Jove
Visualize
联系我们
JoVE
x logofacebook logolinkedin logoyoutube logo
关于 JoVE
概览领导团队博客JoVE 帮助中心
作者
出版流程编辑委员会范围与政策同行评审常见问题投稿
图书馆员
用户评价订阅访问资源图书馆顾问委员会常见问题
研究
JoVE JournalMethods CollectionsJoVE Encyclopedia of Experiments存档
教育
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab Manual教师资源中心教师网站
使用条款与条件
隐私政策
政策

相关概念视频

您也可能阅读

相关文章

通过共同作者、期刊和引用图与本文相关的文章。

排序
Same author

Enhancing pancreatic imaging in CT - prospective comparison of fixed versus individualized post-trigger delay in bolus tracking.

European journal of radiology·2026
Same author

Cardiopulmonary transit time measured on dynamic rubidium cardiac PET/CT is a predictor for pulmonary hypertension.

EJNMMI research·2026
Same author

Estimation of Future Patient Populations Eligible for Radioligand Therapies in the EU and the UK: A Modelling Study.

The Lancet regional health. Europe·2026
Same author

German Society of Nuclear Medicine procedure guideline for brain perfusion SPECT using 99mTc-labelled radiopharmaceuticals (09/2025).

Nuklearmedizin. Nuclear medicine·2026
Same author

The PET tracer [<sup>11</sup>C]MODAG-005 targets alpha-synuclein aggregates in the brain.

Science translational medicine·2026
Same author

Updated Procedure Guideline of the German Society of Nuclear Medicine for Amyloid Brain PET Imaging 3.0.

Nuklearmedizin. Nuclear medicine·2026

相关实验视频

Updated: Jan 7, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K

关于瘤PET/CT成像中全自动化损伤细分的AutoPET挑战,第2部分:域泛化.

Jakob Dexl1,2, Sergios Gatidis3,4, Marcel Früh3

  • 1Department of Radiology, LMU University Hospital, LMU Munich, Munich, Germany; jakob.dexl@med.uni-muenchen.de.

Journal of nuclear medicine : official publication, Society of Nuclear Medicine
|December 30, 2025
PubMed
概括
此摘要是机器生成的。

针对PET/CT病变细分的机器学习模型与域概括性作斗争. 自动PET挑战表明,在一个数据源上训练的模型在不同的临床数据上表现不佳,突出了对各种数据集的需求.

关键词:
聚乙烯/聚乙烯/聚乙烯生物医学图像分析挑战深度学习是一种深度学习.域名通用化域名通用化瘤学 在瘤学方面.细分化 细分化的细分化

更多相关视频

Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images
09:21

Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images

Published on: February 18, 2015

12.6K
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.2K

相关实验视频

Last Updated: Jan 7, 2026

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
04:48

Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography

Published on: November 30, 2022

3.3K
Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images
09:21

Human Brown Adipose Tissue Depots Automatically Segmented by Positron Emission Tomography/Computed Tomography and Registered Magnetic Resonance Images

Published on: February 18, 2015

12.6K
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
10:25

Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping

Published on: September 25, 2019

49.2K

科学领域:

  • 医疗成像医学成像
  • 人工智能在医学中的应用

背景情况:

  • 在正电子发射断层扫描/计算机断层扫描 (PET/CT) 中自动化损伤细分对于癌症诊断和治疗监测至关重要.
  • 自动PET挑战旨在为此任务推进机器学习 (ML) 模型,重点关注现实世界的部署挑战.

研究的目的:

  • 评估基于单一来源的PET/CT数据训练的基于ML的细分模型的域概括能力.
  • 评估跨不同临床变异的模型性能,包括不同的机构,病理,种群和标记物.

主要方法:

  • 第二个自动PET挑战涉及对1014个全身18F-FDG PET/CT扫描进行ML模型训练.
  • 在5个不同的临床领域的200个样本上测试了模型.
  • 使用子相似系数,假阳性体积和假阴性体积量化表现.

主要成果:

  • 从单个数据源进行概括仍然是一个重大挑战,域外性能大幅恶化.
  • 最好的模型获得了0.5038的子得分,但在儿科和PSMA数据上性能下降.
  • 错误分析表明,生理吸收和检测小或低吸收病变存在问题.

结论:

  • 自动PET挑战突出了当前自动化PET/CT细分模型在处理数据变化方面的局限性.
  • 需要多样化,多域公共数据集来提高这些算法的稳定性和临床适用性.