研究显示,在从超声波图像中诊断卵巢癌方面,AI优于人类专家。 AI outperforms human experts in diagnosing ovarian cancer from ultrasound images, study shows.

瑞典Karolinska研究所的研究人员开发了AI模型,这些模型在从超声波图像中识别卵巢癌方面超过了人类专家。 Researchers from Karolinska Institutet in Sweden have developed AI models that outperform human experts in identifying ovarian cancer from ultrasound images. AI使用来自20家医院的病人的17 000多张图像,实现了86.3%的准确率,高于82.6%的专家和非专家的77.7%。 Using over 17,000 images from patients across 20 hospitals, the AI achieved an 86.3% accuracy rate, higher than experts at 82.6% and non-experts at 77.7%. 这将有助于诊断困难案件和在缺乏超声波专家的地区,尽管还需要进一步研究才能充分了解模型的潜力和局限性。 This could help in diagnosing difficult cases and in areas lacking ultrasound experts, though further studies are needed to fully understand the models' potential and limitations.