Polish Journal of Radiology最新文献

Purpose: Hirayama disease (HD) is a rare disease that was commonly mis-diagnosed in the past. The importance of neutral and flexion magnetic resonance imaging (MRI) in its accurate diagnosis has been emphasized along with utility of the inter-segmental angle of flexion. Aim of the study was to observe MRI findings of HD in neutral and flexion position and measure the inter-segmental angle of flexion.

Material and methods: Cervical MR images of 17 patients of suspected HD were evaluated retrospectively for loss of attachment (LOA) of posterior dura, lower cervical cord atrophy, T2 hyperintensity, loss of cervical lordosis, enhancement of posterior epidural venous plexus, and inter-segmental angle of flexion on neutral and flexion MRIs.

Results: Flexion MRI showed LOA of posterior dura (most commonly and maximum at C6 vertebral level) and intense enhancement in posterior epidural space in almost all patients. The mean inter-segmental angle of flexion at C5-C6 was 9.2°, and at C6-C7 it was 6°. Neutral MRI revealed LOA in 64.7%, lower cervical cord atrophy in all patients, T2 hyperintensity in the lower cervical cord in 35.2% of patients, and loss of cervical lordosis in 58.8% of patients.

Conclusions: Flexion MRI is the gold standard for diagnosis of HD; however, certain imaging attributes, i.e. loss of attachment of posterior dura, asymmetrical lower cervical cord atrophy, T2 hyperintensity, and loss of cervical lordosis, can be seen on neutral MRI as well, which subsequently prompts the radiologist to include flexion MRI for confirmation. The inter-segmental angle of flexion is increased in patients with HD, which plays a role in planning timely surgical intervention.

Purpose: Our study aims to assess the role and diagnostic performance of 3 different contrast-enhanced, abbreviated magnetic resonance imaging (MRI) protocols as a screening tool of hepatocellular carcinoma (HCC).

Material and methods: Our retrospective study included 80 patients who were screened for HCC: 47 patients revealed 138 focal hepatic lesions. MRI examinations were performed including full CE-MRI protocols. The MRI was done on a 1.5 T machine. Then 3 different abbreviated contrast-enhanced MRI protocols were analysed separately. The standard dynamic contrast MRI and abbreviated protocols were evaluated following the LI-RADS 2018 lexicon diagnostic features.

Results: A considerable overall kappa (k) agreement between the abbreviated 1, 2, and 3 protocols on LI-RADS classification was noted with k = 0.865. There was almost perfect agreement between all abbreviated protocols and full standard protocol on LI-RADS classification, with k = 0.890. As regards the k agreement on LI-RADS classification, there was a considerable highest agreement between the abbreviated 1 protocol and the full standard protocol, with k = 0.980. The abbreviated 1 and 2 protocols showed high diagnostic performance on LI-RADS classification of lesions, with 100% sensitivity, specificity, PPV, NPV, and accuracy, while the abbreviated 3 protocol showed a lesser but comparable sensitivity 96.9%, NPV 99.4, and accuracy 99.4%.

Conclusions: Abbreviated contrast-enhanced MRI protocols can be used as a screening tool for the detection of HCC, with high sensitivity, specificity, PPV, NPV, and accuracy close to the full protocol. There was a considerable highest agreement between the abbreviated 1 protocol and the full standard protocol. Subsequently, this protocol can be used as a standard protocol for screening high-risk patients.

Purpose: We studied the prevalence of celiac trunk and its anatomical variations on diagnostic computed tomography angiography (CTA) studies and have proposed a new classification to define the celiac artery (CA) variations based on embryology.

Material and methods: We retrospectively assessed the celiac trunk variations in 1113 patients who came to our department for diagnostic CTA for liver and renal donor workup. The patient data were acquired from the Picture Archiving and Communication System of our institutions. We analysed the celiac trunk's origin and branching pattern, including the superior mesenteric artery (SMA) and inferior phrenic artery (IPA).

Results: We evaluated the CTA studies of 1050 patients. A normal trifurcation pattern, the most common type, was observed in 39% of cases. Variation with CA + left IPA was the most common subtype. Other variations noted in the study and their incidences are listed in the table below. We attempted to propose a new classification based on embryo-logy, which comprises 6 main types and their subtypes. We also analysed previous studies from the literature, including cadaveric, post-mortem, CTA, and digital subtraction angiography studies and compared them with the present study.

