Current Medical Imaging - Volume 19, Issue 4, 2023

Deep learning is a prominent method for automatic detection of COVID-19 disease using a medical dataset. This paper aims to give a perspective on the data insufficiency issue that exists in COVID-19 detection associated with deep learning. The extensive study of the available datasets comprising CT and X-ray images is presented in this paper, which can be very much useful in the context of a deep learning framework for COVID-19 detection. Moreover, various data handling techniques that are very essential in deep learning models are discussed in detail. Advanced data handling techniques and approaches to modify deep learning models are suggested to handle the data insufficiency problem in deep learning based on COVID-19 detection.

Aim: This review was conducted to assess the structural and functional brain changes associated with vitamin B12 deficiency in different age groups using MRI. Methods: PubMed, Embase, Medline, CINAHL, Scopus, Web of Science, and Google Scholar were searched for magnetic resonance imaging (MRI) studies that explored structural and functional brain changes associated with vitamin B12 deficiency in different age groups. The inclusion criteria were as follows: (1) the population consisted of people and not animals; (2) patients with known B12 deficiency; (3) English publications; (3) at least one of the following brain MRI techniques had been employed: VBM, DWI, DTI, fMRI, or rs-fMRI. However, case reports, systematic reviews and meta-analyses were excluded. Results: Twelve articles met the inclusion criteria. The results of my review show a connection between vitamin B12 deficiency and abnormal structural and functional brain changes in several brain regions. These changes were observed in different age groups ranging from a mean age of seven years to a mean age in the 70s. The results also highlight the association between brain changes and cognitive decline among affected subjects. Improvements in damaged brain regions post-vitamin B12 treatment were also studied. Conclusion: Structural and functional brain damage was found to be associated with vitamin B12 deficiency in all age groups. Vitamin B12 treatment may lead to partial or complete structural and/or functional recovery, as well as a cognitive recovery.

Background: The lung is one of the major organs affected by the SARS-CoV-2 virus. Lung CT scan and RT-PCR are the most valuable diagnostic methods in the early diagnosis and management of COVID-19. Due to the possible inconsistency of the false-negative results for the RT-PCR test, in our study, we aimed to evaluate the sensitivity and specificity of lung CT-scan as an accurate diagnostic method of COVID-19. Methods: In this cross-sectional study, patients suspected of COVID-19 and referred to Shahid Beheshti Hospital in Qom city from February 26 to April 13, 2020, were enrolled. For a definitive diagnosis of COVID-19, chest CT scan and RT-PCR testing was performed for 644 patients, and both sensitivity and specificity of lung CT scan were evaluated. Results: According to the findings, and comparing to the RT-PCR test as the gold standard, sensitivity, specificity as well as, positive predictive and negative predictive values of lung CT-scan were found as follow; 94.47% (95% CI: 90.73 - 97.02%), 24.71% (95% CI: 20.70 - 29.07%), 40.73% (95% CI: 36.58 - 44.99%), 89.08% (95% CI: 82.4 - 94.05%), respectively. Conclusion: According to the findings, the lung CT scan has a better diagnostic value than RTPCR in symptomatic patients who were referred to the hospital for COVID-19 diagnosis. Performing lung CT-scan in patients with negative RT-PCR tests should be assessed.

Background: The new global pandemic caused by the 2019 novel coronavirus (COVID-19), novel coronavirus pneumonia, has spread rapidly around the world, causing enormous damage to daily life, public health security, and the global economy. Early detection and treatment of COVID-19 infected patients are critical to prevent the further spread of the epidemic. However, existing detection methods are unable to rapidly detect COVID-19 patients, so infected individuals are not detected in a timely manner, which complicates the prevention and control of COVID-19 to some extent. Therefore, it is crucial to develop a rapid and practical COVID-19 detection method. In this work, we explored the application of deep learning in COVID-19 detection to develop a rapid COVID-19 detection method. Methods: Existing studies have shown that novel coronavirus pneumonia has significant radiographic performance. In this study, we analyze and select the features of chest radiographs. We propose a chest X-Ray (CXR) classification method based on the selected features and investigate the application of transfer learning in detecting pneumonia and COVID-19. Furthermore, we combine the proposed CXR classification method based on selected features with transfer learning and ensemble learning and propose an ensemble deep learning model based on transfer learning called COVID-ensemble to diagnose pneumonia and COVID-19 using chest x-ray images. The model aims to provide an accurate diagnosis for binary classification (no finding/pneumonia) and multivariate classification (COVID-19/No findings/ Pneumonia). Results: Our proposed CXR classification method based on selection features can significantly improve the CXR classification accuracy of the CNN model. Using this method, DarkNet19 improved its binary and triple classification accuracies by 3.5% and 5.78%, respectively. In addition, the COVIDensemble achieved 91.5% accuracy in the binary classification task and 91.11% in the multi-category classification task. The experimental results demonstrate that the COVID-ensemble can quickly and accurately detect COVID-19 and pneumonia automatically through X-ray images and that the performance of this model is superior to that of several existing methods. Conclusion: Our proposed COVID-ensemble can not only overcome the limitations of the conventional COVID-19 detection method RT-PCR and provide convenient and fast COVID-19 detection but also automatically detect pneumonia, thereby reducing the pressure on the medical staff. Using deep learning models to automatically diagnose COVID-19 and pneumonia from X-ray images can serve as a fast and efficient screening method for COVID-19 and pneumonia.

Background: In the series of improved versions of U-Net, while the segmentation accuracy continues to improve, the number of parameters does not change, which makes the hardware required for training expensive, thus affecting the speed of training convergence. Objective: The objective of this study is to propose a lightweight U-Net to balance the relationship between the parameters and the segmentation accuracy. Methods: A lightweight U-Net with full skip connections and deep supervision (LFU-Net) was proposed. The full skip connections include skip connections from shallow encoders, deep decoders, and sub-networks, while the deep supervision learns hierarchical representations from full-resolution feature representations in outputs of sub-networks. The key lightweight design is that the number of output channels is based on 8 rather than 64 or 32. Its pruning scheme was designed to further reduce parameters. The code is available at: https://github.com/dengdy22/U-Nets. Results: For the ISBI LiTS 2017 Challenge validation dataset, the LFU-Net with no pruning received a Dice value of 0.9699, which achieved equal or better performance with a mere about 1% of the parameters of existing networks. For the BraTS 2018 validation dataset, its Dice values were 0.8726, 0.9363, 0.8699 and 0.8116 on average, WT, TC and ET, respectively, and its Hausdorff95 distances values were 3.9514, 4.3960, 3.0607 and 4.3975, respectively, which was not inferior to the existing networks and showed that it can achieve balanced recognition of each region. Conclusion: LFU-Net can be used as a lightweight and effective method in the segmentation tasks of two and multiple classification medical imaging datasets.

Background: Airway segmentation is a way to quantify the diagnosis of pulmonary diseases, including chronic obstructive problems and bronchiectasis. Manual analysis by radiologists is a challenging task due to the complex airway structure. Additionally, tree-like patterns, varied shapes, sizes, and intensity make the manual airway segmentation task more complex. Deeper airways are even more difficult to segment as their intensity starts matching the lung parenchyma as the diameter of the airway cross-section gets reduced. Objective: Many earlier works have proposed different deep learning networks for airway segmentation but were unable to achieve the desired performance; hence the task of airway segmentation still possesses challenges in this field. Methods: This work proposes a convolutional neural network based on deep U-Net architecture and employs an attention block technique for airway segmentation. The attention mechanism aids in the extraction of the complicated and multi-sized airways found in the lung region, hence increasing the efficiency of the U-Net architecture. Results: The model has been validated using VESSEL12 and EXACT09 datasets, individually and combined, with and without trachea images. The best DSC scores on EXACT09 and VESSEL12 datasets are 95.21% and 95.80%, respectively. The performance on both datasets combined gave a DSC score of 94.1%, showing that the overall performance of the proposed methodology is quite satisfactory. The generalizability of the model is also confirmed using k-fold cross-validation. The comparison of the proposed model to existing research on airway segmentation found competitive results. Conclusion: The use of an attention unit in the proposed model highlights the relevant information and reduces the irrelevant features, which helps to improve the performance and saves time.

