Median Coefficient Of Variation Research Articles

Estimating the uncertainty of the greenhouse gas emission accounts in global multi-regional input–output analysis

Abstract. Global multi-regional input–output (GMRIO) analysis is the standard tool to calculate consumption-based carbon accounts at the macro level. Recent inter-database comparisons have exposed discrepancies in GMRIO-based results, pinpointing greenhouse gas (GHG) emission accounts as the primary source of variation. A few studies have analysed the robustness of GHG emission accounts, using Monte Carlo simulations to understand how uncertainty from raw data propagates to the final GHG emission accounts. However, these studies often make simplistic assumptions about raw data uncertainty and ignore correlations between disaggregated variables. Here, we compile GHG emission accounts for the year 2015 according to the resolution of EXIOBASE V3, covering CO2, CH4 and N2O emissions. We propagate uncertainty from the raw data, i.e. the United Nations Framework Convention on Climate Change (UNFCCC) and EDGAR inventories, to the GHG emission accounts and then further to the GHG footprints. We address both limitations from previous studies. First, instead of making simplistic assumptions, we utilise authoritative raw data uncertainty estimates from the National Inventory Reports (NIRs) submitted to the UNFCCC and a recent study on uncertainty of the EDGAR emission inventory. Second, we account for inherent correlations due to data disaggregation by sampling from a Dirichlet distribution. Our results show a median coefficient of variation (CV) for GHG emission accounts at the country level of 4 % for CO2, 12 % for CH4 and 33 % for N2O. For CO2, smaller economies with significant international aviation or shipping sectors show CVs as high as 94 %, as seen in Malta. At the sector level, uncertainties are higher, with median CVs of 94 % for CO2, 100 % for CH4 and 113 % for N2O. Overall, uncertainty decreases when propagated from GHG emission accounts to GHG footprints, likely due to the cancelling-out effects caused by the distribution of emissions and their uncertainties across global supply chains. Our GHG emission accounts generally align with official EXIOBASE emission accounts and OECD data at the country level, though discrepancies at the sectoral level give cause for further examination. We provide our GHG emission accounts with associated uncertainties and correlations at https://doi.org/10.5281/zenodo.10041196 (Schulte et al., 2023) to complement the official EXIOBASE emission accounts for users interested in estimating the uncertainties of their results.

Impact of intra-patient variability of tacrolimus on allograft function and CD4 + /CD8 + ratio in kidney transplant recipients: a retrospective single-center study.

Tacrolimus is a critical component of immunosuppressive therapy for kidney transplant recipients. Intra-patient variation (IPV) of tacrolimus levels affects the function of transplanted kidney. This study aimed to investigate the impact of tacrolimus IPV on kidney function, examine its association with post-transplant duration, and assess its effect on the immune status of transplant recipients. This retrospective study was conducted from January 2016 to February 2022. IPV was evaluated using the coefficient of variation (CV) of tacrolimus trough levels from 6 to 48months after transplantation. Patients were divided into low- and high-IPV groups based on the median CV. Significant differences in kidney function, CD4 + /CD8 + ratio, and post-transplant duration between these groups were analyzed. Among 189 patients, tacrolimus IPV showed a strong correlation with serum creatinine clearance rate (Ccr) and estimated glomerular filtration rate (eGFR) (p < 0.05). Tacrolimus IPV was significantly correlated with post-transplant duration in only two patients (p < 0.05). Using a median CV of 15.4% to categorize patients, the high IPV group, compared to the low IPV group, exhibited significantly higher eGFR at 6-9months (p < 0.05), lower Ccr at 9-12months (p < 0.05), and reduced Ccr and eGFR at 15-18months (p < 0.05). Six months after transplantation, the high IPV group had a significantly lower CD4 + /CD8 + ratio than the low IPV group (p < 0.05). This study highlights the significant impact of tacrolimus IPV on transplant kidney function and immune status in transplant patients at various post-transplantation intervals.

Ongoing Laboratory Performance Study on Chemical Analysis of Hydrophobic and Hydrophilic Compounds in Three Aquatic Passive Samplers.

The quality of chemical analysis is an important aspect of passive sampling-based environmental assessments. The present study reports on a proficiency testing program for the chemical analysis of hydrophobic organic compounds in silicone and low-density polyethylene (LDPE) passive samplers and hydrophilic compounds in polar organic chemical integrative samplers. The median between-laboratory coefficients of variation (CVs) of hydrophobic compound concentrations in the polymer phase were 33% (silicone) and 38% (LDPE), similar to the CVs obtained in four earlier rounds of this program. The median CV over all rounds was 32%. Much higher variabilities were observed for hydrophilic compound concentrations in the sorbent: 50% for the untransformed data and a factor of 1.6 after log transformation. Limiting the data to the best performing laboratories did not result in less variability. Data quality for hydrophilic compounds was only weakly related to the use of structurally identical internal standards and was unrelated to the choice of extraction solvent and extraction time. Standard deviations of the aqueous concentration estimates for hydrophobic compound sampling by the best performing laboratories were 0.21 log units for silicone and 0.27 log units for LDPE (factors of 1.6 to 1.9). The implications are that proficiency testing programs may give more realistic estimates of uncertainties in chemical analysis than within-laboratory quality control programs and that these high uncertainties should be taken into account in environmental assessments.

