Kikuchi Yoshitomo, Togao Osamu, Kikuchi Kazufumi, Momosaka Daichi, Obara Makoto, Van Cauteren Marc, Fischer Alexander, Ishigami Kousei, Hiwatashi Akio
Department of Clinical Radiology, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka, 812-8582, Japan.
Department of Molecular Imaging and Diagnosis, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan.
Eur Radiol. 2022 May;32(5):2998-3005. doi: 10.1007/s00330-021-08427-2. Epub 2022 Jan 7.
To develop an automated model to detect brain metastases using a convolutional neural network (CNN) and volume isotropic simultaneous interleaved bright-blood and black-blood examination (VISIBLE) and to compare its diagnostic performance with the observer test.
This retrospective study included patients with clinical suspicion of brain metastases imaged with VISIBLE from March 2016 to July 2019 to create a model. Images with and without blood vessel suppression were used for training an existing CNN (DeepMedic). Diagnostic performance was evaluated using sensitivity and false-positive results per case (FPs/case). We compared the diagnostic performance of the CNN model with that of the twelve radiologists.
Fifty patients (30 males and 20 females; age range 29-86 years; mean 63.3 ± 12.8 years; a total of 165 metastases) who were clinically diagnosed with brain metastasis on follow-up were used for the training. The sensitivity of our model was 91.7%, which was higher than that of the observer test (mean ± standard deviation; 88.7 ± 3.7%). The number of FPs/case in our model was 1.5, which was greater than that by the observer test (0.17 ± 0.09).
Compared to radiologists, our model created by VISIBLE and CNN to diagnose brain metastases showed higher sensitivity. The number of FPs/case by our model was greater than that by the observer test of radiologists; however, it was less than that in most of the previous studies with deep learning.
• Our convolutional neural network based on bright-blood and black-blood examination to diagnose brain metastases showed a higher sensitivity than that by the observer test. • The number of false-positives/case by our model was greater than that by the previous observer test; however, it was less than those from most previous studies. • In our model, false-positives were found in the vessels, choroid plexus, and image noise or unknown causes.
利用卷积神经网络（CNN）和容积各向同性同步交错亮血与黑血检查（VISIBLE）开发一种自动检测脑转移瘤的模型，并将其诊断性能与观察者测试进行比较。
这项回顾性研究纳入了2016年3月至2019年7月期间因临床怀疑脑转移瘤而接受VISIBLE成像的患者，以创建一个模型。使用有和没有血管抑制的图像来训练现有的CNN（DeepMedic）。使用每例的敏感性和假阳性结果（FPs/例）评估诊断性能。我们将CNN模型的诊断性能与12名放射科医生的诊断性能进行了比较。
50例患者（30例男性和20例女性；年龄范围29 - 86岁；平均63.3±12.8岁；共有165个转移瘤）在随访中临床诊断为脑转移瘤，用于训练。我们模型的敏感性为91.7%，高于观察者测试（平均值±标准差；88.7±3.7%）。我们模型中每例的FPs数量为1.5，大于观察者测试的结果（0.17±0.09）。
与放射科医生相比，我们通过VISIBLE和CNN创建的用于诊断脑转移瘤的模型显示出更高的敏感性。我们模型每例的FPs数量大于放射科医生的观察者测试结果；然而，它比大多数先前的深度学习研究中的结果要少。
• 我们基于亮血与黑血检查的卷积神经网络诊断脑转移瘤的敏感性高于观察者测试。• 我们模型每例的假阳性数量大于先前的观察者测试结果；然而，它比大多数先前研究中的结果要少。• 在我们的模型中，假阳性出现在血管、脉络丛以及图像噪声或不明原因处。
