.065 units in average fatigue (95% confidence interval [CI]: −0.121, −0.009), of 0.206 units in average pain (95% CI: −0.355, −0.058), and of 0.103 units in average QoL scores (95% CI: −0.155, −0.051), with emotional problems similar over time.ConclusionExclusive EC users had less health-related symptoms and better QoL than those who were exclusive CC smokers. This should be taken into account when assessing the harm reduction potential of ECs.

Abstract 4122763: Health Status Outcomes of Antianginal Therapies in Patients After Coronary Revascularization: Insights From the GREAT Disease Registry Study

Background: The impact of antianginal therapies on the health status of coronary artery disease (CAD) patients who have undergone coronary revascularization has not been extensively researched. This study aimed to investigate the effect of nicorandil on the health status in patients after coronary revascularization, measured by the Seattle Angina Questionnaire (SAQ). Methods: We analyzed data from 1,556 patients enrolled in the GREAT Disease Registry Study (Clinicaltrial No. NCT05050773) between September 2021 and May 2022. Changes in SAQ summary score (SAQ-SS) from baseline to 3, 6, 9, and 12 months were assessed. Statistical methods included a linear mixed-effects model and propensity score matching (PSM) for comparative analysis. Results: Among the 1,038 patients receiving revascularization analyzed (22.7% female; mean age [SD], 59.9 [10.2] years), 648 (62.4%) received nicorandil, while 390 (37.6%) did not. In the linear mixed-effects model analysis, statistically significant improvements were observed with nicorandil in SAQ-SS across all time points ( Table 1 ), indicating beneficial effects of nicorandil on health status outcomes after revascularization. Sensitive analysis was conducted using the PSM method. After PSM adjustment, both groups comprised 321 patients each. The change in SAQ-SS from baseline to 12 months was significantly greater in the nicorandil group (19.2 ± 14.06 points) compared to the non-nicorandil group (14.91 ± 12.9 points, p < 0.001). Conclusions: These findings indicated that nicorandil is associated with improved health status outcomes in CAD patients who have undergone coronary revascularization.

Abstract 4137636: Cardiac radiation dose and changes in physical activity and quality of life in patients with locally advanced lung cancer

Background: Physical activity is associated with prognosis in cancer. However, in non-small cell lung cancer (NSCLC), a high cardiovascular (CV) population, the relationships between RT dose, physical activity, and CV patient-reported outcomes remain unclear. We determined the associations between cardiac radiation dose-volume metrics and changes in physical activity and quality-of-life (QoL) in patients treated with definitive chemoradiation. Methods: NSCLC participants treated with chemoradiation were enrolled in a multi-center longitudinal cohort study. Self-reported physical activity via Godin–Shephard Leisure-Time Physical Activity Questionnaire (Godin) and QoL via Functional Assessment of Chronic Illness Therapy (FACIT) Fatigue and Dyspnea Scales were collected at baseline, end of RT, 6, and 12 months post-RT. Changes in Godin and FACIT scores over time were examined using linear mixed-effects models. Multivariable linear regression via generalized estimating equations assessed the associations between: 1) mean heart dose (MHD) and longitudinal changes in physical activity and QoL, and 2) concurrent changes in physical activity and FACIT scores. Results: Across 101 participants, the mean age was 66 years, with 57% male, 66% White, and 25% Black. The median (Q1, Q3) MHD was 10.8 Gy (6.1,16.3). Very low physical activity [9 (0, 25)], high dyspnea [4 (1, 9)], and low fatigue [39 (31,46)] were reported at baseline. With chemoRT, there was a significant worsening of physical activity ( P = 0.03), fatigue ( P < 0.001), and dyspnea ( P < 0.001). Each 1-Gy increase in MHD was associated with a modest decline in physical activity (-0.28; 95% CI -0.53, -0.02, P = 0.03), and higher physical activity was associated with a concurrent decrease in fatigue ( P < 0.001) and dyspnea ( P = 0.007) (Table). Conclusions: In locally advanced NSCLC, physical activity levels prior to RT were very low, and higher cardiac radiation dose was associated with a further decline in physical activity. Increased physical activity was associated with concurrent decreases in fatigue and dyspnea. Strategies to promote physical activity may result in improved QoL in this high CV group.

