Alberto De Luca

Background and aim: In memory clinic patients, assessing diagnostic relevance of white matter hyperintensities (WMHs), as a marker to explain cognitive deficits, can be challenging. Here, we aimed to quantify this relevance by measuring the variance in cognitive function they explain, and consider, beyond total WMH volume, WMH location in specific tracts and normative WMH values. Methods: We analyzed previously pooled and harmonized individual patient data (n=3,481, 71±9 years old, 50% females) from 11 international memory clinic cohorts from the Meta-VCI-Map consortium. Cognitive function was assessed in 4 domains: attention and executive function, information processing speed, verbal memory and language. We measured WMH volumes in 20 white matter tracts. These were also norm-adjusted for age and sex, using normative WMH values from 14,876 community-dwelling individuals. Explained variance in cognitive function was measured for demographics only (age, sex and education) and demographics combined with total WMH volume, tract-specific WMH volumes and norm-adjusted and tract-specific WMH z-scores. Results: Beyond demographics, total and tract-specific WMH volumes explained significantly more variance in attention and executive functioning (ΔR2total=3.7%, 95%CI=[3.3%;4.1%], ΔR2tracts=6.2%, 95%CI=[5.6%;6.7%]) and verbal memory (ΔR2total=3.6%, 95%CI=[3.2%;3.9%], ΔR2tracts=8.0%, 95%CI=[7.5%;8.6%]). In processing speed, tract-specific (ΔR2=2.5%, 95%CI=[2.1%;3.0%]) —but not total— WMH volumes explained more variance than demographics. Norm-adjusting tract-specific volumes explained no additional variance. Conclusion: Considering location of WMHs increased their diagnostic relevance, compared to total volume, explaining more variance in attention and executive function, processing speed and verbal memory. This advocates considering WMH location beyond total volume as a diagnostic marker.

Tractography is a key component of efforts to map brain connectivity. As a rapidly-evolving field of neuroscience, current tractography methods are diverse, often varying across research laboratories and different software pipelines. Therefore, it suffers from a lack of standardization leading to inconsistencies in results, which can limit reproducibility, and affect the robustness needed for research and clinical applications of these methods. Variability in data acquisition procedures, inconsistencies in spatial referencing schemes and implementations, and anatomical heterogeneity —at the individual level, across the lifespan, and across species— hinders comparative analyses. Additionally, the lack of consensus on best practices complicates the development of robust automated quality control pipelines and limits the clinical translation of tractography-based procedures. Establishing standardized protocols for acquisition, preprocessing, and tractography reconstruction are critical towards enabling reliable tract-specific analyses, facilitating cross-study harmonization, and supporting replicable large-scale population studies. The present article provides an overview of the current challenges in tractography standardization and identifies the key aspects that require standardization for reliable, reproducible, and robust tractography.

Brain connectivity, quantified with diffusion MRI (structural connectivity, SC) and resting‐state functional MRI (functional connectivity, FC), can offer crucial insights into glioma‐brain network interactions. Currently, no standardized approach exists to integrate information from FC and SC and to identify potential tumor‐induced abnormalities at the single‐patient level. Variational autoencoders (VAEs) have been shown to be promising for learning the distribution of features representing a healthy brain and deviations thereof and can naturally be applicable to multiple modalities. This study explores the potential of VAE to integrate FC and SC and detect multimodal anomalies in brain connectivity in glioma patients. The VAE is trained on concatenated FC‐SC healthy data to learn how to reconstruct normative connectivity patterns. After ad hoc transfer learning, the model parameters are applied to the oncological dataset, to obtain the healthy version of the pathological matrices. Given the healthy, pathological, and reconstructed matrices, a statistic is developed with the goal of identifying specific alterations in SC, FC, and their FC + SC integration in glioma patients. SC, FC, and FC + SC abnormalities are compared with each other to explore their interplay and their link with tumor and surrounding brain tissues. Results show that FC is more sensitive to alterations distant from the tumor, while SC is more affected in its vicinity. Then, the alterations identified by FC are generally more in agreement with the alterations identified by FC + SC compared with those highlighted by SC. Moreover, SC abnormalities never overlap with FC + SC out of the tumor, and FC and SC single impairments partially overlap within the tumor core and never overlie in other brain tissues. This information could facilitate patient stratification, prognostic modeling, and personalized treatment planning.