Abstract 81: High-resolution monitoring of B cell isotype switching for biomarker analysis of adenosine pathway inhibition immunotherapy

Abstract Background B cell isotype switching gives rise to IgG-, IgA-, and IgE-expressing B cells and is an essential step in generating effective humoral responses to antigen challenge. Within the tumor microenvironment, extracellular adenosine may contribute to cancer immune evasion by inhibiting isotype switching through interaction with A2aR, a possibility that has driven investigation of adenosine pathway inhibition as an immunotherapeutic modality. In this context, methods to quantify isotype switching may provide insight into the pharmacodynamics of candidate agent activity and serve as a source of predictive biomarkers of response. We recently clinically validated the Oncomine IGH-LR assay for the measurement of CLL somatic hypermutation (SHM). Given that this assay also detects isotype switching events, we retrospectively analyzed our validation dataset to determine the suitability of the assay as a potential tool for biomarker analysis of adenosine pathway inhibition immunotherapy. Methods The Oncomine IGH-LR assay determines B cell SHM and isotype through targeted next-generation sequencing of IGH chains from RNA. The assay software organizes detected IGH chains into B cell clonal lineages, of which a subset may comprise multi-isotype lineages indicative of isotype switching events. The software further reports secondary repertoire features such as clonality/evenness. Total RNA was extracted from peripheral blood mononuclear cells or bone marrow aspirates from patients identified as having &amp;gt;10% monoclonal B-cell population by flow cytometry. Intra-assay precision was evaluated by processing RNA from 7 samples in triplicate and sequencing in the same run. Inter-assay precision was evaluated by having different operators process 10 samples on different days with different sequencing barcodes. Dominant clone isotype and clonal lineage assignment, total isotype abundance, and repertoire clonality were evaluated across replicates. Results Overall, 100% concordance was achieved for dominant clone isotype and clonal lineage calls for both intra- and inter-assay studies. Intra-assay and inter-assay median coefficient of variation (CV) for reported isotype frequencies ranged from 0.10-0.16 and 0.15-0.28, respectively. Intra-assay and inter-assay median CV for clonality (1 - Normalized Shannon Entropy) was 0.06 and 0.04, respectively. Conclusions We observe high reproducibility and repeatability of the IGH-LR assay for B cell isotyping, clonal lineage analysis, and clonality assessment. Multi-isotype clonal lineages are automatically reported by the software and provide a means to directly monitor isotype switching events in the context of antigen challenge or immunomodulation. Taken together, we anticipate this assay to be a useful tool for biomarker analysis of adenosine pathway inhibition immunotherapy. Citation Format: Ramadevi Prathapam, Kristen Champion, Anne Burkhardt, Chelsey L. Smith, Rajesh Singh, Timothy J. Looney. High-resolution monitoring of B cell isotype switching for biomarker analysis of adenosine pathway inhibition immunotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 81.

High intra-patient variability of tacrolimus within post-operative 1month predicted worse 1-year outcomes in pediatric liver transplant recipients.

The pharmacokinetics of tacrolimus (TAC) show high intra-patient variability (IPV), which is associated with poor long-term outcomes following adult liver transplantation(LT). However, this relationship remains to be confirmed in pediatric liver transplant (PLT) recipients. The present study aimed to investigate the association between TAC IPV and grafts or patient outcomes after pediatric liver transplantion. This retrospective study included 848 PLT recipients (including infants) between January, 2016, and June, 2021. The IPV of TAC concentrations was estimated by calculating the coefficient of variation (CV) of trough concentrations in whole blood within 1month after transplantation. Patients were categorized into two groups, low IPV (CV < 45%) and high IPV (CV ≥ 45%), based on the third quartile of the CV distribution. A total of 848 patients were included in our study. The low CV group included 614 patients, with a mean TAC trough concentration of 8.59 ± 1.65ng/ml and a median CV of 32.37%. In contrast, the high CV group included 214 patients, the mean TAC trough concentration and median CV were 8.81 ± 2.00ng/ml and 54.88%, respectively. The median hospital duration was significantly higher in the high CV group (22days vs. 20days, P = 0.01). Univariate analysis was performed to evaluate the significant differences in 1-year recipient survival (P = 0.041) and 1-year graft survival (P = 0.005) between the high- and low-CV groups. Moreover, high CV (HR 2.316, 95%CI 1.026-5.231, P = 0.043) and persistent EBV viremia (HR 13.165, 95%CI 3.090-56.081, P < 0.001) were identified as independent risk factors for 1- year mortality after PLT. PLT recipients with high TAC trough concentration of CV in the first month were associated with poor 1-year outcomes. This CV calculation provides a valuable strategy to monitor TAC exposure.

