基于多通道三维卷积神经网络的多参数MR成像前列腺癌半自动分类
研究纳入200例患者共318个有组织学对照的病灶,设计并训练三维卷积神经网络,以不同MRI序列组合为输入进行前列腺癌半自动分类。ADC、DWI与K-trans组合平均AUC为0.897,检测活检证实的显著前列腺癌AUC范围0.89至0.91。
要点
- ADC、DWI与K-trans组合的平均AUC为0.897,检测显著前列腺癌的AUC范围为0.89至0.91。
- 模型可直接处理MR序列而无需预先勾画病灶。
- 灌注MR图像K-trans对网络整体性能影响最大,T2w图像几乎不改善性能。
核心结论
- 三维卷积神经网络检测临床显著前列腺癌的AUC良好、敏感度良好且特异度高。
- 前列腺癌分类性能与使用PI-RADS的有经验放射科医师所报告的性能相当。
- 病灶大小和最大直径对网络性能无影响。
研究设计与样本
- 研究为基于深度学习的半自动前列腺癌分类研究,采用三维卷积神经网络。
- 共分析200例患者、318个有组织学对照的病灶。
- 以活检结果作为检测显著前列腺癌的参考标准。
- 网络使用不同MRI序列组合作为输入进行设计、训练和验证,包括T2加权、表观扩散系数(ADC)、扩散加权图像和K-trans。
- 模型采用八折交叉验证进行训练和验证。
诊断标准与关键数值
- ADC、DWI与K-trans输入组合的平均AUC为0.897,AUC范围为0.89(敏感度88.6%、特异度90.0%)至0.91(敏感度81.2%、特异度90.5%)。
- 使用T2w与K-trans时性能差异显著,P值为0.02;使用T2w、ADC与DWI时P值为0.00025。
- 其他输入组合在整体性能和平均AUC方面得分较低。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVE: To present a deep learning-based approach for semi-automatic prostate cancer classification based on multi-parametric magnetic resonance (MR) imaging using a 3D convolutional neural network (CNN).
METHODS: Two hundred patients with a total of 318 lesions for which histological correlation was available were analyzed. A novel CNN was designed, trained, and validated using different combinations of distinct MRI sequences as input (e.g., T2-weighted, apparent diffusion coefficient (ADC), diffusion-weighted images, and K-trans) and the effect of different sequences on the network's performance was tested and discussed. The particular choice of modeling approach was justified by testing all relevant data combinations. The model was trained and validated using eightfold cross-validation.
RESULTS: In terms of detection of significant prostate cancer defined by biopsy results as the reference standard, the 3D CNN achieved an area under the curve (AUC) of the receiver operating characteristics ranging from 0.89 (88.6% and 90.0% for sensitivity and specificity respectively) to 0.91 (81.2% and 90.5% for sensitivity and specificity respectively) with an average AUC of 0.897 for the ADC, DWI, and K-trans input combination. The other combinations scored less in terms of overall performance and average AUC, where the difference in performance was significant with a p value of 0.02 when using T2w and K-trans; and 0.00025 when using T2w, ADC, and DWI. Prostate cancer classification performance is thus comparable to that reported for experienced radiologists using the prostate imaging reporting and data system (PI-RADS). Lesion size and largest diameter had no effect on the network's performance.
CONCLUSION: The diagnostic performance of the 3D CNN in detecting clinically significant prostate cancer is characterized by a good AUC and sensitivity and high specificity.
KEY POINTS: • Prostate cancer classification using a deep learning model is feasible and it allows direct processing of MR sequences without prior lesion segmentation. • Prostate cancer classification performance as measured by AUC is comparable to that of an experienced radiologist. • Perfusion MR images (K-trans), followed by DWI and ADC, have the highest effect on the overall performance; whereas T2w images show hardly any improvement.
引用
- Aldoj N, Lukas S, Dewey M, et al. Semi-automatic classification of prostate cancer on multi-parametric MR imaging using a multi-channel 3D convolutional neural network. Eur Radiol. 2019. PMID: 31468158. doi:10.1007/s00330-019-06417-z
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。