Clinical Hospital Centre Rijeka Research Articles

Infections in Children with Acute Lymphoblastic Leukemia.

Background and Objectives: Infections are the most common and potentially life-threatening complications of the treatment of children with acute lymphoblastic leukemia (ALL). The aim of this study was to determine epidemiological, clinical, and microbiological characteristics of infections in pediatric patients with ALL. Materials and Methods: Twenty-three children (16 males and 7 females, with a mean age of 5.9 years (range of 1.3 to 12.2 years)) with ALL, treated at the Division of Hematology, Oncology, and Clinical Genetics, Department of Pediatrics, Clinical Hospital Center Rijeka, Croatia, from 1 January 2015 to 31 December 2020, were included in the study. Results: One hundred and four infectious episodes (IEs) were reported (an average of 4.5 IE per patient). IEs were more frequent in the intensive phases of antileukemic treatment. Neutropenia was present in 48 IEs (46.2%) with a duration greater than 7 days in 28 IEs (58.3%). The respiratory tract was the most common infection site (48.1%). We documented 49 bacterial (47.1%), 4 viral (3.9%), 4 fungal (3.9%), and 10 mixed isolates (9.6%), while in 37 IEs (35.6%), a pathogen was not isolated. The most common causes of bacteremia were coagulase-positive staphylococci. The most frequent empirical therapy was third- and fourth-generation cephalosporins, followed by piperacillin/tazobactam. The modification of first-line antimicrobial therapy was performed in 56.9% of IEs. Granulocyte-colony stimulating factor was administered in 53.8% of IEs, and intravenous immunoglobulins were administered in 62.5% of IEs. One patient required admission to the intensive care unit. No infection-related mortality was reported. Conclusions: ALL patients have frequent IEs. Close monitoring, the identification of risk factors, the rapid empirical use of antibiotics in febrile neutropenia, and the timely modification of antimicrobial therapy play key roles in reducing infection-related morbidity and mortality in children with ALL.

Building RadiologyNET: an unsupervised approach to annotating a large-scale multimodal medical database

BackgroundThe use of machine learning in medical diagnosis and treatment has grown significantly in recent years with the development of computer-aided diagnosis systems, often based on annotated medical radiology images. However, the lack of large annotated image datasets remains a major obstacle, as the annotation process is time-consuming and costly. This study aims to overcome this challenge by proposing an automated method for annotating a large database of medical radiology images based on their semantic similarity.ResultsAn automated, unsupervised approach is used to create a large annotated dataset of medical radiology images originating from the Clinical Hospital Centre Rijeka, Croatia. The pipeline is built by data-mining three different types of medical data: images, DICOM metadata and narrative diagnoses. The optimal feature extractors are then integrated into a multimodal representation, which is then clustered to create an automated pipeline for labelling a precursor dataset of 1,337,926 medical images into 50 clusters of visually similar images. The quality of the clusters is assessed by examining their homogeneity and mutual information, taking into account the anatomical region and modality representation.ConclusionsThe results indicate that fusing the embeddings of all three data sources together provides the best results for the task of unsupervised clustering of large-scale medical data and leads to the most concise clusters. Hence, this work marks the initial step towards building a much larger and more fine-grained annotated dataset of medical radiology images.

The Type of Surgical Axillary Staging Following Neoadjuvant Systemic Treatment Has No Impact on Breast Cancer Patients’ Oncological Outcomes

Background: Although response-adjusted surgery is a highly recommended strategy following neoadjuvant systemic treatment (NAST), consensus on axillary management in cN+/ycN0 breast cancer patients is still lacking. In this setting, clinical significance of the higher false negative rate of sentinel lymph node biopsy (SLNB) procedure is unknown. The present analysis aims to evaluate the long-term safety of the SLNB in ycN0 patients. Methods: In this study, 60 patients with the operable breast cancer, undergoing surgery after NAST in Clinical Hospital Centre Rijeka, Croatia, from May 2016 to May 2018, were included in the analysis. Following a preliminary retrospective analysis in 2019, follow-up (FU) was extended, and all outcomes were re-evaluated in December 2022. Results: The median FU time was 65 months and 98% of patients had complete FU data. In the ypN0 group, ALND was performed for 15 and SLNB for 20 patients. The median number of LN retrieved in ALND and SLNB was 15 and 3, respectively. The method of surgical axillary staging had no impact on oncological outcomes; Regional Recurrence Free Survival Chi-square=0.5789, P=0.4467; Distant Recurrence Free Survival Chi-square=1.3658, p=0.2425; Breast Cancer Specific Survival Chi-square=0.9755, P=0.3233. Conclusion: Irrespective of a higher FNR following NAST, as compared to the upfront surgery setting, SLNB is a safe procedure and should be considered for all ycN0 patients, regardless of pre-treatment cN status.

