Automated detection and segmentation of pulmonary embolisms on computed tomography pulmonary angiography (CTPA) using deep learning but without manual outlining

We present a novel computer algorithm to automatically detect and segment pulmonary embolisms (PEs) on computed tomography pulmonary angiography (CTPA). This algorithm is based on deep learning but does not require manual outlines of the PE regions. Given a CTPA scan, both intra- and extra-pulmonary arteries were firstly segmented. The arteries were then partitioned into several parts based on size (radius). Adaptive thresholding and constrained morphological operations were used to identify suspicious PE regions within each part. The confidence of a suspicious region to be PE was scored based on its contrast in the arteries. This approach was applied to the publicly available RSNA Pulmonary Embolism CT Dataset (RSNA-PE) to identify three-dimensional (3-D) PE negative and positive image patches, which were used to train a 3-D Recurrent Residual U-Net (R2-Unet) to automatically segment PE. The feasibility of this computer algorithm was validated on an independent test set consisting of 91 CTPA scans acquired from a different medical institute, where the PE regions were manually located and outlined by a thoracic radiologist (>18 years' experience). An R2-Unet model was also trained and validated on the manual outlines using a 5-fold cross-validation method. The CNN model trained on the high-confident PE regions showed a Dice coefficient of 0.676±0.168 and a false positive rate of 1.86 per CT scan, while the CNN model trained on the manual outlines demonstrated a Dice coefficient of 0.647±0.192 and a false positive rate of 4.20 per CT scan. The former model performed significantly better than the latter model (p<0.01). The promising performance of the developed PE detection and segmentation algorithm suggests the feasibility of training a deep learning network without dedicating significant efforts to manual annotations of the PE regions on CTPA scans.

Can Some Viral Respiratory Infections Observed Before the Pandemic Announcement Be Related to SARS-CoV-2?

OBJECTIVE

There have been doubts that SARS-CoV-2 has been circulating before the first case was announced. The aim of this study was to evaluate the possibility of COVID-19 in some cases diagnosed to be viral respiratory tract infection in the pre-pandemic period in our center.

MATERIAL AND METHODS

Patients who were admitted to our hospital's pulmonary diseases, infectious diseases, and intensive care clinics with the diagnosis of viral respiratory system infection within a 6-month period between October 2019 and March 12, 2020, were screened. Around 248 archived respiratory samples from these patients were analyzed for SARS-CoV-2 ribonucleic acid by real-timequantitative polymerase chain reaction. The clinical, laboratory, and radiological data of the patients were evaluated.

RESULTS

The mean age of the study group was 47.5 (18-89 years); 103 (41.5%) were female and 145 (58.4%) were male. The most common presenting symptoms were cough in 51.6% (n = 128), fever in 42.7% (n = 106), and sputum in 27.0% (n = 67). Sixty-nine percent (n = 172) of the patients were pre-diagnosed to have upper respiratory tract infection and 22.0% (n = 55) had pneumonia, one-third of the patients (n = 84, 33.8%) were followed in the service. Respiratory viruses other than SARS-CoV-2 were detected in 123 (49.6%) patients. Influenza virus (31.9%), rhinovirus (10.5%), and human metapneumovirus (6.5%) were the most common pathogens, while none of the samples were positive for SARS-CoV-2 RNA. Findings that could be significant for COVID-19 pneumonia were detected in the thorax computed tomography of 7 cases.

CONCLUSION

The negative SARS-CoV-2 real-time-quantitative polymerase chain reaction results in the respiratory samples of the cases followed up in our hospital for viral pneumonia during the pre-pandemic period support that there was no COVID-19 among our cases during the period in question. However, if clinical suspicion arises, both SARS and non-SARS respiratory viral pathogens should be considered for differential diagnosis.

Transbronchial lung cr onchial lung cryobiopsy for the diagnosis of diffuse obiopsy for the diagnosis of diffuse parenchymal lung disease: Pitfalls and challenges, a single center experience

BACKGROUND

: Transbronchial lung cryobiopsy (TBLC) is a minimally invasive technique of the diagnosis of diffuse parenchymal lung diseases (DPLD). The aim of this study is to determine the clinical-radiological and histopathological characteristics of patients in whom cryobiopsy contributes to the diagnosis.

METHODS

: In this retrospective study, we searched for the medical records of patients who underwent TBLC from July 2015 to March 2020 at the pulmonology department of our university hospital clinic. Radiological images were evaluated by a chest radiologist experienced in DPLD. Prediagnosis was indicated by clinical-radiological findings. The final diagnosis was determined by the contribution of histopathological diagnosis. The agreement of pretest/posttest diagnosis and the diagnostic yield of TBLC were calculated.

RESULTS

Sixty-one patients with female predominance (59.0%) and current or ex-smoker (49.2%) made up the study population. We found the diagnostic yield of TBLC 88.5%. The most common radiological and clinical-radiological diagnosis was idiopathic pulmonary fibrosis (IPF) (n = 12, 19.6%) while the most common multidisciplinary final diagnosis was cryptogenic organizing pneumonia (COP) (n = 14, %22.9). The concordance of pre/posttests was significant (p < 0.001) with a kappa agreement = 0.485. The usual interstitial pneumonia (UIP) diagnosis was detected in six patients among 12 who were prediagnosed as IPF having also a suspicion of other DPLD by clinical-radiological evaluation (p < 0.001). After the contribution of TBLC, the multidisciplinary final diagnosis of 22(36.1) patients changed. The histopathological diagnosis in which the clinical-radiological diagnosis changed the most was nonspecific interstitial pneumonia (NSIP).

