基于置信度和基于体积的域自适应的逐步自训练，用于使用 MRI 图像上的非局部网络进行多数据集深度学习脑转移检测。

Gradual Self-Training via Confidence and Volume Based Domain Adaptation for Multi Dataset Deep Learning-Based Brain Metastases Detection Using Nonlocal Networks on MRI Images.

机构信息

Electrical and Computer Systems Engineering Discipline, School of Engineering, Monash University Malaysia, Bandar Sunway, Malaysia.

Radiology Department, Sunway Medical Centre, Bandar Sunway, Malaysia.

出版信息

J Magn Reson Imaging. 2023 Jun;57(6):1728-1740. doi: 10.1002/jmri.28456. Epub 2022 Oct 8.

DOI:10.1002/jmri.28456

PMID:36208095

Abstract

BACKGROUND

Research suggests that treatment of multiple brain metastases (BMs) with stereotactic radiosurgery shows improvement when metastases are detected early, providing a case for BM detection capabilities on small lesions.

PURPOSE

To demonstrate automatic detection of BM on three MRI datasets using a deep learning-based approach. To improve the performance of the network is iteratively co-trained with datasets from different domains. A systematic approach is proposed to prevent catastrophic forgetting during co-training.

STUDY TYPE

Retrospective.

POPULATION

A total of 156 patients (105 ground truth and 51 pseudo labels) with 1502 BM (BrainMetShare); 121 patients with 722 BM (local); 400 patients with 447 primary gliomas (BrATS). Training/pseudo labels/validation data were distributed 84/51/21 (BrainMetShare). Training/validation data were split: 121/23 (local) and 375/25 (BrATS).

FIELD STRENGTH/SEQUENCE: A 5 T and 3 T/T1 spin-echo postcontrast (T1-gradient echo) (BrainMetShare), 3 T/T1 magnetization prepared rapid acquisition gradient echo postcontrast (T1-MPRAGE) (local), 0.5 T, 1 T, and 1.16 T/T1-weighted-fluid-attenuated inversion recovery (T1-FLAIR) (BrATS).

ASSESSMENT

The ground truth was manually segmented by two (BrainMetShare) and four (BrATS) radiologists and manually annotated by one (local) radiologist. Confidence and volume based domain adaptation (CAVEAT) method of co-training the three datasets on a 3D nonlocal convolutional neural network (CNN) architecture was implemented to detect BM.

STATISTICAL TESTS

The performance was evaluated using sensitivity and false positive rates per patient (FP/patient) and free receiver operating characteristic (FROC) analysis at seven predefined (1/8, 1/4, 1/2, 1, 2, 4, and 8) FPs per scan.

RESULTS

The sensitivity and FP/patient from a held-out set registered 0.811 at 2.952 FP/patient (BrainMetShare), 0.74 at 3.130 (local), and 0.723 at 2.240 (BrATS) using the CAVEAT approach with lesions as small as 1 mm being detected.

DATA CONCLUSION

Improved sensitivities at lower FP can be achieved by co-training datasets via the CAVEAT paradigm to address the problem of data sparsity.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY STAGE: 2.

摘要

背景

研究表明，对于多发性脑转移瘤（BM）的治疗，采用立体定向放射外科治疗时，若能早期发现转移瘤，则效果会有所改善，这就为在小病灶上检测 BM 提供了可能。

目的

使用基于深度学习的方法，在三个 MRI 数据集上演示 BM 的自动检测。通过迭代式的与不同领域的数据集共同训练，以提高网络的性能。提出了一种系统的方法来防止共同训练期间的灾难性遗忘。

研究类型

回顾性。

人群

共有 156 名患者（105 名有明确诊断，51 名带有伪标签），共 1502 个 BM（BrainMetShare）；121 名患者，722 个 BM（local）；400 名患者，447 个原发性脑肿瘤（BrATS）。训练/伪标签/验证数据的分布为 84/51/21（BrainMetShare）。训练/验证数据的分割：121/23（local）和 375/25（BrATS）。

场强/序列：5T 和 3T/T1 自旋回波对比后（T1-梯度回波）（BrainMetShare），3T/T1 磁化准备快速获取梯度回波对比后（T1-MPRAGE）（local），0.5T，1T 和 1.16T/T1 加权液体衰减反转恢复（T1-FLAIR）（BrATS）。

评估

ground truth 由两位（BrainMetShare）和四位（BrATS）放射科医生手动分割，由一位（local）放射科医生手动标注。使用置信度和基于体积的领域自适应（CAVEAT）方法，在一个 3D 非局部卷积神经网络（CNN）架构上对三个数据集进行共同训练，以检测 BM。

统计学检验

使用灵敏度和每位患者的假阳性率（FP/patient）以及七个预定义（1/8、1/4、1/2、1、2、4 和 8）扫描 FP 下的免费接收器操作特征（FROC）分析，对预留数据集进行评估。

结果

使用 CAVEAT 方法，在 BrainMetShare 中，0.811 的灵敏度和 2.952 FP/patient 的 FP/patient，在 local 中为 0.74 和 3.130 FP/patient，在 BrATS 中为 0.723 和 2.240 FP/patient，可检测到 1mm 大小的病灶。