Background: Correct delineation of organs at risk (OARs) is an important step for radiotherapy and it is also a time-consuming process that depends on many factors. Objective: An automatic quality assurance (QA) method based on deep learning (DL) was proposed to improve efficiency for detecting contouring errors of OARs. Materials and Methods: A total of 180 planning CT scan sets at the pelvic site and the corresponding OARs contours from clinics were enrolled in this study. Among them, 140 cases were randomly chosen as the training datasets, 20 cases were used as the validation datasets, and the remaining 20 cases were used as the test datasets. DL-based models were trained through data curation for data cleaning based on the Dice similarity coefficient and the 95th percentile Hausdorff distance between the original contours and the predicted contours. All contouring errors could be classified into two types; minor modification required and major modification required. The pass criteria were established using Bias- Corrected and Accelerated bootstrap on 20 manually reviewed validation datasets. The performance of the QA method was evaluated with the metrics of sensitivity, specificity, the area under the receiving operator characteristic curve (AUC), and detection rate sensitivity on the 20 test datasets. Results: For all OARs, segmentation results after data curation were superior to those without. The sensitivity of the QA method was greater than 0.890 and the specificity was higher than 0.975. The AUCs were 0.948, 0.966, 0.965, and 0.932 for the bladder, right femoral head, left femoral head, and rectum, respectively. Almost all major errors could be detected by the automatic QA method, and the lowest detection rate sensitivity of minor errors was 0.863 for the rectum. Conclusions: QA of OARs is an important step for the correct implementation of radiotherapy. The DL-based QA method proposed in this study showed a high potential to automatically detect contouring errors with high precision. The method can be integrated into the existing radiotherapy procedures to improve the efficiency of delineating the OARs.

Objective: The study aims to conduct lung cancer screening by low-dose CT to identify the nature of the pulmonary nodule. The purpose of this study was to evaluate the role of preoperative medical EC glue localization of pulmonary nodules of uncertain nature by minimally invasive surgical resection. Methods: From December 2017 to December 2019, 18 patients (12 women, 6 men; median age: 54 years)with pulmonary nodules were located using medical EC glue under the guidance of preoperative CT and then resected under video thoracoscopy at Air Force Medical Center of PLA. The clinical characteristics were retrospectively collected to evaluate the effectiveness, safety and feasibility of the operation. Results: The mean value of the maximum diameter of pulmonary nodules on CT images before the operation was 10.8 mm. The average depth was 10.3 mm (1.0-39.5 mm). Among 18 nodules, 8 were pure ground glass nodules, 3 were solid nodules, and 7 were partial solid nodules. The diagnosis rate of medical glue localization under the guidance of CT after the operation was 100%. Postoperative pathological diagnosis showed that there were 10 cases of primary lung adenocarcinoma, 1 case of invasive lung adenocarcinoma, 3 cases of adenocarcinoma in situ, 1 case of metastatic adenocarcinoma, and 3 cases of benign nodules. No obvious serious complications were found after localization. Conclusion: This study suggests that CT-guided percutaneous medical EC glue localization is a reliable, safe, feasible and practical method for undiagnosed pulmonary nodules and can significantly improve the rate of resection of small pulmonary nodules. Furthermore, it was considered to be more reasonable to remove pulmonary nodules and maximize the preservation of lung function.

Background: Behçet's disease is a chronic inflammatory disease that involves systemic organs and vascular structures. Arterial system involvement is rarer than the venous system, and it is one of the major causes of death. Vertebral artery (VA) involvement is rare and there are a limited number of reported cases in the literature. This report aimed to present the imaging findings of a case of Behçet's disease with a giant aneurysm originating from the right VA. Case Report: A 35-year-old woman with a 10-year history of Behçet's disease applied to our center with complaints of pain on the right side of the neck. CT angiography showed a 7.5x6.5 cm sized saccular aneurysm in the superior of the right hemithorax adjacent to the mediastinum, with contrast filling the V1 segment of the VA. In our case, the VA aneurysm had no relation to the subclavian artery and showed filling from the contralateral VA. The described findings were demonstrated in Doppler examination, phase-contrast MRI, and conventional and CT angiography. Conclusion: Behçet's disease is a chronic inflammatory disease that can involve the arterial system. Although VA involvement has not been reported frequently, this vasculitic process may affect all arterial structures.

Background: Varicella-zoster virus (VZV) is a latent virus that can remain in the central nervous system. Reactivation of the virus can cause some neurologic manifestations, and myelitis is one of the rarest of them. In this report, we aimed to present the MRI features of long-segment cervical myelitis after VZV infection, which is rarely reported in the literature. Case Presentation: A 69-year-old patient, who was diagnosed with varicella-zoster two months ago and treated with valacyclovir, was admitted to our clinic with worsening of his complaints and weakness in the right upper extremity. Neurological examination was normal in the left upper and bilateral lower extremities, but the muscle strength in the right upper extremity was evaluated as 4/5 grade. While rare leukocytes (10 leukocytes/mm³, 50 erythrocytes/mm³, high protein level (46 mg/dl, ref.15- 40 mg/dl)) were seen in the cerebrospinal fluid (CSF) analysis, no microorganisms were seen, and no growth was observed in the culture. VZV antibody-immunoglobulin G (Ab-IgG) was positive in CSF, while polymerase chain reaction [PCR] for VZV was negative. On cervical MRI, lesions showing T2 hyperintensity were observed from the C3-4 level to the C7-T1 level, eccentrically located in the right paramedian spinal cord. On post-contrast images, patchy heterogeneous contrast enhancement was noted in these regions of the spinal cord. When the patient's history, CSF features and MRI examinations were evaluated together, the lesions were consistent with VZV myelitis. The patient was started on valacyclovir treatment, and during the follow-up, the patient's complaints decreased, while no progress was observed in neurologic symptoms. Conclusion: As a result, we aimed to report the MRI features of this rare complication of varicellazoster and emphasize the necessity of keeping this in mind in the etiology of myelitis, especially in cases with patchy enhancement, to achieve early diagnosis and treatment.

Background: Umbilical vein catheterization (UVC) is an early venous access route in newborns and is frequently used for delivering total parenteral nutrition (TPN) and medications. Vascular, hepatic parenchymal, and infectious complications of UVC can be seen rarely. Objective: We present preterm neonates' X-ray, US, and MRI findings with parenchymal TPN extravasation and portal vein thrombosis. Our case was the first case about MRI findings of TPNoma. Case Presentation: A 30 week female with a birth weight of 1340 g was born via Cesarean section. Due to the diagnosis of transient tachypnea of the newborn and prematurity, the infant was hospitalized in the neonatal intensive care unit. On the first day of hospitalization, UVC was inserted and TPN with 20% lipid content was started. After 10 days, UVC was removed and TPN treatment was continued with a piccline catheter. Abdominal US and portal venous Doppler examination was applied to the patient whose general condition deteriorated on the 12th day. US and Doppler revealed a lesion of 17x17x18mm in size with lobulated contour, hyperechoic heterogeneous internal structure, and no central or peripheral blood supply was observed in the left lobe of the liver. Moreover, a filling defect compatible with a thrombus was observed in the proximal part of the left portal vein. An abdominal MRI was performed to characterize this liver lesion. Axial T1 weighted and T2- weighted images showed a heterogeneous hyperintense lesion without contrast enhancement. Axial fat saturation T1-weighted and out of phase T1-weighted images showed a reduction in signal intensity. US and MRI examinations showed that the thrombosed umbilical vein ended superior to the lesion. In the differential diagnoses, fat-containing lesions such as lipoma-teratoma and fat-containing collection secondary to extravasation of TPN treatment via UVC were thought. Conclusion: In the differential diagnosis of liver localized lesions in newborns, UVC-related liver injury should be considered and the localization of the catheter tip should be checked. In case of the doubt based on US and X-ray findings, presence of fat on MRI could be diagnosed. Serious complications should be avoided with catheter revision or removal.

