Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Representation Learning-Based Dynamic Trajectory Phenotyping for Acute Respiratory Failure in Medical Intensive Care Units

Alan Wu, Tilendra Choudhary, Pulakesh Upadhyaya, Ayman Ali, Philip Yang, Rishikesan Kamaleswaran

TL;DR

This study addresses the heterogeneity of sepsis-associated ARF by applying a deep representation learning–based trajectory clustering (CRLI) to multivariate time-series data from EMRs collected up to 24 hours before invasive ventilation. The approach identifies four distinct phenotypes—liver dysfunction/heterogeneous, hypercapnia, hypoxemia, and MODS—each with unique mortality and comorbidity patterns, validated by silhouette analysis and Kaplan–Meier survival differences (p < 0.005). The work demonstrates that data-driven trajectory phenotyping can uncover clinically relevant subgroups beyond traditional ARF categorizations, with potential to guide prognosis and personalized treatment in ICUs. External validation and prospective testing are needed to confirm generalizability and to integrate these phenotypes into real-time clinical decision support.

Abstract

Sepsis-induced acute respiratory failure (ARF) is a serious complication with a poor prognosis. This paper presents a deep representation learningbased phenotyping method to identify distinct groups of clinical trajectories of septic patients with ARF. For this retrospective study, we created a dataset from electronic medical records (EMR) consisting of data from sepsis patients admitted to medical intensive care units who required at least 24 hours of invasive mechanical ventilation at a quarternary care academic hospital in southeast USA for the years 2016-2021. A total of N=3349 patient encounters were included in this study. Clustering Representation Learning on Incomplete Time Series Data (CRLI) algorithm was applied to a parsimonious set of EMR variables in this data set. To validate the optimal number of clusters, the K-means algorithm was used in conjunction with dynamic time warping. Our model yielded four distinct patient phenotypes that were characterized as liver dysfunction/heterogeneous, hypercapnia, hypoxemia, and multiple organ dysfunction syndrome by a critical care expert. A Kaplan-Meier analysis to compare the 28-day mortality trends exhibited significant differences (p < 0.005) between the four phenotypes. The study demonstrates the utility of our deep representation learning-based approach in unraveling phenotypes that reflect the heterogeneity in sepsis-induced ARF in terms of different mortality outcomes and severity. These phenotypes might reveal important clinical insights into an effective prognosis and tailored treatment strategies.

Deep Representation Learning-Based Dynamic Trajectory Phenotyping for Acute Respiratory Failure in Medical Intensive Care Units

TL;DR

This study addresses the heterogeneity of sepsis-associated ARF by applying a deep representation learning–based trajectory clustering (CRLI) to multivariate time-series data from EMRs collected up to 24 hours before invasive ventilation. The approach identifies four distinct phenotypes—liver dysfunction/heterogeneous, hypercapnia, hypoxemia, and MODS—each with unique mortality and comorbidity patterns, validated by silhouette analysis and Kaplan–Meier survival differences (p < 0.005). The work demonstrates that data-driven trajectory phenotyping can uncover clinically relevant subgroups beyond traditional ARF categorizations, with potential to guide prognosis and personalized treatment in ICUs. External validation and prospective testing are needed to confirm generalizability and to integrate these phenotypes into real-time clinical decision support.

Abstract

Sepsis-induced acute respiratory failure (ARF) is a serious complication with a poor prognosis. This paper presents a deep representation learningbased phenotyping method to identify distinct groups of clinical trajectories of septic patients with ARF. For this retrospective study, we created a dataset from electronic medical records (EMR) consisting of data from sepsis patients admitted to medical intensive care units who required at least 24 hours of invasive mechanical ventilation at a quarternary care academic hospital in southeast USA for the years 2016-2021. A total of N=3349 patient encounters were included in this study. Clustering Representation Learning on Incomplete Time Series Data (CRLI) algorithm was applied to a parsimonious set of EMR variables in this data set. To validate the optimal number of clusters, the K-means algorithm was used in conjunction with dynamic time warping. Our model yielded four distinct patient phenotypes that were characterized as liver dysfunction/heterogeneous, hypercapnia, hypoxemia, and multiple organ dysfunction syndrome by a critical care expert. A Kaplan-Meier analysis to compare the 28-day mortality trends exhibited significant differences (p < 0.005) between the four phenotypes. The study demonstrates the utility of our deep representation learning-based approach in unraveling phenotypes that reflect the heterogeneity in sepsis-induced ARF in terms of different mortality outcomes and severity. These phenotypes might reveal important clinical insights into an effective prognosis and tailored treatment strategies.
Paper Structure (17 sections, 2 equations, 5 figures, 4 tables)

This paper contains 17 sections, 2 equations, 5 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Data Collection
  4. Data Preprocessing
  5. Clustering Algorithms
  6. Clustering Representation Learning on Incomplete time-series data (CRLI)
  7. K-means and Dynamic Time Warping (DTW)
  8. Optimal Clustering and Phenotype Characterization
  9. Statistical Analysis
  10. Results
  11. Demographics
  12. Silhouette Score Analysis
  13. Phenotyping Results
  14. Kaplan-Meier Analysis
  15. Comorbidity Analysis
  16. ...and 2 more sections

Figures (5)

  • Figure 1: Block diagram depicting an overview of our study design for sepsis-induced ARF trajectory phenotyping.
  • Figure 2: UMAP 2-D representation of clustering latent from CRLI, where identified phenotypes are highlighted in different color codes.
  • Figure 3: Kaplan-Meier curves on short-term mortality, showing survival probabilities for different phenotypes. A multivariate log-rank test on the survival curves was performed with $p<0.005$.
  • Figure 4: Trajectories of CRLI phenotypes over the 24-hour prior to ventilation with 83.4% confidence intervals (dashed lines).
  • Figure :