Pusterla Orso, Willers Corin, Sandkühler Robin, Andermatt Simon, Nyilas Sylvia, Cattin Philippe C, Latzin Philipp, Bieri Oliver, Bauman Grzegorz
Department of Radiology, Division of Radiological Physics, University Hospital Basel, University of Basel, Basel, Switzerland; Department of Biomedical Engineering, University of Basel, Basel, Switzerland; Division of Pediatric Respiratory Medicine and Allergology, Department of Pediatrics, Inselspital, Bern University Hospital, University of Bern, Switzerland.
Division of Pediatric Respiratory Medicine and Allergology, Department of Pediatrics, Inselspital, Bern University Hospital, University of Bern, Switzerland.
Z Med Phys. 2024 Sep 19. doi: 10.1016/j.zemedi.2024.08.001.
To introduce and evaluate TrueLung, an automated pipeline for computation and analysis of free-breathing and contrast-agent free pulmonary functional magnetic resonance imaging.
Two-dimensional time-resolved ultra-fast balanced steady-state free precession acquisitions were transferred to TrueLung, which included image quality checks, image registration, and computation of perfusion and ventilation maps with matrix pencil decomposition. Neural network whole-lung and lobar segmentations allowed quantification of impaired relative perfusion (R) and fractional ventilation (R). TrueLung delivered functional maps and quantitative outcomes, reported for clinicians in concise documents. We evaluated the pipeline using 1.5T data from 75 children with cystic fibrosis by assessing the feasibility of functional MR imaging, average scan time, and the robustness of the functional outcomes. Whole-lung and lobar segmentations were manually refined when necessary, and the impact on R and R was quantified.
Functional imaging was feasible in all included CF children without any dropouts. On average, 7.9 ± 1.8 (mean±SD) coronal slice positions per patient were acquired, resulting in a mean scan time of 6min 20s per patient. The whole pipeline required 20min processing time per subject. TrueLung delivered the functional maps of all the subjects for radiological assessment. Quality controlling maps and segmentations lasted 1min 12s per patient. The automated segmentations and quantification of whole-lung defects were satisfying in 88% of patients (97% of slices) and the lobar quantification in 73% (93% of slices). The segmentations refinements required 16s per patient for the whole-lung, and 2min 10s for the lobe masks. The relative differences in R and R between fully-automated and manually refined data were 0.7% (1.2%) and 2.0% (2.9%) for whole-lung quantification (median, [third quartile]), and excluding two outliers, 1.7% (3.9%) and 1.2% (3.8%) for the lobes, indicating the refinements could be potentially omitted in several patients.
TrueLung quickly delivers functional maps and quantitative outcomes in an objective and standardized way, suitable for radiological and pneumological assessment with minimal manual input. TrueLung can be used for clinical research in cystic fibrosis and might be applied across various lung diseases.
介绍并评估TrueLung，这是一种用于自由呼吸且无需使用造影剂的肺功能磁共振成像计算和分析的自动化流程。
将二维时间分辨超快速平衡稳态自由进动采集数据传输至TrueLung，该流程包括图像质量检查、图像配准以及使用矩阵束分解计算灌注和通气图。神经网络全肺和肺叶分割可实现受损相对灌注（R）和分数通气（R）的量化。TrueLung生成功能图和定量结果，并以简洁文档形式报告给临床医生。我们通过评估功能磁共振成像的可行性、平均扫描时间以及功能结果的稳健性，使用来自75名囊性纤维化儿童的1.5T数据对该流程进行了评估。必要时对全肺和肺叶分割进行手动优化，并对其对R和R的影响进行量化。
在所有纳入的囊性纤维化儿童中，功能成像均可行，无任何退出研究的情况。每位患者平均采集7.9±1.8（均值±标准差）个冠状位切片位置，每位患者的平均扫描时间为6分20秒。整个流程每位受试者需要20分钟的处理时间。TrueLung为所有受试者生成了用于放射学评估的功能图。质量控制图和分割每位患者耗时1分12秒。88%的患者（97%的切片）全肺缺陷的自动分割和量化结果令人满意，73%（93%的切片）肺叶量化结果令人满意。全肺分割优化每位患者需要16秒，肺叶掩码优化需要2分10秒。全肺量化中，全自动数据与手动优化数据之间R和R的相对差异为0.7%（1.2%）和2.0%（2.9%）（中位数，[第三四分位数]），排除两个异常值后，肺叶的相对差异为1.7%（3.9%）和1.2%（3.8%），这表明在一些患者中可能无需进行优化。
TrueLung能够以客观、标准化的方式快速生成功能图和定量结果，适合在极少人工干预的情况下进行放射学和呼吸病学评估。TrueLung可用于囊性纤维化的临床研究，也可能适用于各种肺部疾病。
