In this work, we identify a problem with the process of volume-to-surface mapping of functional Magnetic Resonance Imaging (fMRI) data when interested in investigating local connectivity. We show that neighborhood correlations on the surface of the brain vary spatially with anatomical patterns (gyral structure), even when the underlying volumetric data are uncorrelated noise. This could potentially have impacted studies focusing upon local neighborhood connectivity. We explore the effects of this anomaly across varying data resolutions and surface mesh densities, and propose an approach to mitigate these unwanted effects.
To date, the best-performing blind super-resolution (SR) techniques follow one of two paradigms: (A) train standard SR networks on synthetic low-resolution–high-resolution (LR–HR) pairs or (B) predict the degradations of an LR image and then use these to inform a customised SR network. Despite significant progress, subscribers to the former miss out on useful degradation information and followers of the latter rely on weaker SR networks, which are significantly outperformed by the latest architectural advancements. In this work, we present a framework for combining any blind SR prediction mechanism with any deep SR network. We show that a single lightweight metadata insertion block together with a degradation prediction mechanism can allow non-blind SR architectures to rival or outperform state-of-the-art dedicated blind SR networks. We implement various contrastive and iterative degradation prediction schemes and show they are readily compatible with high-performance SR networks such as RCAN and HAN within our framework. Furthermore, we demonstrate our framework’s robustness by successfully performing blind SR on images degraded with blurring, noise and compression. This represents the first explicit combined blind prediction and SR of images degraded with such a complex pipeline, acting as a baseline for further advancements.
1 September 2026 in IEEE Transactions on Radiation and Plasma Medical Sciences
The BrainPET-7T insert is designed for simultaneous human-brain PET imaging with ultra-high-field MRI. It offers a high sensitivity and signal-to-noise ratio by detector blocks of pixelated LSO scintillators with a height of 2.4cm, digital silicon photomultiplier readout, and enhanced data processing methods. This makes it particulary interesting for +--emitter applications, such as dual-isotope imaging, which rely on the detection of triple coincidences. We investigated the scanner sensitivity using the isotopes 124I and 22Na, in comparison to 18F. Count rate performance was assessed using 124I and 18F for different energy window settings with an analysis following the NEMA NU 2-2018 protocol. For the count rate determination, an adapted strategy for estimating the scatter fraction for +--emitters was applied. The peak double coincidence sensitivity for 18F measured at the center and off-center was between 10.16(8)% and 12.37(10)% for energy windows of 440 to 594keV and 420 to 650keV. The true triple sensitivities obtained at the isocenter were 3.13(9)% and 4.88(12)% for 124I and between 0.8662(23)% and 3.112(5)% for 22Na depending on the selected energy window and rejection rule. The maximum RNEC of the 18F-measurement was 276kcps at 16.3kBqmL-1, while the maximum Rtrue was 780kcps at 22.7kBqmL-1 for an energy window of 440 to 594keV. The scatter fraction ranged from 28% to 33% and the random fraction from 3% to 40%. In the case of 124I, the ratio of estimated true triples to doubles was ≈6%, with a higher random fraction of 27% for triples compared to 2.8% for doubles. These results demonstrate the potential feasibility of employing the BrainPET-7T insert for +--emitter applications.
Objective. Inter-crystal scattering (ICS) degrades spatial resolution and contrast by causing spurious lines of response in small-pitch, staggered three-layer pixelated LSO scintillation detectors used in a high-resolution BrainPET-7T insert. This work aimed to investigate the distribution of ICS events in our BrainPET-7T detector and understand its underlying mechanisms. Approach. Six physical approaches for computing probabilities to identify the first interaction positions of ICS events were studied using Geant4 simulations, focusing specifically on events with two interactions, which represent the most frequent ICS occurrence. Performance was evaluated using ideal interaction positions and energies directly from simulations, ground truth crystal-center positions with ideal energies, as well as crystal-center positions and energies determined by a maximum likelihood positioning (MLP) algorithm. Main results. A method based on Klein-Nishina probabilities, which were computed using geometric information (KN-G) and MLP-derived crystal-center positions and energies, achieved superior accuracy, particularly when the distance between successive interactions in Compton events was shorter than 6 mm and when incident angles were below 15 • . This advantage is partly attributable to crystal-center discretization, which masks a substantial portion of backscattering events. It reached a peak accuracy of 92.2 % at a distance of 19 mm. Significance. These findings demonstrate the feasibility of combining the KN-G method with the MLP algorithm to mitigate the ICS effects, with the potential to improve the image quality of the 7T BrainPET insert.
Background: Most studies evaluating O-(2-[18F]-fluoroethyl)-L-tyrosine (FET) PET for glioblastoma progression rely on single-scan assessments, leaving diagnostic uncertainties in cases with borderline tracer uptake. We evaluated metabolic activity evolution during the transition of glioblastoma progression using serial FET PET to determine whether longitudinal parameter changes improve diagnostic certainty. Patients and Methods: Between 2016 and 2026, 182 patients with glioblastoma undergoing 690 dynamic FET PET scans (median, 3; range, 2-12) were retrospectively evaluated. Eligible cases required documented progressive disease according to RANO 2.0 criteria and serial FET PET at least 2 time points (at suspected progression and within a 6-month prior window). Static PET parameters, ie tumor-to-brain ratios (TBRmean, TBRmax) and metabolic tumor volume (MTV), and their longitudinal changes were evaluated according to PET RANO 1.0 criteria, alongside additional dynamic PET parameters (time-to-peak). Single-scan metrics and serial changes were compared in cases with borderline uptake. Results: Ninety patients with 102 confirmed progression events met all selection criteria (median interscan interval, 2.7 months). At progression, TBRmean, TBRmax, and MTV increased significantly (P < .001). Median percentage increases were modest for TBRmean (+4%) and TBRmax (+17%), whereas MTV exhibited a substantial expansion (+182%). In cases with borderline uptake, applying a longitudinal MTV increase criterion of ≥40% per PET RANO 1.0 substantially improved progression detection to up to 89%, yielding a 14% higher diagnostic detection rate compared to single-scan thresholds. Conclusions: In glioblastomas with documented RANO 2.0 progression, integrating longitudinal MTV changes markedly improves diagnostic certainty over single-scan uptake metrics, particularly in challenging cases with borderline uptake.
To date, the best-performing blind super-resolution (SR) techniques follow one of two paradigms: A) generate and train a standard SR network on synthetic low-resolution - high-resolution (LR - HR) pairs or B) attempt to predict the degradations an LR image has suffered and use these to inform a customised SR network. Despite significant progress, subscribers to the former miss out on useful degradation information that could be used to improve the SR process. On the other hand, followers of the latter rely on weaker SR networks, which are significantly outperformed by the latest architectural advancements. In this work, we present a framework for combining any blind SR prediction mechanism with any deep SR network, using a metadata insertion block to insert prediction vectors into SR network feature maps. Through comprehensive testing, we prove that state-of-the-art contrastive and iterative prediction schemes can be successfully combined with high-performance SR networks such as RCAN and HAN within our framework. We show that our hybrid models consistently achieve stronger SR performance than both their non-blind and blind counterparts. Furthermore, we demonstrate our framework's robustness by predicting degradations and super-resolving images from a complex pipeline of blurring, noise and compression.