Multimodal data fusion using sparse canonical correlation analysis and cooperative learning: a COVID-19 cohort study.

Through technological innovations, patient cohorts can be examined from multiple views with high-dimensional, multiscale biomedical data to classify clinical phenotypes and predict outcomes. Here, we aim to present our approach for analyzing multimodal data using unsupervised and supervised sparse linear methods in a COVID-19 patient cohort. This prospective cohort study of 149 adult patients was conducted in a tertiary care academic center. First, we used sparse canonical correlation analysis (CCA) to identify and quantify relationships across different data modalities, including viral genome sequencing, imaging, clinical data, and laboratory results. Then, we used cooperative learning to predict the clinical outcome of COVID-19 patients: Intensive care unit admission. We show that serum biomarkers representing severe disease and acute phase response correlate with original and wavelet radiomics features in the LLL frequency channel (cor(Xu1, Zv1) = 0.596, p value < 0.001). Among radiomics features, histogram-based first-order features reporting the skewness, kurtosis, and uniformity have the lowest negative, whereas entropy-related features have the highest positive coefficients. Moreover, unsupervised analysis of clinical data and laboratory results gives insights into distinct clinical phenotypes. Leveraging the availability of global viral genome databases, we demonstrate that the Word2Vec natural language processing model can be used for viral genome encoding. It not only separates major SARS-CoV-2 variants but also allows the preservation of phylogenetic relationships among them. Our quadruple model using Word2Vec encoding achieves better prediction results in the supervised task. The model yields area under the curve (AUC) and accuracy values of 0.87 and 0.77, respectively. Our study illustrates that sparse CCA analysis and cooperative learning are powerful techniques for handling high-dimensional, multimodal data to investigate multivariate associations in unsupervised and supervised tasks.

Apical fibrobullous lung disease in spondyloarthritis patients with biologic DMARDs indication.

OBJECTIVE: The rate of pleuroparenchymal involvement in patients with SpA varies widely, from 0% to 85%. The most common form is apical fibrobullous disease (AFLD). The aim of this study was to determine the incidence of AFLD and associated factors in SpA patients under and/or planned to start biologic DMARDs (bDMARDs) therapy. METHODS: The records of 3021 SPA patients registered with HUR-BIO who had indication of bDMARDs between 2010 and 2021 were scanned. The study included 2489 patients with at least one chest radiograph (X-ray). Patient demographics, comorbidities, laboratory data, bDMARDs used, baseline DASs, and purified protein derivative and/or QuantiFERON test results before initiation of bDMARDs were recorded. RESULTS: Of the 2489 patients, 36 (1.4%) were found to have AFLD by X-ray and/or CT. The mean disease duration was 11.7 (7.1) years. Patients with AFLD were more likely to be male [28 (77.8%) vs 1321 (53.9%), P = 0.004], older [56.3 (10.5) years vs 44.8 (11.4) years, P < 0.001], heavy smokers [27 (79.4%) vs 1468 (60.9%), P = 0.028] and have had longer disease duration [17. 7 (9.7) years vs 11.6 (7) years, P = 0.001]. QuantiFERON positivity was higher in the AFLD group [9 (36%) vs 309 (16.1%), P = 0.013]. While treatment with adalimumab was less preferred in those with AFLD, treatment with etanercept was more frequently preferred. CONCLUSION: As the radiological findings of AFLD can be confused with those of tuberculosis, special attention should be paid to differentiating between tuberculosis and the disease in males and in patients who have had long disease duration.

Multimodal Biomedical Data Fusion Using Sparse Canonical Correlation Analysis and Cooperative Learning: A Cohort Study on COVID-19.

Through technological innovations, patient cohorts can be examined from multiple views with high-dimensional, multiscale biomedical data to classify clinical phenotypes and predict outcomes. Here, we aim to present our approach for analyzing multimodal data using unsupervised and supervised sparse linear methods in a COVID-19 patient cohort. This prospective cohort study of 149 adult patients was conducted in a tertiary care academic center. First, we used sparse canonical correlation analysis (CCA) to identify and quantify relationships across different data modalities, including viral genome sequencing, imaging, clinical data, and laboratory results. Then, we used cooperative learning to predict the clinical outcome of COVID-19 patients. We show that serum biomarkers representing severe disease and acute phase response correlate with original and wavelet radiomics features in the LLL frequency channel (corr(Xu1, Zv1) = 0.596, p-value < 0.001). Among radiomics features, histogram-based first-order features reporting the skewness, kurtosis, and uniformity have the lowest negative, whereas entropy-related features have the highest positive coefficients. Moreover, unsupervised analysis of clinical data and laboratory results gives insights into distinct clinical phenotypes. Leveraging the availability of global viral genome databases, we demonstrate that the Word2Vec natural language processing model can be used for viral genome encoding. It not only separates major SARS-CoV-2 variants but also allows the preservation of phylogenetic relationships among them. Our quadruple model using Word2Vec encoding achieves better prediction results in the supervised task. The model yields area under the curve (AUC) and accuracy values of 0.87 and 0.77, respectively. Our study illustrates that sparse CCA analysis and cooperative learning are powerful techniques for handling high-dimensional, multimodal data to investigate multivariate associations in unsupervised and supervised tasks.