Abstract 4122840: Patient activation improves with a multi-component personalized mHealth intervention in older adults at risk of cardiovascular disease

Background: Patient activation, a person’s level of confidence, comprehension, and autonomy toward their health, is an essential goal of cardiovascular disease therapy. Previous research has shown a link between high levels of patient activation and better patient outcomes and experiences. However, whether individualized interventions such as a text-based mobile health (mHealth), improve patient activation is unclear. Research on patient activation in people at risk for cardiovascular disease and influential factors is minimal. Aims: This study aimed to 1) determine the effect of a multi-component mHealth intervention on patient activation and 2) examine its predictors among older adults at risk for cardiovascular disease. Methods: In this pilot randomized controlled trial, community-dwelling older adults with poor eating behaviors and reduced physical activity (n=54) were randomly assigned to the GetFIT (control) or GetFIT+ (intervention) groups, with three- and six-month follow-up periods. The GetFIT group received healthy lifestyle counseling from a licensed health coach, a free commercial mHealth app with push alerts on physical activity, and an activity tracker for physical activity monitoring. The GetFIT + group received the same components but had personalized text messages instead of push alerts. The 13-item Patient Activation Measure assessed patient activation; higher scores indicate better activation. Linear mixed-effects models were used to investigate between-group changes in outcomes across time. Results: The mean age was 65.4 ± 6.0 years; 61% were females and 61% were married. Baseline characteristics were comparable between groups. Significant improvements were observed in the GetFIT+ group at three months (mean 3.53, 95% CI 0.11, 6.96; p=0.043) and six months (mean 4.37, 95% CI 0.91, 7.83; p=0.014), whereas improvements in the GetFIT group were non-significant. Adjusting for age, gender, education, employment, marital status, social support, smartphone confidence, and self-perceived health did not alter the results. Nevertheless, only social support was associated with higher patient activation overall (B=5.14, 95% CI 1.00, 9.27; p=0.015). Conclusions: People at risk for cardiovascular disease can improve their self-management abilities, knowledge, and confidence through personalized text messaging. Possessing social support is critical for maximizing the benefits of mHealth interventions.

Sex-dependent variability of isoniazid and rifampicin serum levels in patients with tuberculosis

Abstract Introduction Drug-sensitive TB (DS-TB) is treated with isoniazid, rifampicin, ethambutol, and pyrazinamide. Factors like fast-metabolizing enzymes, malabsorption, and drug interactions can influence serum drug levels. Current TB treatment guidelines recommend weight-adapted dosing without considering sex differences. This study examines drug levels of isoniazid and rifampicin in TB patients treated between 2019 and 2023 at our center focusing on sex-specific aspects. Methods Patients diagnosed with TB and available serum levels of isoniazid or rifampicin between 2019 and 2023 were retrospectively identified. Serum levels were measured using liquid chromatography–mass spectrometry and high-performance liquid chromatography. Patients were stratified by sex and a linear regression mixed effect model was used to assess predictors for different serum levels. Results The study included 281 single therapeutic drug monitoring (TDM) measurements from 59 patients (28 women, 47.5%). For isoniazid, no sex-specific differences in serum drug levels were identified. On the other hand, female sex was a significant predictor of higher rifampicin plasma levels (coefficient 4.16, 95% CI 0.74–7.59, p = 0.009). Only 38.2% of rifampicin serum level measurements in male patients were within target range, the majority (40/68, 58.8%) were below range and only 2 (2.9%) TDM-levels were above range. Women displayed higher overall rifampicin serum levels than men (median 13.7 mg/l vs. 7.1 mg/l, p = 0.04), although weight adjusted doses were not significantly different (median 10.0 mg/kg vs. 9.8 mg/kg p = 0.56). Adverse effects were noted in 42.9% (42/98) of measurements in women and 29.5% (54/183) of measurements in men (p = 0.03). Discussion Rifampicin levels were significantly lower in men compared to women, despite weight-adjusted dosing. Clinicians should consider TDM and potential sex differences when treating patients with TB.

