Refine
Has Fulltext
- yes (29)
Is part of the Bibliography
- yes (29)
Year of publication
- 2021 (29) (remove)
Document Type
- Journal article (21)
- Doctoral Thesis (7)
- Preprint (1)
Keywords
- deep learning (3)
- CT angiography (2)
- MRI (2)
- Periphere arterielle Verschlusskrankheit (2)
- endovascular (2)
- renal artery (2)
- stroke (2)
- 13C-Methacetin-Atemtest (1)
- 7 T (1)
- 7T (1)
Institute
- Institut für diagnostische und interventionelle Radiologie (Institut für Röntgendiagnostik) (29) (remove)
Purpose
While [\(^{18}\)F]-fluorodeoxyglucose ([\(^{18}\)F]FDG) is the standard for positron emission tomography/computed tomography (PET/CT) imaging of oral squamous cell carcinoma (OSCC), diagnostic specificity is hampered by uptake in inflammatory cells such as neutrophils or macrophages. Recently, molecular imaging probes targeting fibroblast activation protein α (FAP), which is overexpressed in a variety of cancer-associated fibroblasts, have become available and might constitute a feasible alternative to FDG PET/CT.
Methods
Ten consecutive, treatment-naïve patients (8 males, 2 females; mean age, 62 ± 9 years) with biopsy-proven OSCC underwent both whole-body [\(^{18}\)F]FDG and [\(^{68}\)Ga]FAPI-04 (FAP-directed) PET/CT for primary staging prior to tumor resection and cervical lymph node dissection. Detection of the primary tumor, as well as the presence and number of lymph node and distant metastases was analysed. Intensity of tracer accumulation was assessed by means of maximum (SUV\(_{max}\)) and peak (SUV\(_{peak}\) standardized uptake values. Histological work-up including immunohistochemical staining for FAP served as standard of reference.
Results
[\(^{18}\)F]FDG and FAP-directed PET/CT detected all primary tumors with a SUVmax of 25.5 ± 13.2 (FDG) and 20.5 ± 6.4 (FAP-directed) and a SUVpeak of 16.1 ± 10.3 ([\(^{18}\)F]FDG) and 13.8 ± 3.9 (FAP-directed), respectively. Regarding cervical lymph node metastases, FAP-directed PET/CT demonstrated comparable sensitivity (81.3% vs. 87.5%; P = 0.32) and specificity (93.3% vs. 81.3%; P = 0.16) to [\(^{18}\)F]FDG PET/CT. FAP expression on the cell surface of cancer-associated fibroblasts in both primary lesions as well as lymph nodes metastases was confirmed in all samples.
Conclusion
FAP-directed PET/CT in OSCC seems feasible. Future research to investigate its potential to improve patient staging is highly warranted.
Purpose
To evaluate whether a deep learning model (DLM) could increase the detection sensitivity of radiologists for intracranial aneurysms on CT angiography (CTA) in aneurysmal subarachnoid hemorrhage (aSAH).
Methods
Three different DLMs were trained on CTA datasets of 68 aSAH patients with 79 aneurysms with their outputs being combined applying ensemble learning (DLM-Ens). The DLM-Ens was evaluated on an independent test set of 104 aSAH patients with 126 aneuryms (mean volume 129.2 ± 185.4 mm3, 13.0% at the posterior circulation), which were determined by two radiologists and one neurosurgeon in consensus using CTA and digital subtraction angiography scans. CTA scans of the test set were then presented to three blinded radiologists (reader 1: 13, reader 2: 4, and reader 3: 3 years of experience in diagnostic neuroradiology), who assessed them individually for aneurysms. Detection sensitivities for aneurysms of the readers with and without the assistance of the DLM were compared.
Results
In the test set, the detection sensitivity of the DLM-Ens (85.7%) was comparable to the radiologists (reader 1: 91.2%, reader 2: 86.5%, and reader 3: 86.5%; Fleiss κ of 0.502). DLM-assistance significantly increased the detection sensitivity (reader 1: 97.6%, reader 2: 97.6%,and reader 3: 96.0%; overall P=.024; Fleiss κ of 0.878), especially for secondary aneurysms (88.2% of the additional aneurysms provided by the DLM).
Conclusion
Deep learning significantly improved the detection sensitivity of radiologists for aneurysms in aSAH, especially for secondary aneurysms. It therefore represents a valuable adjunct for physicians to establish an accurate diagnosis in order to optimize patient treatment.
