Long-term outcomes of mechanical thrombectomy in acute ischaemic stroke patients with concomitant malignancy.

Introduction: Patients with cancer (CP) need a different approach to acute ischaemic stroke (AIS) treatment as intravenous thrombolysis (IVT) may be contraindicated. Mechanical thrombectomy (MT) is a treatment of choice for otherwise eligible patients, although the literature on its long-term outcomes in CP is limited. Aim: Assessing outcomes of MT-treated AIS patients with concomitant malignancy in a year-long follow-up. Material and methods: The study included 593 MT-treated AIS patients admitted in 2019-2021. The group was divided into CP (defined as a diagnosis of malignancy and undergoing/qualified for cancer treatment within previous 5 years) and a control group. The profile of cardiovascular risk factors, stroke severity and discharge, 90-day and 365-day outcomes were compared between the groups. Results: CP and controls had a similar profile of cardiovascular risk factors and comparable stroke severity. CP were less frequently treated with IVT (25.7% vs. 59.1%, p < 0.001). There were no differences between the groups in the successful reperfusion rate and occurrence of haemorrhagic complications. Discharge and 90-day outcomes were similar. CP had higher 365-day mortality (48.6% vs. 29.9%, p = 0.024) but the percentage of patients achieving good functional outcome in a year-long observation was comparable. Conclusions: Treatment with MT seems beneficial for AIS patients with concomitant malignancy both in short- and long-term observation.

Association of Imaging-based Predictors with Outcome in Different Treatment Options for Intracerebral Hemorrhage.

PURPOSE: Intracerebral hemorrhage is the deadliest form of stroke. This study aimed to enhance the prediction of 30-day mortality in intracerebral hemorrhage patients by integrating computational parameters. METHODS: This study retrospectively analyzed 435 patients with spontaneous intracerebral hemorrhage (ICH). Utilizing the acquired computed tomography (CT) images, we extracted the contour and visual representation of ICH. For the extracted contour, the analysis encompassed factors including compactness, fractal dimension, Fourier factor, and circle factor. For the images depicting ICH, we calculated various factors related to density distribution including mean, coefficient of variance, skewness and kurtosis, as well as texture parameters, such as energy, entropy, contrast and homogeneity. To assess the impact of surgical treatment on 30-day mortality, logistic regression analysis was used. RESULTS: A total of 126 patients (29.09%) died within 30 days. A total of 62 (14.25%) patients underwent surgical treatment. Multivariate logistic regression analysis revealed that surgical treatment was independently associated with a lower risk of 30-day mortality (odds ratio, OR 0.226, 95% confidence interval, CI 0.049-0.85; p = 0.039). Based on the moderated analysis, we found that the volume of ICH (OR 0.905, 95% CI 0.902-0.908; p < 0.001) and ICH energy (OR 1.389, 95%CI 0.884-0.988; p = 0.010) had positive moderating effect on such associations while the presence of intraventricular blood had negative moderating effect (OR 1.154, 95% CI 1.034-1.628; p = 0.010). CONCLUSION: Patients exhibiting a higher volume and energy of ICH might benefit from surgical treatment; however, this efficacy was found to be diminished in cases involving the presence of intraventricular blood.

The influence of embolization of internal carotid artery aneurysms on arterial tortuosity: a prospective cohort study.

