Cross-visit tumor sub-segmentation and registration with outlier rejection for dynamic contrast-enhanced MRI time
G A Buonaccorsi1, C J Rose, J P B O'Connor
1Imaging Science and Biomedical Engineering, School of Cancer and Imaging Sciences, University of Manchester, Manchester, United Kingdom.
Abstract:
Clinical trials of anti-angiogenic and vascular-disrupting agents often use biomarkers derived from DCE-MRI, typically reporting whole-tumor summary statistics and so overlooking spatial parameter variations caused by tissue heterogeneity. We present a data-driven segmentation method comprising tracer-kinetic model-driven registration for motion correction, conversion from MR signal intensity to contrast agent concentration for cross-visit normalization, iterative principal components analysis for imputation of missing data and dimensionality reduction, and statistical outlier detection using the minimum covariance determinant to obtain a robust Mahalanobis distance. After applying these techniques we cluster in the principal components space using k-means. We present results from a clinical trial of a VEGF inhibitor, using time-series data selected because of problems due to motion and outlier time series. We obtained spatially-contiguous clusters that map to regions with distinct microvascular characteristics. This methodology has the potential to uncover localized effects in trials using DCE-MRI-based biomarkers.
Insights
This study introduces a new method for analyzing dynamic contrast-enhanced MRI (DCE-MRI) data, revealing distinct microvascular characteristics within tumors. This approach overcomes limitations of whole-tumor analysis, offering insights into localized treatment effects.
Area of Science:
- Medical Imaging
- Oncology
- Biophysics
Background:
- Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) is crucial for assessing anti-angiogenic and vascular-disrupting agents in clinical trials.
- Current methods often rely on whole-tumor summary statistics, potentially missing spatially heterogeneous microvascular changes.
- Tissue heterogeneity can lead to significant variations in DCE-MRI derived biomarkers, complicating treatment response assessment.
Purpose of the Study:
- To develop and validate a data-driven segmentation method for DCE-MRI analysis that accounts for spatial parameter variations.
- To improve the robustness and accuracy of DCE-MRI biomarkers by addressing motion and data quality issues.
- To identify distinct microvascular characteristics within tumors that correlate with treatment response.
Main Methods:
- A novel data-driven segmentation method was developed, incorporating tracer-kinetic model-driven registration for motion correction.
- Techniques included conversion from MR signal intensity to contrast agent concentration for normalization, iterative principal components analysis for data imputation and dimensionality reduction, and minimum covariance determinant for outlier detection.
- K-means clustering was applied in the principal components space to identify distinct regions.
Main Results:
- The methodology successfully generated spatially contiguous clusters representing regions with unique microvascular properties.
- Applied to a clinical trial of a VEGF inhibitor, the method effectively handled challenging data with motion and outlier time series.
- The identified clusters demonstrated distinct microvascular characteristics, suggesting localized treatment effects.
Conclusions:
- The developed DCE-MRI analysis methodology can uncover localized treatment effects often missed by traditional whole-tumor analysis.
- This approach enhances the utility of DCE-MRI biomarkers in clinical trials, particularly for targeted therapies.
- The technique offers a more nuanced understanding of tumor microvasculature and its response to therapeutic interventions.