Abstract 4143242: Right Ventricular Remodeling During the First Decade After Tetralogy of Fallot Repair: An Echocardiographic Study

Background: After TOF repair (TOFr), pulmonary regurgitation resulting in right ventricular (RV) dilatation and dysfunction is common. We have previously demonstrated that RV function (RVF) improves in the first two years after TOFr, however RV changes in the first decade after TOFr are less known. We investigated RVF and RV size (termed remodeling) using echocardiography during the first decade after TOFr. Methods: Retrospective cohort study of patients who underwent TOFr and were followed at our institution from 3/2012 to 12/2023. Those with at least two echocardiograms separated by one year were included. Two patients that received a pulmonary valve replacement were excluded. After confirming intra- and inter-reader reproducibility, RV remodeling was assessed offline by global longitudinal strain (GLS), free wall strain (FWS), fractional area change (FAC), end systolic area (ESA/BSA) and end diastolic area (EDA/BSA). RVF changes over time were assessed with multivariable linear mixed effects models adjusting for perioperative covariates. Two years after TOFr (Y2) was used as reference given RVF peaks at this point following surgery. Results: Out of 758 echocardiograms from 116 patients, 688 were suitable for analysis. Median age at TOFr was 3.5 months (IQR:1.92, 4.89). Compared to Y2, we identified the largest decline in RVF at Year 3, with no worsening from year to year until Y10. There were no associations of perioperative covariates with RVF except for ventriculotomy (FAC) and residual pulmonary stenosis (FWS). There was no progression in RV remodeling over the first decade after TOFr. (Graph, table). Conclusion: After the initial 2-year improvement following TOFr, RVF worsens but remains stable throughout the first decade with stable RV size accounting for BSA. Our findings augment the knowledge of RV remodeling during childhood and could help guide outpatient follow up, counseling, frequency of echocardiograms, and timing of the first cardiac MR.

Abstract 4115652: Effect of the sodium-glucose co-transporter 2 inhibitor, Dapagliflozin, on ventricular repolarization electrocardiographic parameters: DAPA-ECG study

Background: In patients with type 2 diabetes (T2D), hyperglycemia and glycemic variability lead to prolongation and greater heterogeneity of ventricular repolarization, manifested on the electrocardiogram through an increase in QT, QTc, TpeakTend (TpTe) intervals and the TpTe/QT ratio, increasing the risk of potentially malignant arrhythmias. Dapagliflozin has demonstrated efficacy in reducing cardiovascular events in T2D patients and in the risk of serious ventricular arrhythmias and sudden cardiac death. However, the exact mechanisms by which dapagliflozin confers this protection have not yet been fully elucidated. Goals/Aims: The main objective of the study was to evaluate the impact of dapagliflozin on the TpTe interval of patients with T2D, and secondarily, it examined its impact on various electrocardiographic parameters such as the QT and QTc intervals, the TpTf/QT ratio, QT dispersion, J-T peak interval, QRS-T angle and heart rate. Methods: This randomized, prospective, multicenter and open-label study involved 174 patients with T2D, divided into two groups: one treated with dapagliflozin and the other with optimized medical therapy without iSGLT2. Clinical, electrocardiographic, laboratory and echocardiographic evaluations were carried out at the beginning and after three months. The statistical analysis included means, standard deviations, quartiles, and frequencies, with 95% confidence intervals, using Chi-square (or Fisher) and t-test (or Mann-Whitney) for initial differences, and a linear mixed-effects model to evaluate the results, adopting a significance level of 0.05. Results: The TpTe interval was significantly reduced in the dapagliflozin group from 76.87 ± 12.47 to 67.80 ± 10.34 (p<0.001), and the QTc interval reduced from 424.07 ± 30.76 to 411.26 ± 25.31 (p=0.022). The TpTe/QT ratio in the dapagliflozin group was shortened from 0.20 ± 0.03 to 0.18 ± 0.02 (p<0.001). Both groups showed reductions in mean blood glucose and HbA1c, with a greater decrease observed in the control group. No significant changes were recorded in other variables analyzed. Conclusion: In patients with T2D, dapagliflozin significantly reduced the duration of the TpTe and QTc intervals, in addition to the TpTe/QT ratio. These findings indicate a decrease in ventricular electrical remodeling with a potential cardioprotective effect of dapagliflozin, which would partly explain its benefits in reducing potentially malignant ventricular arrhythmias and sudden cardiac death.

