[论文解读] DaTscan SPECT Image Classification for Parkinson's Disease

Justin Quan, Xu Lin+3 more|arXiv (Cornell University)|Sep 9, 2019

Brain Tumor Detection and Classification参考文献 14被引用 10

一句话总结

本研究提出一种基于InceptionV3架构的深度卷积神经网络（CNN），用于将DaTscan SPECT图像分类为帕金森病（PD）或正常，通过3切片（40–42）RGB拼接和数据增强技术处理小样本数据集。该模型在659幅图像的PPMI数据集上实现了98.48%的测试准确率、99.67%的PR AUC和99.09%的ROC AUC，表现出高敏感性（98.88%）和特异性（97.67%）。

ABSTRACT

Parkinson's Disease (PD) is a neurodegenerative disease that currently does not have a cure. In order to facilitate disease management and reduce the speed of symptom progression, early diagnosis is essential. The current clinical, diagnostic approach is to have radiologists perform human visual analysis of the degeneration of dopaminergic neurons in the substantia nigra region of the brain. Clinically, dopamine levels are monitored through observing dopamine transporter (DaT) activity. One method of DaT activity analysis is performed with the injection of an Iodine-123 fluoropropyl (123I-FP-CIT) tracer combined with single photon emission computerized tomography (SPECT) imaging. The tracer illustrates the region of interest in the resulting DaTscan SPECT images. Human visual analysis is slow and vulnerable to subjectivity between radiologists, so the goal was to develop an introductory implementation of a deep convolutional neural network that can objectively and accurately classify DaTscan SPECT images as Parkinson's Disease or normal. This study illustrates the approach of using a deep convolutional neural network and evaluates its performance on DaTscan SPECT image classification. The data used in this study was obtained through a database provided by the Parkinson's Progression Markers Initiative (PPMI). The deep neural network in this study utilizes the InceptionV3 architecture, 1st runner up in the 2015 ImageNet Large Scale Visual Recognition Competition (ILSVRC), as a base model. A custom, binary classifier block was added on top of this base. In order to account for the small dataset size, a ten fold cross validation was implemented to evaluate the model's performance.

研究动机与目标

• 开发一种客观、自动化的分类方法，将DaTscan SPECT图像分类为PD或非PD，以减少对主观放射科医生读片的依赖。
• 评估深度CNN在PPMI数据库中来自小样本、类别不平衡的DaTscan SPECT图像数据集上的性能。
• 探索数据增强和迁移学习技术，以在训练数据有限的情况下提升模型泛化能力。
• 评估将3D SPECT图像切片作为2D RGB输入用于有效深度学习分类的可行性。

提出的方法

• 本研究采用迁移学习，使用预训练的InceptionV3架构作为特征提取器，并在DaTscan SPECT图像上进行微调。
• 选择在纹状体区域信号强度最高的切片40、41和42，将其拼接为3通道RGB图像，以保留空间上下文信息。
• 通过Keras的ImageDataGenerator在训练过程中实时应用数据增强，包括小幅旋转、平移和翻转，以增加有效数据集规模。
• 采用10折交叉验证评估模型鲁棒性，随后进行最终的80:20训练-测试划分，并采用类别平衡采样。
• 模型训练500个周期，学习率采用余弦退火调度，从1e-3降至1e-6，批量大小为16。
• 性能评估采用标准指标：准确率、AUC-ROC、AUC-PR、敏感性、特异性和精确率。

实验结果

研究问题

• RQ1 深度CNN能否在高准确率下有效将DaTscan SPECT图像分类为PD和非PD类别？
• RQ2 数据增强和迁移学习在小样本、类别不平衡的医学影像数据集上如何提升性能？
• RQ3 与使用完整体积数据相比，选择特定SPECT切片（40–42）在多大程度上影响分类性能？
• RQ4 该模型的敏感性和特异性与临床诊断标准相比如何？

主要发现

• 模型在132张未见的测试图像上实现了98.48%的测试准确率，其中88例为真正例，42例为真负例。
• 敏感性达到98.88%，表明模型正确识别了98.88%的PD病例。
• 特异性为97.67%，表明模型正确分类了97.67%的非PD病例。
• 精确率为98.88%，意味着98.88%的预测PD病例为正确。
• PR曲线下面积为0.9967，表明精确率-召回率性能极佳。
• ROC曲线下面积为0.9909，表明模型在PD与非PD类别之间具有强大的判别能力。