Citations

Abstract Background Managing preanalytical variables is critical to providing quality laboratory test results. However, this can be a challenge for laboratory staff, particularly with manual detection of interfering substances and complicated workflows. For coagulation testing, we assess specimen integrity and acceptability by visual checks and quantification of interfering substance on a chemistry analyzer. To improve this subjective process, we evaluated the technical performance of a new model of coagulation analyzer with automated preanalytical specimen integrity detection. Our objective was to determine if we could reduce the number of specimens requiring visual inspection/off-analyzer interference testing in order to simplify our testing workflow. Methods Citrated plasma specimens received in the Hospital Clinical Laboratory at Mayo Clinic (Rochester, MN) were tested on the TOP ACL 500 or TOP ACL 550 Series Coagulation analyzers (Werfen, Boston, MA) according to manufacturer’s instruction. Specimens tested on the TOP 500 were visually inspected for hemolysis, icterus, and lipemia (HIL). If the technologist detected a possible interferent, an aliquot of plasma was analyzed on our cobas c501 chemistry analyzer (c501) (Roche Diagnostics, Indianapolis, IN) to obtain HIL values. Laboratory-established or manufacturer-defined interference thresholds were applied in determining specimen acceptability. For samples tested on the TOP 550, specimen integrity was assessed by the instrument’s automated interference detection module. The analyzer reported HIL as an “interference cloud” or estimated range, rather than a single value like the c501. HIL values were estimated by visual assessment of the low end of the “interference cloud”. Values were confirmed on the c501 only for specimens with an “interference cloud” at/exceeding the test-specific interference thresholds. We analyzed data from a 2-month period pre- (TOP 500) and post-implementation (TOP 550). The percent of specimens requiring quantitation/confirmation on the c501 was calculated. Statistical significance was calculated using the Chi-squared test where significance was defined as p<0.01. Correlation between c501 and TOP 550 estimated HIL values was evaluated. Results A total of 2.2% (171/7605) and 1.3% (113/8983) of specimens tested on the TOP 500 and TOP 550, respectively, required HIL measurements on the c501 (p<0.001). Of those samples, 8.2% (14/171) and 14.2% (16/113) from the TOP 500 and TOP 550, respectively, had HIL values above the interference thresholds and were rejected by the laboratory. Comparison of the estimated HIL values from the “interference cloud” and from the c501 resulted in slopes (Pearson correlation coefficients) of: H 1.56 (0.77), I 0.67 (0.97), and L 1.35 (0.99). Conclusions In comparison to our workflow with visual assessment of interference, use of the automated interference detection functionality on the TOP 550 instrument resulted in a lower percentage of specimens requiring HIL measurements on the chemistry analyzer. Agreement between interference measurements obtained on the c501 and the TOP 550 was moderate where TOP 550 overestimated the degree of hemolysis and lipemia and underestimated icterus. Implementation of automated interference detection on the TOP 550 analyzer improved our workflow by eliminating the need for tedious and subjective visual checks and reducing the percentage of specimens requiring additional HIL measurements.

