Clinical and MRI Evaluation of Silicone Gel Implants with RFID-M Traceability System: A Prospective Controlled Cohort Study Related to Safety and Image Quality in MRI Follow-Up

BACKGROUND

SmoothSilk implants (SSI) are the first generation of implants to incorporate a radio-frequency identification device (RFID-M), a non-invasive traceability system. Although the RFID-M is considered compatible with magnetic resonance imaging (MRI), the size of the artifact and its influence on breast tissue vary. This prospective study assessed safety and MRI issues in a cohort of breast reconstruction patients.

METHODS

Forty-four SSI were used for breast reconstruction in patients undergoing treatment for breast cancer. All patients were evaluated for magnetic field interactions, MRI-related heating and artifacts in a 1.5-T MRI unit using standard T1/T2-weighted sequences utilized in clinical assessment of breast tissue/implants.

RESULTS

Mean patient age was 41.5 years (27-53ys) and body mass index was 28+-6.44 kg/m². In 18/22 patients (81.8%), mastectomies were unilateral. No patients reported local heat/discomfort. All implants showed RFID-M-related artifacts with an estimated mean volume in T1 of 42.9cm³ (26.2-63.6cm³; SD±8.6 and 95% CI, 40.37-45.45) and in T2 of 60.5cm³ (35.4-97.2cm³; SD±14.7 and 95% CI, 56.29-65.01). Artifact volume was smaller in T1 than in T2, to a statistically significant degree (p <0.001). There were no statistically significant differences in artifact volume according to surgical indication, breast side or implant volume. There were 4/44 (9%) cases of minor rotation (<45°). In all cases, adequate analysis of the breast tissue was performed.

CONCLUSIONS

The results demonstrate that SSI with RFID-M technology presented an artifact volume of 42.9cm³ and 60.5cm³ in T1 and T2 images, respectively. Our findings provide detailed information on the quality and location of MRI artifacts in a reconstructed cohort which can help guide clinical decision-making for patients. To our knowledge, this is the first time RFID-M breast implants have been prospectively evaluated for clinical and MRI issues in a cohort of reconstructive patients.

LEVEL OF EVIDENCE IV

This journal requires that authors assign a level of evidence to each article. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .

Prediction of recurrence after surgery based on preoperative MRI features in patients with pancreatic neuroendocrine tumors

OBJECTIVES

To investigate useful MRI features in pancreatic neuroendocrine tumor (PNET) patients for predicting recurrence and its timing after surgery.

METHODS

A total of 99 patients with PNET who underwent MRI and surgery from 2000 to 2018 were enrolled. Two radiologists independently assessed MRI findings, including size, location, margin, T1 and T2 signal intensity, enhancement patterns, common bile duct (CBD) or main pancreatic duct (MPD) dilatation, vascular invasion, lymph node enlargement, DWI, and ADC value. Imaging findings associated with recurrence and disease-free survival (DFS) were assessed using logistic regression analysis and Cox proportional hazard regression analysis.

RESULTS

The median follow-up period was 40.4 months, and recurrence after surgery occurred in 12.1% (12/99). Among them, 6 patients experienced recurrence within 1 year, and 9 patients experienced recurrence within 2 years after surgery. In multivariate analysis, major venous invasion (OR 10.76 [1.14-101.85], p = 0.04) was associated with recurrence within 1 year, and portal phase iso- to hypoenhancement (OR 51.89 [1.73-1557.89], p = 0.02), CBD or MPD dilatation (OR 10.49 [1.35-81.64], p = 0.03) and larger size (OR 1.05 [1.00-1.10], p = 0.046) were associated with recurrence within 2 years. The mean DFS was 116.4 ± 18.5 months, and the 5-year DFS rate was 85.7%. In multivariate analysis, portal phase iso- to hypoenhancement (HR 21.36 [2.01-197.77], p = 0.01), ductal dilatation (HR 5.22 [1.46-18.68], p = 0.01), major arterial invasion (HR 42.90 [3.66-502.48], p = 0.003), and larger size (HR 1.04 [1.01-1.06], p = 0.01) showed a significant effect on poor DFS.

CONCLUSION

MRI features, including size, enhancement pattern, vascular invasion, and ductal dilatation, are useful in predicting recurrence and poor DFS after surgery in PNET. Key Points • MRI features are useful for predicting prognosis in patients with PNET after surgery. • PV or SMV invasion (OR 10.49 [1.35-81.64], p = 0.04) was significantly associated with 1-year recurrence. • Portal phase iso- to hypoenhancement (HR 21.36), CBD or MPD dilatation (HR 5.22), arterial invasion (HR 42.90), and larger size (HR 1.04) had significant effects on poor DFS (p < 0.05).

