利用迁移学习将超低计数淀粉样蛋白PET/MRI深度学习模型泛化至不同扫描硬件与协议
研究评估深度学习泛化能力:将站点1预训练的超低计数淀粉样蛋白PET/MRI增强网络,直接应用或迁移学习后应用于站点2不同硬件和协议的数据。结果表明迁移学习(方法B)在图像指标和区分阳阴性能力上表现最佳,直接应用(方法A)精度较低,提示应用预训练网络时需考虑数据偏倚。
要点
- 迁移学习(方法B)使合成图像SUVR与真实值差异最小、区分阳阴性效应量最高。
- 直接应用站点1预训练网络(方法A)的准确性和图像质量低于其他方法。
- 应用预训练超低计数淀粉样蛋白PET/MRI网络进行泛化时,应考虑数据偏倚。
核心结论
- 深度学习能够从短帧重建中成功生成可诊断的淀粉样蛋白PET图像。
- 应用预训练超低计数淀粉样蛋白PET/MRI网络进行泛化时,应考虑数据偏倚。
研究设计与样本
- 将注射后90-110分钟的20分钟列表模式PET数据重建为真实值。
- 站点1通过100倍欠采样获得超低计数数据,站点2取PET采集第一分钟数据,分别重建为超低剂量/超短时间(分别为1%剂量和5%时间)PET图像。
- 深度卷积神经网络先用站点1数据预训练,然后(A)直接应用或(B)使用迁移学习在站点2数据上进一步训练;另从零开始基于(C)站点2数据或(D)全部数据训练网络。
- 由认证医师判定淀粉样蛋白摄取(+/-)状态以计算准确性并评分图像质量,计算图像与真实值之间的峰值信噪比、结构相似性和均方根误差,并分析站点2中37例成功FreeSurfer分割的平均区域标准化摄取值比(SUVR,参考区:小脑皮质)。
主要发现
- 所有网络合成图像均比其超低计数重建图像噪声降低。
- 定量上,方法B的图像指标改善最大,其SUVR与真实值的变异性最小,区分阳性与阴性图像的效应量最高。
- 方法A图像的准确性和图像质量低于其他方法。
- 方法B-D合成的图像评分与真实值图像相似或更优。
关键数值与推荐
- 超低计数数据获取:站点1通过100倍欠采样,站点2取PET采集第一分钟。
- 超低剂量/超短时间PET图像对应1%剂量(站点1)和5%时间(站点2)。
- SUVR参考区为小脑皮质。
- 站点2成功FreeSurfer分割37例。
临床怎么用
- 原文未据此给出推荐。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
PURPOSE: We aimed to evaluate the performance of deep learning-based generalization of ultra-low-count amyloid PET/MRI enhancement when applied to studies acquired with different scanning hardware and protocols. METHODS: Eighty simultaneous [18F]florbetaben PET/MRI studies were acquired, split equally between two sites (site 1: Signa PET/MRI, GE Healthcare, 39 participants, 67 ± 8 years, 23 females; site 2: mMR, Siemens Healthineers, 64 ± 11 years, 23 females) with different MRI protocols. Twenty minutes of list-mode PET data (90-110 min post-injection) were reconstructed as ground-truth. Ultra-low-count data obtained from undersampling by a factor of 100 (site 1) or the first minute of PET acquisition (site 2) were reconstructed for ultra-low-dose/ultra-short-time (1% dose and 5% time, respectively) PET images. A deep convolution neural network was pre-trained with site 1 data and either (A) directly applied or (B) trained further on site 2 data using transfer learning. Networks were also trained from scratch based on (C) site 2 data or (D) all data. Certified physicians determined amyloid uptake (+/-) status for accuracy and scored the image quality. The peak signal-to-noise ratio, structural similarity, and root-mean-squared error were calculated between images and their ground-truth counterparts. Mean regional standardized uptake value ratios (SUVR, reference region: cerebellar cortex) from 37 successful site 2 FreeSurfer segmentations were analyzed. RESULTS: All network-synthesized images had reduced noise than their ultra-low-count reconstructions. Quantitatively, image metrics improved the most using method B, where SUVRs had the least variability from the ground-truth and the highest effect size to differentiate between positive and negative images. Method A images had lower accuracy and image quality than other methods; images synthesized from methods B-D scored similarly or better than the ground-truth images. CONCLUSIONS: Deep learning can successfully produce diagnostic amyloid PET images from short frame reconstructions. Data bias should be considered when applying pre-trained deep ultra-low-count amyloid PET/MRI networks for generalization.
引用
- Chen KT, Schürer M, Ouyang J, et al. Generalization of deep learning models for ultra-low-count amyloid PET/MRI using transfer learning. Eur J Nucl Med Mol Imaging. 2020. PMID: 32535655. doi:10.1007/s00259-020-04897-6
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。