Pruthi, Rajiv

Abstract Background Interpretation of multi-analyte laboratory profiles, such as lupus anticoagulant (LAC) testing, requires integration of numerous test results and clinical information. However, this process is time-consuming, requires domain knowledge, and has limited scalability. Given that rule-based systems are rigid and difficult to maintain and large language models pose risks of “hallucinations,” we developed and evaluated a novel retrieval-based approach to automate LAC profile interpretation to improve efficiency while maintaining fidelity and reliability for clinical use. Methods Our method processes laboratory values through ensemble random forest (RF) models to retrieve the most similar expert interpretation from a curated database of 347 previously interpreted LAC profiles. The system's performance was assessed through technical accuracy metrics, a work flow efficiency timing study, and expert review of interpretation accuracy for held-out cases. Results Technical validation demonstrated high accuracy in retrieving correct interpretative elements (median: 93% accuracy, interquartile range [IQR]: 67% to 100%). Retrieval accuracy was highly correlated with the prevalence of a given element (ρ = 0.8307). In a timing study of 93 consecutive LAC profiles, ensemble-assisted interpretation achieved a median time of 22 s per report (IQR: 16 s to 31 s), representing a 78.6% total reduction compared to manual interpretation (median: 118 s, IQR: 96 s to 172 s; t = 18.61, P < 0.001). Furthermore, 92.5% of AI-assisted interpretations were completed within 60 s, exceeding the target of 85%. Expert review confirmed 100% concordance for LAC presence/absence. Conclusions The proposed approach significantly improves efficiency of complex laboratory interpretation for LAC profiles, while maintaining consistency and high fidelity. Future studies will explore the method’s inherent extensibility to the interpretation of other multi-analyte laboratory profiles.

Introduction: Calibrated automated thrombogram assay parameters have been shown to assess hypercoagulation state and, therefore, may independently predict posttrauma, symptomatic venous thromboembolism (VTE). Yet, calibrated automated thrombogram lacks high throughput and standardization, limiting diagnostic potential. This study analyzed thrombin generation (TG) profiles in trauma patients using a high-throughput assay, processing up to 10 samples at once, providing batch testing capabilities. We hypothesized that trauma patients who developed VTE would show distinct TG profiles compared with those who did not. Methods: Trauma patients presenting to a Level I Trauma Center had samples collected within 12 hours postinjury prospectively. Follow-up was conducted to 90 days, time to symptomatic VTE diagnosis or death, confirmed via autopsy or imaging. TG profiles were measured from the ST Genesia assay. Data were presented as median (interquartile ranges) or n (%), with the Wilcoxon rank-sum or chi-squared test. Associations between TG parameters and clinical risk factors were assessed using Spearman correlation analysis or the Wilcoxon rank-sum test. Results: Two hundred and fifty-four trauma patients were analyzed [48.5 years (31.0, 62.0), 72.0% male]: 64 patients with VTE (25.2%) to 190 with non-VTE. VTE pts had a median time to VTE of 8 days, with 29 developing deep venous thrombosis, 24 with pulmonary embolism, and 11 with deep venous thrombosis/pulmonary embolism. No significant differences were found in age, sex, or body mass index; significantly more VTE patients underwent surgery and blood transfusions within 24 hours. TG profiles demonstrated accelerated TG in patients who developed VTE compared with non-VTE, with significant differences in peak height ( P = 0.001), time to peak ( P = 0.021), endogenous thrombin potential ( P < 0.001), and velocity index ( P = 0.001). Conclusion: TG profiles using ST Genesia can differentiate trauma patients at high risk of developing VTE. The platform suits clinical labs needing high throughput by providing batch testing capabilities. Study type: Prospective study with less than large effect and no negative criteria. Level of evidence: II.

AbstractCoagulation is a crucial biological mechanism in human bodies to prevent blood loss. Abnormal coagulation can cause bleeding diathesis or thrombosis, common pathologic conditions in our clinical practice. Many individuals and organizations have dedicated their efforts in the past decades to understanding the biological and pathological mechanisms of coagulation and developing laboratory testing tools and treatment options to help patients with bleeding or thrombotic conditions. Since 1926, the Mayo Clinic coagulation group has made significant contributions to the clinical and laboratory practice, basic and translational research on various hemostatic and thrombotic disorders, and the education and collaboration to share and advance our knowledge in coagulation through a highly integrated team and practice model. We would like to use this review to share our history and inspire medical professionals and trainees to join the efforts to advance our understanding of coagulation pathophysiology and improve our care for patients with coagulation disorders.

AbstractPatients with thrombophilia remain concerned about venous thromboembolism (VTE) risk with COVID‐19 vaccinations. The aim of this study was to examine VTE outcomes in patients with inherited or acquired thrombophilia who were vaccinated for COVID‐19. Vaccinated patients ≥18 years between November 1, 2020 and November 1, 2021 were analyzed using electronic medical records across the Mayo Clinic enterprise. The primary outcome was imaging confirmed acute VTE occurring 90 days before and after the date of the first vaccine dose. Thrombophilia patients were identified through laboratory testing results and ICD‐10 codes. A total of 792 010 patients with at least one COVID‐19 vaccination were identified. Six thousand sixty‐seven of these patients were found to have a thrombophilia, among whom there was a total of 39 VTE events after compared to 51 VTE events before vaccination (0.64% vs. 0.84%, p = .20). In patients with Factor V Leiden or prothrombin gene mutation, VTE occurred in 27 patients before and in 29 patients after vaccination (0.61 vs. 0.65%, p = .79). In patients with antiphospholipid syndrome, VTE occurred in six patients before and four patients after vaccination (0.59% vs. 0.39%, p = .40). No difference was observed in the overall VTE rate when comparing the postvaccination 90 days to the prevaccination 90 days, adjusted hazard ratio 0.81 (95% confidence interval: 0.53–1.23). In this subgroup of COVID‐19 vaccinated patients with thrombophilia, there was no increased risk for acute VTE postvaccination compared to the prevaccination timeframe. These results are consistent with prior studies and should offer additional reassurance to patients with inherited or acquired thrombophilia.

Congenital factor V deficiency (FVD) is a rare bleeding disorder. In this study, we investigated the genetic basis in an African American patient with factor V activity 3%. Custom sequence capture and targeted next-generation (NGS) sequencing of the F5 gene were undertaken followed by PCR and Sanger sequencing. Two novel variants were identified. In silico analyses correlated clinically with the patient's factor V activity and hemorrhagic tendency. A review of the literature regarding these genomic alterations is presented. We described two novel mutations causing moderate FVD. The first, Chr1:g.169483698C>A with cDNA change (F5):c.6529-1G>T, occurred in a conserved nucleotide at the canonical acceptor splice site of intron 24. The second, Chr1:g.169515775C>T with cDNA change (F5):c.1667G>A, was a missense variant of exon 11, affecting a highly conserved amino acid in the A2 domain. Further research into the mechanisms of F5 mutations leading to FVD and residual factor V expression are needed.