Abstract Background Handgrip strength (HGS) in rheumatoid arthritis (RA) is commonly attributed to joint pathology, but may also reflect extra-articular manifestations, including atrophy of motor-related brain regions. We investigated HGS as a marker of peripheral joint status, systemic immune regulation, and central motor integrity. Methods Maximal HGS was assessed using a dynamometer. Joint pathology was evaluated using radiographic and clinical measures, and upper limb disability using questionnaire. Brain volumes were quantified using MRI and MAPER software. Transcriptome sequencing was performed on circulating CD4⁺ and CD14⁺ cells. In a six-month, single-arm pilot trial, a subgroup of patients performed daily hand exercises. Associations and longitudinal changes were analysed using linear mixed-effects models accounting for repeated measurements across hands and timepoints, with variable selection performed using LASSO regression. Results A total of 59 women with established RA were included in the cross-sectional analysis (median age 64 years [range 23–76], DAS28 2.46 [1.1–5.8], disease duration 11 years [0–45]). Lower HGS was associated with greater disability. HGS was also independently associated with premotor and supplementary motor cortex (PMA/SMA) volume after adjustment for age, hand dominance, and joint pathology. Among joint pathology measures, tender joint count showed a significant negative association with HGS. Transcriptome analyses of CD4⁺ and CD14⁺ cells indicated that lower HGS was associated with reduced immune responsiveness and altered cytokine signalling pathways. In a six-month pilot hand exercise trial (n = 12; median age 54 years [28–68], DAS28 2.87 [1.2–3.6], disease duration 14 years [1–40]), HGS increased at 3 months, with a non-significant trend at 6 months. Baseline PMA/SMA volume showed a non-significant trend towards predicting HGS improvement. Longitudinal analyses revealed region-specific brain changes, with a decrease in PMA/SMA volume and an increase in insular volume over time. Conclusions Handgrip weakness in RA may reflect both joint pathology and motor cortex atrophy in the PMA/SMA. Hand exercise improved HGS and induced certain structural changes in the brain, though effects on motor regions remain uncertain and warrant further study. Trial registration Clinical trial registration: ClinicalTrials.gov, NCT04378621. Registration date: May 5, 2020.

OBJECTIVE:Investigate whether enteral supplementation with arachidonic acid (AA) and docosahexaenoic acid (DHA), from birth to term-equivalent age (TEA), promotes brain maturation as a prespecified secondary outcome of a multicentre randomised controlled trial. PARTICIPANTS:206 infants born at 22-28 weeks gestational age (GA) were randomised into intervention or control groups from three university hospitals in Sweden. INTERVENTION:The intervention group received an oil with AA (100 mg/kg/d) and DHA (50 mg/kg/d) starting at birth until 40 weeks postmenstrual age (PMA) in addition to standard nutrition. Standard-of-care infants received standard nutrition according to national guidelines. MAIN OUTCOME AND MEASURES:MRI volumetrics were defined a priori as a secondary outcome of the trial and included total brain, white and cortical grey matter, central structures and cerebellum. Univariable and multivariable linear regression models were used for comparisons. RESULTS:MRI data in 117 infants had sufficient quality for inclusion (n=58 intervention). Birth weight, GA at birth, sex distribution, and PMA at MRI were similar in the groups. Infants receiving intervention had significantly larger white-matter volume at TEA, as compared with standard of care, in models adjusted for GA at birth, sex, study centre and PMA at MRI (β=6.8 cm3, 95% CI 0.7 to 12.9, p=0.028). The contribution of the intervention to white-matter volume corresponded to 10 days of prolonged gestation. CONCLUSION AND RELEVANCE:Our findings in this hypothesis-generating study suggest that AA+DHA promotes white matter growth, which may protect the developing brain in this vulnerable population. TRIAL REGISTRATION NUMBER:NCT03201588.

Abstract Chronic systemic inflammation has been implicated in age-related neurodegeneration, but whether rheumatoid arthritis (RA) is associated with accelerated brain aging remains unclear. We combined structural magnetic resonance imaging (MRI), circulating neurodegeneration biomarkers, and peripheral monocyte transcriptomics to investigate brain aging in RA across two independent cohorts. A brain-age prediction model trained in healthy controls from the IXI imaging dataset was applied to RA patient cohorts from Gothenburg (n = 71) and Glasgow (n = 50). RA was associated with significantly elevated corrected brain-age gap relative to healthy controls (+6.5 years, 95% CI 4.2–8.8 years, p = 1.2 × 10 −7 ), with substantially stronger effects in patients ≥60 years. Older RA patients demonstrated a significant ventricular enlargement together with reduced frontal and parietal lobe volumes. Serum brain-derived tau and glial fibrillary acidic protein levels were elevated in RA. The increased brain-age gap was associated with altered myeloid transcriptional signatures. These findings demonstrate that RA is associated with age-related neurostructural alterations consistent with accelerated brain aging.