Challenges of the Social Workers in the Field of Discharge Planning and Promotion of Patients’ Right to Self-Determination - a Case Report

The aim of this paper was to show the challenges of the social worker in the Clinical Hospital Centre Rijeka in the process of planned discharge and at the same time promoting the right to self-determination of an elderly patient with impaired functioning who has no close family members. The paper presents the tasks of a social worker in the pro- cess of planned discharge, which range from targeted and short analyzes of specific needs, to comprehensive psychosocial assessments and risk assessments of a whole range of social and psychological needs, strengths and stressors. The challenges that were imposed and ultimately overcome were the patient’s lack of social support, different perceptions of the problem and insufficiently transparent communication between stakeholders, the patient’s weakened social capital and physical disability. Therefore, the presentation of the case shows the importance of direct mutual communication of all the mentioned stakeholders, timely exchange and transparency of information. Although, the primary and most important thing is to respect the voice of the patient and enable him to shape the vision of his future life, in accordance with the resources available to him.

Post-COVID Challenges in Occupational Medicine as a Branch of Clinical and Educational Medicine

BACKGROUND: The COVID-19 pandemic has exhausted the employees of hospital center and had a negative impact on medical students. AIM: The aims are to show the efforts of Occupational Medicine to provide support for healthcare workers and to increase student’s interest in O.M. MATERIALS AND METHODS: Out of the total number of employees of the Clinical Hospital Center Rijeka, which is 3500, 275 of them claim for recognition of COVID-19 occupational diseases in 2022. The 6th-year medical students of the Rijeka Faculty of Medicine, 130 in total and 22 sanitary engineers were surveyed by the quality service regarding their satisfaction with the Occupational Medicine classes. Statsoft Statistica 10 was used to calculate the results. RESULTS: All HCW are recognized as having COVID-19 as a professional disease. Older age correlated with greater absenteeism. Hospital employees are dissatisfied, tired, they demand work benefits, contraindications for shift work, night work, work in COVID centers, and requesting maternity leave. Many also quit their jobs the hospital management is looking for and hiring new employees, who quickly leave due to difficult working conditions and low incomes. Occupational medicine also has a problem with students. During the pandemic, classes were held online. Interest in classes declined, a small number of students joined the lectures, but all responded to the survey. The answers showed a lack of interest in the Occupational Medicine course. With subsiding of the pandemic, it would be obligatory to continue “live” classes, take students on a tour of industrial facilities, and cancel the failed evaluation of teachers. CONCLUSION: Easing of working conditions for HCW and changes in teaching is necessary.

Neonatal Intestinal Obstruction Caused by a Perforated Duodenal Web

Aim: To report the clinical courses of two patients with intestinal obstruction caused by duodenal web, who were diagnosed with this rare congenital anomaly at Clinical Hospital Center Rijeka in a span of 5 months. Case report: One of the patients was prenatally suspected to have duodenal obstruction due to ultrasonographic findings, while the other showed signs of obstruction by not tolerating oral intake from the second day of life (DOL). Ceasing enteral feeding, placing a nasogastric tube and introducing intravenous hydration and parenteral nutrition was performed as initial management. Abdominal ultrasonography, abdominal radiography, upper gastrointestinal contrast study and contrast enema were performed but showed no signs of obstruction. Both patients showed improvement in tolerating oral intake over the next few days. However, the first patient did not tolerate meals exceeding 15 mL of infant formula, and the second patient again did not tolerate any oral intake from the 15th day of life. A broad diagnostic workup was repeated but did not show conclusive evidence of obstruction regarding the first patient. Esophagogastroduodenoscopy found duodenal obstruction of unknown origin regarding the second patient. Laparotomy was performed on the 19th and 25th DOL, respectively. Duodenal web was visualized and excised in both patients following duodenotomy. Postoperative recovery was orderly with minor nonsurgical complications. The patients were discharged from hospital on the 21st and 36th postoperative day, respectively. Conclusions: Diagnosing a perforated duodenal web is often challenging because of possible intermittent nature of its symptoms. Exploratory laparotomy is sometimes required to reach diagnosis.

