Abstract
Studies have shown that tumor angiogenesis is an essential process for tumor growth, proliferation and metastasis. Also, tumor angiogenesis is an important prognostic factor of clear cell renal cell carcinoma (ccRCC), as well as a factor in guiding treatment with antiangiogenic agents. Here, we attempted to find the associations between tumor angiogenesis and radiomic imaging features from PET/MRI. Specifically, sparse canonical correlation analysis was conducted on 3 feature datasets (i.e., radiomic imaging features, tumor microvascular density (MVD), and vascular endothelial growth factor (VEGF) expression) from 9 patients with primary ccRCC. In order to overcome the potential bias of intratumoral heterogeneity of angiogenesis, this study investigated the relationship between regional expressions of angiogenesis and VEGF, and localized radiomic features from different parts within the tumors. Our study highlighted the significant strong correlations between radiomic features and MVD, and also demonstrated that the spatiotemporal features extracted from DCE-MRI provided stronger radiomic correlation to MVD than the textural features extracted from Dixon sequences and FDG PET. Furthermore, PET/MRI, which takes advantage of the combined functional and structural information, had higher radiomics correlation to MVD than solely utilizing PET or MRI alone.
Introduction
Clear-cell renal-cell carcinoma (ccRCC) is characterized by significant intratumoral heterogeneity. Different regions within a single tumor can exhibit diverse genetic and vascular profiles. Consequently, traditional single-needle biopsies often suffer from "sampling bias," failing to represent the entire tumor's biological landscape. This limitation complicates the prediction of patient responses to antiangiogenic therapies designed to inhibit tumor blood supply.
This proof-of-concept study explores whether radiomics—the quantitative extraction of high-dimensional features from medical imaging—can serve as a non-invasive proxy for these internal biological markers. Specifically, the researchers aimed to correlate radiomic signatures with microvascular density (MVD) and vascular endothelial growth factor (VEGF) expression.
Methods and Materials
The study involved 9 patients with histologically confirmed primary ccRCC. To address intratumoral heterogeneity, 43 distinct tissue samples were collected. Each physical sample was meticulously matched to a specific "Region of Interest" (ROI) on pre-operative hybrid PET/MRI scans to ensure spatial correlation between imaging and histology.
Imaging Protocol
The hybrid PET/MRI system captured multi-modal data:
- Dynamic Contrast-Enhanced (DCE)-MRI: Used to extract spatiotemporal features by monitoring contrast agent kinetics (wash-in/wash-out).
- Dixon Sequences: Employed to quantify lipid accumulation by separating fat and water signals.
- $^{18}$F-FDG PET: Utilized to measure glucose metabolic activity.
Radiomic and Statistical Analysis
A total of 168 radiomic features (42 spatiotemporal and 126 textural) were extracted. To link these features with biological ground truth (MVD via CD31 staining and VEGF mRNA expression), the team utilized Sparse Canonical Correlation Analysis (SCCA). SCCA is particularly effective for high-dimensional datasets with small sample sizes, as it identifies the strongest linear relationships while filtering out irrelevant variables.
Results
The analysis yielded several key insights into the relationship between imaging and tumor biology:
- Significant MVD Correlation: SCCA demonstrated a statistically significant correlation between combined radiomic features and MVD ($p = 0.044$).
- Superiority of Hybrid Imaging: PET/MRI provided higher correlation coefficients for MVD than standalone DCE-MRI, Dixon MRI, or PET imaging alone, highlighting the synergy of metabolic and functional data.
- Spatiotemporal Dominance: Features describing how a tumor enhances over time (e.g., absolute peak enhancement and wash-in slopes) were stronger predictors of MVD than static textural features.
- VEGF Trends: Although the correlation between radiomics and VEGF expression was high ($\rho_1 = 0.756$), it did not reach statistical significance ($p = 0.169$), likely due to the reduced sample size for genetic analysis (23 samples).
Discussion and Conclusion
This study demonstrates that PET/MRI radiomics can non-invasively characterize the internal vascular landscape of ccRCC. By correlating regional imaging features with localized tissue samples, the authors showed that radiomics can capture the nuances of heterogeneous masses, potentially overcoming the limitations of traditional biopsies.
Limitations: The small cohort ($n=9$) and missing VEGF data for some samples limit generalizability.
Future Work: Validation in larger, multi-center cohorts is necessary. Future research should also focus on developing predictive models to identify patients most likely to benefit from specific antiangiogenic treatments and improving the temporal resolution of DCE-MRI to better capture rapid vascular dynamics.
Summary updated