Abstract Background Fibrinogen plays a critical role in hemostasis and is essential to clot formation. Changes in the quantity or activity of fibrinogen can lead to disorders affecting coagulation. Thus, it is essential to precisely quantify fibrinogen levels to differentiate between hyperfibrinogenemia in hypercoagulable disorders and dysfibrinogenemia, hypo/afibrinogenemia, and irregular bleeding. Current diagnostic tests that use antibodies to detect fibrinogen or activity-based measures are susceptible to interference from heparin contamination, hemolysis, and hyperbilirubinemia. Here, we evaluated the use of targeted proteomics for developing a quantitative assay for fibrinogen in plasma samples. Methods Four peptides each for alpha and gamma chains of fibrinogen were synthesized using FMOC chemistry on a Liberty Blue (CEM Corp) peptide synthesizer. A targeted multiple reaction monitoring (MRM) assay was developed and optimized using Dionex Ultimate 3000 HPLC connected to a TSQ Altis triple quadrupole mass spectrometer (Thermo Scientific). Evaluations were conducted using a range of standards, such as pure protein, WHO plasma and commercially available plasma with known fibrinogen amounts for the accurate quantitation of fibrinogen. Calibration curves were generated using different dilutions in various matrices. The repeatability and precision of the assay were evaluated across five different days. The plasma samples were processed and digested in a 96-well plate and subjected to C18 clean-up using a Bravo automated liquid handler (Agilent). All samples were spiked-in with corresponding heavy-labeled peptides prior to digestion with trypsin. Multiple reaction monitoring assays were performed on tryptic digests of plasma samples. Results Calibration curves using commercial plasma for all peptides showed linearity from 20 mg/dl to 1,308 mg/dl. This covers the entire clinical range including normal (200-350 mg/dl), deficient states (<100 mg/dl) and excess production (>350 mg/dl) that can be used for evaluation of bleeding disorders including consumption coagulopathy and hyperfibrinogenemia. Based on the overall performance in terms of linearity, limit of detection/quantitation, carryover and coefficient of variation (CV) across five different days, two peptides per chain were finally selected for developing the quantitative assay. The intra-assay variability was <5% and intra-day CV was ∼20%. This assay is currently being tested on a variety of clinical samples representing a range of fibrinogen levels. Conclusions Our study focused on assessing whether multiplexed MRM assay could be implemented to measure fibrinogen levels in plasma. Our data indicate the feasibility of MRM-based assays for determining fibrinogen levels as an alternative to existing non-mass spectrometry-based activity or antigen assays.

Abstract Background Thoroughly investigating single patient result errors may identify systemic, large-scale testing errors. The laboratory observed an event where a revision to a Total Bile Acid (TBA, Diazyme Laboratories, Inc.) patient result following a 22S quality control (QC) failure displayed a larger than expected difference (52 to 5mcmol/L). TBA is primarily used for diagnosis and monitoring of intrahepatic cholestasis of pregnancy with results >10mcmol/L considered elevated. Retrospective review of result revisions following QC failures (3/15/2017-5/30/2023) revealed that 50%(20/40) yielded result differences of >10mcmol/L(>20SD based on assay imprecision). The aim of this study was to apply a systematic approach to i) estimate the rate of TBA errors, ii) ensure accurate result reporting during investigation period, iii) perform root cause analysis (RCA), and iv) determine corrective and preventative action(s). Methods Residual samples from TBA testing (6/24-7/5/2023) were retested(n=158), differences >+/-3.0mcmol/L or 15% were confirmed on an alternate analyzer, and reports were revised. Automated repeat testing of TBA samples >10mcmol/L was operationalized 7/6/2023. Initial and repeat TBA results were compared(n=448) and results differing by >+/-20% were remeasured on an alternate analyzer. RCA was conducted using a fishbone diagram. NaOH reagent probe washes were implemented for TBA and the error rate was re-assessed. Assays run prior to TBA samples with errors (n=15) were identified. To assess reagent carryover a residual serum pool (TBA ∼5mcmol/L) was aliquoted into 5 tubes in a sample rack. The first sample in the rack was programmed to run an assay suspected of causing carryover followed by 4 TBA measurements. Mean±standard deviation(SD) TBA concentrations were calculated. TBA was also measured in the liquid reagent for amylase, lipase, acetylcholinesterase (ACE) and fructosamine. Results Initial TBA retesting yielded no errors when initial TBA was ≤10mcmol/L(n=51). For samples with TBA >10mcmol/L(n=107), 9(8.4%) had differences exceeding criteria with 8/9 being revised to ≤10mcmol/L. Analysis of automated patient repeat data showed a 3.8%(17/448) error rate when initial TBA >10mcmol/L. After NaOH reagent probe washes were implemented, the error rate decreased to 0.7%(3/448). Assays run directly before an erroneously high TBA result included: lactate, fructosamine, soluble transferrin receptor(STFR), lipase, and ACE. A serum TBA pool(mean±SD=5.1±0.4 mcmol/L,n=15) measured 37.6±1.2mcmol/L(n=3) and 6.7±0.7mcmol/L(n=3) after lipase and fructosamine, respectively. No other assays demonstrated carryover. TBA in lipase reagent compartment B and C was 653 and 649mcmol/L, respectively, and 653mcmol/L in fructosamine reagent compartment B. Conclusions A single patient TBA result revision was investigated and led to identification of reagent carryover causing erroneously high TBA results on the Roche c502 analyzer. Automatic, real-time, repeat testing was implemented to prevent reporting incorrect patient results until RCA could identify the cause. Occurrence rate of erroneous high TBA results decreased from 3.8% to 0.7% after implementation of NaOH washes on the instruments. Carryover experiments confirmed that lipase and fructosamine assays cause carryover when run prior to a TBA sample due to measurable TBA in the reagent. The lab is pursuing moving the fructosamine and lipase reagent to the Roche Cobas c701 to prevent the issue.