Cerebellar mutism syndrome (CMS) is a potential complication of pediatric posterior fossa tumor (pPFT) resection and may be related to disruption of the dentato‐rubro‐thalamic tract (DRTT). Intraoperative tractography allows assessment of changes to the DRTT during surgery. We evaluated macro‐ and microstructural changes of the DRTT by comparing pre‐ and intraoperative diffusion MRI (dMRI) tractography in patients with pPFT. Pre‐ and intraoperative T1‐weighted and dMRI data were acquired to reconstruct the DRTT, while the arcuate fasciculus (AF) was reconstructed as a control tract. To evaluate macrostructural intraoperative alterations, tract volume and diameter were calculated. Microstructural changes were assessed using fractional anisotropy (FA) and mean diffusivity (MD), both globally and along the tract. Speech disturbances, as a characteristic symptom of CMS, were scored preoperatively and at hospital discharge, based on retrospective neurological reports. In 30 patients (aged > 21 months) with sufficient data quality, intraoperative reductions in tract volumes and diameters were observed ( p < 0.05) in the decussating DRTT (d‐DRTT) (−45.4% to −10%) and AF (−9.9% to −5.1%), while the non‐decussating DRTT (nd‐DRTT) remained stable. Right d‐DRTT volume loss correlated with ventricular enlargement (r = −0.476, p = 0.022). Whole‐tract analyses revealed increased FA in the AF and increased MD in the nd‐DRTT and AF. Along‐tract analyses demonstrated high variance of the MD in the DRTT cerebellar segments. Patients with postoperative speech disturbances showed higher variance of FA and MD along the DRTT, particularly in the left cerebellar segment. pPFT surgery affected the variation of the DRTT microstructure near the resection cavity, whereas the AF remained relatively stable. Although we did not find significant differences in patients with postoperative speech disturbances, we observed a much higher variability in MD in this group, suggesting a potential effect of DRTT disruption in these symptoms.

Tractography is a key component of efforts to map brain connectivity. As a rapidly evolving field of neuroscience, current tractography methods are diverse, often varying across research laboratories and different software pipelines. Therefore, it suffers from a lack of standardization, leading to inconsistencies in results, which can limit reproducibility and affect the robustness needed for research and clinical applications of these methods. Variability in data acquisition procedures, inconsistencies in spatial referencing schemes and implementations, and anatomical heterogeneity—at the individual level, across the lifespan, and across species—hinder comparative analyses. Additionally, the lack of consensus on best practices complicates the development of robust automated quality control pipelines and limits the clinical translation of tractography-based procedures. Establishing standardized protocols for acquisition, preprocessing, and tractography reconstruction is critical toward enabling reliable tract-specific analyses, facilitating cross-study harmonization, and supporting replicable large-scale population studies. The present article provides an overview of the current challenges in tractography standardization and identifies the key aspects that require standardization for reliable, reproducible, and robust tractography.

Surgery plays a key role in treating neuroblastoma. To assist surgical planning, anatomical 3D models derived from the segmentation of anatomical structures on MRI scans are often used. Automation using deep learning can make segmentations less time-consuming and more reliable. We organized the Surgical Planning in PedIatric Neuroblastoma (SPPIN) challenge, to stimulate developments and benchmarking of automatic segmentation of neuroblastoma on MRI. SPPIN is the first segmentation challenge in extracranial pediatric oncology. Nine teams provided a valid submission. Evaluation was based on the Dice similarity coefficient (Dice score), the 95th percentile of the Hausdorff distance (HD95), and the volumetric similarity (VS). A combination of these scores determined the ranking of the teams. The spread in the median evaluation scores per team was large (Dice: 0.21–0.82; HD95: 63.31–7.69; VS: 0.31–0.91). The top-performing team achieved a median Dice score of 0.82 (with an HD95 of 7.69 mm and a VS of 0.91) using a large, pre-trained model. However, in the pre-operative segmentations, significantly lower evaluation scores were observed. Our results indicate that pre-training might be useful in small, pediatric datasets. Although the general results of the winning team were high, they were insufficient to use for surgical planning in small, pre-operative tumors.