Magnetic resonance imaging with intravenous gadoteridol injection based on 3D-real IR sequence of the inner ear in Meniere's disease patient: feasibility in 3.5-h time interval

BACKGROUND

Intravenous gadoteridol injection can be applied to visualize endolymphatic hydrops (EH).

AIMS/OBJECTIVES

To explore whether 3.5-h time interval was feasible for clinical practice.

MATERIALS AND METHODS

We collected 70 unilateral Meniere's disease (MD) patients who were divided into two groups randomly (group A: 3.5-h time interval; group B: 4-h time interval). Among the two groups, the signal intensity (SI) in perilymphatic area of the basal turn of cochlea, the results of visual evaluations in the vestibule, cochlea and semicircular canal and the detection results of EH were compared.

RESULTS

Regarding the SI, no difference was found between A-affected ears and B-affected ears (p=.499), and no difference was found between A-unaffected ears and B-unaffected ears (p=.111). However, a difference was found between A-affected ears and A-unaffected ears (p=.005), and a difference was found between B-affected ears and B-unaffected ears (p=.012). Besides, no difference was found between the visual evaluations in the vestibule, cochlea, and semicircular canal of the two groups. Regarding the detection results of EH, no difference was found between the two groups (all p>.05).

CONCLUSIONS AND SIGNIFICANCE

In the clinical application of gadoteridol for the inner ear, 3.5-h delayed MR imaging is feasible.

Using artificial intelligence to diagnose fresh osteoporotic vertebral fractures on magnetic resonance images

BACKGROUND CONTEXT

Accurate diagnosis of osteoporotic vertebral fracture (OVF) is important for improving treatment outcomes; however, the gold standard has not been established yet. A deep-learning approach based on convolutional neural network (CNN) has attracted attention in the medical imaging field.

PURPOSE

To construct a CNN to detect fresh OVF on magnetic resonance (MR) images.

STUDY DESIGN/SETTING

Retrospective analysis of MR images PATIENT SAMPLE: This retrospective study included 814 patients with fresh OVF. For CNN training and validation, 1624 slices of T1-weighted MR image were obtained and used.

OUTCOME MEASURE

We plotted the receiver operating characteristic (ROC) curve and calculated the area under the curve (AUC) in order to evaluate the performance of the CNN. Consequently, the sensitivity, specificity, and accuracy of the diagnosis by CNN and that of the two spine surgeons were compared.

METHODS

We constructed an optimal model using ensemble method by combining nine types of CNNs to detect fresh OVFs. Furthermore, two spine surgeons independently evaluated 100 vertebrae, which were randomly extracted from test data.

RESULTS

The ensemble method using VGG16, VGG19, DenseNet201, and ResNet50 was the combination with the highest AUC of ROC curves. The AUC was 0.949. The evaluation metrics of the diagnosis (CNN/surgeon 1/surgeon 2) for 100 vertebrae were as follows: sensitivity: 88.1%/88.1%/100%; specificity: 87.9%/86.2%/65.5%; accuracy: 88.0%/87.0%/80.0%.

CONCLUSIONS

In detecting fresh OVF using MR images, the performance of the CNN was comparable to that of two spine surgeons.

Clinical features of Trousseau's syndrome with multiple acute ischemic strokes

BACKGROUND

Trousseau's syndrome or migrating thrombophlebitis can cause venous or arterial thrombosis; however, multiple acute ischemic strokes (MAIS) caused by Trousseau's syndrome are rare. The aim of this study was to analyse the clinical features of Trousseau's syndrome with MAIS and to improve the awareness and the knowledge of this disease.

METHODS

Clinical data from fifteen patients who were diagnosed as Trousseau's syndrome with MAIS in Rizhao People's Hospital from January 2017 to April 2020 were collected and analysed. The clinical data included the following: patients' basic information (including gender, age, underlying diseases, and tumour stage), laboratory results, imaging features, treatment regimens, and short-term prognoses were collected.

RESULTS

The mean age was 65.5 years, with thirteen males and two females. Most patients (11/15) had a history of smoking and (or) drinking. The average score of NIHSS was 2.13. 6 of the 15 patients first presented with ischemic stroke and then found the primary tumour. Most common types of primary tumour was lung cancer (11/15), and other types of primary tumour were gastric adenocarcinoma, renal cell carcinoma, oesophageal adenosquamous carcinoma, and cholangiocarcinoma (one in each). All the 15 patients showed different levels of increase of D-dimer. The increase in CRP appears in 10 of the 15 patients. Various tumour markers were increased in the 15 patients, especially for CYFRA-211, all the patients of which were higher than normal. All of the 15 patients had multiple vascular territory lesions in DWI, and most lesions were near the cortex areas. Only 4 of the 15 patients (26.7%) occurred with peripheral venous thrombosis. Thirteen patients were given low molecular heparin for anticoagulant therapy, of which 9 patients were improved in short-term while 4 patients were not.

