Customer Cases

Industry

Medical

Product

DeepFlow

Copied to clipboard

MLOps speeds medical AI adoption

Case Description

Sepsis is an acute condition that progresses extremely quickly, and the danger is that early symptoms can be easily ignored, and delays in treatment can lead to fatal consequences. Being able to assist medical teams in rapid diagnosis and providing precise treatment at an early stage can not only save more lives, but also significantly reduce mortality.

From infection to worsening of sepsis, there are often only three short stages. According to statistics from Taiwan’s Ministry of Health and Welfare in 2021, sepsis has jumped to the 14th leading cause of death in Taiwan. When a patient develops symptoms of suspected infection, the medical team not only needs to quickly assess the condition, but also needs to consider more complex disease states, such as severe dyspnea, abnormal blood pressure, changes in consciousness, etc.

How to quickly and accurately diagnose patients with sepsis has become a key challenge for current medical teams, and finding a solution will have a significant impact on patients' chances of survival.

Solution

Use the “AutoML” module to quickly generate a sepsis model with an optimal algorithm and deploy it to the API Server to generate a sepsis prediction service to predict the probability of suffering from sepsis and improve the accuracy of interpretation.
Monitor the patient’s vital signs such as blood pressure, blood oxygen, and body temperature, and connect it through the sepsis warning APP to obtain the sepsis prediction results and detect and discover potential sepsis cases early.

Benefit

This allows doctors to focus on medical data analysis and professional applications without writing code, and quickly complete AI model training.
DeepFlow helps record relevant model history. In addition to improving overall efficiency, it also helps model management and monitoring, while taking into account the quantity and quality of the model.