Medical data science in rhinology: Background and implications for clinicians.

BACKGROUND

An important challenge of big data is using complex information networks to provide useful clinical information. Recently, machine learning, and particularly deep learning, has enabled rapid advances in clinical practice. The application of artificial intelligence (AI) and machine learning (ML) in rhinology is an increasingly relevant topic.

PURPOSE

We review the literature and provide a detailed overview of the recent advances in AI and ML as applied to rhinology. Also, we discuss both the significant benefits of this work as well as the challenges in the implementation and acceptance of these methods for clinical purposes.

METHODS

We aimed to identify and explain published studies on the use of AI and ML in rhinology based on PubMed, Scopus, and Google searches. The search string "nasal OR respiratory AND artificial intelligence OR machine learning" was used. Most of the studies covered areas of paranasal sinuses radiology, including allergic rhinitis, chronic rhinitis, computed tomography scans, and nasal cytology.

RESULTS

Cluster analysis and convolutional neural networks (CNNs) were mainly used in studies related to rhinology. AI is increasingly affecting healthcare research, and ML technology has been used in studies of chronic rhinitis and allergic rhinitis, providing some exciting new research modalities.

CONCLUSION

AI is especially useful when there is no conclusive evidence to aid decision making. ML can help doctors make clinical decisions, but it does not entirely replace doctors. However, when critically evaluating studies using this technique, rhinologists must take into account the limitations of its applications and use.