Objective: Intraoperative fiber tractography can help neurosurgeons in localizing eloquent tracts potentially displaced by tumors. In pediatric posterior fossa tumor (pPFT) patients, disturbances to the eloquent dentato-rubro-thalamic tract (DRTT) might contribute to cerebellar mutism syndrome. This study investigates the effect of fiber tractography parameters on DRTT reconstruction in pre- and intraoperative settings. Methods: T1-weighted and diffusion MRI data were acquired from ten pPFT patients and two healthy volunteers. The patients were scanned pre- and intraoperatively. The DRTT was reconstructed using multiple fiber orientation distribution (FOD) and angle thresholds. An expert panel evaluated tract reconstructions to identify optimal parameters in our dataset. The corticospinal tract (CST) served as a control. Relative tract volumes of the DRTT and the CST were calculated. Results: Diffusion MRI data were sufficient for reliable DRTT reconstruction in healthy volunteers. In most pPFT patients, an FOD of 0.01 and a 60° angle threshold were evaluated as optimal for DRTT reconstruction in our dataset. Preoperative DRTT reconstructions showed more reconstructed streamlines and larger relative volumes, particularly in non-decussating tracts. CST reconstructions remained consistent across both timepoints. Discussion: DRTT reconstruction is feasible in pPFT patients before and during surgery. However, inter-subject variability suggests that some patients may require adjusted thresholds for optimal results.

Purpose: Diffusion weighting is achieved by the application of external field gradients typically for tens of milliseconds, during which the signal also substantially decays due to inherent T₂ relaxation. Employing strong gradients will shorten the echo time (TE) and increase the signal‐to‐noise ratio. This study demonstrates a proof‐of‐principle of using an ultrastrong head gradient insert for diffusion MRI at 7 T.Methods: A lightweight (45 kg) single‐axis gradient coil operating in the z‐direction was interfaced with a 7 T system and operated as an additional fourth‐gradient axis to the whole‐body gradient setup. Several diffusion MRI experiments were conducted to compare acquisitions taken with a gradient insert operating at full capacity (200 mT/m and 1300 T/m/s) to those using conventional gradient strengths (40–80 mT/m and 200 T/m/s).Results: A shorter TE of (28.8, 32.4, 38.5, 48.7) ms was achieved using the insert gradient at full capability for readout and diffusion encoding, compared to (49.7, 54.5, 65.9, 85.5) ms (80 mT/m) or (59.9, 68.8, 88.3, 120.9) ms (40 mT/m) for b = (0, 500, 1000, 3000, 10 000) s/mm² using conventional gradient strength. A SNR increase for all b‐value acquisitions was observed. When using strong gradients for readout, TE was shortened by 20 ms, and a lower degree of geometrical distortions was observed.Conclusion: The results illustrate a successful proof‐of‐concept for performing diffusion MRI using a plug‐and‐play head gradient insert at 7 T. Use of low‐cost gradient inserts could make advanced diffusion MRI experiments more widely available.

Tractography involves many steps, and as such can be heavily affected by data quality, operator decisions such as algorithm and parameter selection and/or biological variables - in addition to random effects such as noise and motion. Due to the interplay between sources of variability, even seemingly minor decisions or overlooked assumptions can lead to major variations in output that can potentially confound tractography results. This article summarizes the main sources of variability in tractography to remind both new and experienced users of the importance of making informed decisions at all stages of the tractography pipeline.

Background: The apparent diffusion coefficient (ADC) derived from diffusion-weighted imaging (DWI) is a potential biomarker for treatment response in pediatric rhabdomyosarcoma. Due to its rarity, investigations into this marker require multicenter approaches, which can result in variability in acquisition parameters. Objective: To evaluate the impact of different acquisition parameters on ADC estimates in a multicenter dataset of rhabdomyosarcoma patients. Materials and methods: We included 114 pediatric and adolescent rhabdomyosarcoma patients from 22 treatment centers (195 scans). Median age: 6.0 years (0.3–21.8). We evaluated the impact of voxel size, (number of) b-values, and echo time on tumor ADC values. The effect of the highest b-value was separately investigated on a subset of scans with five or more b-values. Results: We observed a large variability in key acquisition parameters in the overall cohort, and for individual imaging centers. No parameter showed a significant effect on ADC estimates of the whole cohort when corrected for multiple-comparisons. Decreasing the highest b-value within the same acquisition caused ADC to decrease on average by 2.8% per 100 s mm^-2. Differing b-values between scans at diagnosis and treatment response yielded significant changes in the longitudinal ADC for each patient (Pb-value between baseline and follow-up can impact longitudinal ADC estimates. As such, we recommend the highest b-value to remain constant. This retrospective study was reviewed and approved by the Internal Review Board (UMC Utrecht, reference ID: 18-412). Graphical