CONCLUSION

Trousseau's syndrome with MAIS was associated with old-age male, smoking and (or) drinking history, low NIHSS score, increased D-dimer, CRP and tumour markers, and lesions near the cortex areas with multiple vascular territories in DWI. Patients with these features should be alert of malignant tumour. Most common types of primary tumour were lung cancer. Treatment with low molecular heparin may be effective in short term.

Radiologic criteria of retropharyngeal lymph node metastasis in maxillary sinus cancer

Objective

To determine the most appropriate radiologic criteria of metastatic retropharyngeal lymph nodes (RLNs) in patients with maxillary sinus cancer (MSC).

Materials and methods

We retrospectively evaluated 16 consecutive patients who underwent magnetic resonance imaging (MRI) before and after the treatment of locally advanced squamous cell carcinoma of the maxillary sinus. The minimal and maximal diameters of all RLNS were recorded. RLNs were classified as metastatic on the basis of the MRI follow-up (f/u). RLNs were considered non-metastatic if stable disease continued until the final MRI f/u and metastatic in cases with different evaluations (complete response, partial response, progressive disease) determined using Response Evaluation Criteria in Solid Tumours (RECIST) ver. 1.1. The receiver operating characteristic curve (ROC) and area under the curve (AUC) were used to assess the accuracy of various criteria in the diagnosis of metastatic RLNs.

Results

Of the 34 RLNs in 16 cases observed on pretreatment MRI, 7 were classified as metastatic RLNs and 27 as non-metastatic RLNs. Using the radiologic criteria, metastatic RLNs tended to be diagnosed more accurately with the minimal axial diameter than with the maximal axial diameter (AUC; 0.97 vs. 0.73, p = 0.06). The most accurate size criterion of metastatic RLNs was a minimal axial diameter of 5 mm or larger, with an accuracy of 94.1% (32 of 34).

Conclusions

The most appropriate radiologic criterion of metastatic RLNs in MSC is a minimal axial diameter of 5 mm or longer.

Magnetic resonance-guided thermal ablation for small liver malignant tumor located on segment II or IVa abutting the heart: a retrospective cohort study

OBJECTIVE

This study aimed to evaluate the clinical safety and efficacy magnetic resonance (MR)-guided percutaneous thermal ablation for the treatment of small liver malignant tumors of segment II and IVa (≤3.0 cm) abutting the heart.

METHOD

The enrollment of 24 patients with 25 malignant liver lesions located on the II or IVa segment abutting the heart who underwent MRI-guided thermal ablation between August 2010 and February 2020 were retrospectively analyzed. Follow-up MRI was performed to evaluate the curative effect. Local tumor progression-free survival and overall survival rates were also calculated.

RESULTS

The procedures including radiofrequency ablation (RFA) for 15 patients and microwave ablation (MWA) for 9 patients were successfully accomplished (technical success rate of 100%) without major complications. The mean duration time was 78.4 ± 29.4 min (40-140 min), and mean follow-up time was 31.5 ± 22.2 months (6-92 months). The technical efficacy was 100% following one ablation session with MRI assessment after one month. Local tumor progression was observed in one patient with a metastatic lesion located in segment II at 18 months follow-up. The progression-free survival time was 20.1 ± 16.9 months (median: 15 months). The 1-, 3-, and 5-year local tumor progression-free survival rates of this patient were 100%, 94.7%, and 94.7%, respectively. With regards to all the patients, the 1-, 3-, and 5-year estimated overall survival rates were 91.7%, 80.6%, and 50.1%, respectively.

CONCLUSION

MR-guided thermal ablation is safe and effective for the treatment of small liver malignant tumors located on the II or IVa segment abutting the heart.

Predictive signs of peripheral rim instability with magnetic resonance imaging in no-shift-type complete discoid lateral meniscus

PURPOSE

To investigate the associations between the preoperative MRI findings suggestive of meniscal instability and the intraoperative finding of peripheral rim instability (PRI) in patients with complete discoid lateral meniscus (CDLM) of no-shift-type, which was identified as the peripheral portion was not separated from the capsule.

METHODS

The records of 56 patients diagnosed with no-shift-type CDLM who underwent arthroscopic surgery were reviewed. We evaluated MRI findings of anterior parameniscal soft-tissue edema, linear fluid signal at the anterior meniscal margin, bulging of the meniscal margin, absence of popliteomeniscal fascicles, and hiatus widening on routine MRI. The positive predictive value (PPV), sensitivity, and specificity of these findings in predicting PRI were calculated; PRI was further investigated according to anterior and posterior location.

RESULTS

Linear fluid signal at the anterior meniscal margin and bulging had high PPV and specificity (P = .004 and = .029, respectively) for overall of PRI. The presence of either anterior parameniscal soft-tissue edema or linear fluid signal at the anterior meniscal margin predicted anterior PRI with high PPV, sensitivity, and specificity. Bulging of the meniscal margin had high specificity, and either bulging of the meniscal margin or absence of popliteomeniscal fascicle had high sensitivity in predicting posterior PRI.