DISCUSSION

We found the overall diagnostic yield of TBLC high. The pretest clinical-radiological diagnosis was often compatible with the multidisciplinary final diagnosis. However, TBLC is useful for the confirmation of clinical radiological diagnosis as well as clinical entities such as NSIP which is difficult to diagnose clinical-radiological. We also suggest that TBLC should be considered in patients whose clinicopathological IPF diagnosis is not precise.

Prevalence of Gastrointestinal Symptoms and Clinical Outcomes in Hospitalized Coronavirus Disease 2019 Patients: A Single-Center Study from Turkey

BACKGROUND

In patients with coronavirus disease 2019, the gastrointestinal symptoms have been reported increasingly in addition to the respiratory system symptoms. The studies show that the prevalence of gastrointestinal system symptoms and how the gastrointestinal system contributes to the severity and prognosis of the disease is still not clear. This study aims to find the prevalence of gastrointestinal symptoms and the correlation between the gastrointestinal symptoms and the clinical results in hospitalized patients diagnosed with coronavirus disease 2019.

METHODS

This study retrospectively analyzes patients diagnosed with coronavirus disease 2019 and hospitalized in the pandemic unit between March 2020 and August 2020 and compares their demographic and clinical characteristics, laboratory and radiologic findings, coronavirus disease 2019 treatments received, the clinical course of the disease, and the gastrointestinal symptoms.

RESULTS

In our study, we included 322 patients diagnosed with coronavirus disease 2019 and hospitalized; 39 patients (12.1%) were admitted to the hospital with at least one gastrointestinal symptom (nausea and vomiting, diarrhea, abdominal pain, and the loss of taste). Nausea and vomiting are the most common gastrointestinal symptoms with a prevalence of 7.1%, followed by diarrhea with 2.8%, the loss of taste with 2.2%, and abdominal pain with 1.5%. The mean age and D-dimer levels of the patients showing gastrointestinal symptoms were lower than those who did not have any gastrointestinal symptoms. We did not find a significant correlation between the presence of the gastrointestinal symptoms and the severity of the disease, treatment received, risk of acute respiratory distress syndrome and septic shock, admission to the intensive care unit, the need for mechanical ventilation, the mortality rate or the length of hospitalization in the medical floor or the intensive care unit.

CONCLUSION

In this study, we observed that 12.1% of coronavirus disease 2019 patients apply to the hospital due to gastrointestinal symptoms. Furthermore, the gastrointestinal symptoms do not seem to affect the severity and the course of the disease, it is important to identify coronavirus disease 2019 patients showing unusual symptoms such as the gastrointestinal symptoms at an early stage to protect healthcare professionals from infection risk.

Contribution of F-18 fluorodeoxyglucose PET/CT and contrast-enhanced thoracic CT texture analyses to the differentiation of benign and malignant mediastinal lymph nodes

BACKGROUND

Texture analysis and machine learning methods are useful in distinguishing between benign and malignant tissues.

PURPOSE

To discriminate benign from malignant or metastatic mediastinal lymph nodes using F-18 fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) and contrast-enhanced computed tomography (CT) texture analyses with machine learning and determine lung cancer subtypes based on the analysis of lymph nodes.

MATERIAL AND METHODS

Suitable texture features were entered into the algorithms. Features that statistically significantly differed between the lymph nodes with small cell lung cancer (SCLC), adenocarcinoma (ADC), and squamous cell carcinoma (SCC) were determined.

RESULTS

The most successful algorithms were decision tree with the sensitivity, specificity, and area under the curve (AUC) values of 89%, 50%, and 0.692, respectively, and naive Bayes (NB) with the sensitivity, specificity, and AUC values of 50%, 81%, and 0.756, respectively, for PET/CT, and NB with the sensitivity, specificity, and AUC values of 10%, 96%, and 0.515, respectively, and logistic regression with the sensitivity, specificity, and AUC values of 21%, 83%, and 0.631, respectively, for CT. In total, 13 features were able to differentiate SCLC and ADC, two features SCLC and SCC, and 33 features ADC and SCC lymph node metastases in PET/CT. One feature differed between SCLC and ADC metastases in CT.

CONCLUSION

Texture analysis is beneficial to discriminate between benign and malignant lymph nodes and differentiate lung cancer subtypes based on the analysis of lymph nodes.

The Predictive Values of Respiratory Rate Oxygenation Index and Chest Computed Tomography Severity Score for High-Flow Nasal Oxygen Failure in Critically Ill Patients with Coronavirus Disease-2019

Background:

The prediction of high-flow nasal oxygen (HFNO) failure in patients with coronavirus disease-2019 (COVID-19) having acute respiratory failure (ARF) may prevent delayed intubation and decrease mortality.

Aims:

To define the related risk factors to HFNO failure and hospital mortality.

Study Design:

Retrospective cohort study.

Methods:

To this study, 85 critically ill patients (≥18 years) with COVID-19 related acute kidney injury who were treated with HFNO were enrolled. Treatment success was defined as the de-escalation of the oxygenation support to the conventional oxygen therapies. HFNO therapy failure was determined as the need for invasive mechanical ventilation or death. The patients were divided into HFNO-failure (HFNO-F) and HFNO-success (HFNO-S) groups. Electronic medical records and laboratory data were screened for all patients. Respiratory rate oxygenation (ROX) index on the first hour and chest computed tomography (CT) severity score were calculated. Factors related to HFNO therapy failure and mortality were defined.