Alzheimer's disease (AD) causes progressive structural brain changes that precede clinical symptoms by years. Detecting these changes using structural MRI remains challenging, especially in early stages and when relying on visual interpretation alone. Automated semantic segmentation methods offer anatomical precision and objective measurements, but their outputs are rarely used to support human visual assessment. In this study, we explored whether such segmentation outputs can be used to guide a non-expert investigator in developing and applying interpretable diagnostic criteria. We used images from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and implemented a structured, segmentation-informed workflow in which a novice with no prior training in radiology or neuroanatomy developed classification rules based on visual appearance and volumetric readouts through three guided pilot phases. In a fourth phase, the investigator applied these criteria to an independent subset of ADNI images while blinded to the respective ADNI participants' diagnostic labels. Using an anatomical segmentation model (MAPER) with training data from a pre-release version of the Hammers Adult Brain Atlas Database (120 brain regions), the investigator focused on the piriform cortex (PC). The choice of PC was context-driven, reflecting an ongoing quantitative study of PC volume. A binary classification (AD-like versus CN-like) rule based on PC volume (< or > 430 mm 3), supported by assessments of PC shape and global atrophy, yielded an accuracy of 0.71 across 200 cases spanning four diagnostic groups. Accuracy increased to 0.77 when the analysis was restricted to CN and AD cases (with intermediate pathology (MCI) excluded). These results show that segmentation-guided visual workflows can enable non-experts to apply anatomically grounded classification criteria with moderate accuracy. Our framework can be expanded to other regions and promises to be useful for generating interpretable models, for supporting explainable AI, and for accelerating the acquisition of diagnostic skills.

The aim of this study was to evaluate the capability of histogram analysis of weighted and quantitative MRI images to improve preoperative endometrial cancer (EC) risk stratification by providing information about the histological properties of the tumours. In this prospective study, 94 patients with biopsy verified endometrial carcinoma underwent a preoperative MRI examination performed according to the European Society of Urogenital Radiology (ESUR) guidelines with addition of synthetic MRI, dynamic contrast enhancement and diffusion weighted imaging (DWI) with high b-values. Quantitative relaxation maps, perfusion maps and diffusion kurtosis imaging (DKI) maps were generated from the additional sequences. Tumours were segmented on three adjacent slices and histogram properties were compared between tumours with low and high histological risk. Significant differences were found between tumours with low and high histological risk in the histogram properties for the DKI derived apparent diffusion maps (Dapp): mean (p = 0.048), median (p = 0.025), skewness (p < 0.001) and kurtosis (p = 0.003). No significant differences between the groups were observed in histogram properties of quantitative relaxation maps, acquired by synthetic MRI. Histogram analysis of DKI shows better potential to discriminate between EC histological risk groups and histologically determined endometrioid tumour grades than regular DWI, relaxation maps from synthetic MR, perfusion maps and T1 or T2 weighted images.

Even though simultaneous optimization of similarity metrics is a standard procedure in the field of semantic segmentation, surprisingly, this is much less established for image registration. To help closing this gap in the literature, we investigate in a complex multi-modal 3D setting whether simultaneous optimization of registration metrics, here implemented by means of primitive summation, can benefit image registration. We evaluate two challenging datasets containing collections of pre- to post-operative and pre- to intra-operative MR images of glioma. Employing the proposed optimization, we demonstrate improved registration accuracy in terms of TRE on expert neuroradiologists' landmark annotations.