Conclusions: Because variations of CA classifications reported to date do not encompass all CA branching pattern variants, we have proposed a new classification that incorporates most of the variants. We reiterate the clinical importance of anatomical variants of CA, IPA, and SMA in surgical and interventional radiology procedures.

Purpose: Using 3-dimensional (3D) printers, the creation of patient-specific models is possible before and after a therapeutic intervention. There are many articles about replicas for training and simulation of aneurysm clipping. However, no paper has focused on 3D replicas obtained from 3-tesla 3D time of flight (3D-TOF) MR angiography for intrasaccular flow diverter (WEB device) embolization of the cerebral aneurysms. In this paper, we aimed to investigate the feasibility of 3D printing models obtained from 3-tesla 3D-TOF data in the management and training of WEB-assisted embolization procedures.

Case presentation: We presented a longitudinal case report with several 3D-TOF MRA prints over time. Three-tesla 3D-TOF data were converted into STL and G-code files using an open-source (3D-Slicer) program. We built patient-specific realistic 3D models of a patient with a middle cerebral artery trifurcation aneurysm, which were able to demonstrate the entire WEB device treatment procedure in the pre-intervention and post-intervention periods. The aneurysmatic segment was well displayed on the STL files and the 3D replicas. They allowed visualization of the aneurysmatic segment and changes within a 6-year follow-up period. We successfully showed the possibility of fast, cheap, and easy production of replicas for demonstration of the aneurysm, the parent vessels, and post-intervention changes in a simple way using an affordable 3D printer.

Conclusions: 3D printing is useful for training the endovascular team and the patients, understanding the aneurysm/parent vessels, and choosing the optimal embolization technique/device. 3D printing will potentially lead to greater interventionalist confidence, decreased radiation dose, and improvements in patient safety.

Pulmonary venous anomalies occur as a result of failure in normal embryological development. These anomalies may present as a spectrum ranging from normal variation to partial anomalous pulmonary venous connection (PAPVC) and total anomalous pulmonary venous connection (TAPVC). Though not rare, PAPVC is an uncommon anomaly in which some of the pulmonary veins abnormally connect and drain into the vascular compartments other than the left atrium (LA); however, the others drain normally into the LA. The clinical presentation and severity of affected patients depend on the morphological heterogeneity of the disease. PAPVC associated with other complex conge-nital cardiac diseases present early and are more severe than isolated PAPVC-associated atrial septal defect only. This radiological review gives a detailed description of PAPVC in terms of morphological variability and associated anomalies along with a discussion of the role of multidetector dual-source computed tomography scan in the diagnostic assessment.

Purpose: The aim of this study was to evaluate the frequency of apical root resorption in the anterior teeth of the maxilla visible on panoramic images during orthodontic treatment with a fixed appliance.

Material and methods: A total of 194 panoramic radiographs of patients with a fixed appliance in the upper arch were analysed to evaluate the severity of root resorption in maxillary incisors and canines according to Levander and Malmgren classification. The research group included 135 females and 59 males, aged 15-28 years, with a mean 20.6 years.

Results: Of examined patients 75.26% had signs of apical root resorption. The tooth most frequently affected by resorptive changes was the right central upper incisor. The gender and age of the patients were not found to be significant factors. The highest number of teeth had second (II) stage root resorption (53.09%).

Conclusions: Panoramic radiographs can be useful in diagnosing external apical root resorption due to orthodontic treatment.

Purpose: Although assessment of the orbital structures using magnetic resonance imaging (MRI) is well described in the literature, there is no consensus as to which measurement method is the most useful in exophthalmos assessment. The aim of the study was to correlate 2 MRI methods of exophthalmos measurement with exophthalmometry results and to determine a proper technique of exophthalmos measurement.

Material and methods: Fifty-four patients (108 orbits) with exophthalmos in the course of Graves' orbitopathy were enrolled in the study. Two measurements on axial T2W orbital MRI images were performed by 2 independent radiologists: the distance from the interzygomatic line to the anterior surface of the globe (AD) and the distance from the interzygomatic line to the posterior sclera (PD). Within 4 weeks, an exophthalmometry was performed by an ophthalmologist using a Hertel exophthalmometer. The inter-observer variation was assessed using the Pearson correlation coefficient. Values were presented as mean and standard deviation, and the differences in values were explored with Student's t-test.

