Search

Integrating modern machine learning and clinical decision-making has great promise for mitigating healthcare's increasing cost and complexity. We introduce the Enhanced Transformer for Health Outcome Simulation (ETHOS), a novel application of the transformer deep-learning architecture for analyzing high-dimensional, heterogeneous, and episodic health data. ETHOS is trained using Patient Health Timelines (PHTs)-detailed, tokenized records of health events-to predict future health trajectories, leveraging a zero-shot learning approach. ETHOS represents a significant advancement in foundation model development for healthcare analytics, eliminating the need for labeled data and model fine-tuning. Its ability to simulate various treatment pathways and consider patient-specific factors positions ETHOS as a tool for care optimization and addressing biases in healthcare delivery. Future developments will expand ETHOS' capabilities to incorporate a wider range of data types and data sources. Our work demonstrates a pathway toward accelerated AI development and deployment in healthcare.

Study Objective: Machine learning models have advanced medical image processing and can yield faster, more accurate diagnoses. Despite a wealth of available medical imaging data, high-quality labeled data for model training is lacking. We investigated whether a gamified crowdsourcing platform enhanced with inbuilt quality control metrics can produce lung ultrasound clip labels comparable to those from clinical experts. Methods: 2,384 lung ultrasound clips were retrospectively collected from 203 patients. Six lung ultrasound experts classified 393 of these clips as having no B-lines, one or more discrete B-lines, or confluent B-lines to create two sets of reference standard labels (195 training set clips and 198 test set clips). Sets were respectively used to A) train users on a gamified crowdsourcing platform, and B) compare concordance of the resulting crowd labels to the concordance of individual experts to reference standards. Results: 99,238 crowdsourced opinions on 2,384 lung ultrasound clips were collected from 426 unique users over 8 days. On the 198 test set clips, mean labeling concordance of individual experts relative to the reference standard was 85.0% +/- 2.0 (SEM), compared to 87.9% crowdsourced label concordance (p=0.15). When individual experts' opinions were compared to reference standard labels created by majority vote excluding their own opinion, crowd concordance was higher than the mean concordance of individual experts to reference standards (87.4% vs. 80.8% +/- 1.6; p<0.001). Conclusion: Crowdsourced labels for B-line classification via a gamified approach achieved expert-level quality. Scalable, high-quality labeling approaches may facilitate training dataset creation for machine learning model development., Comment: 27 pages total

Diagnostic ultrasound is a versatile and practical tool in the abdomen, and is particularly vital toward the detection and mitigation of early-stage non-alcoholic fatty liver disease (NAFLD). However, its performance in those with obesity -- who are at increased risk for NAFLD -- is degraded due to distortions of the ultrasound as it traverses thicker, acoustically heterogeneous body walls (aberration). Many aberration correction methods for ultrasound require measures of channel data relationships. Simpler, bulk speed of sound optimizations based on the image itself have demonstrated empirical efficacy, but their analytical limitations have not been evaluated. Herein, we assess analytically the bounds of a single, optimal speed of sound correction in receive beamforming to correct aberration, and improve the resulting images. Additionally, we propose an objective metric on the post-sum B-mode image to identify this speed of sound, and validate this technique through in virto phantom experiments and in vivo abdominal ultrasound data collection with physical aberrating layers. We find that a bulk correction may approximate the aberration profile for layers of relevant thicknesses (1 to 3 cm) and speeds of sound (1400 to 1500 m/s). Additionally, through in vitro experiments, we show significant improvement in resolution (average point target width reduced by 60 %) and improved boundary delineation in vivo with bulk speed of sound correction determined automatically from the beamformed images. Together, our results demonstrate the utility of simple, efficient bulk speed of sound correction to improve the quality of diagnostic liver images.

Modern deep learning algorithms geared towards clinical adaption rely on a significant amount of high fidelity labeled data. Low-resource settings pose challenges like acquiring high fidelity data and becomes the bottleneck for developing artificial intelligence applications. Ultrasound images, stored in Digital Imaging and Communication in Medicine (DICOM) format, have additional metadata data corresponding to ultrasound image parameters and medical exams. In this work, we leverage DICOM metadata from ultrasound images to help learn representations of the ultrasound image. We demonstrate that the proposed method outperforms the non-metadata based approaches across different downstream tasks., Comment: Accept as a workshop paper at AI4AH, ICLR 2020