Applications and challenges of neural networks in otolaryngology (Review)

Iulian-Alexandru Taciuc ¹, Mihai Dumitru ², Daniela Vrinceanu ²

⁰Department of Pathology, 'Carol Davila' University of Medicine and Pharmacy, 020021 Bucharest, Romania.

Biomedical Reports +

|

May 20, 2024

View abstract on PubMed

Summary

Artificial Intelligence (AI), specifically neural networks (NNs), offers potential in healthcare. This review explores NN applications in ear, nose, and throat research, highlighting current limitations and future directions for improved AI integration.

Area Of Science

Medical research
Computer science
Otolaryngology

Background

Artificial Intelligence (AI) is increasingly relevant across research fields, including medicine.
Key questions persist regarding AI's practical integration into daily medical routines.
Neural networks (NNs) represent a significant area of AI development with potential medical applications.

Purpose Of The Study

To review available neural network (NN) types for AI development.
To discuss the advantages, disadvantages, and practical applications of NNs.
To summarize NN applications in ear, nose, and throat (ENT) research, from diagnosis to treatment.

Main Methods

Review of existing literature on neural networks (NNs) and their application in medical research.
Focus on studies utilizing NNs in the ear, nose, and throat (ENT) research field.
Analysis of reported accuracies and limitations of AI models in these studies.

Main Results

Neural networks, particularly convolutional NNs, have shown promising diagnostic and treatment management capabilities in ENT studies.
Reported accuracies in AI-assisted ENT studies range from 70% to 98%.
Many studies trained AI models on limited patient cohorts (<100 patients), indicating a need for larger datasets.

Conclusions

Understanding NN types is crucial for advancing AI in medicine.
Combining multiple NN types may overcome current AI limitations in accuracy and data relevance.
Lack of standardized AI research protocols hinders data homogeneity and transparency, impacting AI development in ENT and other medical fields.

Keywords:

artificial intelligence head and neck cancer neural networks otorhinolaryngology