Related Experiment Video
Updated: Aug 1, 2025
A Data-Driven Approach to Quantifying Immune States in Sepsis
Published on: February 7, 2025
Differentiation of Fungal, Viral, and Bacterial Sepsis using Multimodal Deep Learning
Aaron Boussina1, Karthik Ramesh1, Himanshu Arora1
1Division of Biomedical Informatics, University of California San Diego, San Diego, California, USA.
Abstract:
Sepsis is a major cause of morbidity and mortality worldwide, and is caused by bacterial infection in a majority of cases. However, fungal sepsis often carries a higher mortality rate both due to its prevalence in immunocompromised patients as well as delayed recognition. Using chest x-rays, associated radiology reports, and structured patient data from the MIMIC-IV clinical dataset, the authors present a machine learning methodology to differentiate between bacterial, fungal, and viral sepsis. Model performance shows AUCs of 0.81, 0.83, 0.79 for detecting bacterial, fungal, and viral sepsis respectively, with best performance achieved using embeddings from image reports and structured clinical data. By improving early detection of an often missed causative septic agent, predictive models could facilitate earlier treatment of non-bacterial sepsis with resultant associated mortality reduction.
Insights
This study developed a machine learning model to differentiate bacterial, fungal, and viral sepsis using chest X-rays and patient data. Early detection of fungal sepsis can reduce mortality rates in vulnerable patients.
Area of Science:
- Medical informatics
- Machine learning in healthcare
- Infectious disease diagnostics
Background:
- Sepsis is a leading cause of death globally, frequently caused by bacterial infections.
- Fungal sepsis, though less common, has higher mortality rates due to delayed diagnosis and prevalence in immunocompromised individuals.
- Accurate differentiation of sepsis types is crucial for effective treatment and improved patient outcomes.
Approach:
- Utilized the MIMIC-IV clinical dataset, including chest X-rays, radiology reports, and structured patient data.
- Developed a machine learning methodology to distinguish between bacterial, fungal, and viral sepsis.
- Evaluated model performance using Area Under the Curve (AUC) metrics for each sepsis type.
Key Points:
- The machine learning model achieved AUCs of 0.81 for bacterial, 0.83 for fungal, and 0.79 for viral sepsis.
- Optimal performance was obtained by integrating embeddings from radiology reports and structured clinical data.
- The approach highlights the potential of AI in identifying less common but high-mortality sepsis pathogens.
Conclusions:
- Machine learning models can effectively differentiate between bacterial, fungal, and viral sepsis.
- Improved early detection of fungal sepsis through AI can lead to timely, targeted treatments.
- This predictive capability holds promise for reducing sepsis-associated mortality, particularly in high-risk populations.