CONCLUSIONS

A linear fluid signal at the anterior meniscus and anterior parameniscal soft-tissue edema were important signs of anterior PRI, whereas bulging of the margin had high specificity and either bulging of the meniscal margin or absence of popliteomeniscal fascicle had high sensitivity in detecting posterior PRI on routine MRI of no-shift-type CDLM.

LEVEL OF EVIDENCE

Level IV therapeutic case series.

Gradual back-projection residual attention network for magnetic resonance image super-resolution

BACKGROUND AND OBJECTIVE

Magnetic Resonance Image (MRI) analysis can provide anatomical examination of internal organs, which is helpful for diagnosis of the disease. Aiming at the problems of insufficient feature information mining in the process of MRI super-resolution (SR) reconstruction, the difficulty of determining the interdependence between the channels of the feature map, and the reconstruction error when reconstructing high-resolution (HR) images, we propose a SR method to solve these problems.

METHODS

In this work, we propose a gradual back-projection residual attention network for MRI super-resolution (GRAN), which outperforms most of the state-of-the-art methods. Firstly, we use the gradual upsampling method to gradually scale the low-resolution (LR) image to a given magnification to alleviate the high-frequency information loss caused by the upsampling process. Secondly, we merge the idea of iterative back-projection at each stage of gradual upsampling, learn the mapping relationship between HR and LR feature maps and reduce the noise introduced during the upsampling process. Finally, we use the attention mechanism to dynamically allocate attention resources to the feature maps generated at different stages of the gradual back-projection network, so that the network model can learn the interdependence between each feature map.

RESULTS

For the 2 × and 4 × enlargement, the proposed GRAN method shows the superiority over the state-of-the-art methods on the Set5, Set14, and Urban100 benchmark datasets, extensive benchmark experiment and analysis show that the superiority of the GRAN algorithm in terms of peak signal-to-noise ratio and structural similarity index indicators.

CONCLUSION

The MRI results reconstructed by gradual back-projection residual attention network on the public dataset IDI have good image sharpness, rich texture details and good visual experience. In addition, the reconstructed image is the closest to the real image, enabling the medical expert to see the biological tissue structure and its early pathological changes more clearly, providing assistance and support to the medical expert in the diagnosis and treatment of the disease.

The total resection rate of glioma can be improved by the application of US-MRI fusion combined with contrast-enhanced ultrasound

OBJECTIVE

This study was performed to evaluate the diagnostic performance of ultrasound-magnetic resonance imaging (MRI) fusion combined with contrast-enhanced ultrasound and to explore its role in improving the total tumor resection rate.

METHODS

Between January 2018 and December 2018, 16 patients in the observation group and 23 patients in the control group were enrolled in this study. The tumor depth and brain shift distance were analyzed, as well as the peak intensity and microvessel density of different grades of gliomas in the observation group. Finally, we compared the difference in total resection rate between the observation and control groups.

RESULTS

Using ultrasound during operations, we found a significant negative correlation between brain shift distance and tumor depth, with correlation coefficient r=-0.868(P<0.05). In glioma, the peak intensity and microvessel density increased synchronously with glioma grade(r=0.806, P<0.05). The total resection rate of lesions was significantly higher in the observation group than in the control group (P<0.05).

CONCLUSIONS

The application of ultrasound-MRI fusion combined with contrast-enhanced ultrasound can improve the total resection rate of lesions, thus playing an important role in clinical practice.

Fat-saturated image generation from multi-contrast MRIs using generative adversarial networks with Bloch equation-based autoencoder regularization

Obtaining multiple series of magnetic resonance (MR) images with different contrasts is useful for accurate diagnosis of human spinal conditions. However, this can be time consuming and a burden on both the patient and the hospital. We propose a Bloch equation-based autoencoder regularization generative adversarial network (BlochGAN) to generate a fat saturation T2-weighted (T2 FS) image from T1-weighted (T1-w) and T2-weighted (T2-w) images of human spine. To achieve this, our approach was to utilize the relationship between the contrasts using Bloch equation since it is a fundamental principle of MR physics and serves as a physical basis of each contrasts. BlochGAN properly generated the target-contrast images using the autoencoder regularization based on the Bloch equation to identify the physical basis of the contrasts. BlochGAN consists of four sub-networks: an encoder, a decoder, a generator, and a discriminator. The encoder extracts features from the multi-contrast input images, and the generator creates target T2 FS images using the features extracted from the encoder. The discriminator assists network learning by providing adversarial loss, and the decoder reconstructs the input multi-contrast images and regularizes the learning process by providing reconstruction loss. The discriminator and the decoder are only used in the training process. Our results demonstrate that BlochGAN achieved quantitatively and qualitatively superior performance compared to conventional medical image synthesis methods in generating spine T2 FS images from T1-w, and T2-w images.

Cerebral Microbleeds Load and Long-Term Outcomes in Minor Ischemic Stroke

BACKGROUND AND PURPOSE

The association between the cerebral microbleed (CMB) count and outcomes in ischemic stroke has not been fully clarified. The aim of this study was to investigate the relationship between the CMBs count and functional outcomes in patients with a minor ischemic stroke treated with antiplatelet therapy METHODS: Non-cardiogenic minor ischemic stroke (NIHSS score < 4 on admission) patients who were treated with antiplatelet therapy were enrolled. The patients were divided into four groups based on the number of CMBs (absent, 1, 2-4, and > 4), and their clinical outcomes were compared. A poor outcome was defined as a modified Rankin scale (mRS) score of 3-6 90 days after symptom onset. Logistic regression analysis was performed to evaluate whether the CMBs count contributes to poor outcomes with well-known risk factors such as age, NIHSS score on admission, ischemic stroke recurrence, large artery atherosclerosis stroke subtype, and DWMHs.

RESULTS

A total of 240 patients were enrolled, and their pre mRS scores were matched based on CMB presence. A higher burden of CMBs was linearly correlated with the incidence of poor outcomes (4% in the absent group, 8% in the 1 CMB group, 13% in the 2-4 CMB group, and 20% in the > 4 CMB group, P = 0.002). Multivariate logistic regression analysis showed that CMBs count was one of the independent factor associated with poor outcomes (odds ratio 1.07, 95% confidence interval 1.02-1.12, P = 0.003).

CONCLUSION

The CMBs count contributes independently to poor outcomes in minor ischemic stroke patients treated with antiplatelet therapy.

Neurofluid as Assessed by Diffusion-Weighted Imaging

The glymphatic system hypothesis is associated with the circulation of cerebrospinal fluid (CSF) in the skull and interstitial fluid (ISF) in the brain. There are several imaging techniques to visualize the dynamics of CSF and ISF. Magnetic resonance imaging (MRI) is one of the promising modalities for glymphatic imaging and diffusion MRI is expected imaging tool. Several disorders are associated with glymphatic dysfunction or impairment in the dynamics of CSF or ISF. The Central Nervous System interstitial fluidopathy concept has been proposed to encompass diseases with pathologies that are predominantly associated with abnormal ISF/CSF dynamics.

Construction and validation of a cerebral white matter hyperintensity probability map of older Koreans

•

We constructed WPM from healthy elderly Koreans.

•

WPM may serve as a tool to study pathology and normal aging of distribution of WMH.

•

WPM provides a prominent atlas of the age related distribution of WMH.

Background and purpose

Although two white matter hyperintensity (WMH) probability maps of healthy older adults already exist, they have several limitations in representing the distribution of WMH in healthy older adults, especially Asian older adults. We constructed and validated a WMH probability map (WPM) of healthy older Koreans and examined the age-associated differences of WMH.

Methods

We constructed WPM using development dataset that consisted of high-resolution 3D fluid-attenuated inversion recovery images of 5 age groups (60–64 years, 65–69 years, 70–74 years, 75–79 years, and 80+ years). Each age group included 30 age-matched men and women each. We tested the validity of the WPM by comparing WMH ages estimated by the WPM and the chronological ages of 30 healthy controls, 30 hypertension patients, and 30 S patients.

Results

Older age groups showed a higher volume of WMH in both hemispheres (p < 0.001). About 90% of the WMH were periventricular in all age groups. With advancing age, the peak of the distance histogram from the ventricular wall of the periventricular WMH shifted away from the ventricular wall, while that of deep WMH shifted toward the ventricular wall. The estimated WMH ages were comparable to the chronological ages in the healthy controls, while being higher than the chronological ages in hypertension and stroke patients.

Conclusions

This WPM may serve as a standard atlas in research on WMH of older adults, especially Asians.

Convolutional neural network for automatically segmenting magnetic resonance images of the shoulder joint

BACKGROUND

Magnetic resonance imaging (MRI) has been known to replace computed tomography (CT) for bone and skeletal joint examination. The accurate automatic segmentation of bone structure in shoulder MRI is important for the measurement and diagnosis of bone injury and disease. Existing bone segmentation algorithms cannot achieve automatic segmentation without any prior knowledge, and their versatility and accuracy are relatively low. Therefore, an automatic segmentation combining pulse coupled neural network (PCNN) and full convolutional neural networks (FCN) is proposed.