Results:

This study assessed 85 patients (median age 67 years, 69.4% male) who were divided into two groups as HFNO success (n = 33) and HFNO failure (n = 52). The respiratory rate oxygenation (ROX) was measured at 1 hour and the computed tomography (CT) score indicated HFNO failure and intubation, with an area under the receiver operating characteristic of 0.695 for the ROX index and 0.628 for the CT score. A ROX index of <3.81 and a CT score of >15 in the first hour of therapy were the predictors of HFNO failure and intubation. Age, Acute Physiology and Chronic Health Evaluation II score, arterial blood gas findings “(i.e., partial pressure of oxygen [PaO₂], PaO₂ [fraction of inspired oxygen]/SO₂ [oxygen saturation] ratio)”, and D-dimer levels were also associated with HFNO failure; however, based on logistic regression analysis, a calculated ROX on the first hour of therapy of <3.81 (odds ratio [OR] = 4.78, 95% confidence interval [CI] = 1.75–13.02, P = 0.001) and a chest CT score of >15 (OR = 2.83, 95% CI = 1.01–7.88, P = <0.001) were the only independent risk factors. In logistic regression analysis, a ROX calculated on the first hour of therapy of <3.81 (OR = 4.78, [95% CI = 1.75–13.02], P = 0.001) and a chest CT score of >15 (OR 2.83, 95% CI = 1.01–7.88, P = <0.001) were the independent risk factors for the HFNO failure. The intensive care unit and hospital mortality rates were 80.2% and 82.7%, respectively, in the HFNO failure group.

Conclusion:

The early prediction of HFNO therapy failure is essential considering the high mortality rate in patients with HFNO therapy failure. Using the ROX index and the chest CT severity score combined with the other clinical parameters may reduce mortality. Additionally, multi-centre observational studies are needed to define the predictive value of ROX and chest CT score not only for COVID-19 but also other causes of ARF.

Clinical characteristics and risk factors for 28-day mortality in critically ill patients with COVID-19: a retrospective cohort study

Background

To date, the coronavirus disease 2019 (COVID-19) caused more than 2.6 million deaths all around the world. Risk factors for mortality remain unclear. The primary aim was to determine the independent risk factors for 28-day mortality.

Materials and Methods

In this retrospective cohort study, critically ill patients (≥ 18 years) who were admitted to the intensive care unit due to COVID-19 were included. Patient characteristics, laboratory data, radiologic findings, treatments, and complications were analyzed in the study.

Results

A total of 249 patients (median age 71, 69.1% male) were included in the study. 28-day mortality was 67.9% (n = 169). The median age of deceased patients was 75 (66–81). Of them, 68.6% were male. Cerebrovascular disease, dementia, chronic kidney disease, and malignancy were significantly higher in the deceased group. In the multivariate analysis, sepsis/septic shock (OR, 15.16, 95% CI, 3.96–58.11, p < 0.001), acute kidney injury (OR, 4.73, 95% CI, 1.55–14.46,p = 0.006), acute cardiac injury (OR, 9.76, 95% CI, 1.84–51.83, p = 0.007), and chest CT score higher than 15 (OR, 4.49, 95% CI, 1.51-13.38, p = 0.007) were independent risk factors for 28-day mortality.

Conclusion

Early detection of the risk factors and the use of chest CT score might improve the outcomes in patients with COVID-19.

Evaluation of COVID-19 antibody response with using three different tests

Background and Objectives: In this study, the performance of three different commercial antibody assays for COVID-19 was examined and parameters affecting the antibody response were investigated. The correlation of patients’ chest CT results, procalcitonin, CRP, and D-dimer levels with the antibody response were retrospectively evaluated. Materials and Methods: COVID-19 antibodies were detected by three commercially available assays in each patient. Two of the assays were rapid immunochromatographic tests and - one was an ELISA-based IgG assay. SARS-CoV-2 RNA was tested by “COVID-19 RT-qPCR Detection Kit” using nasopharyngeal swab samples. The results of antibody tests were com- pared with each other, RT-qPCR, Biochemical parameters and chest CT findings. Results: RT-qPCR was positive in 46.6% (41/88) of the evaluated patients among which 77.3% (68/88) were healthcare workers. Seventeen (41.4%) of viral RNA positive patients had a positive antibody result with at least two assays. Both of the rapid immunochromatographic tests had identical sensitivity of 36.6% and specificity of 100%, compared to RT-qPCR assay; while the sensitivity of the ELISA based Euroimmune test was 43.9%, and the specificity was 95.7%. The sensitivity and specificity of the immunochromatographic tests were 75% and 100% respectively, compared to ELISA test result. There was a correlation between antibody positivity and old age and male gender. The presence of typical chest CT findings increased the antibody positivity 13.62 times. Antibody positivity was also increased with the decrease in Ct value of the PCR assay. There was no significant relationship between the biochemical parameters (CRP, D-dimer and procalcitonin values) and the antibody or RT-qPCR results. Conclusion: There was a correlation between antibody response and male gender, older age, presence of symptoms, typical chest CT findings and low PCR-Ct value.

Three cases of immunoglobulin G4-related respiratory disease with uncommon imaging findings

Background:

Immunoglobulin G4-related disease (IgG4-RD) is a rare multisystemic idiopathic fibroinflammatory disorder. The rare form of IgG4-RD with isolated thorax involvement is called immunoglobulin G4-related respiratory disease (IgG4-RRD). IgG4-RRD, which is reported in a limited number of cases in the literature, can be categorized into four types on the prevalent chest computed tomography (CCT) findings: solid nodular, round-shaped ground-glass opacity, alveolar interstitial, and bronchovascular. Solid nodular form of IgG4-RRD with mass-like lesions is sporadic and described in the literature with a small number of case reports.

Objectives/Methods:

We aim to present the radiologic, pathologic, and clinical findings of three cases of IgG4-RRD mimicking lung cancer.

Results:

In all three patients, IgG4-RRD occurred with mass-like lesions in the thorax. In case-1 and 2, CCT showed multiple, nodular lesions and multiple mediastinal lymph nodes. On positron emission tomography with 2-deoxy-2-[fluorine-18] fluoro- D-glucose integrated with computed tomography (18F-FDG PET/CT), the masses showed increased 18F-FDG uptake in case-2 and 3. The gold standard histopathological verification for IgG4-RRD was provided for all cases.