Background: Neurological involvement in rheumatoid arthritis (RA) is understudied despite a high prevalence of depression, cognitive deficits, chronic pain, and fatigue reported in RA patients. Furthermore, there is increasing interest in the role of peripheral and brain-resident innate immune cells in aging and neurodegenerative disease. Systemic immune activation has been shown to activate microglia and affect brain region size in models of RA [1,2], but how chronic inflammation affects the aging human brain has not been previously studied. Objectives: We investigate regional brain volumes in RA patients compared to healthy age- and sex-matched controls and relate differences to neuropsychiatric symptoms and peripheral blood CD14+ monocyte phenotypes. Methods: We included 71 female patients (median age 64 years, range 23-76) with established RA (disease duration 10 years, range 0-45) and 268 healthy women (median age of 54 years (21-82)) who served as controls (C). T1-weighted cranial magnetic resonance images were used to measure 120 brain regions. For analysis of brain region differences in relation to age, RA patients and controls were binned into age intervals of 5 years: 0-45y (RA: n=7, C: n=98), 45-50y (RA: n=4, C: n=15), 50-55y (RA: n=11, C: n=26), 55-60y (RA: n=7, C: n=32), 60-65y (RA: n=13, C: n=41), 65-70y (RA: n=21, C: n=25), older than 70y (RA: n=8, C: n=31). Patient-reported neuropsychiatric symptoms were assessed with the Fibromyalgia Impact Questionnaire (FIQ). Peripheral blood CD14+ monocytes were isolated and analysed with RNA sequencing. Genes and pathways associated with lateral ventricle size was identified by comparing patients with lateral ventricle size smaller and larger than 30 cm3. Results: RA patients over 65 had significantly enlarged lateral ventricles of the brain compared to healthy controls of the same age range (Figure 1). While there were no differences in lateral ventricle size between RA patients and controls in participants under 65 years, the lateral ventricles were enlarged by 25% (p=0.020) in patients 65-70, and 42% (p=0.00020) in patients over 70. Next, we analysed regional brain volumes in the 71 RA patients compared to 71 age-matched controls. We found that 7 limbic and 5 cortical regions were significantly different in both the left and right hemisphere. Of these regions, the thalamus, the middle frontal gyrus and the superior frontal gyrus had strong inverse correlations to the lateral ventricle volume (thalamus: Spearman r=-0.46, p=0.0002; middle frontal gyrus: r=-0.56, p<0.0001; superior frontal gyrus: r=-0.36, p=0.0052). Interestingly, the reduced thalamus size was associated with depression reported in the FIQ questionnaire according to a linear model controlling for the patients age (Beta=-21, p=0.0080). Monocytes share characteristics with microglia and can migrate to the brain. We found that CD14+ monocytes from patients with larger lateral ventricles had several activated pathways related to neuroinflammation, including neuroinflammation signalling and multiple-sclerosis signalling. Upregulated genes included CYBB, MSR1, VEGFA, HLA-DRA, A2M and CD9, which implies neurodegenerative microglia phenotypes [3]. Conclusion: RA patients over the age of 65 had larger lateral cerebral ventricles, indicating brain atrophy possibly caused by accelerated aging or neurodegeneration. Enlarged ventricles were likely a consequence of shrinkage of the thalamus as well as frontal cortical regions and were associated with neuropsychiatric symptoms in RA patients. Monocytes of patients with enlarged lateral ventricles had a phenotype associated with neuroinflammation and disease-associated microglia. A limitation to our study is an insufficient sample size for rigorous statistical testing using a split-sample approach. We therefore present our generated hypotheses along with statistical test results as suggestions for future rigorous testing on independent data. REFERENCES: [1] Anderson & Wasén et al., PNAS 2019 [2] Süß et al., Cell Reports 2020 [3] Butovsky & Weiner, Nature Reviews Neuroscience 2018 Acknowledgements: NIL. Disclosure of Interests: None declared.