Purpose
To fully automatically derive quantitative parameters from late gadolinium enhancement (LGE) cardiac MR (CMR) in patients with myocardial infarction and to investigate if phase sensitive or magnitude reconstructions or a combination of both results in best segmentation accuracy.
Methods
In this retrospective single center study, a convolutional neural network with a U-Net architecture with a self-configuring framework (“nnU-net”) was trained for segmentation of left ventricular myocardium and infarct zone in LGE-CMR. A database of 170 examinations from 78 patients with history of myocardial infarction was assembled. Separate fitting of the model was performed, using phase sensitive inversion recovery, the magnitude reconstruction or both contrasts as input channels.
Manual labelling served as ground truth. In a subset of 10 patients, the performance of the trained models was evaluated and quantitatively compared by determination of the Sørensen-Dice similarity coefficient (DSC) and volumes of the infarct zone compared with the manual ground truth using Pearson’s r correlation and Bland-Altman analysis.
Results
The model achieved high similarity coefficients for myocardium and scar tissue. No significant difference was observed between using PSIR, magnitude reconstruction or both contrasts as input (PSIR and MAG; mean DSC: 0.83 ± 0.03 for myocardium and 0.72 ± 0.08 for scars). A strong correlation for volumes of infarct zone was observed between manual and model-based approach (r = 0.96), with a significant underestimation of the volumes obtained from the neural network.
Conclusion
The self-configuring nnU-net achieves predictions with strong agreement compared to manual segmentation, proving the potential as a promising tool to provide fully automatic quantitative evaluation of LGE-CMR.
Background: Neurologic symptom severity and deterioration at 24 hours (h) predict long-term outcomes in patients with acute large vessel occlusion (LVO) stroke of the anterior circulation. We aimed to examine the association of baseline multiparametric CT imaging and clinical factors with the course of neurologic symptom severity in the first 24 h after endovascular treatment (EVT). Methods: Patients with LVO stroke of the anterior circulation were selected from a prospectively acquired consecutive cohort of patients who underwent multiparametric CT, including non-contrast CT, CT angiography and CT perfusion before EVT. The symptom severity was assessed on admission and after 24 h using the 42-point National Institutes of Health Stroke Scale (NIHSS). Clinical and imaging data were compared between patients with and without early neurological deterioration (END). END was defined as an increase in ≥4 points, and a significant clinical improvement as a decrease in ≥4 points, compared to NIHSS on admission. Multivariate regression analyses were used to determine independent associations of imaging and clinical parameters with NIHSS score increase or decrease in the first 24 h. Results: A total of 211 patients were included, of whom 38 (18.0%) had an END. END was significantly associated with occlusion of the internal carotid artery (odds ratio (OR), 4.25; 95% CI, 1.90–9.47) and the carotid T (OR, 6.34; 95% CI, 2.56–15.71), clot burden score (OR, 0.79; 95% CI, 0.68–0.92) and total ischemic volume (OR, 1.01; 95% CI, 1.00–1.01). In a comprehensive multivariate analysis model including periprocedural parameters and complications after EVT, carotid T occlusion remained independently associated with END, next to reperfusion status and intracranial hemorrhage. Favorable reperfusion status and small ischemic core volume were associated with clinical improvement after 24 h. Conclusions: The use of imaging parameters as a surrogate for early NIHSS progression in an acute LVO stroke after EVT reached limited performance with only carotid T occlusion as an independent predictor of END. Reperfusion status and early complications in terms of intracranial hemorrhage are critical factors that influence patient outcome in the acute stroke phase after EVT.
Objectives
Triangular fibrocartilage complex (TFCC) injuries frequently cause ulnar-sided wrist pain and can induce distal radioulnar joint instability. With its complex three-dimensional structure, diagnosis of TFCC lesions remains a challenging task even in MR arthrograms. The aim of this study was to assess the added diagnostic value of radial reformatting of isotropic 3D MRI datasets compared to standard planes after direct arthrography of the wrist.
