Teng Yuen, Ran Xiaoping, Chen Boran, Chen Chaoyue, Xu Jianguo
Department of Neurosurgery, West China Hospital of Sichuan University, Chengdu 610041, China.
Department of Neurosurgery, Ziyang People's Hospital, Ziyang 641300, China.
J Clin Med. 2022 Dec 16;11(24):7481. doi: 10.3390/jcm11247481.
The goal of this study was to develop end-to-end convolutional neural network (CNN) models that can noninvasively discriminate papillary craniopharyngioma (PCP) from adamantinomatous craniopharyngioma (ACP) on MR images requiring no manual segmentation.
A total of 97 patients diagnosed with ACP or PCP were included. Pretreatment contrast-enhanced T1-weighted images were collected and used as the input of the CNNs. Six models were established based on six networks, including VGG16, ResNet18, ResNet50, ResNet101, DenseNet121, and DenseNet169. The area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity were used to assess the performances of these deep neural networks. A five-fold cross-validation was applied to evaluate the performances of the models.
The six networks yielded feasible performances, with area under the receiver operating characteristic curves (AUCs) of at least 0.78 for classification. The model based on Resnet50 achieved the highest AUC of 0.838 ± 0.062, with an accuracy of 0.757 ± 0.052, a sensitivity of 0.608 ± 0.198, and a specificity of 0.845 ± 0.034, respectively. Moreover, the results also indicated that the CNN method had a competitive performance compared to the radiomics-based method, which required manual segmentation for feature extraction and further feature selection.
MRI-based deep neural networks can noninvasively differentiate ACP from PCP to facilitate the personalized assessment of craniopharyngiomas.
本研究的目标是开发端到端卷积神经网络（CNN）模型，该模型能够在无需手动分割的情况下，在磁共振成像（MR）图像上无创地区分乳头型颅咽管瘤（PCP）和造釉型颅咽管瘤（ACP）。
共纳入97例诊断为ACP或PCP的患者。收集治疗前的对比增强T1加权图像，并将其用作CNN的输入。基于六个网络建立了六个模型，包括VGG16、ResNet18、ResNet50、ResNet101、DenseNet121和DenseNet169。采用受试者操作特征曲线（AUC）下面积、准确率、灵敏度和特异性来评估这些深度神经网络的性能。应用五折交叉验证来评估模型的性能。
这六个网络均取得了可行的性能，分类的受试者操作特征曲线（AUC）下面积至少为0.78。基于Resnet50的模型获得了最高的AUC，为0.838±0.062，准确率为0.757±0.052，灵敏度为0.608±0.198，特异性为0.845±0.034。此外，结果还表明，与基于放射组学的方法相比，CNN方法具有竞争性能，基于放射组学的方法需要手动分割来提取特征并进行进一步的特征选择。
基于MRI的深度神经网络可以无创地区分ACP和PCP，以促进颅咽管瘤的个性化评估。