Background: Rheumatoid arthritis (RA) is an inflammatory joint disease that leads to significant impairment in hand function. Chronic systemic inflammation in RA has been shown to change the activity of specific brain structures [1]. The relationship between brain structure and hand dysfunction in RA is underexplored. Objectives: We investigated structural changes in the brain of RA patients in relation to grip strength (GS), aiming to identify peripheral and brain immune responses associated with these changes. Methods: Based on GS measurement using a dynamometer, we analysed 60 RA patients with an mean age 60 years (range 23-73), disease duration 14 years (range 0-45), and mean GS 197 N (range 29-400). Patients were dichotomized by the median GS to compare the groups with weak GS versus strong GS. A subset of 12 patients underwent hand training consisting of 5 simple exercises performed separately for each hand for 10 minutes daily during 6 months. Instructions were given on an individual basis. Each patient wrote a training diary. Patients reported their functional disability using the Disabilities of the Arm, Shoulder, and Hand (DASH) questionnaire, the Health Assessment Questionnaire (HAQ) and the Fibromyalgia Impact Questionnaire (FIQ). Brain tissue was assessed by T1-weighted magnetic resonance imaging (MRI) at the baseline and after 6 months of hand training for the subset of 12 patients. Volumes of 121 brain regions were analysed using MAPER software with respect to grey matter (GM), white matter (WM), and cerebrospinal fluid (CSF) compartments. Levels of IFNγ in serum and supernatants of CD4+ T cells were determined by ELISA. RNA sequencing was used to analyse the transcriptional profile of peripheral blood CD4+ T cells. Results: RA patients with weak GS had significantly higher functional disability according to DASH and HAQ, and experienced more pain according to FIQ scores (all, p<0.001). There was no significant difference in age, disease duration or disease activity by DAS28 between the weak GS and strong GS patients. We found that the weak-GS patients had significantly increased WM volume in the putamen (L, p=0.002; R, p=0.0046) and caudate nucleus (L, p=0.0082; R, p=0.067) and decreased GM volume in the thalamus (L, p=0.038; R, p=0.028), regions which play a role in planning the execution of movement. Weak-GS patients also had significantly smaller WM volumes of the sensorimotor network, including the precentral gyrus (PG) (L, p=0.025; R, p=0.083), middle frontal gyrus (MFG) (L, p=0.0023; R, p=0.0099) and superior temporal gyrus anterior part (STGAP) (R, p=0.0051). Interestingly, after 6 months of hand training, some of the sensorimotor regions were increased in 12 patients who completed the training (MFG: 0.75% increase, p=0.0024; STGAP: 2.7% increase, p=0.033). Furthermore, weak-GS patients had lower levels of IFNγ in both serum (p=0.0006) and CD4+ T cell supernatants (p=0.0018). In the CD4+ T cells of the weak-GS patients, the downregulated genes were involved in T-cell activation and migration, and IFNγ production and response, while the upregulated genes were related to ribosome biogenesis. Further correlation analysis showed that the IFNγ sensitive genes were significantly correlated with the WM volume of the PG and putamen, and genes involved in ribosome biogenesis were correlated with WM volume of the PG and MFG. Conclusion: Hand GS loss in RA patients is associated with measurable changes in the brain volume of motor cortex and the dorsal striatum. These changes can be mitigated by daily hand training. Regional brain changes were associated with dysregulated immune system and impaired IFNγ production and signalling in CD4+ T cells. This study provides new clues for understanding the interaction between the immune and nervous systems in RA, which should be taken into account during treatment of RA patients. REFERENCES: [1] Schrepf et al. Nature Communications. 2018. Acknowledgements: NIL. Disclosure of Interests: None declared.

Machine learning models are typically evaluated by computing similarity with reference annotations and trained by maximizing similarity with such. Especially in the biomedical domain, annotations are subjective and suffer from low inter-and intra-rater reliability. Since annotations only reflect one interpretation of the real world, this can lead to sub-optimal predictions even though the model achieves high similarity scores. Here, the theoretical concept of Peak Ground Truth (PGT) is introduced. PGT marks the point beyond which an increase in similarity with the reference annotation stops translating to better Real World Model Performance (RWMP). Additionally, a quantitative technique to approximate PGT by computing inter- and intra-rater reliability is proposed. Finally, four categories of PGT-aware strategies to evaluate and improve model performance are reviewed.

Accurate image registration is pivotal in biomedical image analysis, where selecting suitable registration algorithms demands careful consideration. While numerous algorithms are available, the evaluation metrics to assess their performance have remained relatively static. This study addresses this challenge by introducing a novel evaluation metric termed Landmark Hit Rate (HitR), which focuses on the clinical relevance of image registration accuracy. Unlike traditional metrics such as Target Registration Error, which emphasize subresolution differences, HitR considers whether registration algorithms successfully position landmarks within defined confidence zones. This paradigm shift acknowledges the inherent annotation noise in medical images, allowing for more meaningful assessments. To equip HitR with label-noise-awareness, we propose defining these confidence zones based on an Inter-rater Variance analysis. Consequently, hit rate curves are computed for varying landmark zone sizes, enabling performance measurement for a task-specific level of accuracy. Our approach offers a more realistic and meaningful assessment of image registration algorithms, reflecting their suitability for clinical and biomedical applications.

