Correction of inhomogeneous magnetic resonance images using multiscale retinex for segmentation accuracy improvement

The purpose of this study was to improve the accuracy of tissue segmentation on brain magnetic resonance (MR) images preprocessed by multiscale retinex (MSR), segmented with a combined boosted decision tree (BDT) and MSR algorithm (hereinafter referred to as the MSRBDT algorithm). Simulated brain MR (SBMR) T1-weighted images of different noise levels and RF inhomogeneities were adopted to evaluate the outcome of the proposed method; the MSRBDT algorithm was used to identify the gray matter (GM), white matter (WM), and cerebral-spinal fluid (CSF) in the brain tissues. The accuracy rates of GM, WM, and CSF segmentation, with spatial features (G, x, y, r, θ), were respectively greater than 0.9805, 0.9817, and 0.9871. In addition, images segmented with the MSRBDT algorithm were better than those obtained with the expectation maximization (EM) algorithm; brain tissue segmentation in MR images was significantly more precise. The proposed MSRBDT algorithm could be beneficial in clinical image segmentation.     

Glioblastomas and brain metastases differentiation following an MRI texture analysis-based radiomics approach

PurposeTo evaluate the potential of 2D texture features extracted from magnetic resonance (MR) images for differentiating brain metastasis (BM) and glioblastomas (GBM) following a radiomics approach.MethodsThis retrospective study included 50 patients with BM and 50 with GBM who underwent T1-weighted MRI between December 2010 and January 2017. Eighty-eight rotation-invariant texture features were computed for each segmented lesion using six texture analysis methods. These features were also extracted from the four images obtained after applying the discrete wavelet transform (88 features × 4 images). Three feature selection methods and five predictive models were evaluated. A 5-fold cross-validation scheme was used to randomly split the study group into training (80 patients) and testing (20 patients), repeating the process ten times. Classification was evaluated computing the average area under the receiver operating characteristic curve. Sensibility, specificity and accuracy were also computed. The whole process was tested quantizing the images with different gray-level values to evaluate their influence in the final results.ResultsHighest classification accuracy was obtained using the original images quantized with 128 gray-levels and a feature selection method based on the p-value. The best overall performance was achieved using a support vector machine model with a subset of 32 features (AUC = 0.896 ± 0.067, sensitivity of 82% and specificity of 80%). Naïve Bayes and k-nearest neighbors models showed also valuable results (AUC ≈ 0.8) with a lower number of features (<13), thus suggesting that these models may be more generalizable when using external validations.ConclusionThe proposed radiomics MRI approach is able to discriminate between GBM and BM with high accuracy employing a set of 2D texture features, thus helping in the diagnosis of brain lesions in a fast and non-invasive way.     

Effect of slice thickness on brain magnetic resonance image texture analysis

Background  

The accuracy of texture analysis in clinical evaluation of magnetic resonance images depends considerably on imaging arrangements and various image quality parameters. In this paper, we study the effect of slice thickness on brain tissue texture analysis using a statistical approach and classification of T1-weighted images of clinically confirmed multiple sclerosis patients.     

A novel framework for spatial normalization of dose distributions in voxel-based analyses of brain irradiation outcomes

PurposeTo devise a novel Spatial Normalization framework for Voxel-based analysis (VBA) in brain radiotherapy. VBAs rely on accurate spatial normalization of different patients’ planning CTs on a common coordinate system (CCS). The cerebral anatomy, well characterized by MRI, shows instead poor contrast in CT, resulting in potential inaccuracies in VBAs based on CT alone.MethodsWe analyzed 50 meningioma patients treated with proton-therapy, undergoing planning CT and T1-weighted (T1w) MRI. The spatial normalization pipeline based on MR and CT images consisted in: intra-patient registration of CT to T1w, inter-patient registration of T1w to MNI space chosen as CCS, doses propagation to MNI.The registration quality was compared with that obtained by Statistical Parametric Mapping software (SPM), used as benchmark. To evaluate the accuracy of dose normalization, the dose organ overlap (DOO) score was computed on gray matter, white matter and cerebrospinal fluid before and after normalization. In addition, the trends in the DOOs distribution were investigated by means of cluster analysis.ResultsThe registration quality was higher for the proposed method compared to SPM (p < 0.001). The DOO scores showed a significant improvement after normalization (p < 0.001). The cluster analysis highlighted 2 clusters, with one of them including the majority of data and exhibiting acceptable DOOs.ConclusionsOur study presents a robust tool for spatial normalization, specifically tailored for brain dose VBAs. Furthermore, the cluster analysis provides a formal criterion for patient exclusion in case of non-acceptable normalization results. The implemented framework lays the groundwork for future reliable VBAs in brain irradiation studies.     