Abstract Background Serum aspartate aminotransferase (AST) is a commonly ordered test useful for diagnosing and monitoring hepatocellular damage. The hemolysis (H) index interference threshold for the Roche Cobas AST reagent is H=20, which is the lowest of all tests on Roche chemistry platforms (Cobas c701, c502, and e801). Previous optimization studies supported the implementation of an AST concentration-specific tiered hemolysis interference threshold schema: AST<100 U/L (H=50), AST=100-200 U/L (H=100), AST=201-300 U/L (H=200) and AST>301 U/L (H=500). This schema significantly reduced redraws; however, a significant number of rejected samples were noted to have AST results within the reference interval (RI). The aim of this study was to establish an H-index threshold for samples with AST results within the RI. Methods Twelve pools of residual serum samples were created with AST concentrations ranging from 17-66 U/L. Each pool was split into 2 aliquots, receiving either 10% of a lab-produced hemolysate solution generating H-indices from 62-508 or 10% of a 0.9% saline control. Aliquots were mixed in equal portions to create a series of 5 samples with increasing hemolysis according to CLSI guideline EP-07 and analyzed in EP Evaluator. Median(interquartile range (IQR)) H-index thresholds were calculated for error limits beginning at +5 U/L and increasing incrementally (+1 U/L) to +10 U/L. Clinical cases with an H-index slightly above the hemolysis interference threshold (50≤H≤70) and results within the RIs (Male:8-48U/L; Female:8-43U/L) were reviewed by the Chemistry Resident/Fellow on-call (n=68) over a period of 8 months. Results were released with a comment if risk of clinical misinterpretation due to a falsely elevated result was low. Previous, non-hemolyzed AST results were compared within 45 days, if available, by paired-Wilcoxon test (Graphpad Prism). Results The median(range) of initial AST results in the serum pools was 20(17-66) U/L. The calculated H-index thresholds (median(IQR)) using error limits of +5, +6, +7, +8, +9, +10 U/L were 49(43-53), 56(52-63), 66(61-78), 78(72-81), 91(84-91) and 101 (96-107), respectively. All slightly hemolyzed AST results (50≤H≤70) within the RI were released with a comment upon chart review. Previously reported AST values (n=23) were compared to current (hemolyzed) concentrations and found to be statistically significantly increased relative to previous results from the same patient (median(range) difference: 3.7(-3.6-11)U/L, p=0.009). However, the median difference was <5U/L or 10%, the laboratory’s criteria used to define a clinically significant difference. Conclusions A median H-index limit of up to 100 can be applied to samples when AST is within the RI. The risk of misinterpretation for results with mild-moderate hemolysis was low in the cases reviewed. Sample rejection rates and reporting delays can be diminished by implementing this practice. Interpretation can be further enhanced by appending the following comment to hemolyzed samples: “Specimen was received with hemolysis slightly above the accepted threshold. This degree of hemolysis is associated with a false increase of approximately 5-10 U/L. Interpret in conjunction with other laboratory and clinical findings.”