PURPOSE: To compare quantitative tortuosity descriptors of Internal Carotid Artery (ICA) on aneurysmal and non-aneurysmal side before and after embolization of aneurysm and to determine possible factors associated with its change. MATERIAL AND METHODS: An analysis of 52 patients with intracranial aneurysm, treated with endovascular procedure, was performed. Based on their Digital Subtraction Angiography images, obtained prior to the procedure and after first follow-up examination, tortuosity of ICA, both on the side of embolization and on the other side was analysed. For each patient tortuosity descriptors were calculated: Relative Length, Sum of Angle Metrics, Triangular Index, Product of Angle Distance, and Inflection Count Metric. To represent changes in tortuosity, for each descriptor delta value (Δ) was defined as value of the descriptor prior to embolization - value of the descriptor on follow-up examination. RESULTS: In the follow-up We found no statistically significant changes in tortuosity on non-embolized side. On the embolized side SOAM (2.89±0.92 vs. 2.38±0.94;p<0.001), PAD (5.01±1.83 vs. 3.95±1.72 ;p<0.001) and ICM (12.18±4.55 vs. 9.76±4.04 vs.;p = 0.006) was significantly higher after embolization than before embolization. Mean ΔRelative Length (-0.02 [-0.045--0.002] vs. -0.01 [-0.02-0.003];p - 0.003),ΔProduct of Angle Distance (0.84 [0.30 - 1.82] vs. 0.10 [-0.001 - 1.10];p<0.001) and ΔInflection Count Metric (2.05 [0.42 - 3.50] vs. 0.27 [0.02 - 2.16];p = 0.004) were significantly higher on the embolized side. CONCLUSION: Following study showed that embolization may increase the tortuosity of ICA.

Sex alters the effect of perfusion deficits on functional outcome in patients with acute ischemic stroke undergoing mechanical thrombectomy.

INTRODUCTION: The discourse surrounding differences in cerebral hemodynamics and clinical outcomes among male and female patients treated with mechanical thrombectomy (MT) for acute ischemic stroke (AIS) remains unresolved. We aimed to elucidate these differences by employing computed tomography perfusion (CTP) imaging before MT and examining the influence of perfusion deficits on the 90-day functional outcome. METHODS: This single-center retrospective analysis involved patients with anterior circulation AIS treated with MT at the Comprehensive Stroke Center, University Hospital, Krakow from January 2019 to July 2023. We compared male and female patients in terms of baseline characteristics, CTP deficits, hypoperfusion intensity ratio (HIR, defined as T10max/T6max) and complications. The endpoints included the 90-day excellent functional outcome, defined as modified Rankin Score (mRS)<2, and the 90-day mortality rate. RESULTS: We included 794 patients, of whom 408 were female (51.4%). Female patients had a smaller early infarct volume (median: 7 mL; IQR: [0-24.8] vs 10 mL [0-33], p=0.004), smaller penumbra volume (77.5 mL [46-117] vs 99.5 mL [59.8-140], p<0.001), lower HIR (0.34 [0.16-0.5] vs 0.37 [0.2-9.53], p=0.043) and were less likely to achieve an excellent functional outcome (55.6% vs 66.1%, p=0.003). For every 10 mL increase in early infarct volume, the odds for achieving an excellent outcome were lower in females (odds ratio (OR): 0.82 [95% confidence interval: 0.73-0.92]) compared to males (OR: 0.96 [0.88-1.04]), whereas the risk of death was higher for females (OR: 1.25 [1.13-1.39] than for males (OR: 1.05 [0.98-1.14]). DISCUSSION: Despite more favorable cerebral hemodynamic profile, female AIS patients have worse outcomes than their male counterparts. This effect seems to be independently mediated by the more pronounced impact of early infarct volume on the prognosis in female patients. These findings underscore the possible explanatory power arising for from sex-specific interpretation of early infarct volume in clinical practice.

Risk factors of acute kidney injury during hospitalization in acute ischaemic stroke patients undergoing mechanical thrombectomy.

