IEEE VIS 2024 Content: PhenoFlow: A Human-LLM Driven Visual Analytics System for Exploring Large and Complex Stroke Datasets

PhenoFlow: A Human-LLM Driven Visual Analytics System for Exploring Large and Complex Stroke Datasets



Jaeyoung Kim - Seoul National University, Seoul, Korea, Republic of

 Sihyeon Lee - Seoul National University, Seoul, Korea, Republic of

 Hyeon Jeon - Seoul National University, Seoul, Korea, Republic of

 Keon-Joo Lee - Korea University Guro Hospital, Seoul, Korea, Republic of

 Bohyoung Kim - Hankuk University of Foreign Studies, Yongin-si, Korea, Republic of

 HEE JOON - Seoul National University Bundang Hospital, Seongnam, Korea, Republic of

 Jinwook Seo - Seoul National University, Seoul, Korea, Republic of

 Download Supplemental Material

 Room: Bayshore I

2024-10-16T16:12:00ZGMT-0600Change your timezone on the schedule page
2024-10-16T16:12:00Z

Exemplar figure, described by caption below — PhenoFlow empowers neurologists to explore large and complex stroke datasets with reduced cognitive load. (A) The cohort construction component allows neurologists to define target cohorts using natural language. (B) The Visual Inspection View provides plain-language explanations and small multiples of relevant fields to debug LLM data wrangler behavior. (C) The Cohort View summarizes (C1) cohort relationships in a node-link diagram and (C2) each patient's blood pressure (BP) trajectories as matrix visualization. (C3) Natural language filtering supports iterative cohort exploration. (D1) Linear bar charts and (D2) slice-and-wrap visualization present BP trajectories as time-series, revealing triangular patterns in irregularly measured BP data.

Fast forward

Full Video

Keywords

Stroke, Irregularly spaced time-series data, Multi-dimensional data, Cohort analysis, Large language models

Abstract

Acute stroke demands prompt diagnosis and treatment to achieve optimal patient outcomes. However, the intricate and irregular nature of clinical data associated with acute stroke, particularly blood pressure (BP) measurements, presents substantial obstacles to effective visual analytics and decision-making. Through a year-long collaboration with experienced neurologists, we developed PhenoFlow, a visual analytics system that leverages the collaboration between human and Large Language Models (LLMs) to analyze the extensive and complex data of acute ischemic stroke patients. PhenoFlow pioneers an innovative workflow, where the LLM serves as a data wrangler while neurologists explore and supervise the output using visualizations and natural language interactions. This approach enables neurologists to focus more on decision-making with reduced cognitive load. To protect sensitive patient information, PhenoFlow only utilizes metadata to make inferences and synthesize executable codes, without accessing raw patient data. This ensures that the results are both reproducible and interpretable while maintaining patient privacy. The system incorporates a slice-and-wrap design that employs temporal folding to create an overlaid circular visualization. Combined with a linear bar graph, this design aids in exploring meaningful patterns within irregularly measured BP data. Through case studies, PhenoFlow has demonstrated its capability to support iterative analysis of extensive clinical datasets, reducing cognitive load and enabling neurologists to make well-informed decisions. Grounded in long-term collaboration with domain experts, our research demonstrates the potential of utilizing LLMs to tackle current challenges in data-driven clinical decision-making for acute ischemic stroke patients.