Background: This study aimed to develop a fully nested, information leakage-free machine-learning workflow to predict the volumetric response of meningioma to Gamma Knife radiosurgery (GKRS) from pre-treatment MRI and to compare the predictive value of radiomic, fractal, lacunarity and clinical/radiosurgical features. GKRS is widely used for treating meningiomas because of its high precision and efficacy. Variability in tumor volumetric response highlights the need for reliable predictors of treatment outcome. Methods: This retrospective cohort study included 204 patients treated with GKRS for grade I meningioma. Radiomic, fractal and lacunarity features were extracted from pre-treatment CE-T1w 3-Tesla MRIs. Feature signatures were generated using a machine-learning workflow incorporating five feature selectors based on a consensus principle to reduce spurious feature selection, followed by five classifiers to predict binary outcome. Results: The models demonstrated consistent predictive performance in the test folds, with AUC values from 0.77 to 0.84. Supplementing radiomic features with clinical, fractal or lacunarity features did not improve predictive performance. Conclusions: Radiomic features showed the strongest predictive value for meningioma volumetric response to GKRS. Darker intratumoral intensity values were associated with a favorable volumetric response, possibly reflecting biologically less active tumor regions. The supplied code enables individual-level prediction for newly encountered patients.

Transthyretin (TTR) is a 127-amino acid protein that forms a homotetramer, whose main function is to transport thyroxine and retinol-binding protein-retinol (vitamin A) complex 1 .Misfolding of TTR leads to transthyretin amyloidosis (ATTR), which is associated with cardiomyopathy (ATTR-CM) and polyneuropathy (ATTR-PN). There are two types of ATTR: a wild-type (previously called senile; ATTRwt) and a variant-type (genetic variation; ATTRvt) 2 . More than 150 pathogenic TTR proteins have been identified in ATTRvt, the most common being Val122Ile (TTRVal122Ile), and the one with the most penetrance, TTRVal30Met 3 . Furthermore, in a recent survey, TTRVal30Met was responsible for more than half of amyloidosis cases, followed by ATTRwt and TTRVal122Ile, and it was noted that ATTRwt was a more common cause of symptomatic amyloidosis in North America (59.5%) 4 .Under these conditions, an association between ATTR and lumbar spinal stenosis (LSS) has been found, due to the presence of TTR deposition in the ligamentum flavum (LF) in patients undergoing LSS decompression alongside a diagnosis of ATTR-CM, which may contribute to LF thickening and canal narrowing. There is no exact epidemiologic data about LSS, but a recent JAMA review highlighted its prevalence of 11%, and it is estimated to afflict more than 200,000 individuals in the US, and it is considered the most common indication for spinal surgery in patients over 65 years of age [5][6][7][8] . The first study hypothesizing about the role of ATTR in lumbar spinal stenosis found 21 out of 26 lumbar LF specimens with amyloid 9 .Another study analyzed the presence of TTR in LF in patients aged over 50 years old undergoing lumbar stenosis (LS) surgery, and found a prevalence of 37% of TTR amyloid deposits in vertebral ligament tissue, but a more recent study showed an incidence of 48% of these deposits 10,11 . In addition, an article highlighted a correlation of 0.54 between ATTR in LF and its thickness 12 . There has been speculation that ligamentous tissue may be the primary place for ATTR fibrillogenesis by yet unknown mechanisms 10 . Although the contribution of ATTR deposition to reoperation risk remains unproven, lumbar decompression carries an overall reoperation rate of up to 22%. This highlights the need to explore adjunctive strategies that may address potential biological contributors to disease progression 13 .In recent years, new medications for ATTR have been developed. One of these is the TTR stabilizer tafamidis meglumine (Vyndaqel), which is currently used for the management of both ATTR-CM (since 2019, after the ATTR-ACT trial in 2018) and ATTR-PN (since 2011, in Europe), independent of the type of TTR (wild or variant) [14][15][16] . It works by preventing the degradation of TTR tetramers into monomers by uniting with thyroxine-binding sites via kinetic stabilization 17 . There are other less studied medications, such as TTR silencers (patisiran, inotersen), and the NSAID-based counterpart of tafamidis, diflunisal, both sharing the same mechanism of action. At least for ATTR-CM, no current clinical data justify the use of combination therapy with these drugs, although an argument that it could enhance TTR stabilization and improve outcomes is under debate 18 .The inquiries of the current article seek to illustrate the possible role that tafamidis may exert in lumbar spinal stenosis. Also, defining whether the drug has a radiologically detectable reduction in ligamentum flavum thickness is pertinent; until now, this has been hypothesized but not demonstrated. This may represent one of the first mechanism-driven pharmacologic approaches targeting the biological substrate of degenerative lumbar spinal stenosis. The repurposing of a previously investigated medication, which has been shown to have an adequate safety profile with treatment-emergent adverse event (TEAE) incidence similar to placebo (independent of the tafamidis dose studied), can have a pivotal role in this patient population 19 . We propose that TTR stabilization may represent a disease-modifying adjunct in a biologically defined subset of lumbar spinal stenosis patients.The etiology of LSS is multifactorial, with contributors such as LF hypertrophy (LFH), or thickening, degenerative disc disease, and facet joint hypertrophy being the main culprits 20 .Chronic inflammation due to angiopoietin-like protein 2 (Angptl2) and interleukin-6 (IL-6) is likely responsible for the reduced elasticity and fibrosis of LF, as well as genetic predisposition 20 . The pathogenesis of LFH is not well understood, but decreased elasticity due to collagen and ATTRwt deposition is reported 21 . The presence of the latter is associated with a higher burden of LFH compared to its absence, elevating the clinical relevance and therapeutic options 22 . A study reported 88.4% of amyloid deposition (total) in LF in males and females of 50 years of age and older, of which 37.2% was ATTR 10 .Moreover, the Cardiac Amyloidosis in Spinal Stenosis (CASS) trial found that 59% of the 82 participants undergoing spinal stenosis surgery had a high prevalence of ATTR amyloid deposits in LF biopsy 23 .Several studies have highlighted an association between ATTR-CM and LSS 24 . It is known that LSS can precede the diagnosis of ATTR-CM on average by 5 to 15 years, in up to 40% of cases 23 . Also, more than half of spinal stenosis patients have screening criteria for cardiac amyloidosis, given they share the same age group 25 .To date, there is no indication for routine LF biopsy in order to detect ATTR earlier, and thus, a preventive approach to ATTR-CM. A clinical trial is currently being executed, with the main objective of evaluating the presence of amyloidosis in patients undergoing surgery for non-congenital spinal stenosis and/or spondyloarthropathy 23 . The way toward LF biopsy for a specific subset of vulnerable patients at risk of developing ATTR-CM can be a leap forward in both diagnostics and...