Neuroimaging Findings in US Government Personnel and Their Family Members Involved in Anomalous Health Incidents

US government personnel stationed internationally have reported anomalous health incidents (AHIs), with some individuals experiencing persistent debilitating symptoms. To assess the potential presence of magnetic resonance imaging (MRI)-detectable brain lesions in participants with AHIs, with respect to a well-matched control group. This exploratory study was conducted at the National Institutes of Health (NIH) Clinical Center and the NIH MRI Research Facility between June 2018 and November 2022. Eighty-one participants with AHIs and 48 age- and sex-matched control participants, 29 of whom had similar employment as the AHI group, were assessed with clinical, volumetric, and functional MRI. A high-quality diffusion MRI scan and a second volumetric scan were also acquired during a different session. The structural MRI acquisition protocol was optimized to achieve high reproducibility. Forty-nine participants with AHIs had at least 1 additional imaging session approximately 6 to 12 months from the first visit. AHIs. Group-level quantitative metrics obtained from multiple modalities: (1) volumetric measurement, voxel-wise and region of interest (ROI)-wise; (2) diffusion MRI-derived metrics, voxel-wise and ROI-wise; and (3) ROI-wise within-network resting-state functional connectivity using functional MRI. Exploratory data analyses used both standard, nonparametric tests and bayesian multilevel modeling. Among the 81 participants with AHIs, the mean (SD) age was 42 (9) years and 49% were female; among the 48 control participants, the mean (SD) age was 43 (11) years and 42% were female. Imaging scans were performed as early as 14 days after experiencing AHIs with a median delay period of 80 (IQR, 36-544) days. After adjustment for multiple comparisons, no significant differences between participants with AHIs and control participants were found for any MRI modality. At an unadjusted threshold (P < .05), compared with control participants, participants with AHIs had lower intranetwork connectivity in the salience networks, a larger corpus callosum, and diffusion MRI differences in the corpus callosum, superior longitudinal fasciculus, cingulum, inferior cerebellar peduncle, and amygdala. The structural MRI measurements were highly reproducible (median coefficient of variation <1% across all global volumetric ROIs and <1.5% for all white matter ROIs for diffusion metrics). Even individuals with large differences from control participants exhibited stable longitudinal results (typically, <±1% across visits), suggesting the absence of evolving lesions. The relationships between the imaging and clinical variables were weak (median Spearman ρ = 0.10). The study did not replicate the results of a previously published investigation of AHIs. In this exploratory neuroimaging study, there were no significant differences in imaging measures of brain structure or function between individuals reporting AHIs and matched control participants after adjustment for multiple comparisons.

Abstract 148: Increased Systolic Blood Pressure Variability During the First 24-hours of Hospitalization Associates With Poor 90-day Outcome After Intracerebral Hemorrhage

Introduction: Prospective studies and secondary analyses from clinical trials have identified increased systolic blood pressure variability (SBPV) as a risk factor for poor outcomes. Evidence of this association in real-world intracerebral hemorrhage (ICH) hospitalization is lacking, however. Methods: Data for adult (≥18) patients with primary ICH were retrieved from the REINAH cerebrovascular research database. Systolic blood pressure measurements from the first 24 hours of admission were retrieved and SBPV was calculated as the Coefficient of Variation (CV) = (standard deviation/mean)*100. Socioeconomic deprivation was assessed using the state Area Deprivation Index (ADI), with high deprivation assessed at ADI ≥ 8. The primary outcome was severe disability or death (SDD; modified Rankin Scale ≥4) at 90-days after discharge. Differences in SBPV across SDD were assessed using the Mann-Whitney U test. Associations between SBPV and SDD were assessed using multivariable logistic regression models adjusted for patient characteristics. Adjusted odds ratios (aOR) and 95% confidence intervals (CI) are reported. Results: Final cohort included 807 patients (median [IQR]: 66 [54-77], 45.6% female), with an ethno-racial distribution of 41.9% White, 25.9% Black, 23.9% Hispanic, 5.7% Asian, and 2.6% Other. The median CV was 12.07 [9.50-15.59] and 485 (60.1%) patients experienced SDD. Patients with SDD showed significantly higher SBPV than non-SDD patients (12.90 [10.33-17.11] vs 10.99 [8.72-13.72]; p&lt;0.001). Patients with 24-hour SBPV in the 3 rd (1.86 [1.16-2.96]) and 4 th quartile (3.10 [1.89-5.09]) had higher independent odds of SDD (vs. 1 st quartile) (Figure 1). Discussion: Patients with high SBPV during the first 24 hours of admission have higher odds of severe disability or death at 90-days after discharge.

Effects of intensive glycemic control on microvascular outcomes in type 2 diabetes mellitus are modified by long-term HbA1c variability: A post hoc analysis of the ACCORD trial