METHODOLOGY

By constructing the block-based AlexNet segmentation model and U-Net-based bone segmentation module, we implemented the humeral segmentation model, articular bone segmentation model, humeral head and articular bone segmentation model synthesis model. We use this four kinds of segmentation models to obtain candidate bone regions, and accurately detect the positions of humerus and articular bone by voting. Finally, we perform an AlexNet segmentation model in the detected bone area in one step to segment accuracy at the pixel level.

RESULTS

The experimental data came from 8 groups of patients in Shengjing Hospital affiliated to China Medical University. The scanning volume of each group is approximately 100 images. Five fold cross-validations and for training were recorded, and five sets of data were carefully separated. After using our technique in the three groups of patients tested, the positive predictive value of dice coefficient (PPV) and the average accuracy of sensitivity were very significant, which reached 0.96±0.02, 0.97±0.02 and 0.94±0.03, respectively.

CONCLUSION

The method used in the experiment in this paper is based on a small amount of patient sample data. The deep learning required for the experiment needs to be performed through 2D medical images. The shoulder segmentation data obtained in this way can be very accurate.

Tumor type detection in brain MR images of the deep model developed using hypercolumn technique, attention modules, and residual blocks

Brain cancer is a disease caused by the growth of abnormal aggressive cells in the brain outside of normal cells. Symptoms and diagnosis of brain cancer cases are producing more accurate results day by day in parallel with the development of technological opportunities. In this study, a deep learning model called BrainMRNet which is developed for mass detection in open-source brain magnetic resonance images was used. The BrainMRNet model includes three processing steps: attention modules, the hypercolumn technique, and residual blocks. To demonstrate the accuracy of the proposed model, three types of tumor data leading to brain cancer were examined in this study: glioma, meningioma, and pituitary. In addition, a segmentation method was proposed, which additionally determines in which lobe area of the brain the two classes of tumors that cause brain cancer are more concentrated. The classification accuracy rates were performed in the study; it was 98.18% in glioma tumor, 96.73% in meningioma tumor, and 98.18% in pituitary tumor. At the end of the experiment, using the subset of glioma and meningioma tumor images, it was determined which at brain lobe the tumor region was seen, and 100% success was achieved in the analysis of this determination. In this study, a hybrid deep learning model is presented to determine the detection of the brain tumor. In addition, open-source software was proposed, which statistically found in which lobe region of the human brain the brain tumor occurred. The methods applied and tested in the experiments have shown promising results with a high level of accuracy, precision, and specificity. These results demonstrate the availability of the proposed approach in clinical settings to support the medical decision regarding brain tumor detection.

Large chondral defect not covered by meniscal allograft is associated with inferior graft survivorship after lateral meniscal allograft transplantation

PURPOSE

This study aimed to evaluate graft survivorship according to the size and location of chondral defects and its effect on clinical outcomes after meniscal allograft transplantation (MAT). It was hypothesized that large chondral defects would be associated with inferior outcomes.

METHODS

Patients who underwent lateral MAT with fresh-frozen allografts between 2007 and 2016 were retrospectively reviewed. The inclusion criteria were patients with femoral or tibial chondral defects (International Cartilage Repair Society grade 4) who were followed up more than 2 years with 3.0-T magnetic resonance imaging (MRI) scans. Maximal lesion diameter and location were assessed on MRI. The patients were divided into two groups, with chondral defects of < 3 and ≥ 3 cm² on the tibial side. Graft survivorship was compared between the two groups. Graft failure was defined as revisional MAT, meniscal tear or meniscectomy greater than one-third of the allograft on MRI. Clinical outcomes were evaluated using the modified Lysholm score.

RESULTS

Twenty-eight knees in 26 patients (mean age 37.4 ± 10.3 years) with a mean follow-up of 3.6 ± 1.0 (range 2.0-5.4) years were identified. Nineteen knees in 17 patients had both femoral and tibial chondral defects, 7 knees in 7 patients had only femoral chondral defects, and 2 knees in 2 patients had only tibial chondral defects. The mean preoperative femoral and tibial chondral defect sizes were 1.7 ± 1.2 and 3.0 ± 1.4 cm², respectively. Among the seven graft failures, no graft failure occurred in the cases with tibial chondral defects of < 3 cm². Tibial chondral defects of ≥ 3 cm² were significantly associated with graft failure (P = 0.004; odds ratio 28.3; 95% confidence interval 2.5-4006.7). Defects of < 3 cm² were located primarily in the posterior aspect of the lateral tibial plateau, and most lesions were covered by allograft (7/9, 77.8%). The modified Lysholm scores significantly improved irrespective of chondral defects size (P < 0.001).

CONCLUSIONS

Larger chondral defects, more than 3 cm² on the tibial side, were associated with inferior graft survivorship but did not influence the clinical outcomes after MAT at the 3.6-year follow-up. Chondral defect location was associated with defect size.

LEVEL OF EVIDENCE

IV.