Reproducibility of liver ADC measurements using first moment optimized diffusion imaging

AbstractPurposeCardiac‐induced liver motion can bias liver ADC measurements and compromise reproducibility. The purpose of this work was to enable motion‐robust DWI on multiple MR scanners and assess reproducibility of the resulting liver ADC measurements.MethodsFirst moment‐optimized diffusion imaging (MODI) was implemented on three MR scanners with various gradient performances and field strengths. MODI‐DWI and conventional Stejskal‐Tanner monopolar (MONO) DWI were acquired in eight (N = 8) healthy volunteers on each scanner, and DWI repetitions were combined using three different averaging methods. For each combination of scanner, acquisition, and averaging method, ADC measurements from each liver segment were collected. Systematic differences in ADC values between scanners and methods were assessed with linear mixed effects modeling, and reproducibility was quantified via reproducibility coefficients.ResultsMODI reduced left–right liver lobe ADC bias from 0.43 × 10−3 mm2/s (MONO) to 0.19 × 10−3 mm2/s (MODI) when simple (unweighted) repetition averaging was used. The bias was reduced from 0.23 × 10−3 mm2/s to 0.06 × 10−3 mm2/s using weighted averaging, and 0.14 × 10−3 mm2/s to 0.01 × 10−3 mm2/s using squared weighted averaging. There was no significant difference in ADC measurements between field strengths or scanner gradient performance. MODI improved reproducibility coefficients compared to MONO: 0.84 × 10−3 mm2/s vs. 0.63 × 10−3 mm2/s (MODI vs. MONO) for simple averaging, 0.66 × 10−3 mm2/s vs. 0.50 × 10−3 mm2/s for weighted averaging, and 0.61 × 10−3 mm2/s vs. 0.47 × 10−3 mm2/s for squared weighted averaging.ConclusionThe feasibility of motion‐robust liver DWI using MODI was demonstrated on multiple MR scanners. MODI improved interlobar agreement and reproducibility of ADC measurements in a healthy cohort.

NIMG-13. CHANGES IN CORTICAL ARCHITECTURE ASSOCIATED WITH GLIOBLASTOMA ASSOCIATED EPILEPSY

Abstract BACKGROUND Emerging evidence suggests both local hyper-excitability promoting epilepsy, and synaptic and paracrine mediated integration of Glioblastoma into brain-wide neural networks. In this study we sought clinical evidence of disease progression, and concurrent radiological evidence of differences in remote cortical architecture, in Glioblastoma patients with and without epilepsy. METHODS 55 patients with biopsy-confirmed Glioblastoma, mean-age 59.4yrs±10.9yrs, were studied. 23 patients had a diagnosis of Glioblastoma-related epilepsy on presentation. We compared clinical markers of disease progression in individuals with and without epilepsy. Further, we explored differences in cortical architecture between both groups. Utilising pre-operative T1-weighted, contrast-enhanced MRI scans (n=45), we performed vertex-wise measurements of cortical thickness (CT) and cortical surface area (CSA) using the Virtual Brain Grafting Toolbox and FreeSurfer v.7.0. Radiologically-identifiable tumour and oedema was masked and excluded from analysis. Scans with right-hemispheric tumours were flipped to the left. Linear mixed-effects models were applied to pre-processed imaging-data. RESULTS Individuals with epilepsy had shorter progression-free survival (median 4.2 months) than those without seizures (9.3 months, p<0.05) with gold-standard treatment. No significant difference between lesion volumes (tumour and oedema) existed between the two groups (unpaired t-test, p>0.05). Controlling for age and tumour location, significant increases (p<0.01) in CT and CSA were observed in patients with Glioblastoma-related epilepsy in lesion-ipsilateral supramarginal gyrus, superior parietal lobule, precuneus/cuneus and temporal lobe compared to those without epilepsy. Interestingly, significant (p<0.05) increases in CT in lesion-contralateral supramarginal gyrus were additionally observed in those with epilepsy compared to those without. CONCLUSIONS Glioblastoma-related epilepsy was associated with poorer survival outcomes and changes in remote cortical architecture. Implicated regions form part of the default mode network, which has previously been identified as a crucial hub for epilepsy. Biological underpinnings of cortical network alterations, and their possible role in promoting not only hyper-excitability but also Glioblastoma progression, require correlation with histopathological characteristics and validation in larger study cohorts.

