Computers in Biology and Medicine, Journal Year: 2025, Volume and Issue: 194, P. 110372 - 110372

Published: June 3, 2025

Artificial Intelligence is transforming medical imaging, particularly in the analysis of bone metastases (BM), a serious complication advanced cancers. Machine learning and deep techniques offer new opportunities to improve detection, recognition, segmentation metastasis. Yet, challenges such as limited data, interpretability, clinical validation remain. Following PRISMA guidelines, we reviewed artificial intelligence methods applications for metastasis across major imaging modalities including CT, MRI, PET, SPECT, scintigraphy. The survey includes traditional machine models modern architectures CNNs transformers. We also examined available datasets their effect developing this field. have achieved strong performance tasks modalities, with Convolutional Neural Network (CNN) Transformer showing efficient different tasks. However, limitations persist, data imbalance, overfitting risks, need greater transparency. Clinical translation challenged by regulatory hurdles. holds potential BM diagnosis streamline radiology workflows. To reach maturity, future work must address diversity, model explainability, large-scale validation, which are critical steps being trusted be integrated into oncology care routines.

Language: Английский