The piriform cortex (PC) is located at the junction of the temporal and frontal lobes. It is involved physiologically in olfaction as well as memory and plays an important role in epilepsy. Its study at scale is held back by the absence of automatic segmentation methods on MRI. We devised a manual segmentation protocol for PC volumes, integrated those manually derived images into the Hammers Atlas Database (n = 30) and used an extensively validated method (multi‐atlas propagation with enhanced registration, MAPER) for automatic PC segmentation. We applied automated PC volumetry to patients with unilateral temporal lobe epilepsy with hippocampal sclerosis (TLE; n = 174 including n = 58 controls) and to the Alzheimer's Disease Neuroimaging Initiative cohort (ADNI; n = 151, of whom with mild cognitive impairment (MCI), n = 71; Alzheimer's disease (AD), n = 33; controls, n = 47). In controls, mean PC volume was 485 mm3 on the right and 461 mm3 on the left. Automatic and manual segmentations overlapped with a Jaccard coefficient (intersection/union) of ~0.5 and a mean absolute volume difference of ~22 mm3 in healthy controls, ~0.40/ ~28 mm3 in patients with TLE, and ~ 0.34/~29 mm3 in patients with AD. In patients with TLE, PC atrophy lateralised to the side of hippocampal sclerosis (p < .001). In patients with MCI and AD, PC volumes were lower than those of controls bilaterally (p < .001). Overall, we have validated automatic PC volumetry in healthy controls and two types of pathology. The novel finding of early atrophy of PC at the stage of MCI possibly adds a novel biomarker. PC volumetry can now be applied at scale.

OBJECTIVES:To define requirements that condition trust in artificial intelligence (AI) as clinical decision support in radiology from the perspective of various stakeholders and to explore ways to fulfil these requirements. METHODS:Semi-structured interviews were conducted with twenty-five respondents-nineteen directly involved in the development, implementation, or use of AI applications in radiology and six working with AI in other areas of healthcare. We designed the questions to explore three themes: development and use of AI, professional decision-making, and management and organizational procedures connected to AI. The transcribed interviews were analysed in an iterative coding process from open coding to theoretically informed thematic coding. RESULTS:We identified four aspects of trust that relate to reliability, transparency, quality verification, and inter-organizational compatibility. These aspects fall under the categories of substantial and procedural requirements. CONCLUSIONS:Development of appropriate levels of trust in AI in healthcare is complex and encompasses multiple dimensions of requirements. Various stakeholders will have to be involved in developing AI solutions for healthcare and radiology to fulfil these requirements. CLINICAL RELEVANCE STATEMENT:For AI to achieve advances in radiology, it must be given the opportunity to support, rather than replace, human expertise. Support requires trust. Identification of aspects and conditions for trust allows developing AI implementation strategies that facilitate advancing the field. KEY POINTS:• Dimensions of procedural and substantial demands that need to be fulfilled to foster appropriate levels of trust in AI in healthcare are conditioned on aspects related to reliability, transparency, quality verification, and inter-organizational compatibility. •Creating the conditions for trust to emerge requires the involvement of various stakeholders, who will have to compensate the problem's inherent complexity by finding and promoting well-defined solutions.

Nowadays, registration methods are typically evaluated based on sub-resolution tracking error differences. In an effort to reinfuse this evaluation process with clinical relevance, we propose to reframe image registration as a landmark detection problem. Ideally, landmark-specific detection thresholds are derived from an inter-rater analysis. To approximate this costly process, we propose to compute hit rate curves based on the distribution of errors of a sub-sample inter-rater analysis. Therefore, we suggest deriving thresholds from the error distribution using the formula: median + delta * median absolute deviation. The method promises differentiation of previously indistinguishable registration algorithms and further enables assessing the clinical significance in algorithm development.