AimsTo assess the impact of long-term visit-to-visit variability in HbA1c on microvascular outcomes in type 2 diabetes mellitus (T2DM), and its influence on the effects of intensive glycemic control. MethodsIncluded were participants with T2DM enrolled in the Action to Control Cardiovascular Risk in Diabetes (ACCORD) who had at least three measurements of HbA1c prior to new-onset microvascular outcomes, namely nephropathy, retinopathy and neuropathy. Variability in HbA1c was defined as the coefficient of variation (CV) across HbA1c measurements obtained from enrollment to the transition from intensive to standard glycemic therapy. ResultsDuring a median of 22,005, 23,121, and 13,080 person-years of follow-up, 2,905 nephropathy, 2,655 retinopathy, and 1,974 neuropathy cases were recorded, respectively. Median CV (IQR) was 7.91 % (5.66 %-10.76 %) in the standard treatment group and 9.79 % (7.32 %-13.35 %) in the intensive treatment group. In the standard treatment group, lower HbA1c-CV (the first versus the second quartile) was associated with a higher risk of all microvascular outcomes, while higher HbA1c-CV (the fourth quartile) was associated with a higher risk of nephropathy only. In the intensive treatment group, only higher HbA1c-CV was associated with a higher risk of developing the microvascular outcomes. Intensive therapy reduced all microvascular outcomes among individuals with lower HbA1c-CV, but increased the risk among those with the highest HbA1c-CV (all P values for interaction < 0.0001). For example, hazard ratios (95 % CI) of retinopathy comparing intensive with standard treatments were 0.65 (0.56–0.75), 0.84 (0.71–0.98), 0.97 (0.82–1.14) and 1.28 (1.08–1.53) across the lowest to the highest quartiles of HbA1c variability. ConclusionsThe effects of intensive glycemic control on microvascular outcomes in T2DM appear to be modified by the variability of HbA1c during the treatment process, suggesting the significance of dynamic monitoring of HbA1c levels and timely adjustments to the therapeutic strategy among individuals with a high HbA1c variability.

Longitudinal evaluation of external quality assessment results for CA 15-3, CA 19-9, and CA 125.

Tumor markers are established laboratory tools that help to diagnose, estimate prognosis, and monitor the course of cancer. For meaningful decision-making in patient care, it is essential that methods and analytical platforms demonstrate high sensitivity, specificity, precision, and comparability. Regular participation at external quality assessment (EQA) schemes is mandatory for laboratories. Here, a longitudinal evaluation of EQA data was performed to assess the performance of tumor marker assays over time. Longitudinal data of the cancer antigens (CA) 15-3 (n = 5,492), CA 19-9 (n = 6,802), and CA 125 (n = 5,362) from 14 INSTAND EQAs conducted between 2019 and 2023 were evaluated. A median of 197, 244 and 191 laboratories participated at the EQAs for CA 15-3, CA 19-9 and CA 125, respectively. Data evaluation encompasses intra- and inter-manufacturer specific variations over time, assay precision, and adherence to the EQA limits of ±24% for CA 15-3, ±27% for CA 19-9 and ±36% for CA 125. The study showed median manufacturer-dependent differences of up to 107% for CA 15-3, 99% for CA 125, and even 549% for CA 19-9 between the highest and the lowest methods over the studied period. Regarding the normalized median of all methods, the values of the most deviant methods were 0.42 for CA 15-3, 7.61 for CA 19-9, and 1.82 for CA 125. Intra-manufacturer variability was generally low, with median coefficients of variation (CV) below 10%. As the methods were evaluated according to method-specific consensus values, most participants passed the EQAs within the acceptance criteria. When the criteria were consistently set at 24%, the central 90% of participants passed the EQAs in 78.6%-100% for CA 15-3 (with exception of AX), 89.3%-100% for CA 125, and 64.3%-100% for CA 19-9. While intra-method precision of most analytical platforms is acceptable for all three tumor markers, considerable inter-method variability was observed over the whole studied period demonstrating the necessity for better standardization and harmonization of the methods, development of international reference materials, and comprehensive commutability studies with patient samples.

Smartwatch-derived sleep and heart rate measures complement step counts in explaining established metrics of MS severity

BackgroundPassive remote monitoring of patients with MS (PwMS) with sensor-based wearable technologies promises near-continuous evaluation with high ecological validity. Step counts correlate strongly with traditional measures of MS severity. We hypothesized that remote monitoring of sleep and heart rate will yield complementary information. MethodsWe recruited 31 PwMS and 31 age- and sex-matched healthy volunteers (HV) as part of the dreaMS feasibility study (NCT04413032). Fitbit Versa 2 smartwatches were worn for 6 weeks and provided a total of 25 features for activity, heart rate, and sleep. Features were selected based on their pairwise intercorrelation (Pearson |r| < 0.6), test-retest reliability (intraclass correlation coefficient ≥ 0.6 or median coefficient of variation < 0.2) and group comparisons between HV and PwMS with moderate disability (expanded disability status scale (EDSS) ≥ 3.5) (rank-biserial |r| ≥ 0.5). These selected features were correlated with clinical reference tests (EDSS, timed 25-foot walk (T25FW), MS-walking scale (MSWS-12)) in PwMS, and multivariate models adjusted for age, sex, and disease duration were compared. ResultsWe analyzed 28 PwMS (68% female, mean age 44 years, median EDSS 3.0) and 26 HV in our primary analysis. The objectively selected features discriminated well between HV and PwMS with moderate disability with rank-biserial r = 0.83 for Total number of steps, 0.51 for Deep sleep proportion, -0.51 for Median heart rate, 0.85 for Proportion very active, and 0.65 for Total number of floors. In PwMS they correlated strongly with the three clinical reference tests EDSS (strongest Spearman ρ = -0.75 for Proportion very active), T25FW (-0.75 for Total number of floors), and MSWS-12 (-0.72 for Total number of floors). Deep sleep proportion and Median heart rate complemented Total number of steps in explaining the variance of reference tests. ConclusionsActivity, deep sleep and heart rate measures can be derived reliably from smartwatches and contain independent clinically meaningful information about MS severity, highlighting their potential for continuous passive monitoring in both clinical trials and clinical care of PwMS.