A Novel Likely Pathogenic Variant in the RUNX1 Gene as the Cause of Congenital Thrombocytopenia.

Heterozygous pathogenic and likely pathogenic sequence variants in the RUNX1 (Runt-related Transcription Factor 1) gene are a common genetic cause of decreased platelet count and/or platelet dysfunction and an increased risk of developing myelodysplasia and acute myeloid leukemia. The majority of causative variants are substitutions, which rarely occur de novo. The aim of this case report is to present a patient with congenital thrombocytopenia caused by a deletion variant in exon 9 in the RUNX1 gene. A one-month-old male infant was admitted to the Clinical Hospital Center Rijeka because of anemia and thrombocytopenia verified in the course of an acute viral infection. During follow-up, he occasionally had petechiae and ecchymoses on the lower extremities after mild trauma, with no other symptoms. The patient had persistent slightly decreased values of platelets with normal morphology, but with pathological aggregation with adrenaline and adenosine diphosphate. Due to the unclear etiology of persistent mild thrombocytopenia, he was referred for genetic testing at the age of five. Genomic DNA was isolated from the patient's peripheral blood and whole-exome sequencing was performed using the next-generation sequencing method. A heterozygous frameshift variant, c.1160delG (NM_001754.4), was identified in exon 9. The variant is classified as likely pathogenic. To the best of our knowledge, the heterozygous variant c.1160delG in the RUNX1 gene was first described in our patient. Although pathogenic variants in the RUNX1 genes are very rare, persistently low platelet counts of unclear etiology should raise suspicion of an underlying genetic disorder.

Serum bicarbonate stability study at room temperature- influence of time to centrifugation and air exposure on bicarbonate measurement reported according to the CRESS checklist.

The aim was to evaluate the stability of serum bicarbonate at room temperature, depending on time to centrifugation and air exposure. Stability study was conducted in the laboratory of Clinical Hospital Centre Rijeka, Croatia in January-February 2022. Nine samples from 10 volunteers were collected in clot activator gel tubes (Greiner Bio-One). Bicarbonate was measured on Beckman Coulter AU480 (Beckman Coulter, Brea, USA). Three tubes were left at room temperature for30min, three tubes for 2h, three tubes for 4h until centrifugation. First tube from first group (baseline) wasmeasured immediately after centrifugation. Other measurements were expressed as percentage deviation (PD%) from baseline. First tube was remeasured after 1 and 2h (OT_0h_1h; OT_0h_2h). Second and third tubes were opened 1 and 2h after centrifugation (C_0h_1h; C_0h_2h). Second group of tubes was processed the same way with 2-hour centrifugation delay (WB_2h; OT_2h_1h; OT_2h_2h; C_2h_1h; C_2h_2h), and third group with 4-hour delay (WB_4h; OT_4h_1h; OT_4h_2h; C_4h_1h; C_4h_2h). PD%wascompared to Maximum Permissible Difference (MPD=5.69%). MedCalc statistical software was used (MedCalc, Ostend, Belgium). Bicarbonate baseline mean value (range) was 27.3 (23.4-29.6)mmol/L. Obtained PD% (95%CI) were: C_0h_1h 0.46 (-1.21, 2.12); C_0h_2h 0.18 (-2.22, 2.57); OT_0h_1h-6.46 (-7.57,-5.36); OT_0h_2h-10.67 (-12.13,-9.21); WB_2h-0.15 (-2.04, 1.74); C_2h_1h 0.01 (-1.52, 1.54); C_2h_2h-0.40 (-2.65, 1.85); OT_2h_1h-5.43 (-7.30,-3.55); OT_2h_2h-11.32 (-13.57,-9.07); WB_4h-0.85 (-3.28, 1.58); C_4h_1h-2.52 (-4.93, 0.11); C_4h_2h-3.02 (-5.62, 0.43); OT_4h_1h-7.34 (-9.64,-5.05); OT_4h_2h-11.85 (-14.38,-9.33). Serum bicarbonate is stable for 4h in closed uncentrifuged tubes, another 2h in closed tubes after centrifugation, and is unstable within 1h in opened tube.