Introduction: Acute kidney injury (AKI) seems to worsen the prognosis of acute ischaemic stroke (AIS) patients treated with mechanical thrombectomy (MT). At the same time, the procedure of MT increases AKI risk by iodinated contrast use. Identification of factors predisposing to AKI after MT is important for recognizing vulnerable patients and successful prevention. Aim: To identify factors associated with the occurrence of AKI during hospitalization in MT-treated AIS patients. Material and methods: The study included all AIS patients treated with MT in the University Hospital in Krakow from 2019 to 2021. The diagnosis of AKI during hospitalisation was based on serum creatinine concentration levels, according to the Kidney Disease Improving Global Outcomes guidelines. We compared patients with and without AKI in terms of age, sex, comorbidities, stroke course and laboratory test results at admission. We identified factors associated with the occurrence of AKI using univariate logistic regression analysis, with significant variables subsequently added to the multivariate analyses. Results: Among 593 MT-treated AIS patients the incidence of AKI during hospitalisation was 12.6%. AKI development was associated with diabetes, chronic kidney disease, total volume of iodinated contrast obtained during hospitalisation, posterior circulation stroke, lack of intravenous thrombolysis, and laboratory test results at admission: haemoglobin, glucose, urea, potassium, and creatinine. Total contrast volume and urea level were the most important independent risk factors associated with occurrence of AKI. Conclusions: AKI is common in MT-treated AIS patients. There is a need to establish a protocol for decreasing the risk of AKI in AIS patients undergoing MT and, in case it occurs, a procedure for its treatment.

Prominent crista terminalis mimicking a right atrial mass: a systematic literature review and meta-analysis.

The crista terminalis is an anatomical structure localized on the posterolateral wall of the right atrium (RA). We performed a systematic review of the literature and meta-analysis concerning cases of unusual prominent crista terminalis mimicking RA mass. Moreover, we described the differential diagnosis of cardiac masses with the use of echocardiography, computed tomography, and cardiac magnetic resonance (CMR). We also emphasize the potential importance of this structure in electrophysiological procedures, including its role in exaggerated arrhythmias. Prominent crista terminalis may be a potential obstacle during invasive cardiac procedures or catheter ablation target. In analyzed cases, the crista terminalis was often erroneously interpreted as pathologic and at first confused with a thrombus or tumor during transthoracic echocardiography examination. The correct final diagnoses were mostly made with used transesophageal echocardiography or CMR. The most important imaging findings suggestive of prominent crista terminalis rather than tumor were a similar echogenicity/intensity with adjacent myocardium, the location on posterolateral wall of the RA, the phasic change in size, and no enhancement after contrast injection. We describe up to date and detailed imaging features for the differential diagnostics of selected intracardiac masses using various imaging techniques, including multimodality cardiac imaging. Familiarity with the anatomy and the imaging findings of the prominent crista terminalis will reduce misdiagnosis and avoid additional tests and unwarranted clinical interventions, while in patients considered for invasive cardiac procedures it might increase their efficacy and safety.

Diffusion tensor imaging metrics as natural markers of multiple sclerosis-induced brain disorders with a low Expanded Disability Status Scale score.

Non-invasive and effective differentiation along with determining the degree of deviations compared to the healthy cohort is important in the case of various brain disorders, including multiple sclerosis (MS). Evaluation of the effectiveness of diffusion tensor metrics (DTM) in 3T DTI for recording MS-related deviations was performed using a time-acceptable MRI protocol with unique comprehensive detection of systematic errors related to spatial heterogeneity of magnetic field gradients. In a clinical study, DTMs were acquired in segmented regions of interest (ROIs) for 50 randomly selected healthy controls (HC) and 50 multiple sclerosis patients. Identical phantom imaging was performed for each clinical measurement to estimate and remove the influence of systematic errors using the b-matrix spatial distribution in the DTI (BSD-DTI) technique. In the absence of statistically significant differences due to age in healthy volunteers and patients with multiple sclerosis, the existence of significant differences between groups was proven using DTM. Moreover, a statistically significant impact of spatial systematic errors occurs for all ROIs and DTMs in the phantom and for approximately 90 % in the HC and MS groups. In the case of a single patient measurement, this appears for all the examined ROIs and DTMs. The obtained DTMs effectively discriminate healthy volunteers from multiple sclerosis patients with a low mean score on the Expanded Disability Status Scale. The magnitude of the group differences is typically significant, with an effect size of approximately 0.5, and similar in both the standard approach and after elimination of systematic errors. Differences were also observed between metrics obtained using these two approaches. Despite a small alterations in mean DTMs values for groups and ROIs (1-3 %), these differences were characterized by a huge effect (effect size ∼0.8 or more). These findings indicate the importance of determining the spatial distribution of systematic errors specific to each MR scanner and DTI acquisition protocol in order to assess their impact on DTM in the ROIs examined. This is crucial to establish accurate DTM values for both individual patients and mean values for a healthy population as a reference. This approach allows for an initial reliable diagnosis based on DTI metrics.