Hands-Free Proteomic Profiling of Urinary Extracellular Vesicles with a High-Throughput Automated Workflow.

Extracellular vesicles (EVs) have emerged as a promising source of disease biomarkers for noninvasive early stage diagnoses, but a bottleneck in EV sample processing restricts their immense potential in clinical applications. Existing methods are limited by a low EV yield and integrity, slow processing speeds, low sample capacity, and poor recovery efficiency. We aimed to address these issues with a high-throughput automated workflow for EV isolation, EV lysis, protein extraction, and protein denaturation. The automation can process clinical urine samples in parallel, resulting in protein-covered beads ready for various analytical methods, including immunoassays, protein quantitation assays, and mass spectrometry. Compared to the standard manual lysis method for contamination levels, efficiency, and consistency of EV isolation, the automated protocol shows reproducible and robust proteomic quantitation with less than a 10% median coefficient of variation. When we applied the method to clinical samples, we identified a total 3,793 unique proteins and 40,380 unique peptides, with 992 significantly upregulated proteins in kidney cancer patients versus healthy controls. These upregulated proteins were found to be involved in several important kidney cancer metabolic pathways also identified with a manual control. This hands-free workflow represents a practical EV extraction and profiling approach that can benefit both clinical and research applications, streamlining biomarker discovery, tumor monitoring, and early cancer diagnoses.

Reproducibility of APT-weighted CEST-MRI at 3T in healthy brain and tumor across sessions and scanners

Amide proton transfer (APT)-weighted chemical exchange saturation transfer (CEST) imaging is a recent MRI technique making its way into clinical application. In this work, we investigated whether APT-weighted CEST imaging can provide reproducible measurements across scan sessions and scanners. Within-session, between-session and between scanner reproducibility was calculated for 19 healthy volunteers and 7 patients with a brain tumor on two 3T MRI scanners. The APT-weighted CEST effect was evaluated by calculating the Lorentzian Difference (LD), magnetization transfer ratio asymmetry (MTRasym), and relaxation-compensated inverse magnetization transfer ratio (MTRREX) averaged in whole brain white matter (WM), enhancing tumor and necrosis. Within subject coefficient of variation (COV) calculations, Bland–Altman plots and mixed effect modeling were performed to assess the repeatability and reproducibility of averaged values. The group median COVs of LD APT were 0.56% (N = 19), 0.84% (N = 6), 0.80% (N = 9) in WM within-session, between-session and between-scanner respectively. The between-session COV of LD APT in enhancing tumor (N = 6) and necrotic core (N = 3) were 4.57% and 5.67%, respectively. There were no significant differences in within session, between session and between scanner comparisons of the APT effect. The COVs of LD and MTRREX were consistently lower than MTRasym in all experiments, both in healthy tissues and tumor. The repeatability and reproducibility of APT-weighted CEST was clinically acceptable across scan sessions and scanners. Although MTRasym is simple to acquire and compute and sufficient to provide robust measurement, it is beneficial to include LD and MTRREX to obtain higher reproducibility for detecting minor signal difference in different tissue types.

Target Re-Delineation in Daily Online Adaptive Radiotherapy of Bladder Cancer: AI vs. Adapters

The purpose was to assess the quality of artificial intelligence (AI) generated targets in cone-beam CT (CBCT) guided online adaptive radiotherapy (oART) of bladder cancer and compare it to corresponding contours edited by adapters (radiotherapy technicians performing oART routinely). Furthermore, to assess the inter-observer delineation variation in daily target re-delineation in oART of bladder cancer. Ten consecutive patients treated with oART for muscle-invasive bladder cancer from Feb 2020 to Feb 2021 were retrospectively included. One CBCT randomly selected from each of the 10 clinical treatment courses was exported to an emulator, where delineations and oART simulations were carried out. The bladder was automatically segmented (clinical target volume (CTV)) using AI. Seven adapters independently reviewed and, if necessary, manually edited the AI-generated structure. A ground truth structure was first delineated blindly by a senior oncologist, then reviewed by an oART experienced medical physicist, and finally edited to a consensus structure. Planning CT and MR images with reference delineations were available during all simulations and delineations. The AI-generated and adapter-edited CTVs were compared to ground truth using dice similarity coefficient (DSC) and volume difference. The inter-observer variation in adapter-edited CTVs was assessed using coefficient of variation (CV) and the generalized conformity index (CIgen). Cigen was calculated as the ratio of the sum of all overlapping volumes between pairs of observers and the sum of all overlapping and non-overlapping volumes between the same pairs. Nine CTVs (7 adapter-edited, 1 AI-generated, 1 ground truth) were generated per patient. Patient cases included three male and seven females; three had catheter. The ground truth volume ranged from 51.5 to 221.6 cc. The median volume difference compared to ground truth was -4.5 (range, -17.8 - 42.4) cc and -15.5 (range, -54.2 - 4.3) cc for adapter-edited and AI-generated structures, respectively. Corresponding DSC values were 0.87 (range, 0.71 - 0.95) and 0.84 (range, 0.64 - 0.95). The AI-generated CTV was smaller than ground truth for all patients except one and smaller than all adapter-edited CTVs. The largest differences among adapter-edited CTVs were observed in cranial, caudal and posterior directions. The median CV was 0.08 (range, 0.05-0.11) and Cigen was 0.78 (range, 0.71-0.88). Target re-delineations in daily CBCT-based oART of bladder cancer demonstrated small inter-observer variation. Manual adjustment of AI-generated structures resulted in improved accuracy in target delineation compared to ground truth. Efforts to reduce the inter-observer variation, e.g., physician support during the initial fractions and review by a second adapter, have been implemented clinically.

