Bagher-Ebadian Hassan, Brown Stephen L, Ghassemi Mohammad M, Acharya Prabhu C, Chetty Indrin J, Movsas Benjamin, Ewing James R, Thind Kundan
Department of Radiation Oncology, Henry Ford Hospital, Detroit, USA.
Department of Radiology, Michigan State University, East Lansing, USA.
Sci Rep. 2025 Jan 13;15(1):1786. doi: 10.1038/s41598-024-83306-6.
Best current practice in the analysis of dynamic contrast enhanced (DCE)-MRI is to employ a voxel-by-voxel model selection from a hierarchy of nested models. This nested model selection (NMS) assumes that the observed time-trace of contrast-agent (CA) concentration within a voxel, corresponds to a singular physiologically nested model. However, admixtures of different models may exist within a voxel's CA time-trace. This study introduces an unsupervised feature engineering technique (Kohonen-Self-Organizing-Map (K-SOM)) to estimate the voxel-wise probability of each nested model. Sixty-six immune-compromised-RNU rats were implanted with human U-251 N cancer cells, and DCE-MRI data were acquired from all the rat brains. The time-trace of change in the longitudinal-relaxivity (ΔR) for all animals' brain voxels was calculated. DCE-MRI pharmacokinetic (PK) analysis was performed using NMS to estimate three model regions: Model-1: normal vasculature without leakage, Model-2: tumor tissues with leakage without back-flux to the vasculature, Model-3: tumor vessels with leakage and back-flux. Approximately two hundred thirty thousand (229,314) normalized ΔR profiles of animals' brain voxels along with their NMS results were used to build a K-SOM (topology-size: 8 × 8, with competitive-learning algorithm) and probability map of each model. K-fold nested-cross-validation (NCV, k = 10) was used to evaluate the performance of the K-SOM probabilistic-NMS (PNMS) technique against the NMS technique. The K-SOM PNMS's estimation for the leaky tumor regions were strongly similar (Dice-Similarity-Coefficient, DSC = 0.774 [CI: 0.731-0.823], and 0.866 [CI: 0.828-0.912] for Models 2 and 3, respectively) to their respective NMS regions. The mean-percent-differences (MPDs, NCV, k = 10) for the estimated permeability parameters by the two techniques were: -28%, + 18%, and + 24%, for v, K, and v, respectively. The KSOM-PNMS technique produced microvasculature parameters and NMS regions less impacted by the arterial-input-function dispersion effect. This study introduces an unsupervised model-averaging technique (K-SOM) to estimate the contribution of different nested-models in PK analysis and provides a faster estimate of permeability parameters.
动态对比增强（DCE）-MRI分析的当前最佳实践是从嵌套模型层次结构中逐体素地进行模型选择。这种嵌套模型选择（NMS）假设体素内造影剂（CA）浓度的观察时间轨迹对应于单个生理嵌套模型。然而，不同模型的混合可能存在于体素的CA时间轨迹中。本研究引入了一种无监督特征工程技术（Kohonen自组织映射（K-SOM））来估计每个嵌套模型的体素级概率。66只免疫受损的RNU大鼠植入了人U-251 N癌细胞，并从所有大鼠大脑中获取了DCE-MRI数据。计算了所有动物脑体素纵向弛豫率变化（ΔR）的时间轨迹。使用NMS进行DCE-MRI药代动力学（PK）分析，以估计三个模型区域：模型1：无渗漏的正常血管系统；模型2：有渗漏但无血管反流的肿瘤组织；模型3：有渗漏和血管反流的肿瘤血管。使用大约23万（229,314）个动物脑体素的归一化ΔR曲线及其NMS结果构建一个K-SOM（拓扑大小：8×8，采用竞争学习算法）和每个模型的概率图。使用K折嵌套交叉验证（NCV，k = 10）来评估K-SOM概率-NMS（PNMS）技术相对于NMS技术的性能。K-SOM PNMS对渗漏肿瘤区域的估计与各自的NMS区域非常相似（模型2和3的Dice相似系数，DSC分别为0.774 [CI：0.731 - 0.823]和0.866 [CI：0.828 - 0.912]）。两种技术估计的渗透参数的平均百分比差异（MPD，NCV，k = 10）分别为：v为-28%，K为+18%，v为+24%。KSOM-PNMS技术产生的微血管参数和NMS区域受动脉输入函数弥散效应的影响较小。本研究引入了一种无监督模型平均技术（K-SOM）来估计不同嵌套模型在PK分析中的贡献，并提供了更快的渗透参数估计。