Vaccination against SARS-CoV-2 of adult allergic patients – experience from the Clinical Hospital Center Rijeka

Vaccination against SARS-CoV-2 of adult allergic patients – experience from the Clinical Hospital Center Rijeka

SARS-CoV-2 Spike and Nucleocapsid Antibody Response in Vaccinated Croatian Healthcare Workers and Infected Hospitalized Patients: A Single Center Cohort Study.

Studies assessing the dynamics and duration of antibody responses following SARS-CoV-2 infection or vaccination are an invaluable tool for vaccination schedule planning, assessment of risk groups and management of pandemics. In this study, we developed and employed ELISA assays to analyze the humoral responses to Nucleocapsid and Spike proteins in vaccinated health-care workers (HCW) and critically ill COVID-19 patients. Sera of more than 1000 HCWs and critically ill patients from the Clinical Hospital Center Rijeka were tested across a one-year period, encompassing the spread of major SARS-CoV-2 variants of concern (VOCs). We observed 97% of seroconversion in HCW cohort as well as sustained anti-Spike antibody response in vaccinees for more than 6 months. In contrast, the infection-induced anti-Nucleocapsid response was waning significantly in a six-month period. Furthermore, a substantial decrease in vaccinees’ anti-Spike antibodies binding to Spike protein of Omicron VOC was also observed. Critically ill COVID-19 patients had higher levels of anti-Spike and anti-Nucleocapsid antibodies compared to HCWs. No significant differences in anti-Spike and anti-Nucleocapsid antibody levels between the critically ill COVID-19 patients that were on non-invasive oxygen supplementation and those on invasive ventilation support were observed. However, stronger anti-Spike, but not anti-Nucleocapsid, antibody response correlated with a better disease outcome in the cohort of patients on invasive ventilation support. Altogether, our results contribute to the growing pool of data on humoral responses to SARS-CoV-2 infection and vaccination.

Presentation, diagnosis and treatment of giant cell arteritis

Objective: To present patients with giant cell arteritis (GCA) in our centre over a 10 year period in order to confirm current understanding of giant cell arteritis (GCA) and observe any possible specifities in our cohort. Patients and Methods: In this retrospective study all patients diagnosed with GCA in the Clinical Hospital Centre Rijeka from 1 Jan 2011 to 31 Dec 2021 were included. Data were collected on disease presentation and diagnostic workup at the initial exam, and treatment. Results: A total of 51 patients were included in the study, of which 72.55% patients were female, and 72.55% were over 70. Of 50 patients with available data in medical documentation, 8 (16.00%) had systemic disease (S-GCA), while the rest had cranial disease (C-GCA). Headache, considered also a pathognomonic sign, was the most common initial symptom (88.00% in whole cohort, or 97.62% if S-GCA is excluded). Of 49 patients, all had increased erthyrocyte sedimentation rate. Temporal artery biopsy was positive in 12/16 patients with C-GCA, while temporal artery ultrasound was positive in 12/16 patients. Biopsy or ultrasound was not performed in 18 patients, of which 8 were patients with S-GCA, and data were not available for one patient. No patients had both a biopsy and ultrasound performed. All patients were treated with glucocorticoids, 9.80% were also treated with methotrexate, 7.84% with toclizumab, and one with both methotrexate and azathioprine. Conclusion: These results are comparable to other centres in Croatia and at least one centre abroad.

An analysis of the vitamin D overtesting in a tertiary healthcare centre.

IntroductionVitamin D testing is excessively used in clinical practice, despite of the clinical guidelines statements against population screening for vitamin D deficiency. This study aimed to assess an annual number of performed 25-hydroxy vitamin D (25(OH)D) tests that were unsupported by the national guidelines for prevention, detection and therapy of vitamin D deficiency in adults and to calculate associated financial burden for the publicly funded healthcare.Materials and methodsA representative sample of requested 25(OH)D tests in 2018 (N = 474) was formed after selection and randomisation of data set (N = 5298) collected from the laboratory information system database of the Clinical Department for Laboratory Diagnostics, the Clinical Hospital Centre Rijeka. Records were classified in two groups depending on associated medical condition(s) according to the national guidelines. An annual cost of the total and group specific vitamin D testing was calculated on the base of a single test price reimbursed by the Croatian Healthcare Insurance Fund (CHIF).ResultsMedical conditions with high-risk for vitamin D deficiency were detected in 43% (206/474) of vitamin D requests (group 1). Conditions not associated with vitamin D deficiency were detected in 57% (268/474) requests (group 2). A total cost of 25(OH)D testing for the CHIF was 58,729.50 EUR (25,523.79 EUR in the group 1 and 33,205.71 EUR in the group 2).ConclusionsMore than half of all 25(OH)D tests performed in the clinical laboratory represent avoidable cost for the public healthcare. Prevention of population screening by vitamin D testing is needed.