Conclusions:

IgG4-RD is an immune-mediated condition comprised of a collection of disorders that share particular pathologic, radiologic, serologic, and clinical features. Isolated IgG4-RRD is rarely seen and is available in the literature as case reports. IgG4-RRD, which can make lung involvement in different patterns, rarely appears with mass-like lesions. Still, IgG4-RRD must be considered in the differential diagnosis of mass lesions detected in CCT. Laboratory, radiological, and histopathological findings of the disease should be evaluated together for an accurate diagnosis.

Abdominal multi-organ segmentation with cascaded convolutional and adversarial deep networks

Abdominal anatomy segmentation is crucial for numerous applications from computer-assisted diagnosis to image-guided surgery. In this context, we address fully-automated multi-organ segmentation from abdominal CT and MR images using deep learning. The proposed model extends standard conditional generative adversarial networks. Additionally to the discriminator which enforces the model to create realistic organ delineations, it embeds cascaded partially pre-trained convolutional encoder-decoders as generator. Encoder fine-tuning from a large amount of non-medical images alleviates data scarcity limitations. The network is trained end-to-end to benefit from simultaneous multi-level segmentation refinements using auto-context. Employed for healthy liver, kidneys and spleen segmentation, our pipeline provides promising results by outperforming state-of-the-art encoder-decoder schemes. Followed for the Combined Healthy Abdominal Organ Segmentation (CHAOS) challenge organized in conjunction with the IEEE International Symposium on Biomedical Imaging 2019, it gave us the first rank for three competition categories: liver CT, liver MR and multi-organ MR segmentation. Combining cascaded convolutional and adversarial networks strengthens the ability of deep learning pipelines to automatically delineate multiple abdominal organs, with good generalization capability. The comprehensive evaluation provided suggests that better guidance could be achieved to help clinicians in abdominal image interpretation and clinical decision making.

Diaphragmatic Ultrasound in Non-Cystic Fibrosis Bronchiectasis: Relationship to Clinical Parameters

The aim of this study was to assess diaphragm thickness (DT) and mobility (DM) and to investigate their relationship to clinical parameters in patients with non-cystic fibrosis (non-CF) bronchiectasis. Thirty-eight patients with non-CF bronchiectasis were enrolled in this cross-sectional study. DT was measured using ultrasound at different lung volumes (at residual volume [DT_RV], functional residual capacity [DT_FRC] and total lung capacity [DT_TLC]). DM was measured using ultrasound during quiet breathing (DM_QB) and deep breathing (DM_DB). Disease severity, pulmonary function, respiratory muscle strength, exercise capacity and physical activity were assessed. DT_RV correlated with disease severity (ρ = 0.332, p = 0.042), FEV1% (r = 0.387, p = 0.016) and FVC% (r = 0.405, p = 0.012). DT_FRC correlated with FVC% (r = 0.331, p = 0.042). DT_TLC correlated with disease severity (r = 0.430, p = 0.007) and total physical activity time (r = 0.379, p = 0.019). DM_DB correlated with disease severity (ρ = -0.380, p = 0.019), FEV1% (r = 0.369, p = 0.023) and FVC% (r = 0.405, p = 0.012). DT is related to disease severity, pulmonary function and physical activity, while DM is related to disease severity and pulmonary function in patients with non-CF bronchiectasis.

COVID-19 S: A new proposal for diagnosis and structured reporting of COVID-19 on computed tomography imaging

PURPOSE

Because of the widespread use of CT in the diagnosis of COVID 19, indeterminate presentations such as single, few or unilateral lesions amount to a considerable number. We aimed to develop a new classification and structured reporting system on CT imaging (COVID-19 S) that would facilitate the diagnosis of COVID-19 in the most accurate way.

METHODS

Our retrospective cohort included 803 patients with a chest CT scan upon suspicion of COVID 19. The patients' history, physical examination, CT findings, RT PCR, and other laboratory test results were reviewed, and a final diagnosis was made as COVID 19 or non-COVID 19. Chest CT scans were classified according to the COVID 19 S CT diagnosis criteria. Cohen's kappa analysis was used.

RESULTS

Final clinical diagnosis was COVID-19 in 98 patients (12%). According to the COVID-19 S CT diagnosis criteria, the number of patients in the normal, compatible with COVID 19, indeterminate and alternative diagnosis groups were 581 (72.3%), 97 (12.1%), 16 (2.0%) and 109 (13.6%). When the indeterminate group was combined with the group compatible with COVID 19, the sensitivity and specificity of COVID-19 S were 99.0% and 87.1%, with 85.8% positive predictive value (PPV) and 99.1% negative predictive value (NPV). When the indeterminate group was combined with the alternative diagnosis group, the sensitivity and specificity of COVID-19 S were 93.9% and 96.0%, with 94.8% PPV and 95.2% NPV.

CONCLUSION

COVID-19 S CT classification system may meet the needs of radiologists in distinguishing COVID-19 from pneumonia of other etiologies and help optimize patient management and disease control in this pandemic by the use of structured reporting.

The effect of bone mineral density on development of Schmorl's nodes in young patients

OBJECTIVE

The aim of this study was to detect the relationship between the development of Schmorl's nodes (SNs) and bone mineral density (BMD) in young patients.

METHODS

Computerized tomography (CT) images of the thoracolumbar vertebral column were retrospectively examined by two experienced radiologists for SNs. The diagnostic criterion for SN was defined as a node size larger than one-third but not more than two-thirds of the relevant vertebral endplate. Considering the eligibility criteria, a total of 74 individuals (60 males and 14 females; mean age: 24.3 years; age range: 18-40 years) with SN at the thoracolumbar vertebrae were included in the patient group, and a total of 38 age- and gender-matched individuals (30 males and 8 females; mean age: 25 years) with no evidence of SN were included in the control group. All these individuals were younger than 40 years. In the patient group, SNs were assessed in terms of the distribution of the thoracolumbar vertebrae, the location of the upper and lower endplates, and the total number of lesions. In all individuals included in the study, BMD was measured from the axial CT sections by quantitative CT and then compared between the two groups.

RESULTS

The distribution of age and gender was comparable between the two groups (p=0.438). A total of 208 SNs were identified in the patient group. Of these, 92 (44%) were located at the thoracic vertebrae and 116 (56%) at the lumbar vertebrae. The mean BMD was 131.6 g/cm3 in the patient group and 140.7 g/cm3 in the control group (p=0.03). There was no significant relationship between the total number of SNs per patient and the mean BMD (p=0.156).

CONCLUSION

Evidence from this study revealed that low BMD may be a predisposing factor for the development of SNs in patients younger than 40 years.

LEVEL OF EVIDENCE

Level III, Diagnostic Study.

Reversible 3D compression of segmented medical volumes: usability analysis for teleradiology and storage

BACKGROUND

DICOM standard does not have modules that provide the possibilities of two-dimensional Presentation States to three-dimensional (3D). Once the final 3D rendering is obtained, only video/image exporting or snapshots can be used. To increase the utility of 3D Presentation States in clinical practice and teleradiology, the storing and transferring the segmentation results, obtained after tedious procedures, can be very effective.

PURPOSE

To propose a strategy for preserving interaction and mobility of visualizations for teleradiology by storing and transferring only binary segmented data, which is effectively compressed by modern adaptive and context-based reversible methods.

MATERIAL AND METHODS

A diverse set of segmented data, which include four abdominal organs (liver, spleen, right, and left kidneys) from 20 T1-DUAL and 20 T2-SPIR MRI, liver from 20 CT, and abdominal aorta with aneurysms (AAA) from 19 computed tomography-angiography datasets, are collected. Each organ is segmented manually by expert physicians, and binary volumes are created. The well-established reversible binary compression methods PNG, JPEG-LS, JPEG-XR, CCITT-G4, LZW, JBIG2, and ZIP are applied to medical datasets. Recently proposed context-based (3D-RLE) and adaptive (ABIC) algorithms are also employed. The performance assessment has been presented in terms of the compression ratio that is a universal compression metric.

RESULTS

Reversible compression of binary volumes results with substantial decreases in file size such as 254 to 2.14 MB for CT-AAA, 56.7 to 0.3 MB for CT-liver. Moreover, compared to the performance of well-established methods (i.e., mean 76.14%), CR is observed to be increased significantly for all segmented organs from both CT and MRI datasets when ABIC (95.49%) and 3D-RLE (94.98%) are utilized. The hypothesis is that morphological coherence of scanning procedure and adaptation between the segmented organs, that is, bi-level images, contributes to compression performance. Although the performance of well-established techniques is satisfactory, the sensitivity of ABIC to modality type and the advantage of 3D-RLE when the spatial coherence between the adjacent slices are high results with up to 10 times more CR performance.

CONCLUSION

Adaptive and context-based compression strategies allow effective storage and transfer of segmented binary data, which can be used to re-produce visualizations for better teleradiology practices preserving all interaction mechanisms.

Comparison of semi-automatic and deep learning-based automatic methods for liver segmentation in living liver transplant donors

PURPOSE

To compare the accuracy and repeatability of emerging machine learning based (i.e. deep) automatic segmentation algorithms with those of well-established semi-automatic (interactive) methods for determining liver volume in living liver transplant donors at computerized tomography (CT) imaging.

METHODS

A total of 12 (6 semi-, 6 full-automatic) methods are evaluated. The semi-automatic segmentation algorithms are based on both traditional iterative models including watershed, fast marching, region growing, active contours and modern techniques including robust statistical segmenter and super-pixels. These methods entail some sort of interaction mechanism such as placing initialization seeds on images or determining a parameter range. The automatic methods are based on deep learning and they include three framework templates (DeepMedic, NiftyNet and U-Net) the first two of which are applied with default parameter sets and the last two involve adapted novel model designs. For 20 living donors (6 training and 12 test datasets), a group of imaging scientists and radiologists created ground truths by performing manual segmentations on contrast material-enhanced CT images. Each segmentation is evaluated using five metrics (i.e. volume overlap and relative volume errors, average/RMS/maximum symmetrical surface distances). The results are mapped to a scoring system and a final grade is calculated by taking their average. Accuracy and repeatability were evaluated using slice by slice comparisons and volumetric analysis. Diversity and complementarity are observed through heatmaps. Majority voting and Simultaneous Truth and Performance Level Estimation (STAPLE) algorithms are utilized to obtain the fusion of the individual results.

RESULTS

The top four methods are determined to be automatic deep models having 79.63, 79.46 and 77.15 and 74.50 scores. Intra-user score is determined as 95.14. Overall, deep automatic segmentation outperformed interactive techniques on all metrics. The mean volume of liver of ground truth is found to be 1409.93 mL ± 271.28 mL, while it is calculated as 1342.21 mL ± 231.24 mL using automatic and 1201.26 mL ± 258.13 mL using interactive methods, showing higher accuracy and less variation on behalf of automatic methods. The qualitative analysis of segmentation results showed significant diversity and complementarity enabling the idea of using ensembles to obtain superior results. The fusion of automatic methods reached 83.87 with majority voting and 86.20 using STAPLE that are only slightly less than fusion of all methods that achieved 86.70 (majority voting) and 88.74 (STAPLE).

CONCLUSION

Use of the new deep learning based automatic segmentation algorithms substantially increases the accuracy and repeatability for segmentation and volumetric measurements of liver. Fusion of automatic methods based on ensemble approaches exhibits best results almost without any additional time cost due to potential parallel execution of multiple models.

Vertebral body bone mineral density in patients with lumbar spondylolysis: a quantitative CT study

PURPOSE

Spondylolysis is known to be a part of a disease process, which describes a defect in the pars interarticularis of vertebra. We aimed to use quantitative computed tomography (QCT) to measure vertebral body bone mineral density (BMD) in patients with lumbar spondylolysis and compare it with readings in controls.

METHODS

Forty symptomatic patients with lumbar spondylolysis aged 18-52 years and 40 matched controls of same sex and approximate age (±2 years) were included in the study. Measurements of BMD were performed by QCT analysis for each vertebral body from T12 to L5 and mean BMD was calculated for each case.

RESULTS

Of 40 patients, 22 (55%) demonstrated L5 spondylolysis, 14 (35%) L4 spondylolysis, three (7.5%) L3 spondylolysis, and one (2.5%) L2 spondylolysis. Spondylolisthesis was found in 29 patients (73%). Patients with spondylolisthesis were significantly older than patients without spondylolisthesis (42±6.9 vs. 37.2±5.4, P = 0.024). Mean BMD value of the patient group was significantly lower than that of the controls (105±24 mg/cm³ vs. 118.7±25.6 mg/cm³, P = 0.015). Subgroup analysis of 19 patients and 19 controls under the age of 40 revealed that the mean BMD value of the patients was significantly lower than that of the controls in the younger age group as well (108.7±23.5 mg/cm³ vs. 130±25.8 mg/cm³, P = 0.009).

CONCLUSION

This study demonstrated that patients with spondylolysis had significantly lower mean vertebral body BMD compared with controls.

Abdominal sarcoidosis: cross-sectional imaging findings

Sarcoidosis is a multisystem inflammatory disease of unknown etiology. The lungs and the lymphoid system are the most commonly involved organs. Extrapulmonary involvement is reported in 30% of patients, and the abdomen is the most common extrapulmonary site with a frequency of 50%-70%. Although intra-abdominal sarcoidosis is usually asymptomatic, its presence may affect the prognosis and treatment options. The lesions are less characteristic and may mimick neoplastic or infectious diseases such as lymphoma, diffuse metastasis, and granulomatous inflammation. The liver and spleen are the most common abdominal sites of involvement. Sarcoidosis of the gastrointestinal system, pancreas, and kidneys are extremely rare. Adenopathy which is most commonly found in the porta hepatis, exudative ascites, and multiple granulomatous nodules studding the peritoneum are the reported manifestations of abdominal sarcoidosis. Since abdominal sarcoidosis is less common and long-standing, unrecognized disease can result in significant morbidity and mortality. Imaging contributes to diagnosis and management of intra-abdominal sarcoidosis. In this report we reviewed the cross-sectional imaging findings of hepatobiliary, gastrointestinal, and genitourinary sarcoidosis.

An Aortoenteric Fistula Arising after Endovascular Management of a Mycotic Abdominal Aortic Aneurysm Complicated with a Psoas Abscess

Mycotic aortic aneurysms account for 1-3% of all aortic aneurysms. The management of this disease is controversial. Since open surgical repair is associated with high morbidity and mortality rates, endovascular aneurysm repair is an alternative treatment method with promising early and midterm outcomes, although its long-term durability is unknown. Secondary aortoenteric fistulas may occur iatrogenically after either aortic reconstructive surgery or endovascular repair. As the number of aneurysms managed with endovascular aneurysm repair has substantially increased, cases of aortoenteric fistulas referred for endovascular repair are augmented. We report the case of an aortoduodenal fistula manifested with duodenal perforation after staged endovascular and surgical treatment of a mycotic aortic aneurysm.

Quantitative Evaluation of the Fetal Cerebellar Vermis Using the Median View on Two-Dimensional Ultrasound

Background

Evaluation of the cerebellum and vermis is one of the integral parts of the fetal cranial anomaly screening.

Objectives

The aim of this study was to create a nomogram for fetal vermis measurements between 17 and 30 gestational weeks.

Patients and Methods

This prospective study was conducted on 171 volunteer pregnant women between March 2013 and December 2014. Measurements of the fetal cerebellar vermis diameters in the sagittal plane were performed by two-dimensional transabdominal ultrasonography.

Results

Optimal median planes were obtained in 117 of the cases. Vermian diameters as a function of gestational age were expressed by regression equations and the correlation coefficients were found to be highly statistically significant (P < 0.001). The normal mean (± standard deviation) for each gestational week was also defined.

Conclusion

This study presents the normal range of the two-dimensional fetal vermian measurements between 17 and 30 gestational weeks. In the absence of a three-dimensional ultrasonography, two-dimensional ultrasonography could also be used confidently with more time and effort.

Prognostic Value of Fetal Thymus Size in Intrauterine Growth Restriction

OBJECTIVES

Our aim was to evaluate the size of the fetal thymus by sonography in pregnancies with intrauterine growth restriction (IUGR) and to search for a possible relationship between a small fetal thymus and adverse perinatal outcomes.

METHODS

The transverse diameter of the fetal thymus was prospectively measured in 150 healthy and 143 IUGR fetuses between 24 and 40 weeks' gestation. The fetuses with IUGR were further divided according to normal or abnormal Doppler assessment of the umbilical and middle cerebral arteries and ductus venosus. Measurements were compared with reference ranges from controls. To determine which perinatal outcomes were independently associated with a small fetal thymus, a multivariate logistic regression analysis was performed.

RESULTS

Thymus size was significantly lower in IUGR fetuses compared to controls (P < .05). Among IUGR fetuses, thymus size was significantly smaller in IUGR fetuses with abnormal Doppler flow compared to normal flow (P < .05). A small thymus in IUGR fetuses was independently associated with early delivery (odds ratio [OR], 1.24; 95% confidence interval [CI], 1.05-1.49; P= .023), respiratory distress syndrome (OR, 1.36; 95% CI, 1.09-1.78; P= .005), early neonatal sepsis (OR, 1.65; 95% CI, 1.11-2.42; P= .001), and a longer stay in the neonatal intensive care unit (OR, 1.33; 95% CI, 1.08-1.71; P = .017).

CONCLUSIONS

Intrauterine growth restriction is associated with fetal thymic involution, and a small fetal thymus is an early indicator of adverse perinatal outcomes in pregnancies complicated by IUGR.

Obstetric and neurodevelopmental outcome in fetal cerebral ventriculomegaly

PURPOSE

The purpose of this study is to establish the obstetric and early neurological outcomes of fetuses diagnosed with intrauterine ventriculomegaly (VM).

MATERIALS AND METHODS

This retrospective study included 27 fetuses with VM diagnosed by ultrasound (US) and referred for in utero magnetic resonance imaging (MRI). US and MRI reports and laboratory test results were obtained including chromosome analysis, congenital infections, and first and second trimester screening tests. Infants were evaluated for clinical outcome for six to 24 months of age.

RESULTS

Twenty (51%) fetuses had mild and 19 (49%) fetuses had severe VM. Accompanying central nervous system (CNS) anomalies were statistically significantly more common in severe VM group. The outcome of mild VM group was statistically significantly better than in the severe VM group.

CONCLUSIONS

The authors conclude that ventricular dimension is a significant prognostic factor to detennine the outcome of fetal cerebral VM. The presence of accompanying CNS anomalies is more common with severe VM and may be considered as an unfavorable indicator for a better outcome.

A case of concomitant intralobar bronchopulmonary sequestration and situs inversus totalis

BACKGROUND AND AIMS

Situs inversus is a rare congenital abnormality involving partial or complete transposition of the thoracic or abdominal viscera. In situs inversus totalis, both the thoracic and abdominal viscera are transposed. The incidence of this condition is 0.01% to 0.02%. Bronchopulmonary sequestration (BPS) is a rare congenital abnormality of the respiratory tract with an incidence of 0.15% to 1.80%. Intralobar sequestration is uncommonly associated with congenital anomalies.

METHODS

A routine chest X-ray of a 41-year-old asymptomatic man showed dextrocardia, a left-sided liver, right-sided stomach, and left paracardiac opacity.

RESULTS

Computed tomography (CT) and CT pulmonary angiography revealed dextrocardia with situs inversus totalis and left paracardiac intralobar BPS in the lingular segment of the left upper lobe.

CONCLUSION

We present a rare case of combined situs inversus totalis and intralobar BPS with an atypical location and feeder artery.

The prognostic role of prenatal MRI volumetric assessment in fetuses with isolated ventriculomegaly

In this prospective study, we aimed to establish the value of volumetric assessment by prenatal brain MRI in determining the prognosis of fetuses with isolated VM. A total of 23 fetuses with isolated VM were included in the study. Supratentorial cerebral parenchyma volume (PV) and ventricular volume (VV) were measured, and supratentorial ventricular/parenchymal volume (VV/PV) ratios were calculated. Pregnancy and postnatal neurodevelopmental outcomes up to two years of age were obtained and correlated with the volumetric measurements. VV was found to be strongly and positively correlated with ventricular dimension. There was a statistically significant difference between the VV/ PV ratios of the good and poor prognosis groups into which the cases had been categorized. The fetuses with a poor prognosis had a significantly higher VV/PV ratio. Volumetric parenchymal and ventricular measurements obtained by fetal brain MRI may contribute to future clinical studies concerning the evaluation of fetuses with VM and provide an important indicator in cases where management dilemmas arise.

Prenatal karyotype results of fetuses with nuchal edema, cystic hygroma, and non-immune hydrops

PURPOSE

Subcutaneous edema detected sonographically in the forms of nuchal edema, cystic hygroma (CH), or non-immune hydrops (NIH) may be a sign of chromosomal abnormalities. The aim of this study was to investigate the chromosome abnormality incidence in fetuses with nuchal edema, CH, or NIH.

MATERIALS AND METHODS

The authors performed cytogenetic analysis of 218 singleton fetuses with ultrasound diagnosis of subcutaneous edema in the forms of nuchal edema in the first and second trimesters.

RESULTS

Chromosomal abnormality rates were 30.4, 10.4, 36.8, 34.1, and 60% in the nuchal translucency (NT), nuchal fold thickness (NF), CH, NIH, and CH with NIH groups, respectively. In 71 cases with detected chromosomal abnormalities, 37%, 44%, 15%, and 4% of the pathologic karyotypes were identified as monosomy X, trisomy 21, trisomy 18, and trisomy 13, respectively.

CONCLUSIONS

This study confirms that subcutaneous edema detected sonographically, in the forms of nuchal edema, CH, or NIH, is a significant indicator of abnormal karyotype and deserves further investigation.

Fetal abdominal wall defects: six years experience at a tertiary center

The authors' aim was to detect the associated anomalies and their effect on the management of the fetuses with omphalocele and gastroschisis. Between the period of 2007-2013, the data of fetuses with abdominal wall defects were analyzed. Chromosomal abnormalities and associated morphologic anomalies diagnosed by ultrasonography and autopsy were evaluated. Of the. 61 fetuses, ten (20.4%) omphalocele cases and nine (75%) gastroschisis cases were isolated. Chromosomal abnormalities were found in seven fetuses with omphalocele cases. All fetuses with abnormal karyotypes had multiple additional anomalies. Termination rate was 65.3% for omphalocele group versus none in the gastroschisis group. To give better counseling about the prognosis and outcome of the fetuses with abdominal wall defects, detection of additional anomalies as well as type of the defect are essential tools even if the karyotype is normal.