A Hybrid CNN-SVM Threshold Segmentation Approach for Tumor Detection and Classification of MRI Brain Images

ObjectiveIn this research paper, the brain MRI images are going to classify by considering the excellence of CNN on a public dataset to classify Benign and Malignant tumors.Materials and MethodsDeep learning (DL) methods due to good performance in the last few years have become more popular for Image classification. Convolution Neural Network (CNN), with several methods, can extract features without using handcrafted models, and eventually, show better accuracy of classification. The proposed hybrid model combined CNN and support vector machine (SVM) in terms of classification and with threshold-based segmentation in terms of detection.ResultThe findings of previous studies are based on different models with their accuracy as Rough Extreme Learning Machine (RELM)-94.233%, Deep CNN (DCNN)-95%, Deep Neural Network (DNN) and Discrete Wavelet Autoencoder (DWA)-96%, k-nearest neighbors (kNN)-96.6%, CNN-97.5%. The overall accuracy of the hybrid CNN-SVM is obtained as 98.4959%.ConclusionIn today's world, brain cancer is one of the most dangerous diseases with the highest death rate, detection and classification of brain tumors due to abnormal growth of cells, shapes, orientation, and the location is a challengeable task in medical imaging. Magnetic resonance imaging (MRI) is a typical method of medical imaging for brain tumor analysis. Conventional machine learning (ML) techniques categorize brain cancer based on some handicraft property with the radiologist specialist choice. That can lead to failure in the execution and also decrease the effectiveness of an Algorithm. With a brief look came to know that the proposed hybrid model provides more effective and improvement techniques for classification.     

In phantom evaluation of targeting accuracy in MRI-based brain radiosurgery

PurposeTo determine the targeting accuracy of brain radiosurgery when planning procedures employing different MRI and MRI + CT combinations are adopted.Materials and methodA new phantom, the BrainTool, has been designed and realized to test image co-registration and targeting accuracy in a realistic anatomical situation. The phantom was created with a 3D printer and materials that mimic realistic brain MRI and CT contrast using a model extracted from a synthetic MRI study of a human brain. Eight markers distributed within the BrainTool provide for assessment of the accuracy of image registrations while two cavities that host an ionization chamber are used to perform targeting accuracy measurements with an iterative cross-scan method. Two procedures employing 1.5 T MRI-only or a combination of MRI (taken with 1.5 T or 3 T scanners) and CT to carry out Gamma Knife treatments were investigated. As distortions can impact targeting accuracy, MR images were preliminary evaluated to assess image deformation extent using GammaTool phantom.ResultsMR images taken with both scanners showed average and maximum distortion of 0.3 mm and 1 mm respectively. The marker distances in co-registered images resulted below 0.5 mm for both MRI scans. The targeting mismatches obtained were 0.8 mm, 1.0 mm and 1.2 mm for MRI-only and MRI + CT (1,5T and 3 T), respectively.ConclusionsProcedures using a combination of MR and CT images provide targeting accuracies comparable to those of MRI-only procedures. The BrainTool proved to be a suitable tool for carrying out co-registration and targeting accuracy of Gamma Knife brain radiosurgery treatments.     

