Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN, United States; Chemical and Physical Biology Program, Vanderbilt University, Nashville, TN, United States.
Institute of Imaging Science, Vanderbilt University Medical Center, Nashville, TN, United States; Department of Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, United States.
Magn Reson Imaging. 2022 Dec;94:144-150. doi: 10.1016/j.mri.2022.10.002. Epub 2022 Oct 6.
It remains a clinical challenge to differentiate brain tumors from radiation-induced necrosis in the brain. Despite significant improvements, no single MRI method has been validated adequately in the clinical setting.
Multi-parametric MRI (mpMRI) was performed to differentiate 9L gliosarcoma from radiation necrosis in animal models. Five types of MRI methods probed complementary information on different scales i.e., T (relaxation), CEST based APT (probing mobile proteins/peptides) and rNOE (mobile macromolecules), qMT (macromolecules), diffusion based ADC (cell density) and SSIFT iAUC (cell size), and perfusion based DSC (blood volume and flow).
For single MRI parameters, iAUC and ADC provide the best discrimination of radiation necrosis and brain tumor. For mpMRI, a combination of iAUC, ADC, and APT shows the best classification performance based on a two-step analysis with the Lasso and Ridge regressions.
A general mpMRI approach is introduced to choosing candidate multiple MRI methods, identifying the most effective parameters from all the mpMRI parameters, and finding the appropriate combination of chosen parameters to maximize the classification performance to differentiate tumors from radiation necrosis.
区分脑肿瘤和脑放射性坏死仍然是临床面临的挑战。尽管有了显著的进步，但在临床环境中，没有一种单一的 MRI 方法得到充分验证。
对动物模型中的 9L 胶质肉瘤和放射性坏死进行多参数 MRI（mpMRI）检查，以区分开来。五种类型的 MRI 方法从不同尺度探测互补信息，即 T（弛豫）、CEST 基于 APT（探测移动蛋白/肽）和 rNOE（移动大分子）、qMT（大分子）、基于扩散的 ADC（细胞密度）和 SSIFT iAUC（细胞大小）以及基于灌注的 DSC（血容量和血流）。
对于单个 MRI 参数，iAUC 和 ADC 可最好地区分放射性坏死和脑肿瘤。对于 mpMRI，使用 Lasso 和 Ridge 回归的两步分析，基于 iAUC、ADC 和 APT 的组合显示出最佳的分类性能。
介绍了一种通用的 mpMRI 方法来选择候选的多种 MRI 方法，从所有 mpMRI 参数中确定最有效的参数，并找到所选参数的适当组合，以最大限度地提高区分肿瘤和放射性坏死的分类性能。
