TL;DR
This study explores a novel MRI technique (GE-SE EPIK) that can simultaneously measure multiple parameters related to brain tumor oxygenation and metabolism in just 2 minutes. By comparing these MRI-derived measures to PET scans using a common tumor tracer (FET), the researchers found that the MRI approach shows promise for characterizing tumors, though artifacts and treatment effects can complicate interpretation.
What Problem Does This Paper Address?
Understanding the oxygenation status of brain tumors is crucial because tissue hypoxia (low oxygen) is linked to tumor aggressiveness, resistance to therapy, and poorer patient outcomes. Traditionally, measuring the oxygen extraction fraction (OEF), a key indicator of oxygen metabolism, has relied on PET scans using short-lived isotopes like O-15. While effective, these PET methods are logistically complex and require specialized equipment. This paper addresses the need for a faster, more accessible method to assess tumor oxygenation and metabolism. The researchers aim to evaluate if a specialized MRI sequence, GE-SE EPIK, can provide comparable information to PET imaging, specifically by correlating its derived parameters with FET-PET uptake, a marker of increased metabolic activity in brain tumors.
Methods
The core of this study is the application of a 10-echo gradient-echo spin-echo echo-planar imaging with keyhole (GE-SE EPIK) sequence. This sequence is designed to acquire multiple contrasts rapidly, allowing for simultaneous quantification of transverse relaxation times (T2 and T2*), venous cerebral blood volume (vCBV), and oxygen extraction fraction (OEF) within a short acquisition time (approximately 2 minutes). The researchers acquired these GE-SE EPIK data on a hybrid MR-PET scanner during FET-PET scans in 57 brain tumor patients.
Key methodological aspects include:
- GE-SE EPIK Sequence: This sequence is novel in its ability to acquire 10 echoes, providing a rich dataset for parameter estimation. It's designed to be contrast agent-free.
- Parameter Calculation: Voxel-wise T2 and T2* maps were generated by fitting the signal decay. From these, R2' (related to susceptibility effects) was calculated. vCBV was derived using a specific formula, and OEF was calculated based on the qBOLD (quantitative Blood Oxygen Level Dependent) theory.
- Comparison with FET-PET: Tumor regions of interest (VOIs) were defined based on FET-PET uptake and FLAIR hyperintensities. Mean MR-derived parameters were calculated within these VOIs and compared to contralateral healthy tissue. Relative measures (tumor/reference) were then analyzed.
- Statistical Analysis: One- and two-sample t-tests were used to compare MR parameters between tumor types and against reference values.
Datasets
The study utilized data from 57 brain tumor patients with 66 histologically confirmed lesions. The patient cohort included various tumor types: suspected glioblastoma, glioblastoma, astrocytoma, metastasis, and oligodendroglioma. The data collected included:
- MRI Data: Acquired using a 3T hybrid MR/PET scanner, including anatomical sequences (MPRAGE, FLAIR) and the novel 10-echo GE-SE EPIK sequence.
- PET Data: O-(2-[18F]fluoroethyl)-L-tyrosine (FET) PET scans were acquired simultaneously with MRI.
No new benchmarks were introduced; the study focused on comparing existing PET tracers with novel MRI acquisition techniques.
Key Insights and Ideas
- MRI's Potential for Rapid Oxygenation Assessment: The GE-SE EPIK sequence demonstrates the feasibility of obtaining multiple quantitative MRI parameters related to oxygen metabolism in a clinically relevant timeframe (2 minutes). This could offer a more accessible alternative to specialized PET scans.
- Tumor Hypoxia Signatures in MRI: The study found significant changes in MR parameters within tumor regions compared to healthy tissue, including increased T2 and T2* relaxation times and decreased vCBV and OEF. These findings align with expectations for hypoxic tumor tissue.
- OEF as a Potential Biomarker: While not directly correlating with FET uptake, the reduced OEF in tumors suggests its potential as an indicator of altered oxygen metabolism, which is a hallmark of many cancers.
- Limitations of Artifacts: Susceptibility artifacts, particularly in T2* and OEF maps, pose a significant challenge for accurate interpretation and can obscure true tumor characteristics. These artifacts are often found at air-tissue interfaces.
- Treatment Effects Complicate Interpretation: The study highlights that treatment-related changes can lead to global alterations in MR parameters that extend beyond the tumor VOI, making it difficult to distinguish between active tumor recurrence and treatment effects.
Key Results
- Increased T2/T2* and Decreased vCBV/OEF: In tumor VOIs (both FET-PET and FLAIR-derived), relative T2 (rT2) and relative T2 (rT2) were significantly increased compared to contralateral reference values, while relative vCBV (rvCBV) and relative OEF (rOEF) were significantly decreased.
- Weak Correlation with FET Uptake: A significant correlation was found between FET uptake and T2/T2* (Pearson correlation coefficients of 0.26 and 0.31, respectively). However, OEF showed no significant correlation with FET uptake (Pearson correlation coefficient of 0.04).
- Tumor Type Differences: Oligodendrogliomas showed significant differences in rR2' and rOEF compared to glioblastomas and astrocytomas. Suspected gliomas had significantly higher T2 compared to astrocytomas.
- Artifact Impact: Susceptibility artifacts in T2*/OEF maps made differentiating tumor relapse from treatment-related changes challenging.
Limitations and Future Work
- Limitations:
- Small Sample Sizes for Subgroups: The number of tumors, particularly for certain subgroup analyses, was small, limiting the statistical power for differentiating tumor types.
- Heterogeneous Patient Cohort: The cohort included patients with varied treatment histories, which can influence MR parameters and complicate the interpretation of treatment-related changes.
- Susceptibility Artifacts: These artifacts are inherent to the imaging technique and can obscure true tumor characteristics, especially in T2* and OEF maps.
- VCBV Sensitivity: The sensitivity of vCBV quantification was limited by the echo sampling within the static dephasing regime.
- Future Work:
- Improve Image Quality: Enhancing SNR, optimizing acquisition parameters, and employing better shimming are suggested to improve image quality and reduce artifacts.
- Automated Analysis: Developing methods for visual classification to differentiate pathological and artifact-related changes is needed for automated analysis.
- Parameter Combinations: Investigating combinations of MR parameters could improve tumor tissue identification and differentiation.
- Clinical Validation: Future studies should validate OEF quantification against reference-standard PET acquisitions, particularly under hypoxic conditions, and compare it with other MR-based OEF quantification methods.
- Hypoxia Correlation: Further research is needed to confirm the correlation of MR-derived OEF with hypoxia.
Summary updated