Automated classification of urinary stones based on microcomputed tomography images using convolutional neural network

PurposeThe classification of urinary stones is important prior to treatment because the treatments depend on three types of urinary stones, i.e., calcium, uric acid, and mixture stones. We have developed an automatic approach for the classification of urinary stones into the three types based on microcomputed tomography (micro-CT) images using a convolutional neural network (CNN).Materials and methodsThirty urinary stones from different patients were scanned in vitro using micro-CT (pixel size: 14.96 μm; slice thickness: 15 μm); a total of 2,430 images (micro-CT slices) were produced. The slices (227 × 227 pixels) were classified into the three categories based on their energy dispersive X-ray (EDX) spectra obtained via scanning electron microscopy (SEM). The images of urinary stones from each category were divided into three parts; 66%, 17%, and 17% of the dataset were assigned to the training, validation, and test datasets, respectively. The CNN model with 15 layers was assessed based on validation accuracy for the optimization of hyperparameters such as batch size, learning rate, and number of epochs with different optimizers. Then, the model with the optimized hyperparameters was evaluated for the test dataset to obtain classification accuracy and error.ResultsThe validation accuracy of the developed approach with CNN with optimized hyperparameters was 0.9852. The trained CNN model achieved a test accuracy of 0.9959 with a classification error of 1.2%.ConclusionsThe proposed automated CNN-based approach could successfully classify urinary stones into three types, namely calcium, uric acid, and mixture stones, using micro-CT images.     

Object-oriented classification approach for bone metastasis mapping from whole-body bone scintigraphy

PurposeWhole-body bone scintigraphy is the most widely used method for detecting bone metastases in advanced cancer. However, its interpretation depends on the experience of the radiologist. Some automatic interpretation systems have been developed in order to improve diagnostic accuracy. These systems are pixel-based and do not use spatial or textural information of groups of pixels, which could be very important for classifying images with better accuracy. This paper presents a fast method of object-oriented classification that facilitates easier interpretation of bone scintigraphy images.MethodsNine whole-body images from patients suspected with bone metastases were analyzed in this preliminary study. First, an edge-based segmentation algorithm together with the full lambda-schedule algorithm were used to identify the object in the bone scintigraphy and the textural and spatial attributes of these objects were calculated. Then, a set of objects (224 objects, ~ 46% of the total objects) were selected as training data based on visual examination of the image, and were assigned to various levels of radionuclide accumulation before performing the data classification using both k-nearest-neighbor and support vector machine classifiers. The performance of the proposed method was evaluated using as metric the statistical parameters calculated from error matrix.ResultsThe results revealed that the proposed object-oriented classification approach using either k-nearest-neighbor or support vector machine as classification methods performed well in detecting bone metastasis in terms of overall accuracy (86.62 ± 2.163% and 86.81 ± 2.137% respectively) and kappa coefficient (0.6395 ± 0.0143 and 0.6481 ± 0.0218 respectively).ConclusionsIn conclusion, the described method provided encouraging results in mapping bone metastases in whole-body bone scintigraphy.     

Differential diagnosis between progression and radionecrosis in brain metastases after stereotactic radiosurgery using hybrid FDG-PET and MRI coregistered images

IntroductionDual phase 18 FDG brain PET is helpful to assess brain metastases (BM) as tracer will build up in metastases or tumor recurrences while its retention remains stable within normal tissue or inflammatory processes. This is useful when MRI can’t discriminate brain tumor recurrence (TR) rom radionecrosis (RN) after stereotaxic radiosurgery (SRS) for BM. Many studies have sought to improve diagnostic performance by associating FDG-PET and MRI with interesting results but many biases, mostly within image post-processing. Coregistered MRI and dual phase FDG-PET images could alleviate these biases and be used to extract prognostic biomarkers.Materials and methodsWe retrospectively evaluated patients treated with SRS for BM which developed a contrast-enhanced MRI lesion with non-conclusive diagnosis for TR or RN. All patients underwent MRI and FDG-PET at least 3 months after their last SRS session. Dual FDG-PET consisted in an “early” and “delayed” acquisition, respectively 30 minutes and 4 h after injection. MRI included permeability and perfusion sequences. PET and MRI data were all coregistered on the contrast enhanced T1 MRI images. Semi-automated Volumes of Interest (VOI) of the tumor were drawn on the BM and a reference contralateral white-matter ROI (WM) was drawn for standardization; every metric was calculated inside these ROIs, in particular the tumor SUVmax and its variation in time. A 20% increase in the tumor SUVmax was in favor of TR while a modification of less than 100% was in favor of RN. Imaging metrics were then evaluated for their association with TR or RN based on histological, radiological and clinical criteria after at least 6 months follow-up.ResultsNine patients were ruled out as TR and 6 as RN. After standardization, there was a significant difference between groups for VP (P = 0.042), Washin (P = 0.035), Peak Enhancement (P = 0.037), standardized delayed SUVmax (P = 0.008) and RI (P = 0.016). Semi-quantitative analysis found respectively for PET and MRI a Sensitivity of 100% and 87.5% and a Specificity of 100% and 85.71%.ConclusionCoregistered PET-MRI images accurately discriminate between TR and RN. With FDG being the most commonly used PET radiotracer, this protocol remains easily transposable and should be encouraged to obtain non-invasive prognostic and clinically relevant biomarkers.     

