第1941期:AI can "safely" read breast cancer screening images

There is growing interest in how AI could transform healthcare. The idea of using it to help with breast cancer screening is one of the major areas of interest.

人们对可以如何运用人工智能来改造医疗保健服务的兴趣日益浓厚。使用人工智能帮助筛查乳腺癌的想法是人们感兴趣的主要领域之一。

Previously, studies have shown benefit, and this latest research, led by a team at Lund University, adds to that by directly comparing the performance of human experts and AI head-to-head for the first time. It found AI could spot cancer at a similar rate to two radiologists and produce the same number of false positives – where a scan is incorrectly diagnosed as abnormal.

此前的研究已显示出人工智能在这方面有用处，而这项由瑞典隆德大学的团队领头的最新研究则通过首次直接对比人类专家与人工智能的表现，进一步证明了这一点。该研究发现，人工智能检查出癌症的效率与两位放射科医生一同评估筛查图像的效率相似，得出假阳性误报的数量也一致，假阳性指检测扫描被误诊为异常。

But the researchers will have to follow the trial participants for another two years to discover whether AI is better than humans and not missing cancers that should be picked up. The lead researcher, Dr Kristina Lang, said the findings certainly suggested AI could potentially have an important role to play.

但是研究人员必须再跟踪试验参与者两年的时间，以发现人工智能是否强过人类，并且不会遗漏本应被发现的癌症病例。首席研究员克里斯蒂娜·朗博士表示，这些研究发现无疑显示出人工智能有可能发挥重要作用。

UK experts say while AI could never replace radiologists, its use offered huge promise, as a radiologist armed with the data insight and accuracy of AI could be a formidable force in patient care.

英国的专家称，尽管人工智能永远无法取代放射科医生，但是它的运用提供了巨大的发展潜力，因为一位具有人工智能带来的数据洞察力和准确性的放射科医生可能会成为患者护理领域的强大力量。

词汇表

screening 检测，筛查
head-to-head 正面交锋，一对一较量
spot 发现，看出
radiologists 放射科医生
false positive 假阳性，实际非阳性但检验结果呈阳性
abnormal 不正常，异常
trial participants 试验参与者，试验对象
picked up 被发现
armed 装备着，具有
data insight 数据洞察，通过分析研究数据而得到的深度理解
formidable 强大的，令人敬畏的