Department of Radiology and Research Institute of Radiological Science and Center for Clinical Imaging Data Science, Yonsei University College of Medicine, 50-1 Yonsei-ro, Seodaemun-gu, Seoul, 120-752, South Korea.
Department of Electrical and Electronic Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul, 03722, South Korea.
Eur Radiol. 2021 Sep;31(9):6686-6695. doi: 10.1007/s00330-021-07783-3. Epub 2021 Mar 18.
To evaluate whether a deep learning (DL) model using both three-dimensional (3D) black-blood (BB) imaging and 3D gradient echo (GRE) imaging may improve the detection and segmentation performance of brain metastases compared to that using only 3D GRE imaging.
A total of 188 patients with brain metastases (917 lesions) who underwent a brain metastasis MRI protocol including contrast-enhanced 3D BB and 3D GRE were included in the training set. DL models based on 3D U-net were constructed. The models were validated in the test set consisting of 45 patients with brain metastases (203 lesions) and 49 patients without brain metastases.
The combined 3D BB and 3D GRE model yielded better performance than the 3D GRE model (sensitivities of 93.1% vs 76.8%, p < 0.001), and this effect was significantly stronger in subgroups with small metastases (p interaction < 0.001). For metastases < 3 mm, ≥ 3 mm and < 10 mm, and ≥ 10 mm, the sensitivities were 82.4%, 93.2%, and 100%, respectively. The combined 3D BB and 3D GRE model showed a false-positive per case of 0.59 in the test set. The combined 3D BB and 3D GRE model showed a Dice coefficient of 0.822, while 3D GRE model showed a lower Dice coefficient of 0.756.
The combined 3D BB and 3D GRE DL model may improve the detection and segmentation performance of brain metastases, especially in detecting small metastases.
• The combined 3D BB and 3D GRE model yielded better performance for the detection of brain metastases than the 3D GRE model (p < 0.001), with sensitivities of 93.1% and 76.8%, respectively. • The combined 3D BB and 3D GRE model showed a false-positive rate per case of 0.59 in the test set. • The combined 3D BB and 3D GRE model showed a Dice coefficient of 0.822, while the 3D GRE model showed a lower Dice coefficient of 0.756.
评估使用三维（3D）黑血（BB）成像和 3D 梯度回波（GRE）成像的深度学习（DL）模型是否可以提高脑转移瘤的检测和分割性能，与仅使用 3D GRE 成像相比。
本研究共纳入 188 例脑转移患者（917 个病灶），这些患者均接受了包括对比增强 3D BB 和 3D GRE 的脑转移 MRI 方案检查。构建了基于 3D U-net 的 DL 模型。该模型在包括 45 例脑转移患者（203 个病灶）和 49 例无脑转移患者的测试集中进行了验证。
联合 3D BB 和 3D GRE 模型的性能优于 3D GRE 模型（敏感度分别为 93.1%和 76.8%，p < 0.001），并且在小转移瘤亚组中这种效果更强（p 交互 < 0.001）。对于 < 3 mm、≥ 3 mm 且 < 10 mm 和 ≥ 10 mm 的转移瘤，敏感度分别为 82.4%、93.2%和 100%。在测试集中，联合 3D BB 和 3D GRE 模型的假阳性率为每个病例 0.59。联合 3D BB 和 3D GRE 模型的 Dice 系数为 0.822，而 3D GRE 模型的 Dice 系数较低，为 0.756。
联合 3D BB 和 3D GRE 的深度学习模型可能会提高脑转移瘤的检测和分割性能，特别是在检测小转移瘤方面。
联合 3D BB 和 3D GRE 模型在检测脑转移瘤方面的性能优于 3D GRE 模型（p < 0.001），敏感度分别为 93.1%和 76.8%。
在测试集中，联合 3D BB 和 3D GRE 模型的假阳性率为每个病例 0.59。
联合 3D BB 和 3D GRE 模型的 Dice 系数为 0.822，而 3D GRE 模型的 Dice 系数较低，为 0.756。