Abstract Background Autoimmunity targeting GQ1b is associated with group of disorders that includes Miller-Fisher Syndrome (MFS), Bickerstaff Brainstem Encephalitis (BBE) and classic Guillain Barre Syndrome (GBS) with ophthalmoplegia. Collectively these are referred to as GQ1B-IgG related syndromes. The prevalence of GQ1b-IgG in this population of patients is high and has been reported to be >80% in well-defined clinical cohorts. Clinical phenotypic presentation is diverse and requires highly accurate antibody testing. The objective of this study was to evaluate a newly developed custom GQ1b-IgG assay, using a semi-quantitative ELISA-based method. Methods A GQ1b-IgG ELISA was developed and validated. Patient sera along with assay controls and a single-point calibrator are diluted and incubated in duplicate wells in a 96-well plate coated with GQ1b antigen. The plate is washed and incubated with anti-human-IgG secondary antibody conjugated with HRP. After a colorimetric reaction, optical densities (OD) are read. The OD of patient samples are compared against the calibrator, generating a cut-off index (COI). Samples ≥1.0 COI are considered positive. During the validation phase, analytical and clinical performance characteristics were evaluated. Studies included assessment of imprecision at multiple concentrations of GQ1b-IgG, accuracy using method comparison, analytical sensitivity and specificity, reference range, common interferences, plate drift, and clinical performance in relevant disease and control cohorts. Results Intra-assay imprecision (coefficient of variation; %CV) ranged from 1.0 - 16% across all concentrations of GQ1b-IgG. Inter-assay imprecision (%CV) ranged from 12.4 - 44.5% but was <20% at or above the assay cut-off. We found perfect qualitative agreement with an independent ELISA reference method (20 positive and 20 negative samples selected based on reference method results [reference interval <1:100]). Limit of detection (LOD) was calculated at a COI of 0.22. To assess analytical specificity, 40 samples with high concentrations of immunoglobulins (hypergammaglobulinemia [N=15], IgG M-protein [N=15] and systemic lupus erythematosus [N=10] were tested and all were negative. A reference interval study was performed on 140 healthy donor serums, all were negative. Interference studies were conducted on 2 positive and 2 negative serum samples, spiked for highly hemolytic (1000 mg/dL), lipemic (2000 mg/dL) and icteric (60 mg/dL) conditions. There was no impact on the assay even with high levels of a given interferent. Plate drift on 2 distinct patient pools tested across the plate (COI range 0.7 - 1.3) had overall plate CV’s ≤10%. ROC analysis was performed on a cohort of 278 patient samples (15 MFS cases and 263 controls [suspected demyelinating neuropathy or disease mimics]). A cut-off value of 1.0 (COI) was selected based on the Youden index. At this cut-off, clinical sensitivity for MFS was 87% while clinical specificity was 99.2%. Conclusions The observed performance of the GQ1b-IgG ELISA was acceptable for clinical laboratory implementation. A COI of ≥ 1.0 was selected for clinical use.

Abstract Background Patients with elevated free glycerol can have falsely elevated triglyceride concentrations known as pseudohypertriglyceridemia. Most clinical laboratory triglycerides methods measure the concentration of glycerol released following the enzymatic hydrolysis of triglycerides to quantitate triglycerides. Thus, increased free glycerol can lead to falsely increased triglyceride concentration. Identifying and measuring the interference is necessary to effectively manage patient care interventions. A glycerol-corrected triglyceride assay was discontinued from Roche Diagnostics in 2021. We created a process to determine glycerol-corrected triglycerides by measuring glycerol directly using an assay from Randox Laboratories. Methods Glycerol (Randox Laboratories) and nonglycerol-blanked triglycerides (Roche Diagnostics) were measured in serum using the Roche Cobas c501. Stability (n=10) and expected values (n=206) were determined using serum from healthy volunteers. Linearity (n=5) and precision (5 replicates on 2 runs over 5 days) were determined using residual clinical serum spiked with exogenous glycerol as necessary to span the analytical measuring range. Glycerol corrected triglyceride concentrations were calculated by subtracting glycerol (mmol/L) from triglyceride (mmol/L) and converting to mg/dL (x88.59). Samples previously measured using the Triglyceride/Glycerol Blanked Reagent (n=53, Roche Diagnostics) were used to assess accuracy. (Figure 1) Results Glycerol measures were stable 3 days refrigerated and 30 days frozen. Precision coefficient of variation was 1.0% within run and 7.1% within lab. Expected values were 10 - 109 mcmol/L, and the method was linear between 2 - 3,971 mcmol/L. The average bias between methods was 11.0% with slope of 1.02 and R2>0.97 (Figure 1). Conclusions Glycerol can be accurately measured using the Randox method and the result can be used to generate an accurate glycerol-corrected triglyceride concentration.

