Using Artificial Intelligence in Fungal Lung Disease: CPA CT Imaging as an Example

Area of Science:

• Medical imaging
• Artificial intelligence
• Pulmonology

Background:

• Fungal lung diseases, particularly chronic pulmonary aspergillosis, present diagnostic challenges.
• Current methods for assessing treatment response in these infections have high uncertainty.
• Artificial intelligence (AI) shows potential to improve diagnosis and treatment monitoring.

Purpose of the Study:

• To discuss the challenges and opportunities of AI in fungal lung disease.
• To highlight the potential of AI in chronic pulmonary aspergillosis using lung imaging.
• To propose recommendations for advancing AI in this field.

Main Methods:

• Review of current challenges in AI for fungal lung disease.
• Discussion of proof-of-concept results using lung imaging and machine learning.
• Formulation of recommendations for AI development.

Main Results:

• AI has the potential to significantly impact the diagnosis and treatment response analysis of fungal lung infections.
• Developing AI for medical imaging, especially in pulmonology, faces specific hurdles.
• Proof-of-concept results demonstrate the feasibility of AI in analyzing lung imaging for fungal diseases.

Conclusions:

• Engaging the medical community is essential for the successful implementation of AI in fungal infection diagnostics.
• Gathering dedicated imaging registries and linking them with non-imaging data are key first steps.
• Harmonizing image-finding annotations is critical for developing robust AI models for fungal lung disease.