Costovertebral and costotransverse joint involvement in spondyloarthritis.

OBJECTIVES: Costovertebral (CV), costotransverse (CTr), sternoclavicular (SC), and manubriosternal (MS) joints are impacted in spondyloarthritis (SpA) patients; however, clinical aspects of these involvements require additional evaluation. METHODS: A total of 281 SpA patients who had undergone chest computed tomography (CT) for any reason between 2010 and 2020 were included in the study. SpA patients were divided into three groups, ankylosing spondylitis (AS), non-radiographic axial SpA, and psoriatic arthritis. Thirty age- and sex-matched rheumatoid arthritis (RA) patients and 30 non-rheumatic disease individuals were selected for comparison. An experienced radiologist reviewed 24 CV, 20 CTr, 2 SC, and 1 MS joints from a thorax CT for each patient. All joints were classified as: normal (0); suspicious (1), mild (2), moderate (3), or severe (4). RESULTS: Total CV and CTr joint scores differed between diseases (p < .001). Male AS patients had higher CV and CTr scores than female AS patients (male CV score: 52 [range 0-96] and CTr score: 22 [range 0-80]; female CV score: 20 [range 0-96] and CTr score: 12 [range 0-79]). Strong negative correlations were detected in AS patients between chest expansion and CV (r = -0.703 p = .007) and CTr (r = -0.763 p = .002) joint involvement; positive correlations between CV and CTr joints, and modified Stoke Ankylosing Spondylitis Spinal Score (p < .05); and no significant association for MS and SC joints. CONCLUSIONS: CV and CTr joint involvement on thorax CT was more severe in AS and negatively affected chest expansion. The use of thorax CT scans performed for other indications in the examination of these joints may be advantageous for the early beginning of rehabilitation programs targeted at maintaining chest mobility.

Outcomes of high-risk breast lesions diagnosed using image-guided core needle biopsy: results from a multicenter retrospective study.

PURPOSE: The clinical management of high-risk lesions using image-guided biopsy is challenging. This study aimed to evaluate the rates at which such lesions were upgraded to malignancy and identify possible predictive factors for upgrading high-risk lesions. METHODS: This retrospective multicenter analysis included 1.343 patients diagnosed with high-risk lesions using an image-guided core needle or vacuum-assisted biopsy (VAB). Only patients managed using an excisional biopsy or with at least one year of documented radiological follow-up were included. For each, the Breast Imaging Reporting and Data System (BI-RADS) category, number of samples, needle thickness, and lesion size were correlated with malignancy upgrade rates in different histologic subtypes. Pearson's chi-squared test, the Fisher-Freeman-Halton test, and Fisher's exact test were used for the statistical analyses. RESULTS: The overall upgrade rate was 20.6%, with the highest rates in the subtypes of intraductal papilloma (IP) with atypia (44.7%; 55/123), followed by atypical ductal hyperplasia (ADH) (38.4%; 144/375), lobular neoplasia (LN) (12.7%; 7/55), papilloma without atypia (9.4%; 58/611), flat epithelial atypia (FEA) (8.7%; 10/114), and radial scars (RSs) (4.6%; 3/65). There was a significant relationship between the upgrade rate and BI-RADS category, number of samples, and lesion size Lesion size was the most predictive factor for an upgrade in all subtypes. CONCLUSION: ADH and atypical IP showed considerable upgrade rates to malignancy, requiring surgical excision. The LN, IP without atypia, pure FEA, and RS subtypes showed lower malignancy rates when the BI-RADS category was lower and in smaller lesions that had been adequately sampled using VAB. After being discussed in a multidisciplinary meeting, these cases could be managed with follow-up instead of excision.

Two in one: Overlapping CT findings of COVID-19 and underlying lung diseases.

