Artificial-Intelligence-Driven Algorithms for Predicting Response to Corticosteroid Treatment in Patients with
Vojtech Myska1, Samuel Genzor2, Anzhelika Mezina1
1Department of Telecommunications, Faculty of Electrical Engineering and Communications, Brno University of Technology, Technicka 12, 616 00 Brno, Czech Republic.
Predicting corticotherapy benefits for COVID-19 patients with pulmonary fibrosis is crucial. Machine learning models can identify individuals likely to recover with steroid treatment, aiding personalized medicine for post-COVID-19 lung complications.
Area of Science:
- Medical research
- Artificial Intelligence in Medicine
- Pulmonology
Background:
- Pulmonary fibrosis is a severe long-term complication of COVID-19.
- Corticosteroid treatment can improve recovery but carries risks.
- Personalized treatment selection is needed to optimize corticotherapy outcomes.
Purpose of the Study:
- To develop and evaluate machine learning models for predicting patient benefit from corticotherapy post-COVID-19.
- To enable personalized treatment decisions for managing COVID-19-related pulmonary fibrosis.
Main Methods:
- Utilized multiple machine learning algorithms (Logistic Regression, k-NN, Decision Tree, XGBoost, Random Forest, SVM, MLP, AdaBoost, LGBM).
- Trained models on data from 281 patients, including physical exams, blood tests, lung function tests, and imaging (X-ray, HRCT) at baseline and 3 months post-treatment.
- Evaluated model performance using metrics like Balanced Accuracy (BA), ROC-AUC, and F1 score.
Main Results:
- The Decision Tree model demonstrated strong performance with BA of 73.52%, ROC-AUC of 74.69%, and F1 score of 71.70%.
- Random Forest and AdaBoost also showed competitive results, indicating the potential of these algorithms.
- The study confirmed that early post-COVID-19 treatment data can predict corticotherapy response.
Conclusions:
- Machine learning models can effectively predict which patients with COVID-19-related pulmonary fibrosis will benefit from corticotherapy.
- These predictive tools can assist clinicians in making personalized treatment decisions.
- Early intervention and data analysis are key to optimizing patient outcomes in post-COVID-19 care.
Related Concept Videos
COPD: Management Using Bronchodilators and Corticosteroids
Antiasthma Drugs: Inhaled Corticosteroids and Glucocorticoids
ICS work through a multifaceted mechanism of action. They suppress the inflammatory response caused by the proliferation of TH cells. They also reduce the transcription of the IL-2 gene, which is involved in the...
Asthma-IV: Diagnostic and Management
Clinical Assessment for Asthma:
This is the first step in diagnosing and managing asthma. It includes:
Drugs for Treatment of Crohn's Disease in IBD Using Glucocorticoids
Asthma: Pathogenesis and Management
Asthma is classified as allergic and non-allergic. Allergens such as dust mites, pollen, and pet dander trigger allergic asthma, while factors like cold air, intense emotions, or exercise can induce non-allergic asthma.
Asthma-II: Pathophysiology and Classification
Additionally, environmental and genetic factors play crucial roles in determining an individual's susceptibility to asthma and the severity of their condition.
Critical processes in asthma pathophysiology include:


