Current Medical Imaging - Volume 17, Issue 12, 2021

Background: Neurocutaneous Melanosis (NCM) is a rare nonfamilial phakomatosis characterized by the presence of congenital melanocytic nevi and abnormal melanocyte infiltration of the leptomeninges. Objective and Methods: This paper shows the importance of early diagnosis and the most important imaging features of the disease on CT and MR scans. PubMed database was searched from January 1972 to September 2020. Papers including imaging findings of NCM, clinical, follow-up, and treatment features were collected, selecting only 89 studies. Discussion: NCM is a term used for the first time by van Bogaert in 1948. It refers to a condition caused by an error during morphogenesis and migration leading to leptomeningeal melanocytic accumulation. Although histological findings are the gold standard for diagnosis confirmation, neuroimaging and clinical features strongly support the suspect of NCM. Localization and extension of the lesions are predictive of neurological manifestations related to increased intracranial pressure, mass lesions, or spinal cord compression. CT demonstrates sites of increased density in the anterior temporal lobe, mainly the amygdala, thalami, cerebellum, and frontal lobes base. However, MRI is the best imaging method to diagnose central nervous system lesions, often appearing as T1-short signal areas of the cerebral parenchyma, indicative of central nervous system melanosis. MRI can also reveal associated intracranial and intraspinal abnormalities. Conclusion: Early imaging, when available, is helpful if NCM suspect is raised and may be of guidance in comparing later studies. NCM requires a multidisciplinary approach since it is a multisystem disease with a genetic component.

Background: This paper provides a systematic review of the application of Artificial Intelligence (AI) in the form of Machine Learning (ML) and Deep Learning (DL) techniques in fighting against the effects of novel coronavirus disease (COVID-19). Objective & Methods: The objective is to perform a scoping review on AI for COVID-19 using preferred reporting items of systematic reviews and meta-analysis (PRISMA) guidelines. A literature search was performed for relevant studies published from 1 January 2020 till 27 March 2021. Out of 4050 research papers available in reputed publishers, a full-text review of 440 articles was done based on the keywords of AI, COVID-19, ML, forecasting, DL, X-ray, and Computed Tomography (CT). Finally, 52 articles were included in the result synthesis of this paper. As part of the review, different ML regression methods were reviewed first in predicting the number of confirmed and death cases. Secondly, a comprehensive survey was carried out on the use of ML in classifying COVID-19 patients. Thirdly, different datasets on medical imaging were compared in terms of the number of images, number of positive samples and number of classes in the datasets. The different stages of the diagnosis, including preprocessing, segmentation and feature extraction were also reviewed. Fourthly, the performance results of different research papers were compared to evaluate the effectiveness of DL methods on different datasets. Results: Results show that residual neural network (ResNet-18) and densely connected convolutional network (DenseNet 169) exhibit excellent classification accuracy for X-ray images, while DenseNet-201 has the maximum accuracy in classifying CT scan images. This indicates that ML and DL are useful tools in assisting researchers and medical professionals in predicting, screening and detecting COVID-19. Conclusion: Finally, this review highlights the existing challenges, including regulations, noisy data, data privacy, and the lack of reliable large datasets, then provides future research directions in applying AI in managing COVID-19.

Background: Tibial slope measurements performed using only the proximal part of tibia ignore the native tibial anatomical axis. Our first aim is to measure the native medial, lateral and total tibial slope angles of gender groups using the whole tibial anatomical axis on computerized tomography-based three-dimensional anatomical models. The second aim is to determine the correlation between proximal and whole tibial anatomical axis for measurement of medial, lateral, and total tibial slope angles. Methods: We randomly selected 100 females and 100 males between 18-60 years of age. Three-dimensional anatomical models of right and left tibia were created. The gender-specific differences of medial, lateral, and total tibial slope angles according to proximal and whole tibial anatomical axis were measured. Correlation coefficients (r) of medial, lateral, and total tibial slope angles measured with proximal and whole tibial anatomical axis were calculated. Results: The mean age was 47.1 years. A statistically significant difference was observed between female (7.1 ± 3) and male (8.2 ± 2.5) groups in terms of mean lateral tibial slope angles according to the whole tibial anatomical axis (p=0.008). A strong correlation between proximal and whole tibial anatomical axis for all tibial slope angle measurements was detected. Conclusion: The method we determined for 3D measurement of medial, lateral and total tibial slope angles using proximal tibial anatomical axis has a strong correlation with slope angles measured in accordance with the whole tibial anatomical axis. Our 3D tibial slope angle measurement method on the proximal tibia has high reliability and could be used in the daily practice.

Aims: This work was focused on the detection of early-stage breast tumors and their location. Background: The most frequently seen disease progression and mortality in women's lives is Breast Cancer. Because of breast cancer/tumors, the risk of mortality for women has risen exponentially. From the 2020 deadline for breast cancer, women who died from carcinoma were 123.8 cases per lac women between 2006-2010. This problem is also overcome by the early identification of the tumor using different detection procedures like X-ray mammography, computerized tomography, ultrasound imaging technique, Positron Emission Tomography (PET), Magnetic Resonance Imaging (MRI), microwave imaging. Researchers carried out multiple studies in these areas. Objective: To detect early-stage breast tumors and their location using SAR analysis. Methods: The major dielectrical difference between cancerous breast tissues and normal tissues in this technique is the microwave frequency range. The term Specific Absorption Rate (SAR) describes the amount of energy which is absorbed (W/kg) in the breast tissue. This segment illustrates the usefulness to diagnose the tumor position in the breast by means of maximum SAR value coordinates. Changes in breast and tumor size are important for the risk of diagnosis. The power absorbed in connecting with a normal breast and a tumor breast is measured and equivalent for different breast masses. The maximum SAR is also analyzed at distinct tumor locations at various frequency ranges. Results: It is observed that max SAR coordinates are very close to the actual tumor location. So, the maximal value of SAR coordinates indicates the existence of a tumor in the breast phantom. Conclusion: The simulated data above strongly suggests that the Max SAR values were higher in the breast phantom with tumor as compared to the breast without tumor. With different tumor radius (3 mm and 5 mm) analyzed with different resonant frequencies like 3GHz, 4GHz and 5GHz at the actual tumor location of (0, 0, 35). Even though a model representing the real properties of breast tissue is required to assess the validation of any imaging process, so the real-time development of an equivalent breast phantom and its execution is needed.

Objectives: This study aimed to investigate the accuracy of conventional Sagittal Turbo spin Echo T2-weighted (Sag TSE-T2W), variable flip angle 3D TSE (VFA-3D-TSE) and high-resolution 3D heavily T2W (HR-3D-HT2W) sequences in the diagnosis of primary aqueductal stenosis (PAS) and Superior Medullary Velum Stenosis (SMV-S), and the effect of stenosis localization on diagnosis. Methods: Seventy-seven patients were included in the study. The diagnosis accuracy of the HR-3D-HT2W, Sag TSE-T2W and VFA-3D-TSE sequences, was classified into three grades by two experienced neuroradiologists: grade 0 (the sequence has no diagnostic ability), grade 1 (the sequence diagnoses stenosis but does not show focal stenosis itself or membrane formation), and grade 2 (the sequence makes a definitive diagnosis of stenosis and shows focal stenosis itself or membrane formation). Stenosis localizations were divided into three as Cerebral Aquaduct (CA), Superior Medullary Velum (SMV) and SMV+CA. In the statistical analysis, the grades of the sequences were compared without making a differentiation based on localization. Then, the effect of localization on diagnosis was determined by comparing the grades for individual localizations. Results: In the sequence comparison, grade 0 was not detected in the VFA-3D-TSE and HR-3DHT2W sequences, and these sequences diagnosed all cases. On the other hand, 25.4% of grade 0 was detected with the Sag TSE-T2W sequence (P<0.05). Grade 1 was detected by VFA-3D-TSE in 23% of the cases, while grade 1 (12.5%) was detected by HRH-3D-T2W in only one case, and the difference was statistically significant (P<0.05). When the sequences were examined according to localizations, the rate of grade 0 in the Sag TSE-T2W sequence was statistically significantly higher for the SMV localization (33.3%) compared to CA (66.7%) and SMV+CA (0%) (P<0.05). Localization had no effect on diagnosis using the other sequences. Conclusion: In our study, we found that the VFA-3D-TSE and HR-3D-HT2W sequences were successful in the diagnosis of PAS and SMV-S contrary to the Sag TSE-T2W sequence and especially SMV localization decreases the diagnostic accuracy of Sag TSE-T2W sequence.

Introduction: Cervical cancer is the fourth most common cancer in women. In 2018, it was estimated 570000 women were diagnosed with cervical cancer worldwide and about 311000 women died from the disease. An efficient technique is essential for solving the complication in the diagnosis of cervical cancer images. Materials and Methods: In this research, a new method is developed for cervical cancer image segmentation. First, the RGB image is converted into an HSI color model. The thresholding is applied to the saturation and intensity components to get binary images. These binary images are combined to get a new mask. Using the connected component concept nucleus and cytoplasm are segmented accurately. Results: For the performance evaluation, peak signal-to-noise ratio (PSNR), mean square error (MSE), structural similarity index (SSIM), image quality index (IQI), structural content (SC), normalized cross-correlation (NK), precision (PR), recall (RC), the Average Difference (AD) and image fidelity (IF) are taken. The proposed techniques’ highest PSNR values are 44.2341, 46.7953, 60.5925, and 61.4862, respectively. The proposed segmentation technique can attain a high PSNR (>40db) value at a threshold value equal to 0.1. Also proposed approach attains good precision, recall, and SSIM values. Conclusion: The lowest MSE values using proposed segmentation techniques are 0.0454, 0.0351, 0.0924, and 0.0271 individually. The AD, NK, SC, NAE, and LMSE values for the implemented approach are low, showing the segmented image’s quality is very good. Thus, the proposed model performed better compared to other methods.

Background: Former studies showed that magnetic resonance cholangiopancreatography (MRCP) is useful in diagnosing the presence of santorinicele; however, few studies have evaluated MRCP in diagnosing wirsungocele and the association between pancreatitis and santorinicele or wirsungocele. The purpose of the study was to explore the performance of MRCP in diagnosing santorinicele and wirsungocele and investigate the potential association among pancreatitis, pancreas divisum, and santorinicele or wirsungocele. Methods: Sixty-five patients (mean age, 55.68 years; range, 11-82 years) with santorinicele or wirsungocele were included and sorted into two groups: the santorinicele group (n = 48) and the wirsungocele group (n = 17). All patients underwent MRCP. The images were evaluated for the appearance and size of santorinicele or wirsungocele. The diagnostic sensitivity of MRCP was assessed. Additionally, whether two groups are correlated with pancreas divisum or pancreatitis were investigated. Results: The sensitivity of MRCP in detecting santorinicele and wirsungocele showed no difference (68.8% and 76.5%, respectively). The proportion of patients who developed pancreatitis in santorinicele and wirsungocele groups were 60.4% and 11.8%, respectively (p < 0.05). Pancreas divisum accounted for 77.1% and 11.8% of the patients in the santorinicele and wirsungocele groups, respectively (p < 0.05). Patients with santorinicele and pancreas divisum tended to be older when they acquired pancreatitis. Conclusion: MRCP could be an alternative imaging method to detect cystic dilation of the pancreatic duct. Pancreatitis is more common in patients with santorinicele than in those with wirsungocele. Moreover, santorinicele is more closely associated with pancreatitis than with pancreas divisum.

Aims: To prevent Alzheimer’s disease (AD) from progressing to dementia, early prediction and classification of AD are important and they play a crucial role in medical image analysis. Background: In this study, we employed a transfer learning technique to classify magnetic resonance (MR) images using a pre-trained convolutional neural network (CNN). Objective: To address the early diagnosis of AD, we employed a computer-assisted technique, specifically the deep learning (DL) model, to detect AD. Methods: In particular, we classified Alzheimer’s disease (AD), mild cognitive impairment (MCI), and normal control (NC) subjects using whole slide two-dimensional (2D) images. To illustrate this approach, we made use of state-of-the-art CNN base models, i.e., the residual networks Res- Net-101, ResNet-50, and ResNet-18, and compared their effectiveness in identifying AD. To evaluate this approach, an AD Neuroimaging Initiative (ADNI) dataset was utilized. We also showed uniqueness by using MR images selected only from the central slice containing left and right hippocampus regions to evaluate the models. Results: All three models used randomly split data in the ratio of 70:30 for training and testing. Among the three, ResNet-101 showed 98.37% accuracy, better than the other two ResNet models, and performed well in multiclass classification. The promising results emphasize the benefit of using transfer learning, specifically when the dataset is low. Conclusion: From this study, we know that transfer learning helps to overcome DL problems mainly when the data available is insufficient to train a model from scratch. This approach is highly advantageous in medical image analysis to diagnose diseases like AD.

Background and Aims: In the midst of this pandemic, planning the prioritization of hospital admissions for patients affected with COVID-19 should be of prime concern, particularly in healthcare settings with limited resources. Thus, in this study, we aimed to develop a novel approach to triage COVID-19 patients and attempt to prioritize their hospital admission using Lung Ultrasonography (LUS). The efficacy of LUS in triaging suspected COVID-19 patients and assessing the severity of COVID-19 pneumonia was evaluated; the findings were then compared with those obtained by chest computed tomography (CT). Methods: This multicenter, cross-sectional study comprised 243 COVID-19 patients who presented to the emergency department in 3 major university hospitals in Egypt. LUS was performed by an experienced emergency or chest physician, according to the local protocol of each hospital. Demographic, clinical, and laboratory data were then collected from each patient. Each patient was subjected to chest CT scans and LUS. Results: The mean age of the 243 patients was 46.7 ± 10.4 years. Ground-glass opacity, subpleural consolidation, translobar consolidation, and crazy paving were reported in the chest CT scans of 54.3%, 15.2%, 11.1%, and 8.6% of the patients, respectively. B-line artifacts were observed in 81.1% of the patients (confluent pattern, 18.9%). The LUS findings completely coincided with the CT findings (Kappa agreement value, 0.77) in 197 patients (81.1%) and offered a diagnostic sensitivity of 74%, diagnostic specificity of 97.9%, positive predictive value of 90.2%, and negative predictive value of 93.6% for the COVID-19 patients. Following the addition of O2 saturation to the lung imaging findings, the ultrasound method was able to demonstrate 100% sensitivity and specificity in accurately differentiating between severe and non-severe lung diseases. Conclusion: LUS with oxygen saturation might prove to be effective in prioritizing the hospital admission of COVID-19 patients, particularly in healthcare settings with limited resources.

Background: During the COVID-19 pandemic, the workload of all radiologists and radiology residents, as well as other clinical physicians, has increased. Introduction: This study aims to determine the diagnostic performance of radiology residents, who effectively contribute to the diagnosis of COVID-19. Methods: The thoracic Computed Tomography (CT) images of 135 patients aged 20-83 diagnosed with COVID-19 were evaluated retrospectively by five radiology residents and a radiologist with 10 years of experience. The diagnostic performance of the radiology residents in evaluating COVID-19 was assessed according to their year of residency and the patients’ age and gender. Receiver Operating Characteristic (ROC) curve analysis was performed to determine the sensitivity and specificity of radiology residents. Results: The radiology residents’ performance in determining COVID-19 using CT findings was evaluated as follows: sensitivity 97.22%, specificity 88.89%, positive predictive value 90.91%, negative predictive value 96.55%, and accuracy 93.33%. According to the year of residency, the sensitivity and specificity of the radiology residents in determining COVID-19 using CT images were between 92.3% and 100%, and 71.43% and 100%, respectively. Conclusion: The high sensitivity and specificity of radiology residents in evaluating thoracic CT images for COVID-19 diagnosis indicate that radiologists are as important as clinical physicians in the diagnosis of COVID-19.

Purpose: The purpose of this study was to investigate the influencing factors for chest CT hysteresis and severity of coronavirus disease 2019 (COVID-19). Methods: The chest CT data of patients with confirmed COVID-19 in 4 hospitals were retrospectively analyzed. An independent assessment was performed by one clinician using the DEXIN FACT Workstation Analysis System, and the assessment results were reviewed by another clinician. Furthermore, the mean hysteresis time was calculated according to the median time from progression to the most serious situation to improve chest CT in patients after fever relief. The optimal scaling regression analysis was performed by including variables with statistical significance in univariate analysis. In addition, a multivariate regression model was established to investigate the relationship of the percentage of lesion/total lung volume with lymphocyte and other variables. Results: In the included 166 patients with COVID-19, the average value of the most serious percentage of lesion/total lung volume was 6.62, of which 90 patients with fever had an average hysteresis time of 4.5 days after symptom relief, with a similar trend observed in those without fever. Multivariate analysis revealed that lymphocyte count in peripheral blood and transcutaneous oxygen saturation decreased with the increase of the percentage of lesion/total lung volume. Conclusion: There is a hysteresis effect in the improvement of chest CT image relative to fever relief in patients with COVID-19. The pulmonary lesions may be related to the severity as well as decreased lymphocyte count or percutaneous oxygen saturation.

Background: Until now few studies have specially validated whether the sex, body mass index, or imaging projections of pediatric patients undergoing chest Digital Radiography (DR) affect the radiation dose and image quality. Introduction: To investigate the impact of different photography positions on radiation dose for and image quality of chest DR for 3-4-year-old children. Methods: One-hundred twenty 3-4-year-old patients who required chest DR were included. The patients were divided into 3 groups, with 40 patients in each group: supine Anterior-Posterior Projection (APP), standing APP and posterior-anterior projection (PAP). The Dose Area Product (DAP) and Entrance Surface Dose (ESD) values for every patient were recorded after each exposure. The Visual Grading Analysis Score (VGAS) was used to evaluate image quality. Results: The DAP and ESD values for the standing PAP and APP groups were significantly lower than those for the supine APP group (0.19 ± 0.04 dGy cm² and 0.05 ± 0.01 mGy vs. 0.25 ± 0.05 dGy cm² and 0.08 ± 0.01 mGy, P<0.05, respectively). Additionally, the VGAS for the standing APP group was significantly lower than those for the standing PAP and supine APP groups (28.58 ± 0.96 vs. 29.08 ± 0.94 and 29.03 ± 0.80, P<0.05, respectively), whereas the pulmonary field area for the standing PAP group was significantly higher than those for the standing and supine APP groups (118.95 ± 16.81 cm2 vs. 105.65 ± 14.76 cm² and 105.24 ± 16.32 cm², P<0.05, respectively). However, there were no statistically significant differences in DAP, ESD, VGAS, pulmonary field area and body mass index between the male and female patients in the three groups (P>0.05, respectively). Conclusion: The standing PAP should be the first projection choice for chest DR for 3-4-year-old children; compared with the supine and standing APP, the standing PAP may improve image quality and decrease the required radiation dose.

Background: This study aims to describe the radiological findings in chest radiographs (CXRs) of patients with coronavirus disease (COVID-19) and the Middle East respiratory syndrome (MERS) and to compare the radiological patterns of the two diseases. Methods: A retrospective cohort study was conducted in a tertiary care hospital in the Southern region of Saudi Arabia. The CXRs were evaluated for the presence and distribution of ground-glass opacities (GGO), consolidation, pleural effusions, pneumothorax, cavitation, and nodules. Results: A total of 120 CXRs from 15 MERS and six COVID-19 patients were assessed. In the first available CXR, GGO was present in 11 (73%) of the MERS patients (18% located in the middle zones and 67% located peripherally) and in 100% of COVID-19 patients (peripheral and middle zone involvement in all patients). In the first available CXRs, consolidation was present in 10 MERS patients (67%) and in all six COVID-19 patients, while it was present in all patients of both groups when all the available CXRs were considered. Confluent consolidation was present in 47% of the MERS patients and 50% of COVID-19 patients. All the fatal cases had confluent consolidation and radiological signs of acute respiratory distress syndrome (ARDS). Pneumothorax was present in 13% of patients, with 100% mortality, and pleural effusion was present in 47% of patients, with 57% mortality. Lung nodules were present in one patient of each group. Conclusion: COVID-19 patients showed earlier development of GGO and consolidation with greater involvement of the middle and peripheral lung zones. Confluent consolidation and ARDSlike radiological findings were associated with high mortality. Pneumothorax and pleural effusions were more frequent in MERS patients, with high associated mortality.

Background: COVID-19 generally affects the lungs by causing pneumonic infiltration. Typical Computed Tomography (CT) findings are bilateral, multifocal, peripheral, and patchy ground-glass opacities. These CT findings can also be seen in drug toxicities. Rapid and accurate differential diagnosis of COVID-19 pneumonia from Chemotherapy-Induced Pulmonary Toxicity (CPT) is essential because the treatment strategies are different. In this case report, we define the clinical and lung CT findings of two cases of CPT mimicking COVID-19 pneumonia. Case Reports: The first patient was a 39-year-old male who was under a 5-fluorouracil + folinicacid + oxaliplatin + panitumumab (FOLFOX) treatment of regimen because of metastatic rectal cancer. The second patient was a 53-year-old male with testicular seminoma who was treated with bleomycin, etoposide, and cisplatin (BEP) protocol. Both patients presented to the emergency department with increased dyspnea after chemotherapy. Thorax CT showed similar findings to COVID-19 pneumonia. On follow-up CTs, these findings disappeared with steroid treatment as well as withdrawal of the chemotherapeutic agents and this enabled us to make a differential diagnosis. Also, two consecutive COVID-19 real-time polymerase chain reaction tests were negative in both patients. Conclusion: When lung CT findings of the oncologic patients are similar to COVID-19 pneumonia, chemotherapy-induced pulmonary toxicity should also be considered in the differential diagnosis.

Introduction: Nephrocolic fistula is a pathological connection between the kidney and colon. Percutaneous tumour ablation therapy is a rare cause of iatrogenic nephrocolic fistulain in the literature. Case Presentation: Interventional radiologist should be careful especially in patients with repeated ablation. Granulation tissue may lead to unexpected results secondary to tissue fragility and impedance changes. Conclusion: In addition, we should keep in mind that there is decreasing hydrodissection benefit in cases with previous ablation. As far as we know, this is the first case report of an iatrogenic ephrocolic fistula after microwave ablation for recurrence renal cell carcinoma.