Fast standard error estimation for joint models of longitudinal and time-to-event data based on stochastic EM algorithms

Summary Maximum likelihood inference can often become computationally intensive when performing joint modeling of longitudinal and time-to-event data, due to the intractable integrals in the joint likelihood function. The computational challenges escalate further when modeling HIV-1 viral load data, owing to the nonlinear trajectories and the presence of left-censored data resulting from the assay’s lower limit of quantification. In this paper, for a joint model comprising a nonlinear mixed-effect model and a Cox Proportional Hazards model, we develop a computationally efficient Stochastic EM (StEM) algorithm for parameter estimation. Furthermore, we propose a novel technique for fast standard error estimation, which directly estimates standard errors from the results of StEM iterations and is broadly applicable to various joint modeling settings, such as those containing generalized linear mixed-effect models, parametric survival models, or joint models with more than two submodels. We evaluate the performance of the proposed methods through simulation studies and apply them to HIV-1 viral load data from six AIDS Clinical Trials Group studies to characterize viral rebound trajectories following the interruption of antiretroviral therapy (ART), accounting for the informative duration of off-ART periods.

EPCO-48. RESOLVING INTRATUMORAL TRANSCRIPTIONAL HETEROGENEITY ACROSS THE CORE-EDGE SPATIAL AXIS IN GLIOBLASTOMA

Abstract Intratumoral heterogeneity promotes tumor progression, treatment resistance, and recurrence in glioblastoma (GBM), and the cellular tumor core and infiltrating edge represent transcriptionally distinct cellular ecosystems. In this study, we characterize and validate the molecular differences across the core-edge spatial axis in GBM by applying whole-transcriptome GeoMx Digital Spatial Profiling (DSP) to 5 IDH-wildtype tumors. Three 3-micron tissue cores were extracted from each tumor and arranged into a tissue microarray (TMA). Core and edge regions of interest (ROIs) were chosen through coregistration of H&E and immunofluorescent TMA scans. Following DSP, data for 44 ROIs (30 core and 14 edge) and 8530 RNA targets were bioinformatically analyzed. Single-sample gene set enrichment analysis aligned the DSP core and edge regional signatures to those of publicly annotated datasets, substantiating our profiling methodology. Unsupervised analyses revealed ROI clusters unique to each tumor specimen for both pre- and post-normalized data. To mitigate a potential patient-level effect in subsequent differential gene expression and pathway analyses, we leveraged a linear mixed-effect model using core versus edge as a fixed effect and tumor specimen as a random effect. Our analyses showed upregulation of epigenomic markers in core ROIs and neuronal markers in edge ROIs. Further, core and edge signatures were incongruent with the canonical GBM subtypes (Mesenchymal, Classical, Proneural) and malignant cell states (OPC-like, NPC-like, AC-like, MES-like). Cellular deconvolution analysis highlighted differential enrichment of neoplastic cells in core ROIs and neuroglia (astrocytes and oligodendrocytes) in edge ROIs. In summary, we leverage spatial profiling to further validate the geographic molecular heterogeneity of GBM, which is characterized by divergent neuronal programs within the infiltrating edge.