Within-subject variation of HbA1c: A systematic review and meta-analysis.

Glycosylated haemoglobin (HbA1c) measurement is used to diagnose and to guide treatment of diabetes mellitus. Within-subject variability in measured HbA1c affects its clinical utility and interpretation, but no comprehensive systematic review has described within-subject variability. A systematic review and meta-analysis was performed of within-subject variability of HbA1c. Multiple databases were searched from inception to November 2022 for follow-up studies of any design in adults or children, with repeated measures of HbA1c or glycosylated haemoglobin. Title and abstract screening was performed in duplicate, full text screening and data extraction by one reviewer and verified by a second. Risk of bias of included papers was assessed using a modified consensus-based standards for the selection of health measurement Instruments (COSMIN) tool. Intraclass correlation coefficient (ICC) results were pooled with a meta-analysis and coefficient of variation (CV) results were described by median and range. Of 2675 studies identified, 111 met the inclusion criteria. Twenty-five studies reported variability data in healthy patients, 19 in patients with type 1 diabetes and 59 in patients with type 2 diabetes. Median within-subject coefficient of variation (CV) was 0.070 (IQR 0.034 to .09). For healthy subjects the median CV for HbA1c % was 0.017 (IQR 0.013 to 0.022), for patients with type 1 diabetes 0.084 (IQR 0.067 to 0.89) and for type 2 diabetes 0.083 (IQR 0.06 to 0.10). CV increased with mean population HbA1c. Assessment of variability was not the main aim of many of the included studies and some relevant papers may have been missed. Many included papers had few participants or few repeated measurements. Within-subject variability of HbA1c is higher for patients with than without diabetes and increases with mean population HbA1c. This may confound observed relationships between HbA1c variability and health outcomes. Because of its importance in clinical decision-making there is a need for better estimates and understanding of factors associated with of HbA1c variability.

1821-PUB: Relationship between Time in Range and Glucose Management Indicator in Adolescents and Children with Type 1 Diabetes Mellitus

Objective: To investigate the relationship between time in range (TIR, 3.9-10.0mmol/L) and glucose management indicator (GMI) in adolescents and children with type 1 diabetes mellitus (T1DM) and explore the impact of coefficient of variation (CV) on their relationship. Methods: Data derived from continuous glucose monitoring (CGM) and other clinical data (including age, duration of T1DM, and laboratory-measured HbA1c) were obtained from the annual follow-up of the Guangdong T1DM Translational Medicine Study. The patients under 18 years old were included. The patients were divided into CV≤36% group and CV&amp;gt;36% group by the attainment of CV. The relationship between TIR and GMI of both groups was assessed with correlation coefficient. Further, patients were divided into 4 groups by the interquartile range of CV. The linear regression model was used to calculate the TIR predicted value corresponding to the same GMI in 4 groups. Results: The 56 eligible datasets collected from May 2014 to August 2021 were included. The median age, duration of T1DM, and laboratory-measured HbA1c were 14.00 (12.00, 16.00) years, 4.15 (1.66, 5.29) years, and 9.00 (7.65, 11.23) %, respectively. The median TIR, GMI, CV, and valid number of days the CGM device was worn were 60.73 (42.59, 77.72) %, 7.37 (6.78, 8.56) %, 30.38 (24.87, 34.94) %, and 3.39 (2.97, 4.84) days, respectively. TIR and GMI were highly linear correlated (R2=0.89, p&amp;lt;0.001), and a significantly higher Spearman’s correlation coefficient was observed in the CV≤36% group than in CV&amp;gt;36% group ((R2=0.92, p&amp;lt;0.001) vs. (R2=0.63, p=0.004)). When the GMI was 7%, the corresponding TIR predicted values gradually decreased with the increase of CV, which were 75.66% (CV≤24.87%), 73.48% (24.87%&amp;lt;cv≤30.38%), 69.16% (30.38%34.94%). Conclusions: TIR and GMI were highly linear correlated in adolescents and children with T1DM. With constant GMI, the less the glycemic fluctuation, the higher the TIR.&amp;lt;/cv≤30.38%), 69.16% (30.38% Disclosure M.Lei: None. J.Lv: None. X.Mai: None. H.Deng: None. C.Wang: None. D.Yang: None. X.Yang: None. W.Xu: None. J.Yan: None. Funding Science and Technology Planning Project of Guangzhou (202102010154)