INTRODUCTION:Cranial computed tomography (CT) is an affordable and widely available imaging modality that is used to assess structural abnormalities, but not to quantify neurodegeneration. Previously we developed a deep-learning-based model that produced accurate and robust cranial CT tissue classification. MATERIALS AND METHODS:We analyzed 917 CT and 744 magnetic resonance (MR) scans from the Gothenburg H70 Birth Cohort, and 204 CT and 241 MR scans from participants of the Memory Clinic Cohort, Singapore. We tested associations between six CT-based volumetric measures (CTVMs) and existing clinical diagnoses, fluid and imaging biomarkers, and measures of cognition. RESULTS:CTVMs differentiated cognitively healthy individuals from dementia and prodromal dementia patients with high accuracy levels comparable to MR-based measures. CTVMs were significantly associated with measures of cognition and biochemical markers of neurodegeneration. DISCUSSION:These findings suggest the potential future use of CT-based volumetric measures as an informative first-line examination tool for neurodegenerative disease diagnostics after further validation. HIGHLIGHTS:Computed tomography (CT)-based volumetric measures can distinguish between patients with neurodegenerative disease and healthy controls, as well as between patients with prodromal dementia and controls. CT-based volumetric measures associate well with relevant cognitive, biochemical, and neuroimaging markers of neurodegenerative diseases. Model performance, in terms of brain tissue classification, was consistent across two cohorts of diverse nature. Intermodality agreement between our automated CT-based and established magnetic resonance (MR)-based image segmentations was stronger than the agreement between visual CT and MR imaging assessment.