Fully automated 3D segmentation of MR-imaged calf muscle compartments: Neighborhood relationship enhanced fully convolutional network

Automated segmentation of individual calf muscle compartments from 3D magnetic resonance (MR) images is essential for developing quantitative biomarkers for muscular disease progression and its prediction. Achieving clinically acceptable results is a challenging task due to large variations in muscle shape and MR appearance. In this paper, we present a novel fully convolutional network (FCN) that utilizes contextual information in a large neighborhood and embeds edge-aware constraints for individual calf muscle compartment segmentations. An encoder-decoder architecture is used to systematically enlarge convolution receptive field and preserve information at all resolutions. Edge positions derived from the FCN output muscle probability maps are explicitly regularized using kernel-based edge detection in an end-to-end optimization framework. Our method was evaluated on 40 T1-weighted MR images of 10 healthy and 30 diseased subjects by fourfold cross-validation. Mean DICE coefficients of 88.00-91.29% and mean absolute surface positioning errors of 1.04-1.66 mm were achieved for the five 3D muscle compartments.

Detecting pathological brain via ResNet and randomized neural networks

Brain disease is one of the leading causes of death nowadays. Medical imaging is the most effective method for brain disease diagnosis, which provides a clear view of the interior brain. However, manual interpretation requires too much time and effort because medical images contain a large volume of information. Computer aided diagnosis is playing a more and more significant role in the clinic, which can help doctors and physicians to analyze medical images automatically. In this study, a novel pathological brain detection system was proposed for brain magnetic resonance images based on ResNet and randomized neural networks. Firstly, a ResNet was employed as the feature extractor, which was a famous convolutional neural network structure. Then, we used three randomized neural networks, i.e., the Schmidt neural network, the random vector functional-link net, and the extreme learning machine. The weights and biases in the three networks were trained by the chaotic bat algorithm. The three proposed methods achieved similar results based on five runs, and they yielded comparable performance in comparison with state-of-the-art approaches.

Tumor type detection in brain MR images of the deep model developed using hypercolumn technique, attention modules, and residual blocks

Brain cancer is a disease caused by the growth of abnormal aggressive cells in the brain outside of normal cells. Symptoms and diagnosis of brain cancer cases are producing more accurate results day by day in parallel with the development of technological opportunities. In this study, a deep learning model called BrainMRNet which is developed for mass detection in open-source brain magnetic resonance images was used. The BrainMRNet model includes three processing steps: attention modules, the hypercolumn technique, and residual blocks. To demonstrate the accuracy of the proposed model, three types of tumor data leading to brain cancer were examined in this study: glioma, meningioma, and pituitary. In addition, a segmentation method was proposed, which additionally determines in which lobe area of the brain the two classes of tumors that cause brain cancer are more concentrated. The classification accuracy rates were performed in the study; it was 98.18% in glioma tumor, 96.73% in meningioma tumor, and 98.18% in pituitary tumor. At the end of the experiment, using the subset of glioma and meningioma tumor images, it was determined which at brain lobe the tumor region was seen, and 100% success was achieved in the analysis of this determination. In this study, a hybrid deep learning model is presented to determine the detection of the brain tumor. In addition, open-source software was proposed, which statistically found in which lobe region of the human brain the brain tumor occurred. The methods applied and tested in the experiments have shown promising results with a high level of accuracy, precision, and specificity. These results demonstrate the availability of the proposed approach in clinical settings to support the medical decision regarding brain tumor detection.

MRI guided ROLL/SNOLL in breast cancer patients treated with neoadjuvant chemotherapy

PURPOSE

To describe the results of MRI (magnetic resonance image) guided ROLL (radioguided occult lesion localization) and SNOLL (sentinel node occult lesion localization) in the localization of residual disease after neoadjuvant chemotherapy for breast cancer, as well as assessing the surgical results obtained and disease free survival.

METHODS

Prospective observational analysis of 132 patients with 136 tumors, treated with neoadjuvant chemotherapy at our hospital between 2011-2017. Residual disease was located presurgically with MRI guided ROLL/SNOLL technique. We analyzed technical aspects of localization, and variables corresponding to surgical procedures and events occurred during follow-up.

RESULTS

The median tumor size was of 20.5mm (interquartilic range [IQR]: 14-28). The majority (96.3%) were invasive ductal carcinomas. Sentinel lymph node detection rate was 98.9%. Complete pathological response (CPR) in the breast was achieved in 58.1% of cases. The rate of affected margins in 89 cases operated by conservative surgery was 2.2%. With a median follow-up of 50 months (IQR: 37-61) we found a 7.4% rate of relapses. Of these, seven were loco-regional and three at distant sites. The estimated mean of disease-free survival time was 83.2 months (Confidence Interval [CI] 95%: 79.6-86.6).