COMPARISON OF BLOOD PRESSURE AND PULSE PRESSURE VALUES IN THE BRACHIAL ARTERY AND IN CENTRAL AORTA MEASURED BY OUTPATIENT NON-INVASIVE METHODS IN DIFFERENT PATIENT POPULATION GROUP

Objective: The aim of our study was to compare the values of blood pressure (BP) and pulse pressure (PP) in the brachial artery and in the central aorta by outpatient non-invasive methods in different patients age groups and according to the presence of chronic renal disease and diabetes. Design and method: The study included 110 outpatients older than 18 years of age with diagnosed and treated arterial hypertension (AH), under the care of Clinical Hospital Centre Rijeka in the period from June 1 to December 1, 2021. Peripheral BP, PP, and central aortic pressure (CAP) were measured in all patients on the same day. The average values of peripheral arterial pressure and the corresponding PP were obtained by the method of continuous blood pressure monitoring (CBPM). CAP and associated PP were measured on the same day by the method of radial arterial tonometry. Patients were divided into three groups according to the age (18–44 years, group I; 45–64 years, group II, and > 65 years, group III). Systolic and diastolic peripheral blood pressure (SPBP, DPBP), systolic CAP and diastolic CAP (SCAP, DCAP), and both peripheral and central PP (PPP, CPP) were nalysed among groups and according to the presence of chronic kidney disease (CKD). Results: In total, 98 patients were nalysed (mean age 55,08 ± 15,52 years; 79,6% essential hypertension; diabetics 12,23%; 26,53% CKD; 48,98% men). There were significant positive correlations between SPBP and SCAP, DPBP and DCAP, and PPP and CPP in group I and group III. There was no significant correlation between SPBP and SCAP (r = 0,268), and there were significant correlations between DPBP and DCAP, and PPP and CPP in group II. According to the presence of CKD, there was no significant correlation between SPBP and SCAP (r = 0,104), and there were positive correlations between DPBP and DCAP, and PPP and CPP. Conclusions: There was no significant correlation between SPBP and SCAP in 45–65 years old and CKD group, which tells us in favor of the atherosclerosis variability in these patient groups, which was expected. Further research on greater number of patients should be done.

Epidemiologic Characteristics of Children with Diabetic Ketoacidosis Treated in a Pediatric Intensive Care Unit in a 10-Year-Period: Single Centre Experience in Croatia.

Background and Objectives: The incidence of severe and moderate forms of DKA as the initial presentation of type 1 diabetes mellitus (T1D) is increasing, especially during the COVID-19 pandemic. This poses a higher risk of developing cerebral edema as a complication of diabetic ketoacidosis (DKA), as well as morbidity and mortality rates. The aim of this study was to determine the trend and clinical features of children treated in the last 10 years in the Pediatric Intensive Care Unit (PICU) due to the development of DKA. Materials and Methods: This retrospective study was performed in the PICU, Clinical Hospital Centre Rijeka, in Croatia. All children diagnosed with DKA from 2011–2020 were included in this study. Data were received from hospital medical documentation and patient paper history. The number of new cases and severity of DKA were identified and classified using recent International Society for Pediatric and Adolescent Diabetes (ISPAD) guidelines. Results: In this investigation period, 194 children with newly diagnosed T1D were admitted to our hospital: 58 of them were treated in the PICU due to DKA; 48 had newly diagnosed T1D (48/58); and ten previously diagnosed T1D (10/58). DKA as the initial presentation of T1D was diagnosed in 24.7% (48/194). Moderate or severe dehydration was present in 76% of the children at hospital admission. Polyuria, polydipsia, and Kussmaul breathing were the most common signs. Three patients (5.2%) developed cerebral edema, of whom one died. Conclusions: During the investigation period a rising trend in T1D was noted, especially in 2020. About one quarter of children with T1D presented with DKA at initial diagnosis in western Croatia, most of them with a severe form. Good education of the general population, along with the patients and families of children with diabetes, is crucial to prevent the development of DKA and thus reduce severe complications.