Background Existing bio fluid and imaging biomarkers used in research and clinical diagnostics of neurodegenerative diseases are often expensive or invasive and are mainly available in specialised care centres. CT is an affordable and widely available imaging modality predominantly used to evaluate structural abnormalities, but not for the volumetric quantification of neurodegeneration. Previously, we developed a deep learning model trained on MRI segmentations from individuals with paired CT and MR scans, which achieved high accuracy and robust tissue classification based on brain CT images.Purpose To explore the diagnostic utility of deep-learning-derived CT-based atrophy measures and study their association with relevant cognitive, biochemical and other imaging markers of neurodegenerative diseases.Materials and methods In this retrospective study, we analysed 917 CT and 744 MR scans from cognitively healthy participants of the Gothenburg H70 Birth Cohort (70.4 ± 2.6 years) and 204 CT and 241 MR scans from participants of the Memory Clinic Cohort, Singapore (73 Alzheimer’s disease, 20 vascular dementia, 22 cognitively normal; 74.0 ± 8.2 years). We tested associations between six CT-derived volumetric measures with clinical diagnosis, fluid and imaging biomarkers and cognition.Results In the Memory Clinic Cohort, deep-learning-derived CT-based atrophy measures differentiated cognitively healthy individuals from Alzheimer’s disease (AUC 0.88; 95% CI: 0.79-0.96) and vascular dementia (AUC 0.91; 95% CI: 0.81-1.00) patients with high accuracy levels comparable to MR-derived measures. Additionally, CT-based measures distinguished early, prodromal Alzheimer’s disease (AUC= 0.73, 95% CI: 0.62, 0.85) and prodromal vascular dementia patients from healthy individuals (CT-GM: AUC= 0.7, 95% CI: 0.51, 0.81). CT-derived volumes were significantly associated with measures of cognition and biochemical markers of neurodegeneration, notably plasma-derived neurofilament light (ρ=-0.43, p<0.001, in the Memory Clinic Cohort).Conclusion Our findings provide strong evidence for the potential of deep-learning-derived CT-based atrophy measures in aiding neurodegenerative disease diagnostics in primary care settings.### Competing Interest StatementKB has served as a consultant, at advisory boards, or at data monitoring committees for Abcam, Axon, BioArctic, Biogen, JOMDD/Shimadzu. Julius Clinical, Lilly, MagQu, Novartis, Ono Pharma, Pharmatrophix, Prothena, Roche Diagnostics, and Siemens Healthineers, and is a co-founder of Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part of the GU Ventures Incubator Program, outside the work presented in this paper. SK has served at scientific advisory boards and / or as consultant for Geras Solutions and Biogen. MS has served at an advisory board for Servier Pharmaceuticals, outside of the present work. HZ has served at scientific advisory boards and/or as a consultant for Abbvie, Acumen, Alector, Alzinova, ALZPath, Annexon, Apellis, Artery Therapeutics, AZTherapies, CogRx, Denali, Eisai, Nervgen, Novo Nordisk, Passage Bio, Pinteon Therapeutics, Prothena, Red Abbey Labs, reMYND, Roche, Samumed, Siemens Healthineers, Triplet Therapeutics, and Wave, has given lectures in symposia sponsored by Cellectricon, Fujirebio, Alzecure, Biogen, and Roche, and is a co-founder of Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part of the GU Ventures Incubator Program (outside submitted work).### Funding StatementHZ is a Wallenberg Scholar supported by grants from the Swedish Research Council (#2018-02532), the European Research Council (#681712), Swedish State Support for Clinical Research (#ALFGBG-720931), the Alzheimer Drug Discovery Foundation (ADDF), USA (#201809-2016862), the AD Strategic Fund and the Alzheimers Association (#ADSF-21-831376-C, #ADSF-21-831381-C and #ADSF-21-831377-C), the Olav Thon Foundation, the Erling-Persson Family Foundation, Stiftelsen f ör Gamla Tjänarinnor, Hjärnfonden, Sweden (#FO2019-0228), the European Unions Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 860197 (MIRIADE), and the UK Dementia Research Institute at UCL. KB is supported by the Swedish Research Council (#2017-00915), the Alzheimer Drug Discovery Foundation (ADDF), USA (#RDAPB-201809-2016615), the Swedish Alzheimer Foundation (#AF-742881), Hjärnfonden, Sweden (#FO2017-0243), the Swedish state under the agreement between the Swedish government and the County Councils, the ALF-agreement (#ALFGBG-715986), and European Union Joint Program for Neurodegenerative Disorders (JPND2019-466-236). KB is supported by the Swedish Research Council (#2017-00915), the Alzheimer Drug Discovery Foundation (ADDF), USA (#RDAPB-201809-2016615), the Swedish Alzheimer Foundation (#AF-930351, #AF-939721 and #AF-968270), Hjärnfonden, Sweden (#FO2017-0243 and #ALZ2022-0006), the Swedish state under the agreement between the Swedish government and the County Councils, the ALF-agreement (#ALFGBG-715986 and #ALFGBG-965240), the European Union Joint Program for Neurodegenerative Disorders (JPND2019-466-236), the National Institute of Health (NIH), USA, (grant #1R01AG068398-01), and the Alzheimers Association 2021 Zenith Award (ZEN-21-848495). SK was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG-965923, ALFGBG-81392, ALF GBG-771071). The Alzheimerfonden (AF-842471, AF-737641, AF-15939825). The Swedish Research Council (2019-02075), Psykiatriska Forskningsfonden, Stiftelsen Demensfonden, Stiftelsen Hjalmar Svenssons Forskningsfond, Stiftelsen Wilhelm och Martina Lundgrens vetenskapsfond. MS is supported by the Knut and Alice Wallenberg Foundation (Wallenberg Centre for Molecular and Translational Medicine; KAW 2014.0363), the Swedish Research Council (#2017-02869), the Swedish state under the agreement between the Swedish government and the County Councils, the ALF-agreement (#ALFGBG-813971), and the Swedish Alzheimer Foundation (#AF-740191). Image analysis computations were in part carried out with resources provided by the Swedish National Infrastructure for Computing (SNIC), partially funded by the Swedish Research Council through grant agreement no. 2018-05973.### Author DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.YesThe details of the IRB/oversight body that provided approval or exemption for the research described are given below:The Memory Clinic Cohort Study, Singapore (study protocol number DEM4333) was approved by the National Healthcare Group Domain Specific Review Board (Reference number: NHG DSRB 2018/01098-SRF0004). The H70 study was approved by the Regional Ethical Review Board in Gothenburg (Approval Numbers: 869-13, T076-14, T166-14, 976-13, 127-14, T936-15, 006-14, T703-14, 006-14, T201-17, T915-14, 959-15, T139-15), and by the Radiation Protection Committee (Approval Number: 13-64).I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesThe Gothenburg H70 Birth cohort and Memory Clinic Cohort, Singapore cannot openly share data according to existing ethical and data sharing approvals, however, relevant data can and will be shared with research groups after submitting a research proposal which has to be approved by the respective study coordinators.

Existing imaging biomarkers for neurodegenerative diseases such as magnetic resonance imaging (MRI) and positron emission tomography are costly or have limited accessibility. Computed tomography (CT) and novel blood tests have greater potential as screening modalities for suspected neurodegenerative diseases and to potentially exclude other treatable causes of neurodegeneration.