Coronavirus disease 2019 (COVID-19) is associated with pneumonia and has various pulmonary manifestations on computed tomography (CT). Although COVID-19 pneumonia is usually seen as bilateral predominantly peripheral ground-glass opacities with or without consolidation, it can present with atypical radiological findings and resemble the imaging findings of other lung diseases. Diagnosis of COVID-19 pneumonia is much more challenging for both clinicians and radiologists in the presence of pre-existing lung disease. The imaging features of COVID-19 and underlying lung disease can overlap and obscure the findings of each other. Knowledge of the radiological findings of both diseases and possible complications, correct diagnosis, and multidisciplinary consensus play key roles in the appropriate management of diseases. In this pictorial review, the chest CT findings are presented of patients with underlying lung diseases and overlapping COVID-19 pneumonia and the various reasons for radiological lung abnormalities in these patients are discussed.

Admission chest CT findings and risk assessment for stroke-associated pneumonia.

INTRODUCTION: Stroke-associated pneumonia (SAP) is a significant cause of morbidity and mortality after stroke. Various factors, including dysphagia and stroke severity, are closely related to SAP risk; however, the contribution of the baseline pulmonary parenchymal status to this interplay is an understudied field. Herein, we evaluated the prognostic performance of admission chest computed tomography (CT) findings in predicting SAP. METHODS: We evaluated admission chest CT images, acquired as part of a COVID-19-related institutional policy, in a consecutive series of acute ischemic stroke patients. The pulmonary opacity load at baseline was quantified using automated volumetry and visual scoring algorithms. The relationship between pulmonary opacities with risk of pneumonia within 7 days of symptom onset (i.e., SAP) was evaluated by bivariate and multivariate analyses. RESULTS: Twenty-three percent of patients in our cohort (n = 100) were diagnosed with SAP. Patients with SAP were more likely to have atrial fibrillation, COPD, severe neurological deficits, and dysphagia. The visual opacity score on chest CT was significantly higher among patients who developed SAP (p = 0.014), while no such relationship was observed in terms of absolute or relative opacity volume. In multivariate analyses, admission stroke severity, presence of dysphagia and a visual opacity score of ≥ 3 (OR 6.37, 95% CI 1.61-25.16; p = 0.008) remained significantly associated with SAP risk. CONCLUSIONS: Pulmonary opacity burden, as evaluated on admission chest CT, is significantly associated with development of pneumonia within initial days of stroke. This association is independent of other well-known predisposing factors for SAP, including age, stroke severity, and presence of dysphagia.

Nerves of the Thorax: Anatomy, Clinical Signs, and Imaging Findings of Pathological Conditions.

BACKGROUND: Radiological diagnosis of thoracic nerve diseases is difficult because they are rare, and nerves cannot be seen directly on radiological images. The major nerves of the thorax can be listed as the phrenic, vagus, recurrent laryngeal, long thoracic nerve pairs, sympathetic chains, costal nerves, and brachial plexus. Diseases of thoracic nerves have various causes, including traumatic injury, neuromuscular diseases, infection, compression, radiation, drugs, and tumors. OBJECTIVE: This pictorial review aims to describe the anatomic locations of the major thoracic nerves on radiological images, comprehensively describe the causes of thoracic nerve diseases and define the clinical signs and primary and secondary imaging findings of dysfunction of the thoracic nerves. METHODS: This paper was designed to illustrate primary and secondary imaging findings of nerve diseases. Firstly, the normal anatomy of nerves is shown with diagrams. Secondly, we explained primary and secondary imaging features with variable radiological methods, including chest X-Ray, magnetic resonance imaging, and computed tomography. CONCLUSION: Primary findings of nerve diseases can be detected if radiologists are familiar with the courses of the nerves on radiological images. Knowledge of the normal functions of the nerves can aid in diagnosing thoracic nerve diseases identified from secondary imaging findings such as diaphragmatic elevation, muscular atrophy, and winged scapula. It is essential to know the normal anatomy, function, and possible causes of thoracic nerve diseases to make a correct diagnosis and apply the prompt treatment.

A screening study for COVID-19-associated pulmonary aspergillosis in critically ill patients during the third wave of the pandemic.

BACKGROUND: COVID-19-associated pulmonary aspergillosis (CAPA) has been reported as an important cause of mortality in critically ill patients with an incidence rate ranging from 5% to 35% during the first and second pandemic waves. OBJECTIVES: We aimed to evaluate the incidence, risk factors for CAPA by a screening protocol and outcome in the critically ill patients during the third wave of the pandemic. PATIENTS/METHODS: This prospective cohort study was conducted in two intensive care units (ICU) designated for patients with COVID-19 in a tertiary care university hospital between 18 November 2020 and 24 April 2021. SARS-CoV-2 PCR-positive adult patients admitted to the ICU with respiratory failure were included in the study. Serum and respiratory samples were collected periodically from ICU admission up to CAPA diagnosis, patient discharge or death. ECMM/ISHAM consensus criteria were used to diagnose and classify CAPA cases. RESULTS: A total of 302 patients were admitted to the two ICUs during the study period, and 213 were included in the study. CAPA was diagnosed in 43 (20.1%) patients (12.2% probable, 7.9% possible). In regression analysis, male sex, higher SOFA scores at ICU admission, invasive mechanical ventilation and longer ICU stay were significantly associated with CAPA development. Overall ICU mortality rate was higher significantly in CAPA group compared to those with no CAPA (67.4% vs 29.4%, p < .001). CONCLUSIONS: One fifth of critically ill patients in COVID-19 ICUs developed CAPA, and this was associated with a high mortality.

