The application of artificial intelligence (AI) for automated diagnosis of electrocardiograms (ECGs) can improve care in remote settings but is limited by the reliance on infrequently available signal-based data. We report the development of a multilabel automated diagnosis model for electrocardiographic images, more suitable for broader use. A total of 2,228,236 12-lead ECGs signals from 811 municipalities in Brazil are transformed to ECG images in varying lead conformations to train a convolutional neural network (CNN) identifying 6 physician-defined clinical labels spanning rhythm and conduction disorders, and a hidden label for gender. The image-based model performs well on a distinct test set validated by at least two cardiologists (average AUROC 0.99, AUPRC 0.86), an external validation set of 21,785 ECGs from Germany (average AUROC 0.97, AUPRC 0.73), and printed ECGs, with performance superior to signal-based models, and learning clinically relevant cues based on Grad-CAM. The model allows the application of AI to ECGs across broad settings.
Objectives: The intravenous administration of a bolus dose of endotoxin to healthy human subjects triggers acute systemic inflammatory responses that include cytokine production and dynamic changes in gene expression in peripheral blood leukocytes. This study sought to determine the state of clock gene expression in human peripheral blood leukocytes, and leukocyte subpopulations, challenged with in vivo endotoxin at two circadian/diurnal phases of the clock. Design: Clinical and laboratory investigation. Setting: University-based research laboratory and clinical research center. Subjects: Human volunteers. Interventions: Human subjects were administered a standard dose of endotoxin (2 ng/kg) or saline at either 0900 or 2100 hrs. Blood samples were collected at selected time points pre- and postinfusion. Measurements and Main Results: Clock gene expression was determined in human peripheral blood leukocytes, neutrophils, and monocytes by quantitative real-time polymerase chain reaction. The fold change for each gene was determined by use of the 2−ΔΔCT method. We show that endotoxin causes profound suppression of circadian clock gene expression, clearly manifested in human peripheral blood leukocytes, neutrophils, and monocytes. Clock, Cry1-2, Per3, CSNK1 ε, Rora, and Rev-erb gene expression were all reduced by 80% to 90% with the nadir between 3 and 6 hrs postinfusion. Per1 and Per2 reached an expression nadir between 13 and 17 hrs postinfusion. The levels of plasma interleukin-6 and tumor necrosis factor peaked and then returned to baseline within 6 hrs. In contrast, clock gene expression remained suppressed for up to 17 hrs irrespective of the phase of the clock at the time of the endotoxin challenge. Endotoxin did not perturb the melatonin secretory rhythm. Conclusions: Circadian clock gene expression in peripheral blood leukocytes is dramatically altered and possibly uncoupled from the activity of the central clock during periods of acute systemic inflammation. The realignment of the central and peripheral clocks may constitute a previously unappreciated key factor affecting recovery from disease in humans.
More than 65 years ago, complex clinical diagnostic reasoning cases were introduced as the gold standard for the evaluation of expert medical computing systems, a standard that has held ever since. In this study, we report the results of a physician evaluation of a large language model (LLM) on challenging clinical cases across five experiments with a baseline of hundreds of physicians. We then report a real-world study comparing human expert and artificial intelligence (AI) second opinions in randomly selected patients in the emergency room of a major tertiary academic medical center. In all experiments, the LLM outperformed physician baselines and displayed continued improvement from prior generations of AI clinical decision support. Our study suggests that LLMs have eclipsed most benchmarks of clinical reasoning, motivating the urgent need for prospective trials.
Diagnostic errors in health care pose significant risks to patient safety and are disturbingly common. In the emergency department (ED), the chaotic and high‐pressure environment increases the likelihood of these errors, as emergency clinicians must make rapid decisions with limited information, often under cognitive overload. Artificial intelligence (AI) offers promising solutions to improve diagnostic errors in three key areas: information gathering, clinical decision support (CDS), and feedback through quality improvement. AI can streamline the information‐gathering process by automating data retrieval, reducing cognitive load, and providing clinicians with essential patient details quickly. AI‐driven CDS systems enhance diagnostic decision making by offering real‐time insights, reducing cognitive biases, and prioritizing differential diagnoses. Furthermore, AI‐powered feedback loops can facilitate continuous learning and refinement of diagnostic processes by providing targeted education and outcome feedback to clinicians. By integrating AI into these areas, the potential for reducing diagnostic errors and improving patient safety in the ED is substantial. However, successfully implementing AI in the ED is challenging and complex. Developing, validating, and implementing AI as a safe, human‐centered ED tool requires thoughtful design and meticulous attention to ethical and practical considerations. Clinicians and patients must be integrated as key stakeholders across these processes. Ultimately, AI should be seen as a tool that assists clinicians by supporting better, faster decisions and thus enhances patient outcomes.
A wavelet transform of the DNA "walk" constructed from a genomic sequence offers a direct visualization of short and long-range patterns in nucleotide sequences. We study sequences that encode diverse biological functions, taken from a variety of genomes. Pattern irregularities in the transform are frequently associated with sequences of biological interest. Exonic regions, for example, visualize differently under wavelet analysis than introns, and ribosomal RNA regions display distinct universal signatures. DNA walk wavelet analysis can provide a sensitive and rapid assessment of the putative biological significance of genomic DNA.