BACKGROUND AND OBJECTIVES: Penetrating brain injury (PBI) is a highly lethal form of traumatic brain injury, with substantial mortality and marked differences across high-income countries and low- to middle-income countries. The aim of the Comparative Effectiveness Research in the Americas on Penetrating Brain Injury (COMPAS-PBI) study was to evaluate management variability, compare outcomes across resource settings, and assess the feasibility of real-time multinational neurocritical care data collection. METHODS: COMPAS-PBI is a multinational, multicenter observational comparative effectiveness study. It includes a retrospective cohort (2021-2025) and a prospective cohort (2026-2028) of adults aged 18 years or older with confirmed PBI, defined by radiographic evidence of dural penetration on head computed tomography, presenting within 24 hours of injury. Eleven trauma centers across 8 countries in the Americas will participate, with the University of Oklahoma Health Sciences Campus serving as the coordinating center. The principal management strategies compared are intracranial pressure monitoring vs no monitoring and neurosurgical intervention, primarily early decompressive craniectomy, vs medical management alone. Comparative analyses will use propensity score–based methods to address baseline confounding. EXPECTED OUTCOMES: The primary outcome is in-hospital mortality. Secondary outcomes include disposition at discharge, Glasgow Outcome Scale-Extended score (at discharge, 6, and 12 months), and timing of intracranial pressure monitoring and decompressive craniectomy. Feasibility outcomes will assess completeness, timeliness, and quality of prospective real-time data capture, particularly in resource-limited settings. DISCUSSION: COMPAS-PBI will generate the first multinational comparative effectiveness data sets focused on civilian PBI within the Americas. The study is expected to provide pragmatic evidence to inform context-sensitive care strategies, reduce practice differences, and support future collaborative neurotrauma research.

Background/Objectives: Deep learning may support the prediction of meningioma response after Gamma Knife radiosurgery (GKRS), but small clinical cohorts increase the risk of overfitting and unstable performance. We developed leakage-controlled code adjusted for small clinical datasets. Methods: Pretreatment contrast-enhanced T1 MRI from 190 patients was analysed using one tumour-centred axial slice per patient. Response was defined as a tumour volume decrease ≥ 30% per log₁₀ (months). ResNet-10t, ResNet-18, ResNet-50 and DenseNet-121 with ImageNet or RadImageNet pretraining were evaluated in three workflows: (1) frozen backbone CNN, head-only training; (2) logistic regression or naive Bayes classification of features extracted from frozen CNNs; and (3) restricted CNN fine-tuning. Five partially overlapping chronological development/test splits were assessed, with all data-dependent fitting, feature selection and model optimisation restricted to development data. Results: The smallest RadImageNet ResNet-10t CNN performed best when fine-tuned (peak AUCs 0.953 and 1.000) but showed the widest split-to-split variation (range 0.738–1.000). When RadImageNet features extracted from ResNet-10t were classified by naive Bayes and logistic regression, they reached lower median and peak AUCs but within narrower ranges. Predictive performance did not consistently depend on CNN size. Conclusions: To our knowledge, this is the first deep learning study predicting volumetric meningioma response after GKRS. We provide openly available, ready-to-use code for three major workflows in small clinical datasets, with safeguards against information leakage. The benefit of fine-tuning and of RadImageNet pretraining depended on architecture. Robust small-cohort modelling requires attention to performance consistency across chronological splits, not only to the highest AUC in a single split.