From periphery immunity to central domain through clinical interview as a new insight on schizophrenia.

Identifying disease predictors through advanced statistical models enables the discovery of treatment targets for schizophrenia. In this study, a multifaceted clinical and laboratory analysis was conducted, incorporating magnetic resonance spectroscopy with immunology markers, psychiatric scores, and biochemical data, on a cohort of 45 patients diagnosed with schizophrenia and 51 healthy controls. The aim was to delineate predictive markers for diagnosing schizophrenia. A logistic regression model was used, as utilized to analyze the impact of multivariate variables on the prevalence of schizophrenia. Utilization of a stepwise algorithm yielded a final model, optimized using Akaike's information criterion and a logit link function, which incorporated eight predictors (White Blood Cells, Reactive Lymphocytes, Red Blood Cells, Glucose, Insulin, Beck Depression score, Brain Taurine, Creatine and Phosphocreatine concentration). No single factor can reliably differentiate between healthy patients and those with schizophrenia. Therefore, it is valuable to simultaneously consider the values of multiple factors and classify patients using a multivariate model.

Safety and Efficacy of Low-Dose Eptifibatide for Tandem Occlusions in Acute Ischemic Stroke.

OBJECTIVES: The optimal treatment strategy for ischemic stroke in patients presenting with tandem occlusions of the internal carotid artery remains controversial. Several studies have demonstrated better clinical outcomes after eptifibatide, which is a short half-life antiplatelet agent. This retrospective analysis focused on the safety and efficacy of low-dose eptifibatide administration in stroke patients with tandem lesions. METHODS: We evaluated the results of endovascular treatment in 148 stroke patients with tandem lesions. Patients in whom balloon angioplasty alone resulted in satisfactory cerebral flow did not receive eptifibatide (33 patients); others received this drug together with stent implantation (115 patients). Eptifibatide was given as an intravenous bolus of 180 µg/kg and then in a modified low dose of 1 µg/kg/min for 24 hours. RESULTS: There were no statistically significant differences between both groups regarding 30-day mortality, frequency of thrombotic events, or hemorrhagic complications. An analysis of clinical status at 30-day follow-up revealed that the administration of eptifibatide was associated with a statistically significant better outcome: a higher rate of either no neurological symptoms or only mild symptoms (4 NIHSS points maximally). CONCLUSIONS: The administration of eptifibatide in stroke patients presenting with tandem lesions is relatively safe. Moreover, treatment with this drug can improve clinical outcomes in these challenging patients.

Predictors of good functional outcome in ischaemic stroke patients without delayed neurological improvement after mechanical thrombectomy.

