Utilization of artificial intelligence for tumor segmentation in head and neck cancers: A systematic review and meta-analysis comparing PET with PET/CT imaging techniques - Summary - MDSpire
Advertisement
Utilization of artificial intelligence for tumor segmentation in head and neck cancers: A systematic review and meta-analysis comparing PET with PET/CT imaging techniques
To systematically assess the current literature on PET-only and PET/CT-based AI segmentation of head and neck cancers, highlighting the potential improvements in segmentation accuracy.
Key Findings:
PET/CT fusion imaging enhances tumor segmentation accuracy compared to PET alone, which is crucial for treatment planning.
Deep learning models, particularly CNNs and attention-based architectures, improve tumor delineation, leading to better patient outcomes.
Federated learning frameworks enhance the generalizability of segmentation algorithms, addressing data privacy issues.
Interpretation:
The integration of AI in PET/CT imaging significantly improves the precision of tumor segmentation, which is crucial for effective treatment planning and ultimately enhances patient outcomes in head and neck cancers.
Limitations:
Variability in methodologies, datasets, and evaluation criteria across studies complicates comparisons, potentially skewing results.
Few comprehensive reviews exist that focus specifically on PET and PET/CT modalities, limiting the understanding of their comparative effectiveness.
Conclusion:
AI-driven segmentation techniques, particularly in PET/CT imaging, show promise in improving diagnostic accuracy and treatment planning for head and neck cancers, emphasizing the need for further research in clinical applications.
