Robust and accurate pulmonary nodule detection with self-supervised feature learning on domain adaptation.

Frontiers in radiology Pub Date : 2022-12-15 eCollection Date: 2022-01-01 DOI:10.3389/fradi.2022.1041518
Jingya Liu, Liangliang Cao, Oguz Akin, Yingli Tian
{"title":"Robust and accurate pulmonary nodule detection with self-supervised feature learning on domain adaptation.","authors":"Jingya Liu, Liangliang Cao, Oguz Akin, Yingli Tian","doi":"10.3389/fradi.2022.1041518","DOIUrl":null,"url":null,"abstract":"<p><p>Medical imaging data annotation is expensive and time-consuming. Supervised deep learning approaches may encounter overfitting if trained with limited medical data, and further affect the robustness of computer-aided diagnosis (CAD) on CT scans collected by various scanner vendors. Additionally, the high false-positive rate in automatic lung nodule detection methods prevents their applications in daily clinical routine diagnosis. To tackle these issues, we first introduce a novel self-learning schema to train a pre-trained model by learning rich feature representatives from large-scale unlabeled data without extra annotation, which guarantees a consistent detection performance over novel datasets. Then, a 3D feature pyramid network (<i>3DFPN</i>) is proposed for high-sensitivity nodule detection by extracting multi-scale features, where the weights of the backbone network are initialized by the pre-trained model and then fine-tuned in a supervised manner. Further, a High Sensitivity and Specificity (<i>HS</i><math><msup><mi></mi><mn>2</mn></msup></math>) network is proposed to reduce false positives by tracking the appearance changes among continuous CT slices on Location History Images (LHI) for the detected nodule candidates. The proposed method's performance and robustness are evaluated on several publicly available datasets, including LUNA16, SPIE-AAPM, LungTIME, and HMS. Our proposed detector achieves the state-of-the-art result of <math><mn>90.6</mn><mtext>%</mtext></math> sensitivity at <math><mn>1</mn><mrow><mo>/</mo></mrow><mn>8</mn></math> false positive per scan on the LUNA16 dataset. The proposed framework's generalizability has been evaluated on three additional datasets (i.e., SPIE-AAPM, LungTIME, and HMS) captured by different types of CT scanners.</p>","PeriodicalId":73101,"journal":{"name":"Frontiers in radiology","volume":null,"pages":null},"PeriodicalIF":0.0000,"publicationDate":"2022-12-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10365286/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Frontiers in radiology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.3389/fradi.2022.1041518","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2022/1/1 0:00:00","PubModel":"eCollection","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

Medical imaging data annotation is expensive and time-consuming. Supervised deep learning approaches may encounter overfitting if trained with limited medical data, and further affect the robustness of computer-aided diagnosis (CAD) on CT scans collected by various scanner vendors. Additionally, the high false-positive rate in automatic lung nodule detection methods prevents their applications in daily clinical routine diagnosis. To tackle these issues, we first introduce a novel self-learning schema to train a pre-trained model by learning rich feature representatives from large-scale unlabeled data without extra annotation, which guarantees a consistent detection performance over novel datasets. Then, a 3D feature pyramid network (3DFPN) is proposed for high-sensitivity nodule detection by extracting multi-scale features, where the weights of the backbone network are initialized by the pre-trained model and then fine-tuned in a supervised manner. Further, a High Sensitivity and Specificity (HS2) network is proposed to reduce false positives by tracking the appearance changes among continuous CT slices on Location History Images (LHI) for the detected nodule candidates. The proposed method's performance and robustness are evaluated on several publicly available datasets, including LUNA16, SPIE-AAPM, LungTIME, and HMS. Our proposed detector achieves the state-of-the-art result of 90.6% sensitivity at 1/8 false positive per scan on the LUNA16 dataset. The proposed framework's generalizability has been evaluated on three additional datasets (i.e., SPIE-AAPM, LungTIME, and HMS) captured by different types of CT scanners.

Abstract Image

Abstract Image

Abstract Image

利用域适应性自监督特征学习实现稳健准确的肺结节检测
医学影像数据标注既昂贵又耗时。如果使用有限的医疗数据进行训练,有监督的深度学习方法可能会遇到过拟合的问题,并进一步影响计算机辅助诊断(CAD)对不同扫描仪供应商收集的 CT 扫描数据的稳健性。此外,肺结节自动检测方法的高假阳性率也阻碍了其在日常临床诊断中的应用。为了解决这些问题,我们首先引入了一种新颖的自学模式,通过从大规模无标注数据中学习丰富的特征代表来训练预训练模型,无需额外标注,从而保证了在新数据集上的一致检测性能。然后,提出了一种三维特征金字塔网络(3DFPN),通过提取多尺度特征进行高灵敏度结核检测,其中骨干网络的权重由预训练模型初始化,然后以监督方式进行微调。此外,还提出了一种高灵敏度和高特异性(HS2)网络,通过跟踪位置历史图像(LHI)上连续 CT 切片之间的外观变化来减少检测到的结节候选者的假阳性。我们在 LUNA16、SPIE-AAPM、LungTIME 和 HMS 等多个公开数据集上评估了所提方法的性能和鲁棒性。我们提出的检测器在 LUNA16 数据集上达到了 90.6% 的灵敏度,每次扫描的误报率为 1/8。我们还在由不同类型 CT 扫描仪采集的另外三个数据集(即 SPIE-AAPM、LungTIME 和 HMS)上评估了所提出框架的通用性。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
CiteScore
1.20
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信