1312-P: Relationship between Progression of Diabetic Nephropathy and Seasonal Fluctuation of Estimated GFR

Introduction: The seasonal fluctuation of estimated glomerular filtration rate (GFR) in type2 diabetes mellitus (T2DM) patients remains unknown. The present study is first to examine seasonal fluctuation of eGFR and its relationship to the progression of diabetic nephropathy (DN). Methods: This retrospective study initially analyzed data on 21,307 T2DM patients with hypertension (HT). Seasonal fluctuation of eGFR were evaluated using the intrapersonal coefficient of variation (CV) of 20 or more serial measurements of body weight, Blood pressure and HbA1c during a 5-year period from 2014-2018. The primary objective was to determine if there is seasonal fluctuation of eGFR or not. Secondary objective is to examine the relationship between progression of DN and seasonal fluctuation of eGFR. DN was assessed by eGFR and urinary albumin excretion (UAE) changes. Results: The figure shows the changes of eGFR monthly averages. eGFR is significantly lower in summer each year (vs winter, P&amp;lt;0.001). Classified into two groups based on median CV of eGFR, eGFR decrease was obvious in the group with large CV of eGFR (vs group with small CV, P&amp;lt;0.001). In addition, UAE increase was also obvious in the group with large CV of eGFR (vs group with small CV, P&amp;lt;0.001). Conclusion: This study suggests that eGFR is fluctuated by season in T2DM patients with HT. The seasonal fluctuation of eGFR was associated with the progression of DN. Disclosure M.Yuki: None. T.Yamazaki: None. K.Miyashita: None. M.Sakamoto: None.

Clinical pharmacokinetic study of tacrolimus in continuous intravenous administration for lung transplantation.

Tacrolimus is a cornerstone of immunosuppression after lung transplantation. However, there are no clear guidelines on how to administer the drug and the duration to achieve the required therapeutic range in the early phase of lung transplantation. This is a single-center cohort study of adult patients who had lung transplantation. Tacrolimus was administered beginning with a low dose of 0.01 mg/kg/day immediately after transplantation. In addition, the designated clinical pharmacist conducted a daily intervention with trough concentrations to achieve the target of 10-15 ng/mL. Time in the therapeutic range (TTRin, %), time to the therapeutic range (TTRto, days), and coefficient of variation (CoV) of tacrolimus were evaluated for the 2-week post-transplant period. A total of 67 adult patients who had received first-time lung transplantation were included in the analysis. The median percentage of tacrolimus TTRin was 35.7% (21.4-42.9%) for the 2-week postoperative period. The median day of TTRto was 7 days (5-9 days), and the median tacrolimus trough concentration was 10.02 ng/mL (7.87-12.26 ng/mL) for the 2-week postoperative period. The median CoV of tacrolimus was 49.7% (40.8-61.6%). Acute kidney injury following tacrolimus infusion occurred in 23 (34.3%) patients, but there was no neurotoxicity or acute cellular rejection within 1 month of the postoperative period. In conclusion, continuous intravenous administration with the daily measure and dose titration of tacrolimus trough concentrations allowed the therapeutic range of tacrolimus to be reached within 1 week without significant adverse events, although the pharmacokinetic parameters were highly variable over time.

Abstract 3373: Deep learning for circulating tumor cell (CTC) identification with the CELLSEARCH system: achieving top human-level performance

Abstract Background: Enumeration of CTCs in blood samples with the CELLSEARCH® system is a prognostic biomarker in metastatic breast, prostate and colorectal cancer (full intended use documents.cellsearchctc.com). Currently, the final CTC identification is performed by human reviewers through visual assessment, which is a time-consuming procedure that could be affected by subjective interpretations. Modern Artificial Intelligence (AI) methods can provide a solution by automating the process of CTC identification, thus eliminating subjectivity and producing faster and more reproducible results. We have recently presented an automated algorithm (for research use only) based on deep learning models for cell segmentation and classification which allows the automated identification of CTCs in CELLSEARCH® fluorescent images (Ansaloni et al., EACR 2022). The purpose of this study was to assess inter-rater variability among human reviewers and to compare the AI performance against them. Methods: Blood samples derived from metastatic cancer patients were processed with CELLTRACKS® AUTOPREP®. The AI algorithm was trained on fluorescent images of 7255 CTCs and 32876 non-CTC events (training set) acquired by CELLTRACKS ANALYZER II® (CTAII) from 90 breast, 122 prostate and 54 colorectal cancer samples. 10 human reviewers, qualified for CTC image analysis, performed blind labeling of a separate dataset independently (test set: 55 samples taken from 27 breast and 28 prostate cancer). We studied inter-rater agreement and variability using Fleiss Kappa and Coefficient of Variation (CV). The labels provided by 5 experienced reviewers (average experience ≅ 15 years) were used to generate the ground truth (GT) by majority voting. The CTC detection performance of the AI was ranked against the other 5 reviewers (average experience ≅ 5 years). Results: In the test dataset, 2990 out of 9554 events were identified as CTCs by the GT review on CTAII images. In the task of classifying each event as CTC or non-CTC, the inter-rater agreement among the 10 human reviewers measured by Fleiss Kappa was 0.85. These classification results provided the number of CTCs per patient counted by each reviewer. The inter-rater variability on the CTC enumeration among the 10 reviewers was evaluated for each sample (median CV = 17.9% and interquartile range = 17.6% , excluding 27 samples with mode = 0 CTC). Measured against the GT classification, the AI performance (accuracy = 94.7%; F1 = 91.6%) was at the top of the human reviewers’ ranking (accuracy = 92.4 - 93.6 - 93.8 - 94.0 - 95.8%; F1 = 86.8 - 89.2 - 89.2 - 89.5 - 93.1%). Conclusion: The AI algorithm reached top-ranking performance in CTC identification, surpassing 4 out of 5 expert reviewers. These results show the applicability of AI to cell classification to remove human subjectivity from the review process and to maximize standardization among different research centers. Citation Format: Luca Biasiolli, Pietro Ansaloni, Nicolò Gentili, Ramona Miserendino, Chiara Rossi, Giulio Signorini, Gianni Medoro. Deep learning for circulating tumor cell (CTC) identification with the CELLSEARCH system: achieving top human-level performance [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3373.