A COVID-19 first evaluation clinic at a university hospital in Turkey.

BACKGROUND: We aimed to analyze the usefulness of such a reserved area for the admission of the patients' symptoms suggesting COVID-19 and compare the demographic and clinical characteristics of the patients with COVID-19 and without COVID-19 who were admitted to C1 during the first month of the COVID-19 outbreak in our hospital. METHODS: A new area was set up in Hacettepe University Adult Hospital to limit the contact of COVID-19 suspicious patients with other patients, which was named as COVID-19 First Evaluation Outpatient Clinic (C1). C1 had eight isolation rooms and two sampling rooms for SARS-CoV-2 polymerase-chain-reaction (PCR). All rooms were negative-pressurized. Patients who had symptoms that were compatible with COVID-19 were referred to C1 from pretriage areas. All staff received training for the appropriate use of personal protective equipment and were visited daily by the Infection Prevention and Control team. RESULTS: One hundred and ninety-eight (29.4%) of 673 patients who were admitted to C1were diagnosed with COVID-19 between March 20, 2020, and April 19, 2020. SARS-CoV-2 PCR was positive in 142 out of 673 patients. Chest computerized tomography (CT) was performed in 421 patients and COVID-19 was diagnosed in 56 of them based on CT findings despite negative PCR. Four hundred and ninety-three patients were tested for other viral and bacterial infections with multiplex real-time reverse-transcriptase PCR (RTPCR). Blood tests that included complete blood count, renal and liver functions, d-dimer levels, ferritin, C- reactive protein, and procalcitonin were performed in 593 patients. Only one out of 44 healthcare workers who worked at C1 was infected by SARS-CoV-2. DISCUSSION: Early diagnosis of infected patients and ensuring adequate isolation are very important to control the spread of COVID-19. The purpose of setting up the COVID-19 first evaluation outpatient clinic was to prevent the overcrowding of ER due to mild or moderate infections, ensure appropriate distancing and isolation, and enable emergency services to serve for real emergencies. A wellplanned outpatient care area and teamwork including internal medicine, microbiology, and radiology specialists under the supervision of infectious diseases specialists allowed adequate management of the mild-to-moderate patients with suspicion of COVID-19.

Environmental asbestos exposure and nonmalignant pleural findings: a retrospective evaluation of a five-year chest CT repository.

This registry-based case-control study aimed to assess the association between asbestos deposits in the birthplace and/or residence and nonmalignant pleural findings, namely pleural plaques (PPs) and pleural thickening (PT), on chest CT scans. In total, 39,472 CT scans obtained over five years in a tertiary referral hospital in Ankara, Turkey, were evaluated. Cases involving patients with PP (n = 537), PT (n = 263), PP&PT (n = 69), and controls (n = 543) from the same study base without those conditions were included. Each case group was compared to controls using unconditional logistic regression. The presence of asbestos deposits in the district of birthplace (adjusted OR = 2.13, 95% CI: 1.35-3.37) and both birthplace and residence (aOR = 4.32, 95% CI: 2.26-8.27) was significantly related to the PPs. As the importance of environmental asbestos exposure in Turkey continues, future prospective studies could contribute to developing screening strategies.

Imaging of congenital lung diseases presenting in the adulthood: a pictorial review.

Congenital lung diseases in adults are rare diseases that can present with symptoms or be detected incidentally. Familiarity with the imaging features of different types of congenital lung diseases helps both in correct diagnosis and management of these diseases. Congenital lung diseases in adults are classified into three main categories as bronchopulmonary anomalies, vascular anomalies, and combined bronchopulmonary and vascular anomalies. Contrast-enhanced computed tomography, especially 3D reconstructions, CT, or MR angiography, can show vascular anomalies in detail. The tracheobronchial tree, parenchymal changes, and possible complications can also be defined on chest CT, and new applications such as quantitative 3D reconstruction CT images, dual-energy CT (DECT) can be helpful in imaging parenchymal changes. In addition to the morphological assessment of the lungs, novel MRI techniques such as ultra-short echo time (UTE), arterial spin labeling (ASL), and phase-resolved functional lung (PREFUL) can provide functional information. This pictorial review aims to comprehensively define the radiological characteristics of each congenital lung disease in adults and to highlight differential diagnoses and possible complications of these diseases.

