Radiomics and deep learning for myocardial scar screening in hypertrophic cardiomyopathy.

Radiomics and deep learning for myocardial scar screening in hypertrophic cardiomyopathy.
复制标题

DOI:
10.1186/s12968-022-00869-x
复制
发表时间:
2022-06-27
影响因子:
6.4
通讯作者:
Nezafat, Reza
Nezafat, Reza
中科院分区:
医学2区
文献类型:
--
作者:
Fahmy, Ahmed S.;Rowin, Ethan J.;Arafati, Arghavan;Al-Otaibi, Talal;Maron, Martin S.;Nezafat, Reza

文献摘要

参考文献

被引文献

相似文献

使用晚期钆增强 (LGE) 心血管磁共振 (CMR) 量化的心肌疤痕负荷对于肥厚型心肌病 (HCM) 具有重要的预后价值。然而,近 50% 的 HCM 患者没有疤痕,但在其一生中反复接受基于钆的 CMR。我们试图开发一种基于人工智能 (AI) 的筛查模型,使用从平衡稳态自由进动 (bSSFP) 电影序列中提取的放射组学和深度学习 (DL) 特征来识别无疤痕的 HCM 患者。我们使用放射组学、DL 或组合 DL-Radiomics 提取的 bSSFP 电影图像特征评估了三种基于 AI 的筛选模型。在一项多中心/供应商研究中,使用 759 名 HCM 患者(50±16 岁,66% 男性)的图像来开发和测试模型性能。使用 100 名 HCM 患者(53±14 岁,70% 男性)的外部数据集来评估模型的普遍性。使用接收器工作曲线下面积(AUC)评估模型性能。与 DL 和 Radiomics 相比,DL-Radiomics 模型在内部(0.83 vs 0.77,p = 0.006 和 0.78,p = 0.05;n = 159)和外部(0.74 vs 0.64,p = 0.006 和 0.71)中表现出更高的 AUC, p = 0.27;n = 100)数据集。 DL-Radiomics 模型在内部和外部数据集中正确识别了 43% 和 28% 的无疤痕患者,而 Radiomics 模型的准确率分别为 42% 和 16%,DL 模型的准确率分别为 42% 和 23%。使用 bSSFP 电影图像的 DL-Radiomics AI 模型作为钆给药前的疤痕筛查工具,其性能优于单独的 DL 或放射组学模型。尽管具有潜力,但该模型的临床实用性仍然有限,需要进一步研究以提高准确性和普遍性。在线版本包含可在 10.1186/s12968-022-00869-x 获取的补充材料。
Myocardial scar burden quantified using late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR), has important prognostic value in hypertrophic cardiomyopathy (HCM). However, nearly 50% of HCM patients have no scar but undergo repeated gadolinium-based CMR over their life span. We sought to develop an artificial intelligence (AI)-based screening model using radiomics and deep learning (DL) features extracted from balanced steady state free precession (bSSFP) cine sequences to identify HCM patients without scar. We evaluated three AI-based screening models using bSSFP cine image features extracted by radiomics, DL, or combined DL-Radiomics. Images for 759 HCM patients (50 ± 16 years, 66% men) in a multi-center/vendor study were used to develop and test model performance. An external dataset of 100 HCM patients (53 ± 14 years, 70% men) was used to assess model generalizability. Model performance was evaluated using area-under-receiver-operating curve (AUC). The DL-Radiomics model demonstrated higher AUC compared to DL and Radiomics in the internal (0.83 vs 0.77, p = 0.006 and 0.78, p = 0.05; n = 159) and external (0.74 vs 0.64, p = 0.006 and 0.71, p = 0.27; n = 100) datasets. The DL-Radiomics model correctly identified 43% and 28% of patients without scar in the internal and external datasets compared to 42% and 16% by Radiomics model and 42% and 23% by DL model, respectively. A DL-Radiomics AI model using bSSFP cine images outperforms DL or Radiomics models alone as a scar screening tool prior to gadolinium administration. Despite its potential, the clinical utility of the model remains limited and further investigation is needed to improve the accuracy and generalizability. The online version contains supplementary material available at 10.1186/s12968-022-00869-x.
DOI: 10.1148/radiology.180.3.1871277
发表时间: 1991-09-01
期刊: RADIOLOGY
影响因子: 19.7
作者:
BALABAN, RS;CHESNICK, S;HEINEMAN, FW
通讯作者: HEINEMAN, FW
DOI: 10.1016/j.ejrad.2017.04.024
发表时间: 2017-07-01
影响因子: 3.3
作者:
Larroza, Andres;Materka, Andrzej;Moratal, David
通讯作者: Moratal, David
DOI: 10.1038/s41598-021-83593-3
发表时间: 2021-02-17
期刊: Scientific reports
影响因子: 4.6
作者:
Lee J;Steinmann A;Ding Y;Lee H;Owens C;Wang J;Yang J;Followill D;Ger R;MacKin D;Court LE
通讯作者: Court LE
磁场强度对脑成像,体外和体内研究中磁共振成像放射素学特征的影响。
DOI: 10.3389/fonc.2020.541663
发表时间: 2020
影响因子: 4.7
作者:
Ammari S;Pitre-Champagnat S;Dercle L;Chouzenoux E;Moalla S;Reuze S;Talbot H;Mokoyoko T;Hadchiti J;Diffetocq S;Volk A;El Haik M;Lakiss S;Balleyguier C;Lassau N;Bidault F
通讯作者: Bidault F
DOI: 10.1016/s1474-4422(17)30158-8
发表时间: 2017-07-01
期刊: LANCET NEUROLOGY
影响因子: 48
作者:
Gulani, Vikas;Calamante, Fernando;Reeder, Scott B.
通讯作者: Reeder, Scott B.