Methods
Ninety-three patients underwent wrist MRI after fluoroscopy-guided multi-compartment arthrography. Two radiologists collectively analyzed two datasets of each MR arthrogram for TFCC injuries, with one set containing standard reconstructions of a 3D thin-slice sequence in axial, coronal and sagittal orientation, while the other set comprised an additional radial plane view with the rotating center positioned at the ulnar styloid. Surgical reports (whenever available) or radiological reports combined with clinical follow-up served as a standard of reference. In addition, diagnostic confidence and assessability of the central disc and ulnar-sided insertions were subjectively evaluated.
Results
Injuries of the articular disc, styloid and foveal ulnar attachment were present in 20 (23.7%), 10 (10.8%) and 9 (9.7%) patients. Additional radial planes increased diagnostic accuracy for lesions of the styloid (0.83 vs. 0.90; p = 0.016) and foveal (0.86 vs. 0.94; p = 0.039) insertion, whereas no improvement was identified for alterations of the central cartilage disc. Readers' confidence (p < 0.001) and assessability of the ulnar-sided insertions (p < 0.001) were superior with ancillary radial reformatting.
Conclusions
Access to the radial plane view of isotropic 3D sequences in MR arthrography improves diagnostic accuracy and confidence for ulnar-sided TFCC lesions.
Colorectal cancer (CRC) is the third most common malignancy worldwide. Most patients with metastatic CRC develop liver or lung metastases, while a minority suffer from brain metastases. There is little information available regarding the presentation, treatment, and overall survival of brain metastases (BM) from CRC. This systematic review and meta-analysis includes data collected from three major databases (PubMed, Cochrane, and Embase) based on the key words “brain”, “metastas*”, “tumor”, “colorectal”, “cancer”, and “malignancy”. In total, 1318 articles were identified in the search and 86 studies matched the inclusion criteria. The incidence of BM varied between 0.1% and 11.5%. Most patients developed metastases at other sites prior to developing BM. Lung metastases and KRAS mutations were described as risk factors for additional BM. Patients with BM suffered from various symptoms, but up to 96.8% of BM patients were asymptomatic at the time of BM diagnosis. Median survival time ranged from 2 to 9.6 months, and overall survival (OS) increased up to 41.1 months in patients on a multimodal therapy regimen. Several factors including age, blood levels of carcinoembryonic antigen (CEA), multiple metastases sites, number of brain lesions, and presence of the KRAS mutation were predictors of OS. For BM diagnosis, MRI was considered to be state of the art. Treatment consisted of a combination of surgery, radiation, or systemic treatment.
Purpose
Repeat surgery in patients with primary hyperparathyroidism (pHPT) is associated with an increased risk of complications and failure. This stresses the need for optimized strategies to accurately localize a parathyroid adenoma before repeat surgery is performed. However, evidence on the extent of required diagnostics for a structured approach is sparse.
Methods
A retrospective single-center evaluation of 28 patients with an indication for surgery due to pHPT and previous thyroid or parathyroid surgery was performed. Diagnostic workup, surgical approach, and outcome in terms of complications and successful removement of parathyroid adenoma with biochemical cure were evaluated.
Results
Neck ultrasound, sestamibi scintigraphy, C11-methionine PET-CT, and selective parathyroid hormone venous sampling, but not MRI imaging, effectively detected the presence of a parathyroid adenoma with high positive predictive values. Biochemical cure was revealed by normalization of calcium and parathormone levels 24-48h after surgery and was achieved in 26/28 patients (92.9%) with an overall low rate of complications. Concordant localization by at least two diagnostic modalities enabled focused surgery with success rates of 100%, whereas inconclusive localization significantly increased the rate of bilateral explorations and significantly reduced the rate of biochemical cure to 80%.
Conclusion
These findings suggest that two concordant diagnostic modalities are sufficient to accurately localize parathyroid adenoma before repeat surgery for pHPT. In cases of poor localization, extended diagnostic procedures are warranted to enhance surgical success rates. We suggest an algorithm for better orientation when repeat surgery is intended in patients with pHPT.
