基于放射组学的人工智能方法在乳腺癌诊断和预测方面的应用价值

Radiomics-Based Artificial Intelligence in Breast Cancer: Application Value in Diagnosis and Prediction

摘要: 乳腺癌是全球女性中最常见的恶性肿瘤，早期发现对降低死亡率至关重要。包括放射影像学检查、临床评估和活检等的传统诊断方法在早期乳腺癌识别中发挥核心作用，但由于影像学筛查的敏感性和阳性预测值不足，同时侵入性组织活检则引发安全性、患者不适及采样偏差等问题的存在使得乳腺癌的早期诊断存在一定的局限性。人工智能与放射组学作为新兴技术，能够从医学影像中提取超越人类视觉感知的高维量化特征。通过整合这些先进计算技术，可构建更精准且可重复的诊断与预测模型。日益增多的证据表明，基于AI和放射组学的方法在提升乳腺癌检测率、预后评估及治疗反应评价方面具有巨大潜力，为推进精准肿瘤学开辟了新路径。

Abstract: Breast cancer is the most commonly diagnosed malignancy among women worldwide, and early detection remains essential for reducing mortality. Conventional diagnostic approaches, including radiological imaging, clinical evaluation, and biopsy, play a central role in early breast cancer identification; however, these methods face notable limitations. Imaging-based screening often demonstrates suboptimal sensitivity and positive predictive value, while invasive tissue biopsy raises concerns regarding safety, patient discomfort, and sampling bias. Artificial intelligence and radiomics have emerged as innovative approaches capable of extracting high-dimensional quantitative features from medical images that exceed human visual perception. By integrating these advanced computational techniques, it is possible to construct more accurate and reproducible diagnostic and predictive models. Growing evidence indicates that AI- and radiomics-based methods hold significant promise in improving breast cancer detection, prognostic assessment, and treatment response evaluation, offering new opportunities to advance precision oncology.