Objectives To determine the reproducibility and replicability of studies that develop and validate segmentation methods for brain tumours on MRI and that follow established reproducibility criteria; and to evaluate whether the reporting guidelines are sufficient. Methods Two eligible validation studies of distinct deep learning (DL) methods were identified. We implemented the methods using published information and retraced the reported validation steps. We evaluated to what extent the description of the methods enabled reproduction of the results. We further attempted to replicate reported findings on a clinical set of images acquired at our institute consisting of high-grade and low-grade glioma (HGG, LGG), and meningioma (MNG) cases. Results We successfully reproduced one of the two tumour segmentation methods. Insufficient description of the preprocessing pipeline and our inability to replicate the pipeline resulted in failure to reproduce the second method. The replication of the first method showed promising results in terms of Dice similarity coefficient (DSC) and sensitivity (Sen) on HGG cases (DSC=0.77, Sen=0.88) and LGG cases (DSC=0.73, Sen=0.83), however, poorer performance was observed for MNG cases (DSC=0.61, Sen=0.71). Preprocessing errors were identified that contributed to low quantitative scores in some cases. Conclusions Established reproducibility criteria do not sufficiently emphasise description of the preprocessing pipeline. Discrepancies in preprocessing as a result of insufficient reporting are likely to influence segmentation outcomes and hinder clinical utilisation. A detailed description of the whole processing chain, including preprocessing, is thus necessary to obtain stronger evidence of the generalisability of DL-based brain tumour segmentation methods and to facilitate translation of the methods into clinical practice.

Brain tissue segmentation plays a crucial role in feature extraction, volumetric quantification, and morphometric analysis of brain scans. For the assessment of brain structure and integrity, CT is a non-invasive, cheaper, faster, and more widely available modality than MRI. However, the clinical application of CT is mostly limited to the visual assessment of brain integrity and exclusion of copathologies. We have previously developed two-dimensional (2D) deep learning-based segmentation networks that successfully classified brain tissue in head CT. Recently, deep learning-based MRI segmentation models successfully use patch-based three-dimensional (3D) segmentation networks. In this study, we aimed to develop patch-based 3D segmentation networks for CT brain tissue classification. Furthermore, we aimed to compare the performance of 2D- and 3D-based segmentation networks to perform brain tissue classification in anisotropic CT scans. For this purpose, we developed 2D and 3D U-Net-based deep learning models that were trained and validated on MR-derived segmentations from scans of 744 participants of the Gothenburg H70 Cohort with both CT and T1-weighted MRI scans acquired timely close to each other. Segmentation performance of both 2D and 3D models was evaluated on 234 unseen datasets using measures of distance, spatial similarity, and tissue volume. Single-task slice-wise processed 2D U-Nets performed better than multitask patch-based 3D U-Nets in CT brain tissue classification. These findings provide support to the use of 2D U-Nets to segment brain tissue in one-dimensional (1D) CT. This could increase the application of CT to detect brain abnormalities in clinical settings.

MRI is a cornerstone in presurgical evaluation of epilepsy. Despite guidelines, clinical practice varies. In light of the E-PILEPSY pilot reference network, we conducted a systematic review and meta-analysis on the diagnostic value of MRI in the presurgical evaluation of epilepsy patients. We included original research articles on diagnostic value of higher MRI field strength and guideline-recommended and additional MRI sequences in detecting an epileptogenic lesion in adult or paediatric epilepsy surgery candidates. Lesion detection rate was used as a metric in meta-analysis. Eighteen studies were included for MRI field strength and 25 for MRI sequences, none were free from bias. In patients with normal MRI at lower-field strength, 3T improved lesion detection rate by 18% and 7T by 23%. Field strengths higher than 1.5T did not have higher lesion detection rates in patients with hippocampal sclerosis (HS). The lesion detection rate of epilepsy-specific MRI protocols was 83% for temporal lobe epilepsy (TLE) patients. Dedicated MRI protocols and evaluation by an experienced epilepsy neuroradiologist increased lesion detection. For HS, 3DT1, T2, and FLAIR each had a lesion detection rate at around 90%. Apparent diffusion coefficient indices had a lateralizing value of 33% for TLE. DTI fractional anisotropy and mean diffusivity had a localizing value of 8% and 34%. A dedicated MRI protocol and expert evaluation benefits lesion detection rate in epilepsy surgery candidates. If patients remain MRI negative, imaging at higher-field strength may reveal lesions. In HS, apparent diffusion coefficient indices may aid lateralization and localization more than increasing field strength. DTI can add further diagnostic information. For other additional sequences, the quality and number of studies is insufficient to draw solid conclusions. Our findings may be used as evidence base for developing new high-quality MRI studies and clinical guidelines.