Cone-beam computed tomography is a powerful tool for 3D imaging of the appendicular skeleton, facilitating detailed visualization of bone microarchitecture. This study evaluated various combinations of acquisition and reconstruction parameters for the cone-beam CT mode of a twin robotic x-ray system in cadaveric wrist and elbow scans, aiming to define the best possible trade-off between image quality and radiation dose. Images were acquired with different combinations of tube voltage and tube current–time product, resulting in five scan protocols with varying volume CT dose indices: full-dose (FD; 17.4 mGy), low-dose (LD; 4.5 mGy), ultra-low-dose (ULD; 1.15 mGy), modulated low-dose (mLD; 0.6 mGy) and modulated ultra-low-dose (mULD; 0.29 mGy). Each set of projection data was reconstructed with three convolution kernels (very sharp [Ur77], sharp [Br69], intermediate [Br62]). Five radiologists subjectively assessed the image quality of cortical bone, cancellous bone and soft tissue using seven-point scales. Irrespective of the reconstruction kernel, overall image quality of every FD, LD and ULD scan was deemed suitable for diagnostic use in contrast to mLD (very sharp/sharp/intermediate: 60/55/70%) and mULD (0/3/5%). Superior depiction of cortical and cancellous bone was achieved in FD\(_{Ur77}\) and LD\(_{Ur77}\) examinations (p < 0.001) with LD\(_{Ur77}\) scans also providing favorable bone visualization compared to FD\(_{Br69}\) and FD\(_{Br62}\) (p < 0.001). Fleiss’ kappa was 0.618 (0.594–0.641; p < 0.001), indicating substantial interrater reliability. In this study, we demonstrate that considerable dose reduction can be realized while maintaining diagnostic image quality in upper extremity joint scans with the cone-beam CT mode of a twin robotic x-ray system. Application of sharper convolution kernels for image reconstruction facilitates superior display of bone microarchitecture.
Background and Purpose
To provide real-world data on outcome and procedural factors of late thrombectomy patients.
Methods
We retrospectively analyzed patients from the multicenter German Stroke Registry. The primary endpoint was clinical outcome on the modified Rankin scale (mRS) at 3 months. Trial-eligible patients and the subgroups were compared to the ineligible group. Secondary analyses included multivariate logistic regression to identify predictors of good outcome (mRS ≤ 2).
Results
Of 1917 patients who underwent thrombectomy, 208 (11%) were treated within a time window ≥ 6–24 h and met the baseline trial criteria. Of these, 27 patients (13%) were eligible for DAWN and 39 (19%) for DEFUSE3 and 156 patients were not eligible for DAWN or DEFUSE3 (75%), mainly because there was no perfusion imaging (62%; n = 129). Good outcome was not significantly higher in trial-ineligible (27%) than in trial-eligible (20%) patients (p = 0.343). Patients with large trial-ineligible CT perfusion imaging (CTP) lesions had significantly more hemorrhagic complications (33%) as well as unfavorable outcomes.
Conclusion
In clinical practice, the high number of patients with a good clinical outcome after endovascular therapy ≥ 6–24 h as in DAWN/DEFUSE3 could not be achieved. Similar outcomes are seen in patients selected for EVT ≥ 6 h based on factors other than CTP. Patients triaged without CTP showed trends for shorter arrival to reperfusion times and higher rates of independence.
Background
The effect of smoking on coronary vasomotion has been investigated in the past with various imaging techniques in both short- and long-term smokers. Additionally, coronary vasomotion has been shown to be normalized in long-term smokers by L-Arginine acting as a substrate for NO synthase, revealing the coronary endothelium as the major site of abnormal vasomotor response. Aim of the prospective cohort study was to investigate coronary vasomotion of young healthy short-term smokers via magnetic resonance cold pressor test with and without the administration of L-Arginine and compare obtained results with the ones from nonsmokers.
Methods
Myocardial blood flow (MBF) was quantified with first-pass perfusion MRI on a 1.5 T scanner in healthy short-term smokers (N = 10, age: 25.0 ± 2.8 years, 5.0 ± 2.9 pack years) and nonsmokers (N = 10, age: 34.3 ± 13.6) both at rest and during cold pressor test (CPT). Smokers underwent an additional examination after administration of L-Arginine within a median of 7 days of the naïve examination.
Results
MBF at rest turned out to be 0.77 ± 0.30 (smokers with no L-Arginine; mean ± standard deviation), 0.66 ± 0.21 (smokers L-Arginine) and 0.84 ± 0.08 (nonsmokers). Values under CPT were 1.21 ± 0.42 (smokers no L-Arginine), 1.09 ± 0.35 (smokers L-Arginine) and 1.63 ± 0.33 (nonsmokers). In all groups, MBF was significantly increased under CPT compared to the corresponding rest examination (p < 0.05 in all cases). Additionally, MBF under CPT was significantly different between the smokers and the nonsmokers (p = 0.002). MBF at rest was significantly different between the smokers when L-Arginine was given and the nonsmokers (p = 0.035).
