EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2284. Diagnostic advancements: machine learning in identifying nontuberculous mycobacteria (NTM) in non-cystic fibrosis bronchiectasis (NCFBr) population

Invited abstract in session TB-6: Advancements of OR-Analytics in Statistics, Machine Learning and Data Science 13, stream Advancements of OR-analytics in statistics, machine learning and data science.

Tuesday, 10:30-12:00
Room: 1013 (building: 202)

Authors (first author is the speaker)

1. Chen Hajaj
Ariel university

Abstract


In the realm of pulmonary health, diagnosing Nontuberculous mycobacteria (NTM) infections presents a formidable challenge, particularly in patients with non-cystic fibrosis bronchiectasis (NCFBr). These infections are notoriously elusive, with their clinical presentations and radiological findings mirroring a plethora of other pulmonary conditions, thereby complicating accurate diagnosis. Our study embarked on an ambitious journey to harness the potential of machine learning in predicting NTM infections within this patient demographic. Through a comprehensive retrospective analysis of 771 NCFBr patients, our investigation illuminated the presence of NTM in 12.2% of cases, as confirmed by bronchoscopy cultures. Our findings advocate a paradigm shift towards incorporating machine learning as an integral component of the diagnostic toolkit for NTM infections in NCFBr patients
Delving into the intricacies of machine learning, we meticulously evaluated seven distinct models, with the Naive Bayes and Random Forest models emerging as frontrunners, boasting AUCs of 0.81 and 0.71, respectively. This comparative analysis not only underscored the viability of machine learning in the clinical prediction landscape but also highlighted the critical role of radiological and clinical parameters—most notably, patient age and the presence of GERD—in refining predictive accuracy.

Keywords

Status: accepted

Back to the list of papers