Conservation Rates of Breast and Axilla Indicate the Quality of Surgical Management of the Early-Stage Breast Cancer Patient

Background: Multidisciplinary care in quality-assured specialized Brest Centre (BC) is nowadays considered optimal management of breast cancer and is associated with better patient outcomes. Modern breast cancer surgery should provide optimal oncological outcomes and preserve the quality of life. European Society of Breast Cancer Specialists (EUSOMA) set up the minimum requirements for a specialized BC and a set of 15 mandatory quality indicator (Q.I.) benchmarks for BC certification purposes. Six Q.I. are intended for breast cancer surgery quality assessment. Methods: In BC of Clinical Hospital Centre Rijeka, data were collected retrospectively for 2019 and prospectively thereafter in a clinical register, encrypted according to the EUSOMA instructions and uploaded into collective EUSOMA database. Following database validation, all Q.I. were calculated for our BC for 3 consecutive years. In addition, a comprehensive on-site audit was performed in 2021 for all services included in breast cancer management in Rijeka. Results: All mandatory surgical Q.I. were above the EUSOMA benchmarks in all 3 years. Non-compliance with EUSOMA recommendations were reported as major, minor, recommendations and observations. For BC surgical department, no major or minor non-conformities were reported. At the national level, mastectomy rates were above the permitted EUSOMA benchmark in 4 consecutive years. Conclusion: We have voluntarily initiated and performed quality control of our BC. The certificate obtained is a confirmation of the high quality of care. Concerning the mastectomy rate in Croatia, other centres should consider quality evaluation to determine the status of detected suboptimal surgical management.

Fast, Reliable, and Simple Point-of-Care-like Adaptation of RT-qPCR for the Detection of SARS-CoV-2 for Use in Hospital Emergency Departments.

During COVID-19 pandemics, the availability of testing has often been a limiting factor during patient admissions into the hospital. To circumvent this problem, we adapted an existing diagnostic assay, Seegene Allplex SARS-CoV-2, into a point-of-care-style direct qPCR (POC dqPCR) assay and implemented it in the Emergency Department of Clinical Hospital Center Rijeka, Croatia. In a 4-month analysis, we tested over 10,000 patients and demonstrated that POC-dqPCR is robust and reliable and can be successfully implemented in emergency departments and similar near-patient settings and can be performed by medical personnel with little prior experience in qPCR.

ESTIMATION OF CEREBROVASCULAR INSULT FREQUENCY IN EMERGENCY PATIENTS AT THE CLINICAL HOSPITAL CENTRE RIJEKA

Purpose: Cerebrovascular stroke is defi ned as a sudden focal neurologic defi cit caused by cerebrovascular disease lasting longer than 24 hours. Cerebrovascular insult (CVI) is divided into ischemic and hemorrhagic. Along with heart disease and malignancies, cerebrovascular disease is the most common cause of mortality and disability in the modern world, attributing to this disease being the leading cause of disability across the globe. Diagnosis and identifi cation of the type of stroke requires imaging examination with computed tomography. A detailed history and imaging usually rule out the impersonation of a stroke. Computed tomography (CT) is one of the fi rst diagnostic modalities used to image the brain to identify the diff erence between ischemic and hemorrhagic stroke. The aim of this retrospective analysis was to estimate the incidence of cerebrovascular insult in patients who underwent CT brain examination in the Clinical Department of Radiology at Clinical Hospital Centre (CHC) Rijeka in the period from January 1, 2020, to December 31, 2020. Materials and methods: During this period, 255 data sets from patients' medical records were anonymously collected. The data search included patients who underwent CT of the brain with a clinical question about CVI. Results: The data analysis at CHC Rijeka showed that the incidence of CVI is slightly higher in the male population. It is commonly diagnosed in the population of 81-90 years. About half of the patients with suspected CVI had normal CT fi ndings. Conclusion: These results could serve as a guideline for targeting groups in preventive healthcare interventions and education. Keywords: computed tomography, cerebrovascular insult, ischemic stroke, hemorrhagic stroke

Improved Deep Convolutional Neural Network to Classify Osteoarthritis from Anterior Cruciate Ligament Tear Using Magnetic Resonance Imaging.