Modélisation des flux cérébraux par une analogie électrique : validation par vélocimétrie IRM

The cine Phase-Contrast Magnetic Resonance (PCMR) sequence is the only noninvasive technique for the study of cerebrospinal fluid (CSF) oscillations. It can provide CSF and blood flow measurements throughout the cardiac cycle. To study cerebral hydro-hemodynamic, models have been developed; nevertheless the majority of these models did not take into account the CSF oscillations. The objective of this study was to establish reference values for cerebral hydro-hemodynamic and propose a new electrical model of the brain dynamics.Material and methodsCSF and blood flows were measured in 19 control subjects by PCMR imaging. Dynamic flow images were analyzed on dedicated software to reconstruct the flow curves during the cardiac cycle. An electrical analogue was realized. The inputs of the model were fed by PCMR arterial and venous flows to simulate CSF oscillations. The simulated CSF oscillations were compared to the measured CSF oscillations to validate the model.ResultsThe key parameters of the CSF and blood flow curves were obtained, e.g. total cerebral blood flow was 688 ± 115 mL/min, ventricular CSF oscillatory volume was 0.05 ± 0.02 mL/cardiac cycle, and the subarachnoid CSF oscillatory volume was 0.55 ± 0.15 mL/cardiac cycle. A close agreement was found between measured and simulated cerebral CSF oscillations.ConclusionThis study established the main values characterizing cerebral hydrodynamics in a control population. It provided a better understanding of the mechanisms of intracranial volumes regulation during the cardiac cycle. Our results are now used in clinical practice and the model proposed is effective to study cerebral hydro-hemodynamic.     

Widespread sensorimotor and frontal cortical atrophy in Amyotrophic Lateral Sclerosis

Background  

Widespread cortical atrophy in Amyotrophic Lateral Sclerosis (ALS) has been described in neuropathological studies. The presence of cortical atrophy in conventional and scientific neuroimaging has been a matter of debate. In studies using computertomography, positron emission tomography, proton magnetic resonance spectroscopy and conventional T2-weighted and proton-weighted images, results have been variable. Recent morphometric studies by magnetic resonance imaging have produced conflicting results regarding the extent of grey and white matter involvement in ALS patients.     

Lung tumors on multimodal radiographs derived from grating-based X-ray imaging – A feasibility study

PurposeThe purpose of this study was to assess whether grating-based X-ray imaging may have a role in imaging of pulmonary nodules on radiographs.Materials and methodsA mouse lung containing multiple lung tumors was imaged using a small-animal scanner with a conventional X-ray source and a grating interferometer for phase-contrast imaging. We qualitatively compared the signal characteristics of lung nodules on transmission, dark-field and phase-contrast images. Furthermore, we quantitatively compared signal characteristics of lung tumors and the adjacent lung tissue and calculated the corresponding contrast-to-noise ratios.ResultsOf the 5 tumors visualized on the transmission image, 3/5 tumors were clearly visualized and 1 tumor was faintly visualized in the dark-field image as areas of decreased small angle scattering. In the phase-contrast images, 3/5 tumors were clearly visualized, while the remaining 2 tumors were faintly visualized by the phase-shift occurring at their edges. No additional tumors were visualized in either the dark-field or phase-contrast images. Compared to the adjacent lung tissue, lung tumors were characterized by a significant decrease in transmission signal (median 0.86 vs. 0.91, p = 0.04) and increase in dark-field signal (median 0.71 vs. 0.65, p = 0.04). Median contrast-to-noise ratios for the visualization of lung nodules were 4.4 for transmission images and 1.7 for dark-field images (p = 0.04).ConclusionLung nodules can be visualized on all three radiograph modalities derived from grating-based X-ray imaging. However, our initial data suggest that grating-based multimodal X-ray imaging does not increase the sensitivity of chest radiographs for the detection of lung nodules.     

