在MRI上进行前列腺癌检测的少数镜头学习:与放射科医生的性能进行比较分析
Yosuke Yamagishi1,2, Yasutaka Baba3, Jun Suzuki3
1Department of Diagnostic Radiology, Saitama Medical University International Medical Center, Saitama Medical University International Medical Center, Hidaka, Japan. yamagishi-yosuke0115@g.ecc.u-tokyo.ac.jp.
Journal of imaging informatics in medicine
|June 25, 2025
概括
少数拍摄的深度学习模型显示,使用最小的MRI数据来检测前列腺癌具有前景. 这些模型实现了与放射科医生可比的性能,解决了临床环境中的数据限制和领域转移问题.
科学领域:
- 医疗成像医学成像
- 人工智能的人工智能
- 在瘤学瘤学.
背景情况:
- 对于前列腺癌检测而言,深度学习需要大量的数据集,而这些数据集往往因跨机构的领域转移问题而受到阻碍.
- 数据的有限可用性限制了当前深度学习模型的临床适用性.
研究的目的:
- 在多参数核磁共振 (MRI) 上开发一个用于前列腺癌检测的几次学习深度学习模型.
- 为了评估这个模型的诊断性能,与经验丰富的放射科医生进行对比.
- 解决人工智能驱动的癌症检测中的数据稀缺和领域转移挑战.
主要方法:
- 一个2D变压器模型被训练在T2加权,扩散加权和99个活检确认的前列腺癌病例的明显扩散系数 (ADC) 地图图像上.
- 该模型使用了几次射击的学习方法,训练数据最小 (20个案例).
- 使用马修斯相关系数 (MCC) 和F1得分来评估性能,与两个放射科医生进行比较,并在前列腺158数据集上进行外部验证.
主要成果:
- 几次拍摄的模型实现了0.297的MCC和0.707的F1得分,相当于放射科医生1 (MCC:0.276,F1:0.741).
- 放射科医生2的表现优于模型 (MCC:0.504,F1:0.871).
- ImageNet预训练显著改善了模型性能,研究水平ROC-AUC从0.464增加到0.636和PR-AUC从0.637增加到0.773.
结论:
- 少数拍摄的深度学习模型,特别是预训练过的变压器架构,可以实现前列腺癌检测的临床相关性能.
- 这种方法提供了一种可行的解决方案,以克服数据限制和多机构环境中的域移动问题.
- 这些发现支持人工智能的潜力,以提高前列腺癌成像诊断能力.
相关概念视频
Imaging Studies IV: Magnetic Resonance Imaging
60
Introduction:Magnetic Resonance Imaging, or MRI, can include a specialized imaging technique of the urinary system known as Magnetic Resonance Urography (MRU). This radiation-free technique uses strong magnetic fields and radio waves to produce detailed images with the help of a computer. MRU is particularly effective for visualizing fluid-filled structures like the kidneys, ureters, and bladder.Applications of MRI in the Genitourinary SystemKidneys and Ureters: MRI detects tumors, cysts,...
60
Magnetic Resonance Imaging
7.3K
Magnetic resonance imaging (MRI) is a noninvasive medical imaging technique based on a phenomenon of nuclear physics discovered in the 1930s, in which matter exposed to magnetic fields and radio waves was found to emit radio signals. In 1970, a physician and researcher named Raymond Damadian noticed that malignant (cancerous) tissue gave off different signals than normal body tissue. He applied for a patent for the first MRI scanning device in clinical use by the early 1980s. The early MRI...
7.3K


