TL;DR
This study proposes a sequential approach combining Dynamic Susceptibility Contrast (DSC)-MR Perfusion (PWI) and dynamic [<sup>18</sup>F]FET PET imaging to accurately distinguish between tumor progression (TP) and treatment-related changes (TRC) in glioma patients. The sequential strategy, starting with PWI and then using [<sup>18</sup>F]FET PET for ambiguous cases, achieved an overall diagnostic accuracy of 87%, offering a more cost-effective and efficient diagnostic pathway.
What Problem Does This Paper Address?
Diagnosing glioma progression is critical for timely treatment adjustments. However, distinguishing true tumor progression (TP) from treatment-related changes (TRC), such as radiation necrosis or inflammation, is a significant challenge in neuro-oncology. Conventional MRI, relying on morphological changes, often struggles with this differentiation. While both DSC-MR Perfusion (PWI) and dynamic [<sup>18</sup>F]FET PET are known to be valuable tools, their combined diagnostic performance, especially when implemented sequentially, has not been thoroughly investigated. This paper aims to address this gap by evaluating the efficacy of a step-by-step approach using these two imaging modalities.
Methods
The study retrospectively analyzed data from 104 patients with suspected recurrent glioma. The core methodology involved a sequential imaging strategy:
- DSC-MR Perfusion (PWI): This technique measures blood flow in the brain. The key parameter extracted was the maximum relative cerebral blood volume (rCBVmax), calculated by comparing tumor blood volume to that of normal brain tissue.
- Dynamic [<sup>18</sup>F]FET PET: This radiotracer accumulates in cells with high amino acid transporter expression, which is often elevated in tumors. Parameters analyzed included static measures like the maximum tumor-to-brain ratio (TBRmax) and dynamic measures like the slope of the uptake curve over time.
The diagnostic performance of each modality and their sequential combination was evaluated using Receiver Operating Characteristic (ROC) analysis to determine optimal cutoffs for differentiating TP from TRC.
Datasets
The study utilized data from 104 patients with suspected recurrent glioma who underwent both PWI and dynamic [<sup>18</sup>F]FET PET imaging. The final diagnosis for these patients was established through histopathology (n=42) or clinico-radiological follow-up (n=62). No new benchmark datasets were introduced.
Key Insights and Ideas
- Sequential Power: The most significant insight is the effectiveness of a sequential approach. Starting with PWI and then employing [<sup>18</sup>F]FET PET for cases where PWI is inconclusive significantly enhances diagnostic accuracy and efficiency.
- PWI as a Triage Tool: A high rCBVmax value on PWI (above 2.85) was highly specific for tumor progression, allowing for a definitive diagnosis in a substantial portion of patients (42%) without the need for further, more expensive PET imaging.
- [<sup>18</sup>F]FET PET's Complementary Role: When PWI results were ambiguous, [<sup>18</sup>F]FET PET parameters, particularly the combination of TBRmax and Slope, proved highly effective in differentiating TP from TRC, correctly classifying an additional 45% of patients.
- IDH Mutation Impact: The study observed an interesting trend where PWI appeared more reliable than [<sup>18</sup>F]FET PET in IDH-mutant gliomas, a finding that warrants further investigation.
- Cost-Effectiveness: The sequential strategy is proposed as a resource-sparing and cost-effective method, reducing the overall burden of diagnostic imaging.
Key Results
- Overall Accuracy: The sequential combination of PWI and [<sup>18</sup>F]FET PET achieved an overall diagnostic accuracy of 87%.
- PWI's Initial Contribution: Abnormal PWI (rCBVmax > 2.85) correctly identified TP in 42% of patients with 100% specificity and positive predictive value.
- [<sup>18</sup>F]FET PET's Additional Contribution: For the remaining 58% of patients where PWI was inconclusive, the combined use of [<sup>18</sup>F]FET PET parameters (TBRmax and Slope) correctly classified an additional 45% of cases, leading to an overall accuracy of 87%.
- Individual Parameter Performance: While individual parameters showed moderate diagnostic performance (AUC 0.69-0.75), their sequential and combined use significantly improved outcomes.
- IDH-Wild Type vs. Mutant: In IDH-wild-type tumors, PWI parameters showed slightly improved ROC curves, while in IDH-mutant gliomas, rCBVmax performance increased significantly.
Limitations and Future Work
- Inhomogeneous Cohort: The study included patients with various glioma grades and treatment regimens, making the cohort inhomogeneous.
- Selection Bias: Patients referred for [<sup>18</sup>F]FET PET imaging might represent more challenging or ambiguous cases, potentially biasing the results.
- Non-Simultaneous Imaging: The PWI and [<sup>18</sup>F]FET PET scans were not performed simultaneously, with a median interval of 11.5 days, which could theoretically influence results.
- MR Scanner Variability: The use of different MR scanners and protocols for PWI might introduce some variability, although efforts were made to standardize analysis.
- IDH Mutation Further Study: The finding regarding the differential performance of PWI and [<sup>18</sup>F]FET PET in IDH-mutant gliomas is surprising and requires further validation.
- Future Research: The authors suggest that future research should focus on validating the findings in IDH-mutant gliomas and potentially implementing the new consensus PWI protocol for greater reproducibility. They also note that while quantitative parameters are valuable, integrating morphologic appearance and tracer distribution, as considered by experienced clinicians, could further enhance diagnostic accuracy.
Summary updated