Results: The mean AD measured on MRI by the first observer was 20.6 ± 3 mm, and 20.6 ± 2.9 mm by the second observer. PD values were 2.9 ± 2.8 mm and 3.4 ± 2.8 mm, respectively. The mean exophthalmometry result was 21 ± 3.3 mm. The correlation was very high between observers for AD measurements (r = 0.98, p = 0.01) and high for PD measurements (r = 0.95, p = 0.01). AD measurements on MRI and exophthalmometry results were strongly correlated (r = 0.9, p = 0.01).

Conclusions: The AD measurement has better reproducibility and is directly correlated with Hertel exophthalmometry. This method could be sufficient in routine practice.

Purpose: To evaluate clot size and stenotic degree on conventional computed tomographic pulmonary angiography (CTPA) with perfusion defect.

Material and methods: Fifty-two pulmonary embolism (PE) patients with 144 PE locations underwent dual-energy CTPA with an iodine distribution map. Each PE location was rated as to whether there was a perfusion defect. Clot size, stenotic degree, and other associated PE findings were evaluated. These findings were then correlated with whether the perfusion defect was present.

Results: There were no associations between demographics, clinical characteristics, anatomical data, and perfusion defect. The median iodine concentration ratio was 0.11. Imaging interpretation by 2 thoracic radiologists had excellent agreement. The clot size and stenotic degree in PE were significant predictors of perfusion defect on conventional CTPA. Lesions with higher degrees of stenosis had higher percentages of perfusion defect. The generalized estimating equation (GEE) logistic regression confirmed that clot size and stenotic degree could predict PE perfusion defects on conventional CTPA.

Conclusions: The 2 significant predictors of perfusion defect were occluded vessels in both small and large branches together, or complete occlusion of the pulmonary artery.

Purpose: To verify whether deep learning can be used to differentiate between carcinosarcomas (CSs) and endometrial carcinomas (ECs) using several magnetic resonance imaging (MRI) sequences.

Material and methods: This retrospective study included 52 patients with CS and 279 patients with EC. A deep-learning model that uses convolutional neural networks (CNN) was trained with 572 T2-weighted images (T2WI) from 42 patients, 488 apparent diffusion coefficient of water maps from 33 patients, and 539 fat-saturated contrast-enhanced T1-weighted images from 40 patients with CS, as well as 1612 images from 223 patients with EC for each sequence. These were tested with 9-10 images of 9-10 patients with CS and 56 images of 56 patients with EC for each sequence, respectively. Three experienced radiologists independently interpreted these test images. The sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) for each sequence were compared between the CNN models and the radiologists.

Results: The CNN model of each sequence had sensitivity 0.89-0.93, specificity 0.44-0.70, accuracy 0.83-0.89, and AUC 0.80-0.94. It also showed an equivalent or better diagnostic performance than the 3 readers (sensitivity 0.43-0.91, specificity 0.30-0.78, accuracy 0.45-0.88, and AUC 0.49-0.92). The CNN model displayed the highest diagnostic performance on T2WI (sensitivity 0.93, specificity 0.70, accuracy 0.89, and AUC 0.94).

Conclusions: Deep learning provided diagnostic performance comparable to or better than experienced radiologists when distinguishing between CS and EC on MRI.

Purpose: Magnetic resonance imaging (MRI) has a special place in the evaluation of orbital and periorbital lesions. Segmentation is one of the deep learning methods. In this study, we aimed to perform segmentation in orbital and periorbital lesions.

Material and methods: Contrast-enhanced orbital MRIs performed between 2010 and 2019 were retrospectively screened, and 302 cross-sections of contrast-enhanced, fat-suppressed, T1-weighted, axial MRI images of 95 patients obtained using 3 T and 1.5 T devices were included in the study. The dataset was divided into 3: training, test, and validation. The number of training and validation data was increased 4 times by applying data augmentation (horizontal, vertical, and both). Pytorch UNet was used for training, with 100 epochs. The intersection over union (IOU) statistic (the Jaccard index) was selected as 50%, and the results were calculated.

Results: The 77^th epoch model provided the best results: true positives, 23; false positives, 4; and false negatives, 8. The pre-cision, sensitivity, and F1 score were determined as 0.85, 0.74, and 0.79, respectively.

Conclusions: Our study proved to be successful in segmentation by deep learning method. It is one of the pioneering studies on this subject and will shed light on further segmentation studies to be performed in orbital MR images.