当前位置: X-MOL 学术Eur. J. Nucl. Med. Mol. Imaging › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Automatic classification of dopamine transporter SPECT: deep convolutional neural networks can be trained to be robust with respect to variable image characteristics.
European Journal of Nuclear Medicine and Molecular Imaging ( IF 9.1 ) Pub Date : 2019-08-31 , DOI: 10.1007/s00259-019-04502-5
Markus Wenzel 1 , Fausto Milletari 2, 3 , Julia Krüger 4 , Catharina Lange 5 , Michael Schenk 6 , Ivayla Apostolova 6 , Susanne Klutmann 6 , Marcus Ehrenburg 7 , Ralph Buchert 6
Affiliation  

PURPOSE This study investigated the potential of deep convolutional neural networks (CNN) for automatic classification of FP-CIT SPECT in multi-site or multi-camera settings with variable image characteristics. METHODS The study included FP-CIT SPECT of 645 subjects from the Parkinson's Progression Marker Initiative (PPMI), 207 healthy controls, and 438 Parkinson's disease patients. SPECT images were smoothed with an isotropic 18-mm Gaussian kernel resulting in 3 different PPMI settings: (i) original (unsmoothed), (ii) smoothed, and (iii) mixed setting comprising all original and all smoothed images. A deep CNN with 2,872,642 parameters was trained, validated, and tested separately for each setting using 10 random splits with 60/20/20% allocation to training/validation/test sample. The putaminal specific binding ratio (SBR) was computed using a standard anatomical ROI predefined in MNI space (AAL atlas) or using the hottest voxels (HV) analysis. Both SBR measures were trained (ROC analysis, Youden criterion) using the same random splits as for the CNN. CNN and SBR trained in the mixed PPMI setting were also tested in an independent sample from clinical routine patient care (149 with non-neurodegenerative and 149 with neurodegenerative parkinsonian syndrome). RESULTS Both SBR measures performed worse in the mixed PPMI setting compared to the pure PPMI settings (e.g., AAL-SBR accuracy = 0.900 ± 0.029 in the mixed setting versus 0.957 ± 0.017 and 0.952 ± 0.015 in original and smoothed setting, both p < 0.01). In contrast, the CNN showed similar accuracy in all PPMI settings (0.967 ± 0.018, 0.972 ± 0.014, and 0.955 ± 0.009 in mixed, original, and smoothed setting). Similar results were obtained in the clinical sample. After training in the mixed PPMI setting, only the CNN provided acceptable performance in the clinical sample. CONCLUSIONS These findings provide proof of concept that a deep CNN can be trained to be robust with respect to variable site-, camera-, or scan-specific image characteristics without a large loss of diagnostic accuracy compared with mono-site/mono-camera settings. We hypothesize that a single CNN can be used to support the interpretation of FP-CIT SPECT at many different sites using different acquisition hardware and/or reconstruction software with only minor harmonization of acquisition and reconstruction protocols.

中文翻译:

多巴胺转运蛋白SPECT的自动分类:可以训练深度卷积神经网络,使其在可变图像特征方面具有较强的鲁棒性。

目的这项研究调查了深卷积神经网络(CNN)在具有可变图像特征的多站点或多摄像机设置中对FP-CIT SPECT自动分类的潜力。方法该研究包括来自帕金森氏病进展指标计划(PPMI)的645位受试者的FP-CIT SPECT,207位健康对照和438位帕金森氏病患者。用各向同性的18毫米高斯核对SPECT图像进行平滑处理,从而得到3种不同的PPMI设置:(i)原始(不平滑),(ii)平滑和(iii)包括所有原始图像和所有平滑图像的混合设置。对于每个设置,使用10个随机分割分别对训练/验证/测试样本分配了60/20/20%的深层CNN,分别训练,验证和测试了具有2,872,642个参数的深层CNN。使用在MNI空间(AAL图集)中预定义的标准解剖ROI或使用最热的体素(HV)分析,可以计算出肠壁特异性结合率(SBR)。使用与CNN相同的随机分割训练两种SBR量度(ROC分析,Youden准则)。在混合的PPMI环境中训练的CNN和SBR也从临床常规患者护理的独立样本中进行了测试(149例非神经退行性疾病和149例神经退行性帕金森综合症)。结果与纯PPMI设置相比,混合PPMI设置中两种SBR测量的性能均较差(例如,AAL-SBR精度在混合设置中为0.900±0.029,而原始设置和平滑设置为0.957±0.017和0.952±0.015,两者p <0.01 )。相比之下,CNN在所有PPMI设置中显示出相似的准确性(0.967±0.018、0.972±0.014和0.955±0。009(混合,原始和平滑设置)。在临床样品中获得了相似的结果。在混合PPMI设置中训练后,只有CNN在临床样本中提供了可接受的性能。结论这些发现提供了概念证明,即与单个站点/单个摄像机设置相比,可以训练深层CNN在可变站点,摄像机或扫描特定的图像特征方面具有较强的鲁棒性,而不会大大降低诊断准确性。 。我们假设单个CNN可以用于支持使用不同的采集硬件和/或重建软件在许多不同站点进行的FP-CIT SPECT的解释,而仅对采集和重建协议进行较小的协调。在混合PPMI设置中训练后,只有CNN在临床样本中提供了可接受的性能。结论这些发现提供了概念证明,即与单个站点/单个摄像机设置相比,可以对深度CNN进行训练,使其具有可变的站点,摄像机或扫描特定图像特征的鲁棒性,而不会大大降低诊断准确性。 。我们假设单个CNN可以用来支持使用不同的采集硬件和/或重建软件在许多不同站点进行的FP-CIT SPECT的解释,而仅对采集和重建协议进行较小的协调。在混合PPMI设置中训练后,只有CNN在临床样本中提供了可接受的性能。结论这些发现提供了概念证明,即与单个站点/单个摄像机设置相比,可以训练深层CNN在可变站点,摄像机或扫描特定的图像特征方面具有较强的鲁棒性,而不会大大降低诊断准确性。 。我们假设单个CNN可以用来支持使用不同的采集硬件和/或重建软件在许多不同站点进行的FP-CIT SPECT的解释,而仅对采集和重建协议进行较小的协调。或特定于扫描的图像特征,与单站点/单摄像机设置相比,不会大大降低诊断准确性。我们假设单个CNN可以用来支持使用不同的采集硬件和/或重建软件在许多不同站点进行的FP-CIT SPECT的解释,而仅对采集和重建协议进行较小的协调。或特定于扫描的图像特征,与单站点/单摄像机设置相比,不会大大降低诊断准确性。我们假设单个CNN可以用来支持使用不同的采集硬件和/或重建软件在许多不同站点进行的FP-CIT SPECT的解释,而仅对采集和重建协议进行较小的协调。
更新日期:2019-08-31
down
wechat
bug