Impact of pelvic biometric measurements, visceral and subcutaneous adipose tissue areas on trifecta outcome and surgical margin status after open radical retropubic prostatectomy

OBJECTIVE

There is no sufficient evidence to demonstrate whether more challenging pelvic anatomy affects the functional and oncological outcomes of radical retropubic prostatectomy (RP). The aim of this study was to investigate the impact of pelvic biometric measurements, visceral and subcutaneous adipose tissue areas on trifecta outcome (cancer control, continence and potency) and surgical margin status after open RP.

MATERIALS AND METHODS

A retrospective study was performed on 270 patients who were diagnosed with clinically localized prostate cancer between 2005 and 2011 and underwent computed tomography imaging before RP. Pelvic bony and soft-tissue measurements, and visceral and subcutaneous adipose tissue area calculations were performed. Patients were evaluated for trifecta outcome and surgical margin status on univariate and multivariate analyses. Subgroup analysis was performed for prostate volume, body mass index (BMI) and D'Amico risk classification.

RESULTS

Univariate analysis revealed that patients with trifecta were younger, and had lower prostate-specific antigen (PSA) levels, wider symphysis angle, and narrower prostate width and soft-tissue width (p < 0.05). Patients with positive surgical margins were older, and had higher PSA levels, lower prostate volume and narrower prostate width (p < 0.05). Visceral adipose tissue area was lower in patients with trifecta when BMI was below 25 kg/cm² (p < 0.05). In multivariate analysis, PSA level and symphysis angle were statistically significant for trifecta while PSA level was the only statistically significant variable for positive surgical margins (p < 0.05). Interfemoral index was significant in the low-risk group of D'Amico classification for trifecta outcome.

CONCLUSION

Pelvic biometric measurements and adipose tissue area calculations may help in preoperative planning and the management of RP.

Management of free-floating thrombus within the arcus aorta and supra-aortic arteries

INTRODUCTION

Clinical management strategy and the best treatment option of free floating thrombus (FFT), detected in patients with acute ischemic cerebrovascular diseases is still uncertain due to the rarity of these conditions. Recent studies emphasize that FFT within cerebral vessels may not be a surgical emergency since complete dissolution without any further neurologic progression occurred in 86% of patients treated medically. The aim of this study was to investigate the effect of medical treatment on the fate of thrombus and the clinical status in acute cerebrovascular event patients with detected FFT in the artery feeding the compromised brain territory at the time of diagnosis.

MATERIALS AND METHODS

We have retrospectively reviewed the initial and follow up computed tomography angiography, diffusion-weighted magnetic resonance imaging examinations of 37 acute ischemic stroke patients with detected FFT and treated medically. The patients were evaluated for any change of the FFT, residual stenosis after the FFT shrunk or disappeared and neurologic outcome.

RESULTS

FFT was located in cervical carotid artery, intracranial artery and arcus aorta in 23, 6 and 5 of the patients respectively. Dissection accompanied the FFT in 3 patients. Recanalization was maintained in 34 patients however, three of them suffered from new infarctions. A total occlusion was detected in three patients. Two of them had new infarcts while one was stable during clinical follow-up period.

CONCLUSION

The results of our retrospective study have shown that FFT in the proximal aorta and supra-aortic arteries tend to resolve with anticoagulant and/or antiplatelet treatment without causing recurrent adverse events in most of the cases. Furthermore, resolution of the clot seems to provide a safer ground for a recanalization procedure like stenting or endarterectomy if needed.

Evaluation of Neoadjuvant Chemotherapy Response with Dynamic Contrast Enhanced Breast Magnetic Resonance Imaging in Locally Advanced Invasive Breast Cancer

OBJECTIVE

The reliability of traditional methods such as physical examination, ultrasonography (US) and mammography is limited in determining the type of treatment response in patients with neoadjuvant chemotherapy (NAC) application for locally advanced breast cancer (LABC). Dynamic contrast-enhanced magnetic resonance imaging (MRI) is gaining popularity in the evaluation of NAC response. This study aimed to compare NAC response as determined by dynamic contrast-enhanced breast MRI in patients with LABC to histopathology that is the gold standard; and evaluate the compatibility of MRI, mammography and US with response types.

MATERIALS AND METHODS

The US, mammography and MRI findings of 38 patients who received NAC with a diagnosis of locally advanced breast cancer and surgical treatment were retrospectively analyzed and compared to histopathology results. Type of response to treatment was determined according to the "Criteria in Solid Tumors Response Evolution 1.1" by mammography, US and MRI criteria. The relationship between response types as defined by all three imaging modalities and histopathology were evaluated, and the correlation of response type as detected by MRI and pathological response and histopathological type of breast cancer was further determined. For statistical analysis, the chi-square, paired t test, correlation and kappa tests were used.

RESULTS

There is a statistical moderate positive correlation between response type according to pathology and MRI (kappa: 0.63). There was a weak correlation between response type according to mammography or US and according to pathology (kappa: 0.2). When the distribution of treatment response by MRI is stratified according to histopathological types, partial response was higher in all histopathological types similar to the type of pathologic response. When compared with pathology MRI detected treatment response accurately in 84.2% of the patients.

CONCLUSION

Dynamic contrast-enhanced breast MRI appears to be a more effective method than mammography or US in the evaluation of response to neoadjuvant chemotherapy. MRI evaluation of LABC is accepted as the appropriate radiological approach.