AI-Based Quantitative CT Analysis of Temporal Changes According to Disease Severity in COVID-19 Pneumonia.

OBJECTIVE: To quantitatively evaluate computed tomography (CT) parameters of coronavirus disease 2019 (COVID-19) pneumonia an artificial intelligence (AI)-based software in different clinical severity groups during the disease course. METHODS: From March 11 to April 15, 2020, 51 patients (age, 18-84 years; 28 men) diagnosed and hospitalized with COVID-19 pneumonia with a total of 116 CT scans were enrolled in the study. Patients were divided into mild (n = 12), moderate (n = 31), and severe (n = 8) groups based on clinical severity. An AI-based quantitative CT analysis, including lung volume, opacity score, opacity volume, percentage of opacity, and mean lung density, was performed in initial and follow-up CTs obtained at different time points. Receiver operating characteristic analysis was performed to find the diagnostic ability of quantitative CT parameters for discriminating severe from nonsevere pneumonia. RESULTS: In baseline assessment, the severe group had significantly higher opacity score, opacity volume, higher percentage of opacity, and higher mean lung density than the moderate group (all P ≤ 0.001). Through consecutive time points, the severe group had a significant decrease in lung volume (P = 0.006), a significant increase in total opacity score (P = 0.003), and percentage of opacity (P = 0.007). A significant increase in total opacity score was also observed for the mild group (P = 0.011). Residual opacities were observed in all groups. The involvement of more than 4 lobes (sensitivity, 100%; specificity, 65.26%), total opacity score greater than 4 (sensitivity, 100%; specificity, 64.21), total opacity volume greater than 337.4 mL (sensitivity, 80.95%; specificity, 84.21%), percentage of opacity greater than 11% (sensitivity, 80.95%; specificity, 88.42%), total high opacity volume greater than 10.5 mL (sensitivity, 95.24%; specificity, 66.32%), percentage of high opacity greater than 0.8% (sensitivity, 85.71%; specificity, 80.00%) and mean lung density HU greater than -705 HU (sensitivity, 57.14%; specificity, 90.53%) were related to severe pneumonia. CONCLUSIONS: An AI-based quantitative CT analysis is an objective tool in demonstrating disease severity and can also assist the clinician in follow-up by providing information about the disease course and prognosis according to different clinical severity groups.

Comparison of clinical and magnetic resonance imaging findings of triple-negative breast cancer with non-triple-negative tumours.

PURPOSE: Triple-negative breast cancer (TNBC) has some distinctive features. The aim of the study was to compare clinical and breast magnetic resonance imaging (MRI) findings of TNBC with non-triple-negative breast cancer (nTNBC) in molecular subtypes such as ADC (apparent diffusion coefficient) values, T2-weighted (T2W) image intensity, shape, margin, lymph node involvement, grade, multifocality, multicentricity, bilaterality, and enhancement pattern differences between tumour subtypes. MATERIAL AND METHODS: A total of 141 patients who underwent breast biopsy at our institution between January 2010 and June 2018 were included in this study. Patients were divided into molecular subtypes according to hormone receptor status, and Ki-67 index. Tumour grade, enhancement patterns, age, lymph node involvement, ADC values, breast imaging reporting and data system (BI-RADS) category, bilaterality, multifocality, multicentricity, margin, shape, and T2W image intensity were evaluated for these subtypes. RESULTS: ADC values were higher in triple-negative tumours than in luminal A and luminal B tumours (p = 0.010 and p = 0.002, respectively). Circumscribed margin, type 2 enhancement curve, and rim enhancement were significantly higher in triple-negative tumours (p < 0.001). No significant difference was found between the groups in terms of other MRI findings including bilaterality, multifocality, multicentricity, shape, and T2W image intensity (p > 0.05). CONCLUSIONS: ADC values, circumscribed margin, and rim enhancement can provide important information about the tumour's biological behaviour and the course of the disease.

Errors, discrepancies and underlying bias in radiology with case examples: a pictorial review.

Interpretation differences between radiologists and diagnostic errors are significant issues in daily radiology practice. An awareness of errors and their underlying causes can potentially increase the diagnostic performance and reduce individual harm. The aim of this paper is to review both the classification of errors and the underlying biases. Case-based examples are presented and discussed for each type of error and bias to provide greater clarity and understanding.