Conclusion
Short-term smokers showed a reduced response to cold both with and without the administration of L-Arginine. However, absolute MBF values under CPT were lower compared to nonsmokers independently of L-Arginine administration.
Background
Functional lung MRI techniques are usually associated with time-consuming post-processing, where manual lung segmentation represents the most cumbersome part. The aim of this study was to investigate whether deep learning-based segmentation of lung images which were scanned by a fast UTE sequence exploiting the stack-of-spirals trajectory can provide sufficiently good accuracy for the calculation of functional parameters.
Methods
In this study, lung images were acquired in 20 patients suffering from cystic fibrosis (CF) and 33 healthy volunteers, by a fast UTE sequence with a stack-of-spirals trajectory and a minimum echo-time of 0.05 ms. A convolutional neural network was then trained for semantic lung segmentation using 17,713 2D coronal slices, each paired with a label obtained from manual segmentation. Subsequently, the network was applied to 4920 independent 2D test images and results were compared to a manual segmentation using the Sørensen–Dice similarity coefficient (DSC) and the Hausdorff distance (HD). Obtained lung volumes and fractional ventilation values calculated from both segmentations were compared using Pearson’s correlation coefficient and Bland Altman analysis.
To investigate generalizability to patients outside the CF collective, in particular to those exhibiting larger consolidations inside the lung, the network was additionally applied to UTE images from four patients with pneumonia and one with lung cancer.
Results
The overall DSC for lung tissue was 0.967 ± 0.076 (mean ± standard deviation) and HD was 4.1 ± 4.4 mm. Lung volumes derived from manual and deep learning based segmentations as well as values for fractional ventilation exhibited a high overall correlation (Pearson’s correlation coefficent = 0.99 and 1.00). For the additional cohort with unseen pathologies / consolidations, mean DSC was 0.930 ± 0.083, HD = 12.9 ± 16.2 mm and the mean difference in lung volume was 0.032 ± 0.048 L.
Conclusions
Deep learning-based image segmentation in stack-of-spirals based lung MRI allows for accurate estimation of lung volumes and fractional ventilation values and promises to replace the time-consuming step of manual image segmentation in the future.
Background
Lipoblastoma is a rare benign mesenchymal neoplasm of infancy that most commonly occurs on the extremities and trunk but can arise at variable sites of the body. Retroperitoneal lipoblastomas are particularly rare but can grow to enormous size, and preoperative diagnosis is difficult with diverse, mostly malignant differential diagnoses that would lead to aggressive therapy. Since lipoblastoma is a benign tumor that has an excellent prognosis after resection, correct diagnosis is crucial.
Case presentation
A case of a large retroperitoneal tumor of a 24-month old infant that was clinically suspicious of a malignant tumor is presented. Due to proximity to the right kidney, clinically most probably a nephroblastoma or clear cell sarcoma of the kidney was suspected. Radiological findings were ambiguous. Therefore, the mass was biopsied, and histology revealed an adipocytic lesion. Although mostly composed of mature adipocytes, in view of the age of the patient, the differential diagnosis of a (maturing) lipoblastoma was raised, which was supported by molecular analysis demonstrating a HAS2-PLAG1 fusion. The tumor was completely resected, and further histopathological workup led to the final diagnosis of a 13 cm large retroperitoneal maturing lipoblastoma. The child recovered promptly from surgery and showed no evidence of recurrence so far.
Conclusion
Although rare, lipoblastoma should be included in the differential diagnoses of retroperitoneal tumors in infants and children, and molecular diagnostic approaches could be a helpful diagnostic adjunct in challenging cases.
Background
Diagnosis of subscapularis (SSC) tendon lesions on magnetic resonance imaging (MRI) can be challenging. A small coracohumeral distance (CHD) has been associated with SSC tears. This study was designed to define a specific threshold value for CHD to predict SSC tears on axial MRI scans.
Methods
This retrospective study included 172 shoulders of 168 patients who underwent arthroscopic surgery for rotator cuff tear or glenohumeral instability. Diagnostic arthroscopy confirmed an SSC tear in 62 cases (36.0%, test group a), rotator cuff tear tears other than SSC in 71 cases (41.3%, control group b) and glenohumeral instability without any rotator cuff tear in 39 cases (22.7%, zero-sample group c). All patients had a preoperative MRI of the shoulder (1.5T or 3T). Minimum CHD was measured on axial fat-suppressed proton density-, T2-, or T1-weigthed sequences. Receiver operating characteristics analysis was used to determine the threshold value for CHD, and sensitivity and specificity were calculated.
Results
CHD measurement had a good interobserver reliability (Intraclass correlation coefficient 0.799). Mean CHD was highly significantly (P < .001) less for test group a (mean 7.3 mm, standard deviation ± 2.2) compared with control group b (mean 11.1 mm, standard deviation ± 2.3) or zero-sample group c (mean 13.6 mm, standard deviation ± 2.9). A threshold value of CHD <9.5 mm had a sensitivity of 83.6% and a specificity of 83.9% to predict SSC tears.
Conclusion
A CHD <9.5 mm on MRI is predictive of SSC lesions and a valuable tool to diagnose SSC tears.
Background: Endovascular therapy is the gold standard in patients with hemodynamic relevant renal artery stenosis (RAS) resistant to medical therapy. The severity grading of the stenosis as well as the result assessment after endovascular approach is predominantly based on visible estimations of the anatomic appearance. We aim to investigate the application of color-coded DSA parameters to gain hemodynamic information during endovascular renal artery interventions and for the assessment of the procedures technical success.
Methods: We retrospectively evaluated 32 patients who underwent endovascular renal artery revascularization and applied color-coded summation imaging on selected monochromatic DSA images. The differences in time to peak (dTTP) of contrast enhancement in predefined anatomical measuring points were analyzed. Furthermore, differences in systolic blood pressure values (SBP) and serum creatinine were obtained. The value of underlying diabetes mellitus as a predictor for clinical outcome was assessed. Correlation analysis between the patients gender as well as the presence of diabetes mellitus and dTTP was performed.
Results: Endovascular revascularization resulted in statistically significant improvement in 4/7 regions of interest. Highly significant improvement of perfusion in terms of shortened TTP values could be found at the segmental artery level and in the intrastenotical segment (p<0.001), significant improvement prestenotical and in the apical renal parenchyma (p<0.05). In the other anatomic regions, differences revealed not to be significant. Differences between SBP and serum creatinine levels before and after the procedure were significant (p=0.004 and 0.0004). Patients ' gender as well as the presence of diabetes mellitus did not reveal to be predictors for the clinical success of the procedure. Furthermore, diabetes and gender did not show relevant correlation with dTTP in the parenchymal measuring points.
Conclusions: The supplementary use of color-coding DSA and the data gained from parametric images may provide helpful information in the evaluation of the procedures ' technical success. The segmental artery might be a particularly suitable vascular territory for analyzing differences in blood flow characteristics. Further studies with larger cohorts are needed to further confirm the diagnostic value of this technique.
Purpose
Artificial neural networks show promising performance in automatic segmentation of cardiac MRI. However, training requires large amounts of annotated data and generalization to different vendors, field strengths, sequence parameters, and pathologies is limited. Transfer learning addresses this challenge, but specific recommendations regarding type and amount of data required is lacking. In this study, we assess data requirements for transfer learning to experimental cardiac MRI at 7T where the segmentation task can be challenging. In addition, we provide guidelines, tools, and annotated data to enable transfer learning approaches by other researchers and clinicians.
Methods
A publicly available segmentation model was used to annotate a publicly available data set. This labeled data set was subsequently used to train a neural network for segmentation of left ventricle and myocardium in cardiac cine MRI. The network is used as starting point for transfer learning to 7T cine data of healthy volunteers (n = 22; 7873 images) by updating the pre-trained weights. Structured and random data subsets of different sizes were used to systematically assess data requirements for successful transfer learning.
Results
Inconsistencies in the publically available data set were corrected, labels created, and a neural network trained. On 7T cardiac cine images the model pre-trained on public imaging data, acquired at 1.5T and 3T, achieved DICE\(_{LV}\) = 0.835 and DICE\(_{MY}\) = 0.670. Transfer learning using 7T cine data and ImageNet weight initialization improved model performance to DICE\(_{LV}\) = 0.900 and DICE\(_{MY}\) = 0.791. Using only end-systolic and end-diastolic images reduced training data by 90%, with no negative impact on segmentation performance (DICE\(_{LV}\) = 0.908, DICE\(_{MY}\) = 0.805).
Conclusions
This work demonstrates and quantifies the benefits of transfer learning for cardiac cine image segmentation. We provide practical guidelines for researchers planning transfer learning projects in cardiac MRI and make data, models, and code publicly available.
Parathyroid carcinoma (PC) is an orphan malignancy accounting for only ~1% of all cases with primary hyperparathyroidism. The localization of recurrent PC is of critical importance and can be exceedingly difficult to diagnose and sometimes futile when common sites of recurrence in the neck and chest cannot be confirmed. Here, we present the diagnostic workup, molecular analysis and multimodal therapy of a 46-year old woman with the extraordinary manifestation of abdominal lymph node metastases 12 years after primary diagnosis of PC. The patient was referred to our endocrine tumor center in 2016 with the aim to localize the tumor causative of symptomatic biochemical recurrence. In view of the extensive previous workup we decided to perform [18F]FDG-PET-CT. A pathological lymph node in the liver hilus showed slightly increased FDG-uptake and hence was suspected as site of recurrence. Selective venous sampling confirmed increased parathyroid hormone concentration in liver veins. Abdominal lymph node metastasis was resected and histopathological examination confirmed PC. Within four months, the patient experienced biochemical recurrence and based on high tumor mutational burden detected in the surgical specimen by whole exome sequencing the patient received immunotherapy with pembrolizumab that led to a biochemical response. Subsequent to disease progression repeated abdominal lymph node resection was performed in 10/2018, 01/2019 and in 01/2020. Up to now (12/2020) the patient is biochemically free of disease. In conclusion, a multimodal diagnostic approach and therapy in an interdisciplinary setting is needed for patients with rare endocrine tumors. Molecular analyses may inform additional treatment options including checkpoint inhibitors such as pembrolizumab.
Whole-Body [\(^{18}\)F]FDG PET/CT Can Alter Diagnosis in Patients with Suspected Rheumatic Disease
(2021)
The 2-deoxy-d-[\(^{18}\)F]fluoro-D-glucose (FDG) positron emission tomography/computed tomography (PET/CT) is widely utilized to assess the vascular and articular inflammatory burden of patients with a suspected diagnosis of rheumatic disease. We aimed to elucidate the impact of [\(^{18}\)F]FDG PET/CT on change in initially suspected diagnosis in patients at the time of the scan. Thirty-four patients, who had undergone [\(^{18}\)F]FDG PET/CT, were enrolled and the initially suspected diagnosis prior to [18F]FDG PET/CT was compared to the final diagnosis. In addition, a semi-quantitative analysis including vessel wall-to-liver (VLR) and joint-to-liver (JLR) ratios was also conducted. Prior to [\(^{18}\)F]FDG PET/CT, 22/34 (64.7%) of patients did not have an established diagnosis, whereas in 7/34 (20.6%), polymyalgia rheumatica (PMR) was suspected, and in 5/34 (14.7%), giant cell arteritis (GCA) was suspected by the referring rheumatologists. After [\(^{18}\)F]FDG PET/CT, the diagnosis was GCA in 19/34 (55.9%), combined GCA and PMR (GCA + PMR) in 9/34 (26.5%) and PMR in the remaining 6/34 (17.6%). As such, [\(^{18}\)F]FDG PET/CT altered suspected diagnosis in 28/34 (82.4%), including in all unclear cases. VLR of patients whose final diagnosis was GCA tended to be significantly higher when compared to VLR in PMR (GCA, 1.01 ± 0.08 (95%CI, 0.95–1.1) vs. PMR, 0.92 ± 0.1 (95%CI, 0.85–0.99), p = 0.07), but not when compared to PMR + GCA (1.04 ± 0.14 (95%CI, 0.95–1.13), p = 1). JLR of individuals finally diagnosed with PMR (0.94 ± 0.16, (95%CI, 0.83–1.06)), however, was significantly increased relative to JLR in GCA (0.58 ± 0.04 (95%CI, 0.55–0.61)) and GCA + PMR (0.64 ± 0.09 (95%CI, 0.57–0.71); p < 0.0001, respectively). In individuals with a suspected diagnosis of rheumatic disease, an inflammatory-directed [\(^{18}\)F]FDG PET/CT can alter diagnosis in the majority of the cases, particularly in subjects who were referred because of diagnostic uncertainty. Semi-quantitative assessment may be helpful in establishing a final diagnosis of PMR, supporting the notion that a quantitative whole-body read-out may be useful in unclear cases.
