Oncology

A new artificial intelligence system significantly improves MRI-based breast cancer diagnosis, reducing false-positive rates and inter-reader variability while enhancing risk stratification of BI-RADS 4 lesions.

Elevated copper levels in tumour tissues, along with lymph node status and TNM stage, emerged as important discriminators of disease characteristics, highlighting potential links between metal imbalances and CRC progression. 

A comparative study shows that lightweight convolutional neural networks can accurately classify lung cancer histopathology with strong generalisation and reduced computational complexity.

Deep learning integration of FDG PET-CT and clinical data significantly improves survival prediction in non-metastatic breast cancer, offering a transparent and more precise approach to prognostic stratification.

Sarcopenia, myosteatosis, and systemic immune inflammatory index were independently associated with reduced oesophageal cancer survival in patients treated with immunotherapy, supporting their role in prognostic assessment.

A large English registry analysis shows survival in multiple myeloma declines with each treatment line and is significantly longer in patients receiving autologous stem cell transplant.

Machine learning models accurately predict survival after surgery for upper tract urothelial cancer, supporting personalised follow up and adjuvant treatment decisions.

Low baseline ctDNA and early clearance during treatment are strongly associated with improved response and survival in advanced breast cancer, highlighting ctDNA as a promising predictive and surrogate biomarker.