INTRODUCTION: This study aimed to identify predictors of 90-day good functional outcome (GFO) in patients with acute ischaemic stroke (AIS) who were treated with mechanical thrombectomy but did not achieve a delayed neurological improvement (DNI). CLINICAL RATIONALE FOR THE STUDY: In-hospital neurological improvement in patients with AIS is consistently associated with long- -term GFO. Patients who experience neither early nor delayed neurological improvement can still achieve long-term GFO, but predictors of such an outcome have not been studied. MATERIAL AND METHODS: This single-centre retrospective study involved 307 patients with anterior circulation AIS treated with mechanical thrombectomy. Multiple clinical, biochemical, radiological, and treatment-related variables were collected and analysed. DNI on day 7 was defined as at least a 10-point reduction in the National Institutes of Health Stroke Scale (NIHSS) score or NIHSS score < 2. GFO on day 90 was defined as a modified Rankin Scale (mRS) score ≤ 2. We compared the characteristics of patients with and without DNI, with special attention paid to patients who achieved 90-GFO despite a lack of DNI. Multivariate analyses were then performed to establish independent predictors of 90-day GFO among patients without DNI. RESULTS: DNI occurred in 150 out of 307 patients (48.7%) and significantly increased the odds for 90-day GFO (odds ratio [OR]: 13.99; p < 0.001). Among patients without DNI, 41.4% achieved 90-day GFO. Younger age (OR: 0.96; 95% confidence interval [CI]: 0.93-0.99; p = 0.008), lower baseline NIHSS score (OR: 0.80; 95% CI: 0.73-0.89; p < 0.001), treatment with intravenous thrombolysis (OR: 3.06; 95% CI: 1.25-7.49; p = 0.014), lack of an undetermined aetiology (OR: 0.40; 95% CI: 0.16-0.998; p = 0.050), lack of pneumonia (OR: 0.08; 95% CI: 0.02-0.31; p < 0.001), and higher haemoglobin concentration on admission (OR: 1.31; 95% CI: 1.04-1.69; p = 0.024) were identified as predictors of 90-day GFO in this subgroup. CONCLUSION: Almost half of patients with AIS in anterior circulation treated with mechanical thrombectomy experience DNI, which is a good predictor of 90-day GFO. Furthermore, 40% of patients without DNI achieve 90-day GFO which can be independently predicted by younger age, lower baseline NIHSS score, treatment with intravenous thrombolysis, higher haemoglobin concentration on admission, lack of undetermined ischaemic stroke aetiology, and lack of pneumonia.

Association of Blood Metabolomics Biomarkers with Brain Metabolites and Patient-Reported Outcomes as a New Approach in Individualized Diagnosis of Schizophrenia.

Given its polygenic nature, there is a need for a personalized approach to schizophrenia. The aim of the study was to select laboratory biomarkers from blood, brain imaging, and clinical assessment, with an emphasis on patients' self-report questionnaires. Metabolomics studies of serum samples from 51 patients and 45 healthy volunteers, based on the liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI-MS/MS), led to the identification of 3 biochemical indicators (cortisol, glutamate, lactate) of schizophrenia. These metabolites were sequentially correlated with laboratory tests results, imaging results, and clinical assessment outcomes, including patient self-report outcomes. The hierarchical cluster analysis on the principal components (HCPC) was performed to identify the most homogeneous clinical groups. Significant correlations were noted between blood lactates and 11 clinical and 10 neuroimaging parameters. The increase in lactate and cortisol were significantly associated with a decrease in immunological parameters, especially with the level of reactive lymphocytes. The strongest correlations with the level of blood lactate and cortisol were demonstrated by brain glutamate, N-acetylaspartate and the concentrations of glutamate and glutamine, creatine and phosphocreatine in the prefrontal cortex. Metabolomics studies and the search for associations with brain parameters and self-reported outcomes may provide new diagnostic evidence to specific schizophrenia phenotypes.

Unruptured intracranial aneurysms: Why should we focus on small aneurysms? A comprehensive update of recent findings.

Intracranial aneurysms (IAs) are a significant public health concern because they have the potential to cause deva-stating consequences, including death and disability. Despite advances in diagnostic and treatment modalities, the outcomes for patients with aneurysmal subarachnoid haemorrhage (aSAH) remain poor, with high rates of rebleeding, vasospasm, and cerebral ischaemia. IAs are a significant risk factor for aSAH, and it is estimated that up to 3% of the general population have IAs. Recent studies using novel imaging modalities have shown that the prevalence of IAs may be much higher, with 6.6% of adults aged 40-84 years having intradural saccular IAs ≥ 2 mm. The risk of rupture for IAs is difficult to predict, and the decision to treat them invasively is based on a balance between the estimated rupture risk and the procedural risks of the treatment. However, the mortality and morbidity rates among patients treated for IAs can be as high as 5%. There is a need for clear guidelines on the treatment of IAs, and this review aims to provide an update on recent findings in this area. To achieve this goal, the authors identified and summarized recent, high-impact studies on IAs. The review focuses on the diagnostic and treatment options for IAs, as well as the risks associated with these interventions. The authors also provide an overview of the natural history of IAs and discuss the challenges and uncertainties in managing these patients.