The occurrence of different subtypes of endogenous Cushing’s syndrome (CS) in single individuals is extremely rare. We here present the case of a female patient who was successfully cured from adrenal CS 4 years before being diagnosed with Cushing’s disease (CD). The patient was diagnosed at the age of 50 with ACTH-independent CS and a left-sided adrenal adenoma, in January 2015. After adrenalectomy and histopathological confirmation of a cortisol-producing adrenocortical adenoma, biochemical hypercortisolism and clinical symptoms significantly improved. However, starting from 2018, the patient again developed signs and symptoms of recurrent CS. Subsequent biochemical and radiological workup suggested the presence of ACTH-dependent CS along with a pituitary microadenoma. The patient underwent successful transsphenoidal adenomectomy, and both postoperative adrenal insufficiency and histopathological workup confirmed the diagnosis of CD. Exome sequencing excluded a causative germline mutation but showed somatic mutations of the β-catenin protein gene (CTNNB1) in the adrenal adenoma, and of both the ubiquitin specific peptidase 8 (USP8) and the glucocorticoid receptor (NR3C1) genes in the pituitary adenoma. In conclusion, our case illustrates that both ACTH-independent and ACTH-dependent CS may develop in a single individual even without evidence for a common genetic background.
Morbidität und Mortalität der Lebererkrankung im Rahmen der Cystischen Fiborse (Cystic fibrosis liver disease, CFLD) sind vornehmlich von Ausmaß und Progredienz der Leberfibrose abhängig. Aufgrund der fehlenden Genauigkeit der bisherigen diagnostischen Verfahren werden viele der an CF erkrankten Menschen erst in fortgeschrittenen Stadien diagnostiziert. Schwere Komplikationen einer Leberzirrhose treten häufig bereits im Kindesalter auf. Die Quantitative Sonoelastographie, hier die Acoustic Radiation Force Impulse (ARFI)-Elastographie, ist ein vielversprechendes, nicht-invasives und strahlenfreies Verfahren zur Messung der Gewebesteifigkeit.
Anhand dieser retrospektiven, monozentrischen Studie soll die ARFI-Elastographie im Hinblick auf den klinischen Einsatz bei der CFLD-Diagnostik untersucht werden. Es wurde eruiert, ob sich mittels ARFI-Elastographie Rückschlüsse auf eine CFLD und deren Schweregrade ziehen lassen.
Hierfür wurden die ARFI-Messungen verschiedener Lebersegmente von 62 an CF erkrankten und 19 lebergesunden Kindern und Jugendlichen verglichen. Zudem erfolgte die Korrelation der Ergebnisse mit zwei etablierten klinischen Leberfibrose-Scores (APRI, Williams-Score).
Im Patientenkollektiv konnten tendenziell erhöhte Scherwellengeschwindigkeiten, entsprechend einer fibrotischen Aktivität, gemessen werden. Die transkostale Messposition in Segment VII/VIII (TC VII/VIII) erwies sich als zuverlässigste Position zur Differenzierung zwischen einer CF-Hepatopathie und einem gesunden Leberparenchym. Hingegen war das Errechnen von Cut-off Werten zur Graduierung von Fibrosestadien nicht möglich. Auch war keine Korrelation zu Leberfiborsescores feststellbar.
Insgesamt zeigt sich, dass die Diagnosestellung einer CFLD aktuell nur in Kombination mit bisherigen Messmethoden wie der klinischen Untersuchung, der Laboranalytik und der Sonographie möglich ist. Die Interpretation der ARFI-cut-off Werte bleibt aufgrund mangelnder Sensitivität und Spezifität und vor dem Hintergrund der CF-typischen heterogenen Leberpathologie erschwert. Die ARFI-Elastographie kann als zusätzlicher Baustein in der Diagnostik der CFLD, bei unklaren Befundkonstellationen oder zum Therapie-Monitoring herangezogen werden. Um einen klaren klinischen Einsatz in der Routinediagnostik zu definieren bedarf es weiterer großer, multizentrischer und prospektiver Studien.