Computer generated transcript

Warning!
The following transcript was generated automatically from the content and has not been checked or corrected manually.

Artificial Intelligence in Breast Cancer Imaging for Dense Breast Tissue Aphrodite P . Pascoe (Year 3, European University Cyprus) ABSTR ACT AIMS Dense breast tissue poses a significant challenge in breast cancer To explore how artificial intelligence (AI) enhances the detection of screening due to its ability to obscure tumours on mammograms, leading breast cancer in women with dense breast tissue. to delayed diagnoses and reduced survival rates. Traditional imaging methods, while essential, are often insufficient for accurate detection in To evaluate the effectiveness of key AI models (e.g. DeepMind, these cases. This poster explores how artificial intelligence (AI) is Transpara, ProFound AI) in improving imaging outcomes. transforming diagnostic imaging by enhancing pattern recognition, reducing both false positives and false negatives, and supporting radiologists through advanced computational analysis. AI-integrated To identify ethical and practical considerations in integrating AI into systems such as DeepMind, Transpara, and ProFound AI demonstrate clinical breast cancer screening workflows. improved sensitivity in detecting malignancies within dense tissue, without increasing clinical workload. Furthermore, AI offers a pathway to more personalised screening strategies by factoring in individual risk profiles. Despite these advancements, issues around data bias, consent, and METHODS integration remain key barriers to widespread adoption. This review highlights both the promise and the limitations of AI as a tool to improve This review synthesises information from peer-reviewed scientific early detection and outcomes for women with dense breast tissue. literature, official clinical trial data, and published reports on AI tools used in breast imaging. AI model analysis was based on four key stages: Data Input – Imaging data (e.g. mammograms, MRIs) are fed INTRODUCTION into the system. Feature Extraction – The AI breaks down images into Breast cancer is the most common cancer affecting women globally . mathematical features like texture, margins, and density . Early detection is key to improving survival rates. Dense breast tissue is present in ~40% of women: Model Training – The system learns from millions of annotated images (benign vs malignant). Contains more fibroglandular than fatty tissue. Prediction Output – AI assigns malignancy risk scores and Both dense tissue and tumours appear white on mammograms. Creates a “masking effect” → harder to detect cancer. highlights suspicious areas for radiologist review. Standard imaging (mammography, ultrasound, MRI) has reduced Case studies and comparative performance metrics were assessed to determine diagnostic accuracy and real-world applicability. accuracy in dense tissue cases. This leads to: Missed diagnoses (false negatives) Unnecessary biopsies (false positives) RESULTS Artificial intelligence (AI) offers a new approach: 1) AI Enhances Detection in Dense Tissue Enhances image interpretation. AI improves tumour visibility by analysing pixel-level features Supports radiologists. undetectable to the human eye. Increases detection accuracy, especially in dense tissue. In clinical trials, AI-assisted mammography increased detection This poster examines how AI improves imaging and the rates of early-stage cancers in women with dense breasts. ethical/practical issues surrounding its use. 2) Reduces False Positives and Negatives Deep learning models distinguish subtle differences between benign and malignant lesions. This leads to fewer unnecessary biopsies and less diagnostic delay. 3) Radiologist Support and Workflow Efficiency AI systems (e.g. Transpara, ProFound AI) act as second readers, increasing diagnostic confidence. AI integration reduces reading time while maintaining or improving accuracy. 4) Real-World Implementation Google DeepMind: Trained on thousands of mammograms, now used in research settings global.y Transpara: CE-marked, used in 40+ countries. ProFound AI: Shown to cut radiologist workload while preserving diagnostic qualit. CONCLUSION Artificial intelligence is transforming breast cancer imaging, particularly for women with dense breast tissue—a population traditionally underserved by standard screening methods. By improving image analysis, reducing diagnostic errors, and supporting radiologists, AI enhances early detection and personalises screening protocols. Models like DeepMind, Transpara, and ProFound AI have demonstrated improved sensitivity without increasing clinical burden. However, successful implementation requires attention to bias in training data, patient consent, and integration into clinical workflows. AI is not a replacement for radiologists but a tool to augment their expertise. As adoption grows, AI has the potential to deliver more accurate, equitable, and efficient breast cancer screening—especially in challenging cases involving dense tissue. REFERENCES & ARTICLES Figure 1 ] Kettering Health, 2025 CAMeSM 2025