Abstract Background Drugs use during pregnancy is a major social and medical issue, which may lead to significant pre- and post-natal complications, such as preterm birth and newborn mortality. Exposed newborns may have long-term behavioral, cognitive, and developmental problems. Accurate detection of in utero drug exposure is crucial for timely medical intervention for the newborn and the mother. Meconium is the ‘gold standard' matrix for detection of drugs in newborn which reflects drug exposure during pregnancy. This study aims to develop and optimize a clinical assay for simultaneous screening of 21 drugs in meconium using a custom biochip array. Methods Custom biochip array (Randox Laboratories Ltd.) included 21 discrete test regions (DTRs) for amphetamine, methamphetamine, opioids/opiates (targeting morphine, oxymorphone, oxycodone, hydrocodone, 6-monoacetyl morphine, tramadol, buprenorphine, fentanyl, and methadone), phencyclidine (PCP), cocaine (targeting benzoylecgonine), benzodiazepines (targeting oxazepam, lorazepam, and clonazepam), zolpidem, barbiturates (targeting phenobarbital), ethyl glucuronide, methylphenidate, and marijuana metabolite (targeting Delta-9-Carboxy-Tetrahydrocannabinol [THC-COOH]). Patient samples were obtained from the residual meconium samples that had been previously used for other tests. Blank meconium samples were spiked at in-house cutoff, negative and positive assay control levels. All samples were extracted in specimen diluent using Biotage® Lysera and centrifuged. The supernatant was analyzed on a fully automated Randox Evidence+® Analyzer. A robotic arm pipetted 60 µL sample on to a biochip and then the analyzer performed a competitive chemiluminescent immunoassay. The results were generated as relative light units (RLUs). Results The sample results were determined to be “positive” or “negative” based on average RLU of cutoff controls for each analyte on an assay. The intra-assay (CV ≤ 15%) and inter-assay (CV ≤ 18%) precision was 100% agreement for positive and negative controls and there was no observed carryover for any of the analytes. The sample recovery was excellent with minimal to no matrix suppression for all analytes. Accuracy was performed by comparing spiked analyte results from Randox Evidence+® Analyzer with the expected results and all the spiked samples matched (positive or negative). When authentic patient or spiked results were compared with external laboratory test results, some discordant results were obtained. However, the discrepancies could be explained based on the different cutoffs (e.g., the Randox assay had lower cutoffs) or the known assay cross reactivities within the analytes of same class. These deviations did not affect clinical utility of the assay. Conclusions We validated a novel and simple qualitative chemiluminescent immunoassay in meconium using a custom biochip array. Unlike other routine screening/ELISA methods, which require separate assays for detection of similar drug targets, the custom biochip array offers a convenient, rapid and efficient method that utilizes only a single sample preparation for the screening of all 21 drug targets. Therefore, this technology has potential to be utilized as high-throughput screening assay. Like all qualitative assays, the limitation of this test is that all positive results will need to be assessed quantitatively for a definitive interpretation.