Possible association between ß-blocker use and a risk of intracranial aneurysm rupture.

INTRODUCTION: Aneurysmal subarachnoid hemorrhage is a devastating type of stroke, associated with high mortality and morbidity. One of modifiable risk factors of aneurysm rupture is hypertension, however, it is still not clear whether any particular antihypertensive drugs play a significant role in the prevention of aneurysm rupture. OBJECTIVES: We decided to investigate whether there is any association between acetylsalicylic acid, α-blockers, ß­blockers, angiotensin­converting enzyme inhibitors, angiotensin II receptor blockers, calcium channel blockers, diuretics, statins, and anticoagulants and a risk of intracranial aneurysm rupture. PATIENTS AND METHODS: We retrospectively analyzed 334 patients with ruptured and unruptured intracranial aneurysm. Based on logistic regression models, we obtained unadjusted and adjusted odds ratios (ORs) of subarachnoid hemorrhage associated with the use of vasoactive medications and with indices of tortuosity. RESULTS: We found that ß­blocker intake was significantly related to higher tortuosity of the cerebral arteries. Also, the intake of ß­blockers (OR, 0.41; 95% CI, 0.21-0.77; P = 0.01) and statins (OR, 0.23; 95% CI, 0.05-0.68; P = 0.01) significantly decreased the risk of aneurysm rupture, a result driven by a decreased rupture risk of anterior circulation aneurysms. No such association was found for the posterior part of the cerebral circulation. CONCLUSIONS: Aneurysm located in the anterior cerebral circulation might be less likely to rupture if patients receive ß­blockers or statins.

Association of Arterial Tortuosity with Hemodynamic Parameters-A Computational Fluid Dynamics Study.

BACKGROUND: Tortuosity of intracranial arteries has been proven to be associated with the risk of intracranial aneurysm development. We decided to analyze which aspects of tortuosity are correlated with hemodynamics parameters promoting intracranial aneurysm development. METHODS: We constructed 73 idealized models of tortuous artery (length: 25 mm, diameter: 2.5 mm) with single bifurcation. For each model, on the course of segment before bifurcation, we placed 1-3 angles with measures 15, 30, 45, 60, or 75 degrees and arc lengths 2, 5, 7, 10, or 15 mm. We performed computational fluid dynamics analysis. Blood was modeled as Newtonian fluid. We have set velocity wave of 2 cardiac cycles. After performing simulation we calculated following hemodynamic parameters at the bifurcation: time average wall shear stress (TAWSS), time average wall shear stress gradient (TAWSSG), oscillatory shear index (OSI), and relative residence time (RRT). RESULTS: We found a significant positive correlation with number of angles and TAWSS (R = 0.329; P < 0.01), TAWSSG (R = 0.317; P < 0.01), and negative with RRT (R = -0.335; P < 0.0.01). Similar results were obtained in terms of arcs lengths. On the other hand, mean angle measure was negatively correlated to TAWSS (R = -0.333; P < 0.01), TAWSSG (R = -0.473 P < 0.01), OSI (R = -0.463; P < 0.01), and positively to RRT (R = 0.332; P < 0.01). On the basis of the obtained results, we developed new tortuosity descriptor, which considered angle measures normalized to its arc length and distance from bifurcation. For such descriptor we found strong negative correlation with TAWSS (R = -0.701; P < 0.01), TAWSSG (R = 0.778; P < 0.01), OSI (R = -0.776; P < 0.01), and positive with RRT (R = 0.747; P < 0.01). CONCLUSIONS: Hemodynamic parameters promoting aneurysm development are correlated with larger number of smaller angles located on larger arcs.

Benefits and Meaning of Lipids Profile in Relation to Oxidative Balance and Brain Morphology in Schizophrenia.