Introduction: Transthyretin amyloid cardiomyopathy was considered a rare pathology. However, recent studies show a significant prevalence in patients with degenerative aortic stenosis and heart failure with preserved ejection fraction. Case Presentation: An 85-year-old woman presented with a four-month history of pain in the rib cage with a history of diffuse large B-cell lymphoma of the oral cavity, essential thrombocytosis and dyslipidemia. She had no significant family history. A transthoracic echocardiogram showed degenerative aortic stenosis and normal systolic function with preserved left ventricular ejection fraction of 70%. Bone-avid tracer cardiac scintigraphy with technetium-99m-labeled hydroxymethylene diphosphonate with SPECT-CT documented grade two myocardial uptake according to the Perugini scale. MRI evidenced late patchy enhancement in the myocardium associated with diffuse subendocardial enhancement. Laboratory tests showed the absence of mutation in the transthyretin (TTR) gene, serum and urine immunofixation electrophoresis (IFE) negative for monoclonal protein and serum-free light chain (sFLC) assay with a normal kappa/lambda (K/L) ratio. All these findings were compatible with a non-invasive diagnosis of wild-type cardiac amyloidosis. Conclusion: The accepted criteria for the definitive non-invasive diagnosis of amyloid cardiomyopathy are based on myocardial uptake by scintigraphy (with SPECT), serum and urine immunofixation electrophoresis, serum-free light chain assay and suggestive findings on echocardiography and/or MRI. Genetic testing should differentiate between ATTRv (v for variant) and ATTRwt (wt for wild type) forms.

Introduction: Extrauterine endometrial stromal sarcoma (EESS) arising from Deep pelvic endometriosis (DPE) has a poor life quality and is difficult to diagnose pre-operatively. However, the patient’s quality of life can be improved when it is diagnosed precisely and managed successfully. Case Report: A 35-year-old woman presented to our hospital with a 3-month history of hematochezia and anal pain. Initially, she was misdiagnosed as a rectal stromal tumor and then precisely diagnosed as having EESS from DPE following several multidisciplinary team (MDT) meetings. The lesion was shrunk by gonadotrophin-releasing hormone agonist (GnRH-α) treatment and then resected with minimal trauma. Conclusion: MDT is crucial in the treatment of the patient. It can promote individualized treatment and improve patient’s quality of life.

Introduction: Anomalous pulmonary venous connection is a rare congenital variant of the pulmonary veins drained into the right atrium. The left-side partial anomalous pulmonary connections are usually detected in adulthood and occasionally drain into the left brachiocephalic vein. Case Description: An asymptomatic 63-year-old woman with a known history of left carotid body paraganglioma was admitted to our institution to evaluate this tumor by computed tomography angiography of the neck and brain. As an incidental finding, CTA demonstrated a supracardiac partial anomalous pulmonary venous connection between the left upper pulmonary vein and the left innominate via the vertical vein. Additionally, dilation of the azygos vein was observed. Conclusion: Supracardiac left-sided partial anomalous pulmonary venous connection is a rare vascular variant, usually asymptomatic and found incidentally in adults.