Anterior cruciate ligament (ACL) tear is caused by partially or completely torn ACL ligament in the knee, especially in sportsmen. There is a need to classify the ACL tear before it fully ruptures to avoid osteoarthritis. This research aims to identify ACL tears automatically and efficiently with a deep learning approach. A dataset was gathered, consisting of 917 knee magnetic resonance images (MRI) from Clinical Hospital Centre Rijeka, Croatia. The dataset we used consists of three classes: non-injured, partial tears, and fully ruptured knee MRI. The study compares and evaluates two variants of convolutional neural networks (CNN). We first tested the standard CNN model of five layers and then a customized CNN model of eleven layers. Eight different hyper-parameters were adjusted and tested on both variants. Our customized CNN model showed good results after a 25% random split using RMSprop and a learning rate of 0.001. The average evaluations are measured by accuracy, precision, sensitivity, specificity, and F1-score in the case of the standard CNN using the Adam optimizer with a learning rate of 0.001, i.e., 96.3%, 95%, 96%, 96.9%, and 95.6%, respectively. In the case of the customized CNN model, using the same evaluation measures, the model performed at 98.6%, 98%, 98%, 98.5%, and 98%, respectively, using an RMSprop optimizer with a learning rate of 0.001. Moreover, we also present our results on the receiver operating curve and area under the curve (ROC AUC). The customized CNN model with the Adam optimizer and a learning rate of 0.001 achieved 0.99 over three classes was highest among all. The model showed good results overall, and in the future, we can improve it to apply other CNN architectures to detect and segment other ligament parts like meniscus and cartilages.

Is Coronal Restoration a Predictor of Posttreatment Apical Periodontitis?

Objectives To investigate the posttreatment apical periodontitis (AP) in endodontically treated teeth through a multivariate approach and to analyze the relative importance of quality and type of coronal restoration as predictors of periapical disease. Materials and Methods The present study sample was drawn within 2-year period from the 1,072 consecutive patients older than 18 years, first time attending the Dental Clinic of the Clinical Hospital Center Rijeka, Croatia. A total of 1,350 endodontically treated teeth were included in the study. For each tooth, the periapical status, root filling quality, intracanal post, separated file presence, marginal bone loss, and quality and type of coronal restoration were recorded. Statistical Analysis Chi-square tests were used to analyze the variations in the periapical status, quality of root canal filling, and quality of coronal restoration in different tooth types. The effect of explanatory variables on periapical status was explored using univariate and multivariate logistic regression models. The outcome variable was set as the presence versus absence of AP in the tooth. Results Multivariate logistic regression analysis revealed statistically significant associations and increased risk for AP presence in molars (odds ratio [OR] = 2.15; p < 0.001), teeth positioned in mandible (OR = 1.49; p = 0.007), teeth with short length of root filling (OR = 4.08; p < 0.001), overfilled teeth (OR = 2.99; p = 0.001), and teeth with inadequate density of root filling (OR = 4.14; p < 0.001). Considering variables related to coronal restoration, neither coronal restoration type nor quality was found to be predictive for posttreatment AP. Merely, the presence of intracanal post significantly increased the odds of AP presence (OR = 1.57; p = 0.009). Conclusion The results of the present study did not indicate that type or quality of coronal restoration may be predictors of posttreatment AP. Periapical disease was significantly associated with molars, mandibular teeth, substandard quality of root fillings, and intracanal post presence.

Deep Semi-Supervised Algorithm for Learning Cluster-Oriented Representations of Medical Images Using Partially Observable DICOM Tags and Images.

The task of automatically extracting large homogeneous datasets of medical images based on detailed criteria and/or semantic similarity can be challenging because the acquisition and storage of medical images in clinical practice is not fully standardised and can be prone to errors, which are often made unintentionally by medical professionals during manual input. In this paper, we propose an algorithm for learning cluster-oriented representations of medical images by fusing images with partially observable DICOM tags. Pairwise relations are modelled by thresholding the Gower distance measure which is calculated using eight DICOM tags. We trained the models using 30,000 images, and we tested them using a disjoint test set consisting of 8000 images, gathered retrospectively from the PACS repository of the Clinical Hospital Centre Rijeka in 2017. We compare our method against the standard and deep unsupervised clustering algorithms, as well as the popular semi-supervised algorithms combined with the most commonly used feature descriptors. Our model achieves an NMI score of with respect to the anatomic region, and an NMI score of with respect to the modality. The results suggest that DICOM data can be used to generate pairwise constraints that can help improve medical images clustering, even when using only a small number of constraints.