A New Index of Insulin Sensitivity from Glucose Sensor and Insulin Pump Data: In Silico and In Vivo Validation in Youths with Type 1 Diabetes.

Background: Estimation of insulin sensitivity (SI) and its daily variation are key for optimizing insulin therapy in patients with type 1 diabetes (T1D). We recently developed a method for SI estimation from continuous glucose monitoring (CGM) and continuous subcutaneous insulin infusion (CSII) data in adults with T1D (SISP) and validated it under restrained experimental conditions. Herein, we validate in vivo a new version of SISP performing well in daily life unrestrained conditions. Methods: The new SISP was tested in both simulated and real data. The simulated dataset consists of 100 virtual adults of the UVa/Padova T1D Simulator monitored during an open-loop experiment, whereas the real dataset consists of 10 youths with T1D monitored during a hybrid closed-loop meal study. In both datasets, participants underwent two consecutive meals (breakfast and lunch, at 7 and 11 am) with the same carbohydrate content (70 g). Plasma glucose and insulin were measured during each meal to estimate the oral glucose minimal model SI (SIMM). CGM and CSII data were used for SISP calculation, which was then validated against the gold standard SIMM. Results: SISP was estimated with good precision (median coefficient of variation <20%) in 100% of the real and 91% of the simulated meals. SISP and SIMM were highly correlated, both in the simulated and real datasets (R = 0.82 and R = 0.83, P < 0.001), and exhibited a similar intraday pattern. Conclusions: SISP is suitable for estimating SI in both closed- and open-loop settings, provided that the subject wears a CGM sensor and a subcutaneous insulin pump.

Abstract TP122: Systolic Blood Pressure Variability Increases During The Initial 24 Hours After Intravenous Nicardipine Administration Among Patients With Primary Intracerebral Hemorrhage

Background: Systolic blood pressure variability (SBPV) has been associated with poor outcomes among patients with intracerebral hemorrhage (ICH). Intravenous nicardipine (IVN) is frequently used for blood pressure management among hospitalized ICH patients, however SBPV in response to IVN has not been characterized. Methods: Data for primary ICH patients who received IVN were retrieved from a stroke-specific bioinformatics pipeline. SBPV was quantified as the coefficient of variation (CV) of cuff-measured systolic blood pressure, and CV was calculated over the time periods 0-24 hrs prior to (Pre24), 0-24 hrs after (Post24), and 24-48 hrs after (Post48) initial IVN administration. Group differences were assessed using Friedman’s Test and Bonferroni-corrected Wilcoxon Signed-rank tests. Contributions of SBPV at each timepoint to poor in-hospital mortality or discharge to hospice (poor outcome) were assessed in separate multivariable logistic regression models adjusted for patient characteristics. Adjusted Odds ratios (aOR) and 95% confidence intervals (CI) are reported. Results: Data included 370 ICH patients with a median [interquartile range] age of 65 [54 - 76]. Patients included were 46.0% female, 41.1% Non-Hispanic White, 25.1% Non-Hispanic Black, 23.8% Hispanic, 8.1% Asian, and 1.9% other. Patients had a median CV of 9.2 [6.1-13.6] in the Pre24 period, 11.6 [9.46 - 13.9] in Post24, and 9.5 [7.6 - 12.1] in Post48. SPBV showed significant differences across timepoints (p=0.000), with Post24 showing elevated SBPV (vs. Pre24 (p=0.000) or Post48 (p=0.000), Figure 1A, B). In multivariable modelling, patients in the highest quartile of SPBV during the Post24 period showed increased risk of poor outcome: aOR (CI) 2.91 (1.06-8.01), Figure 1C. Conclusions: SBPV increases over the first 24 hours of IVN administration and patients with higher SPBV during this time are at higher risk for in-hospital mortality. SPBV management protocols need to be evaluated.