Abstract Background Live cell-based assays (CBA) are considered the gold-standard methodology for detecting pathogenic antibodies targeting neural cell surface antigens. However, the adoption of these methods in clinical laboratories has been limited by their manual nature and low throughput. Here, we describe the development and validation of an automated system utilizing a live flow cytometry-based CBA to detect AQP4-IgG. Establishing AQP4-IgG serostatus is a critical component of the diagnostic criteria for Neuromyelitis optica spectrum disorders (NMOSD) as it enables the differentiation of NMOSD from Multiple Sclerosis where there may be significant clinical overlap. The accurate diagnosis of the correct disease state is vital, as the prognosis and treatment approach differ significantly for these two conditions. Two redundant systems were built that automatically receive reagents (live cells in suspension, buffers, diluents, secondary antibodies, and patients’ samples) and generate flow cytometer ready 96-well plates (each containing up to 45 serum or CSF samples tested in duplicate). Methods We assessed the analytical performance of an AQP4-IgG live flow cytometry-based CBA on these new automated systems. We measured intra- and inter-assay imprecision at three levels in 5x5 format and for end-point titer assessment, accuracy (20 positive and 20 negative sera; 5 positive and 5 negative CSF; established on the current manual assay), reference range (20 negative sera and 20 negative CSF from disease controls), analytical sensitivity (LOB and LOD) and specificity (high IgG containing serum specimens), within-plate draft (1 positive sample near the diagnostic cut-off [IgG-binding index of 2.0] that was tested across the entire plate), instrument to instrument correlation (42 positive samples/dilutions tested simultaneously on both systems), and total throughput for both CSF and Serum. Results Imprecision testing yielded the following results: Serum Intra/Inter precision level 1 (IBI = 2) CV% (current method:4.6%/20.7%), (system 1: 3.1%/18.4%), (system 2: 2.6%/22.5%), serum Intra/Inter precision level 2 (IBI=4) CV% (current method:4.9%/23.3%), (system 1: 3.2%/21.9%), (system 2: 2.6%/24.1%), and serum Intra/Inter precision level 3 (IBI =10) pos CV% (current method:6.0%/31.1%), (system 1: 3.8%/27.4%), (system 2: 4.2%/29.4%), and CSF Intra/Inter precision (IBI =2) CV% (System 1: 11.7%/16.8%), (system 2: 6.1%/10.7%). The median end-point titers for all four samples assessed matched within one end-point titer. For accuracy, 100% qualitative agreement between current and both automated systems, and titration accuracy (5 positive serums, 1 positive CSF) all matched within one titer between current and both automated systems. All reference range samples were negative. LOB/LOD in serum (Current: 0.85/1.11), (System 1: 0.82/0.95), (system 2: 0.82/0.96) and CSF (Current: 1.06/1.52), (System 1: 1.02/1.38), (system 2: 1.02/1.64) were comparable. All high IgG samples were negative. Plate drift analysis showed improved whole plate %CV: (Current:6.52%), (system 1: 4.51%), (system 2: 3.46%). Instrument to instrument comparison utilizing linear regression indicated strong agreement across the two systems (slope = 0.93 and R2= 0.95). Assessment of throughput based on real-time timings indicated a maximum output of approximately >700 patient samples per 24hrs per system. Conclusions An automated system for performing live CBAs provides equivalent to superior analytical performance compared to pre-existing manual methods with improved testing capacity.