CONCLUSIONS

MRI guided ROLL/SNOLL is a great tool for breast cancer residual disease localization following neoadjuvant chemotherapy. In addition, this technique attains good loco-regional control of the diseases and has excellent surgical results.

Prediction of pseudoprogression and long-term outcome of vestibular schwannoma after Gamma Knife radiosurgery based on preradiosurgical MR radiomics

BACKGROUND AND PURPOSE

Gamma Knife radiosurgery (GKRS) is a safe and effective treatment modality with a long-term tumor control rate over 90% for vestibular schwannoma (VS). However, numerous tumors may undergo a transient pseudoprogression during 6-18 months after GKRS followed by a long-term volume reduction. The aim of this study is to determine whether the radiomics analysis based on preradiosurgical MRI data could predict the pseudoprogression and long-term outcome of VS after GKRS.

MATERIALS AND METHODS

A longitudinal dataset of patients with VS treated by single GKRS were retrospectively collected. Overall 336 patients with no previous craniotomy for tumor removal and a median of 65-month follow-up period after radiosurgery were finally included in this study. In total 1763 radiomic features were extracted from the multiparameteric MRI data before GKRS followed by the machine-learning classification.

RESULTS

We constructed a two-level machine-learning model to predict the long-term outcome and the occurrence of transient pseudoprogression after GKRS separately. The prediction of long-term outcome achieved an accuracy of 88.4% based on five radiomic features describing the variation of T2-weighted intensity and inhomogeneity of contrast enhancement in tumor. The prediction of transient pseudoprogression achieved an accuracy of 85.0% based on another five radiomic features associated with the inhomogeneous hypointensity pattern of contrast enhancement and the variation of T2-weighted intensity.

CONCLUSION

The proposed machine-learning model based on the preradiosurgical MR radiomics provides a potential to predict the pseudoprogression and long-term outcome of VS after GKRS, which can benefit the treatment strategy in clinical practice.

Percutaneous radiofrequency ablation for small hepatocellular carcinoma in hepatic dome under MR-guidance: clinical safety and efficacy

Purpose: To evaluate the clinical safety and efficacy of percutaneous radiofrequency ablation (RFA) using multitined expandable electrodes under magnetic resonance imaging (MRI) guidance in the treatment of small hepatocellular carcinomas (HCCs) in the hepatic dome.Materials and methods: The data of 49 patients with 50 HCC lesions in the hepatic dome who underwent MRI-guided RFA from April 2010 to January 2018 were retrospectively analyzed. Planning, targeting, and controlling were performed under MR-guidance during the procedure. The complications after RFA were observed. Follow-up MRI was performed to evaluate the curative effect. The local progression-free survival, recurrence-free survival, and overall survival rates were calculated using the Kaplan-Meier survival curve.Results: The procedures were successfully accomplished in all patients without major complications. The mean follow-up time was 36.9 ± 25.8 months (range, 3-99 months). Technical success was 100% after one RFA session with MRI assessment after 1 month. Local tumor progression was observed in one patient (2%) with the lesion located in the hepatic dome at 4 months on a subsequent follow-up MRI. The progression-free survival time was 25.0 ± 22.7 months (median, 17.0 months). The 1-,3-, and 5-year local tumor progression-free survival rates were all 98.0%. The 1-,3-, and 5-year recurrence-free survival rates were 68.1%, 39.9%, and 28.5%, respectively, and the estimated overall survival rates were 93.7%, 76.3%, and 54.3%, respectively.Conclusion: Planning, targeting, and controlling of RFA were well supported by MRI with acceptable time. MRI-guided RFA for small HCCs in the hepatic dome is safe and effective with fewer RF sessions.

Osteochondral lesion of talus with gout tophi deposition: A case report

BACKGROUND

Osteochondral lesion of talus is a broad term used to describe an injury or abnormality of the talar articular cartilage and adjacent bone. It arises from diverse causes, and although trauma is implicated in many cases, it does not account for the etiology of every lesion. Gout is a chronic arthritic disease caused by excess levels of uric acid in blood. Intraosseous deposition of monosodium urate in the clavicle, femur, patella and calcaneus was reported previously. Gout is common disease but rare at a young age, especially during teenage years. Osteochondral lesion caused by intra-articular gouty invasion is very rare.

CASE SUMMARY

We encountered a rare case of a 16-year-old male who has osteochondral lesion of the talus (OLT) with gout. He had fluctuating pain for more than 2 years. We could see intra-articular tophi with magnetic resonance image (MRI) and arthroscopy. We performed arthroscopic exploration, debridement and microfracture. Symptoms were resolved after operation, and bony coverage at the lesion was seen on postoperative images. We had checked image and uric acid levels for 18 mo.

CONCLUSION

It is rare to see OLT with gouty tophi in young adults. While it is challenging, the accuracy of diagnosis can be improved through history taking, MRI and arthroscopy.