Differentiating true progression from pseudoprogression in glioblastomas: Dynamic contrast-enhanced MRI habitat analysis with parsimonious pharmacokinetic model.
Authors
Affiliations (4)
Affiliations (4)
- Institute for Advancing Intelligence, TCG Centres for Research and Education in Science and Technology, Kolkata, India.
- Academy of Scientific and Innovative Research (AcSIR), Ghaziabad, India.
- Department of Radiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA.
- Centre for Preclinical Imaging, University of Liverpool, Liverpool, UK.
Abstract
Distinguishing true progression (TP) from pseudoprogression (PsP) in glioblastoma (GBM) remains a clinical challenge as both often manifest as indeterminate contrast-enhancing lesions on post-treatment MRIs. However, integrating dynamic contrast-enhanced (DCE)-MRI with analytical methods such as clustering and machine learning offers a promising approach for distinguishing TP from PsP. Quantitative analysis of DCE-MRI data from 82 patients with TP (<i>n</i> = 53) and PsP (<i>n</i> = 29) was performed using 5 pharmacokinetic models. Subsequently, a parsimonious selection method was used to identify the best fit for each voxel. In the next step, clustering analysis using K-means++ was performed to classify pixels into low-, medium-, and high-intensity regions. Tumors were segmented into enhancing, non-enhancing, and edema regions using automated nnU-Net algorithm, and clusters were extracted from these regions. Significant features derived from clustering analysis and descriptive statistics (mean, median, and skewness) of DCE-MRI parameters and tumor volume were incorporated as input variables for classifier development. Finally, 5 machine learning classifiers were employed to differentiate between TP and PsP. Univariable clustering analysis of DCE-MRI data identified K<sup>trans</sup>, V<sub>e</sub>, V<sub>p</sub>, τ<sub>i</sub>, and tumor volumes as significant parameters for distinguishing TP from PsP. Multivariable clustering revealed that combined high-K<sup>trans</sup>-high-V<sub>e</sub> is a key distinctive variable. Among machine learning classifiers tested using these parameters, Random Forest demonstrated the best diagnostic performance, achieving an accuracy of 83% and an AUC of 0.85. Parsimonious analysis of DCE-MRI data in conjunction with clustering and machine learning could serve as a valuable method for accurately distinguishing TP from PsP in GBMs.