Abstract Background Myasthenia gravis (MG) is caused by disrupting the function of postsynaptic acetylcholine receptors (AChRs) at the neuromuscular junction. 85% of MG patients have autoantibodies targeting AChRs, while 6% of MG patients have autoantibodies targeting the muscle-specific receptor tyrosine kinase (MuSK). MuSK is located on the muscle membrane and is activated by the low-density lipoprotein receptor-related protein 4 (LRP4) upon the binding with ligand Agrin. Activated MuSK undergoes autophosphorylation and stimulates downstream signaling to cluster AChRs, which is important for the formation and maintenance of neuromuscular synapses. IgG4 antibodies that bind MuSK are thought to directly disrupt the ability of MuSK to cluster AChR. The gold standard methodology for MuSK antibody testing is radioimmunoprecipitation assay (RIA) using anti-human IgG secondary antibodies. Performing RIA methods in the clinical laboratory is a burden due to radioactive biohazard concerns and the analytical limitations of these assays. This study is to assess the analytical and clinical performance of a live cell-based flow cytometry assay as a potential replacement for the MuSK RIA. Methods HEK293T cells were transiently transfected with a vector co-expressing human MuSK protein and GFP. Patient sera were incubated with a mix of HEK293 cells (GFP-positive cells that express MuSK protein on the cell surface and GFP-negative cells that do no express MuSK) followed by incubation with anti-human IgG4 secondary antibody conjugated with alexa fluor 647 (AF647). An IgG binding index (IBI) was calculated for each sample as the ratio of the median AF647 fluorescence intensity (MFI) of the MuSK-expressing cells (GFP+) to the MFI of the non-MuSK expressing cells (GFP-). An arbitrary positive IBI cut-off of 2.0 was selected. A preliminary assessment of assay imprecision was performed for the flow cytometry assay, as well as a method comparison study to compare this assay to RIA utilizing a cohort of 130 (34 positive and 96 negative) residual samples previously tested by the gold-standard assay. 133 disease control samples were also tested to assess the clinical specificity of the flow cytometry-based assay. Results To assess the imprecision, eight samples with varying IBIs were tested with 5 replicates in each assay over five days. The coefficient of variation (CV) for intra-assay replicates ranged from 1% to 7% for all samples, while CV for inter-assay replicates ranged from 2% to 15%. Based on the 130 samples previously tested by RIA, the Positive Percentage Agreement (PPA) between these two assays was 76.5% and the negative Percentage Agreement (NPA) was 100%. Of the discordant samples (n=8), clinical information was available for only 1 sample (positive by RIA, negative by flow). This patient had a final non-MG diagnosis (ALS). All control samples were negative on the flow cytometry-based assay. Conclusions The live cell-based flow cytometry MuSK IgG4 assay demonstrated encouraging preliminary analytical performance that warrants further development and validation.

Abstract Background Laboratory identification of lupus anticoagulant (LAC) is critical for diagnosing antiphospholipid antibody syndrome, a common cause of unprovoked thromboembolic events and pregnancy morbidity. Commonly used clot-based assays for LAC include the dilute Russell’s viper venom time (DRVVT) and activated partial thromboplastin time (APTT), but these are susceptible to anticoagulant interferences. Despite standardization efforts, LAC interpretation remains challenging due to pre- and post-analytical complexity and limited expertise. This study explored the use of deep neural networks (DNN) to facilitate LAC diagnosis. Methods Retrospective data for LAC profiling with 13 features performed at Mayo Clinic between 2021 and 2022 was collected. The profiles of 7,202 patients were randomly divided into training (n=4608), validation (n=1153), and test cohorts (n=1441). Diagnostic outputs including LAC positivity (determined by DRVVT or APTT) and the presence of warfarin (WAR) and heparin (HEP), were confirmed by expert assessment. Three DNN architectures were developed and evaluated (Figure 1): (1) four single output DNN models with a domain knowledge driven selection of input features, (2) four single output DNN models using the same 13 inputs across all models, and (3) a multi-output DNN model. Model performance was evaluated using the F1 score and the index of balanced accuracy (IBA) for each output label. Results Architecture 1 achieved high diagnostic accuracy with IBA values of 0.989-1.000 for all four outputs and F1 scores of 0.945-0.991. Architecture 2 exhibited F1 scores of 0.902-0.973 and IBA values of 0.935-0.992. Architecture 3 achieved F1 scores of 0.933-0.967 and IBA values of 0.941-0.983. Conclusions All three architectures accurately classify LAC and common anticoagulation effects. The comparable performance of architectures 1 and 2 suggests that manual feature extraction is unnecessary. Given its relative simplicity and versatility, architecture 3 is the preferred choice for implementation in a clinical laboratory to standardize LAC diagnosis.