Schizophrenia is characterized by complex metabolic dysregulations and their consequences. Until now, numerous theories have explained its pathogenesis, using a spectrum of available technologies. We focused our interest on lipid profile-periphery high-density cholesterol level and lipoproteins in the human brain and compared magnetic resonance imaging (MRI) scans of patients with schizophrenia and the healthy group. Detailed analysis of biochemical parameters was performed using magnetic resonance spectroscopy. Our study aimed to reveal correlations between periphery high-density lipoproteins levels and lipoproteins in the brain, depicted in MRI scans, and parameters of peripheral oxidative stress expressed as paraoxonase. Patients with schizophrenia have decreased levels of high-density lipoproteins, low paraoxonase activity, and slightly raised sodium in the blood. Positive significant correlations between serum high-density cholesterol and anterior cingulate cortex, unique brain area for schizophrenia pathophysiology, MR spectroscopy signals, and diffusion have been revealed. To our knowledge, this is the first study to describe the effect of an anterior cingulate disorder on high-density cholesterol levels on the development of schizophrenia.

Multiparametric Cardiac Magnetic Resonance and Arrhythmias in Myocarditis.

Myocarditis is an inflammatory disease of the myocardium with a wide range of potential etiological factors, including a variety of infectious agents (mainly viral), systemic diseases, drugs, and toxins.

POLCOVID: a multicenter multiclass chest X-ray database (Poland, 2020-2021).

The outbreak of the SARS-CoV-2 pandemic has put healthcare systems worldwide to their limits, resulting in increased waiting time for diagnosis and required medical assistance. With chest radiographs (CXR) being one of the most common COVID-19 diagnosis methods, many artificial intelligence tools for image-based COVID-19 detection have been developed, often trained on a small number of images from COVID-19-positive patients. Thus, the need for high-quality and well-annotated CXR image databases increased. This paper introduces POLCOVID dataset, containing chest X-ray (CXR) images of patients with COVID-19 or other-type pneumonia, and healthy individuals gathered from 15 Polish hospitals. The original radiographs are accompanied by the preprocessed images limited to the lung area and the corresponding lung masks obtained with the segmentation model. Moreover, the manually created lung masks are provided for a part of POLCOVID dataset and the other four publicly available CXR image collections. POLCOVID dataset can help in pneumonia or COVID-19 diagnosis, while the set of matched images and lung masks may serve for the development of lung segmentation solutions.

Pathological changes or technical artefacts? The problem of the heterogenous databases in COVID-19 CXR image analysis.

BACKGROUND: When the COVID-19 pandemic commenced in 2020, scientists assisted medical specialists with diagnostic algorithm development. One scientific research area related to COVID-19 diagnosis was medical imaging and its potential to support molecular tests. Unfortunately, several systems reported high accuracy in development but did not fare well in clinical application. The reason was poor generalization, a long-standing issue in AI development. Researchers found many causes of this issue and decided to refer to them as confounders, meaning a set of artefacts and methodological errors associated with the method. We aim to contribute to this steed by highlighting an undiscussed confounder related to image resolution. METHODS: 20 216 chest X-ray images (CXR) from worldwide centres were analyzed. The CXRs were bijectively projected into the 2D domain by performing Uniform Manifold Approximation and Projection (UMAP) embedding on the radiomic features (rUMAP) or CNN-based neural features (nUMAP) from the pre-last layer of the pre-trained classification neural network. Additional 44 339 thorax CXRs were used for validation. The comprehensive analysis of the multimodality of the density distribution in rUMAP/nUMAP domains and its relation to the original image properties was used to identify the main confounders. RESULTS: nUMAP revealed a hidden bias of neural networks towards the image resolution, which the regular up-sampling procedure cannot compensate for. The issue appears regardless of the network architecture and is not observed in a high-resolution dataset. The impact of the resolution heterogeneity can be partially diminished by applying advanced deep-learning-based super-resolution networks. CONCLUSIONS: rUMAP and nUMAP are great tools for image homogeneity analysis and bias discovery, as demonstrated by applying them to COVID-19 image data. Nonetheless, nUMAP could be applied to any type of data for which a deep neural network could be constructed. Advanced image super-resolution solutions are needed to reduce the impact of the resolution diversity on the classification network decision.