Cinétiques de la fixation du 18F-FDG dans les structures cérébrales : faisabilité et résultats préliminaires

ObjectDynamic positron emission tomography (dyn-PET) acquisitions using the radiotracer ¹⁸F-deoxyglucose (¹⁸F-FDG) are mainly developed in research studies of brain PET in kinetic modelling to determine the local glucose consumption rate. This procedure is difficult to establish, due to its requirement for blood sampling. Here, we propose a simple approach to constructing time–activity curves (TACs) for four different brain structures (the arterial & venous regions and grey & white matter) based on direct image measurements on chronologically reconstructed image volumes of regions of interest.Materials and methodsWe applied our processing on 14 control subjects to extract their physiological state. We defined the reference ¹⁸F-FDG kinetic curves as a “population averaged TAC” for four structures. To increase the curves accuracy, our method included the evaluation of two normalization based on the integral of the activity curve in the arteries and the veins.ResultsThe method showed discrimination between artery, venous, grey and white matters. The two normalization methods significantly reduce the dispersion for the grey and white matter curves and that venous normalization showed the best overall efficiency.ConclusionWe have designed and evaluated an approach for directly defining PopAv_TACs which are representative of given anatomical structures.     

Strategic Lesions in the Anterior Thalamic Radiation and Apathy in Early Alzheimer's Disease

BackgroundBehavioural disorders and psychological symptoms of Dementia (BPSD) are commonly observed in Alzheimer’s disease (AD), and strongly contribute to increasing patients'' disability. Using voxel-lesion-symptom mapping (VLSM), we investigated the impact of white matter lesions (WMLs) on the severity of BPSD in patients with amnestic mild cognitive impairment (a-MCI).MethodsThirty-one a-MCI patients (with a conversion rate to AD of 32% at 2 year follow-up) and 26 healthy controls underwent magnetic resonance imaging (MRI) examination at 3T, including T2-weighted and fluid-attenuated-inversion-recovery images, and T1-weighted volumes. In the patient group, BPSD was assessed using the Neuropsychiatric Inventory-12. After quantitative definition of WMLs, their distribution was investigated, without an a priori anatomical hypothesis, against patients’ behavioural symptoms. Unbiased regional grey matter volumetrics was also used to assess the contribution of grey matter atrophy to BPSD.ResultsApathy, irritability, depression/dysphoria, anxiety and agitation were shown to be the most common symptoms in the patient sample. Despite a more widespread anatomical distribution, a-MCI patients did not differ from controls in WML volumes. VLSM revealed a strict association between the presence of lesions in the anterior thalamic radiations (ATRs) and the severity of apathy. Regional grey matter atrophy did not account for any BPSD.ConclusionsThis study indicates that damage to the ATRs is strategic for the occurrence of apathy in patients with a-MCI. Disconnection between the prefrontal cortex and the mediodorsal and anterior thalamic nuclei might represent the pathophysiological substrate for apathy, which is one of the most common psychopathological symptoms observed in dementia.     

Neuroanatomical Correlates of Theory of Mind Deficit in Parkinson’s Disease: A Multimodal Imaging Study

BackgroundParkinson’s disease (PD) patients show theory of mind (ToM) deficit since the early stages of the disease, and this deficit has been associated with working memory, executive functions and quality of life impairment. To date, neuroanatomical correlates of ToM have not been assessed with magnetic resonance imaging in PD. The main objective of this study was to assess cerebral correlates of ToM deficit in PD. The second objective was to explore the relationships between ToM, working memory and executive functions, and to analyse the neural correlates of ToM, controlling for both working memory and executive functions.MethodsThirty-seven PD patients (Hoehn and Yahr median = 2.0) and 15 healthy controls underwent a neuropsychological assessment and magnetic resonance images in a 3T-scanner were acquired. T1-weighted images were analysed with voxel-based morphometry, and white matter integrity and diffusivity measures were obtained from diffusion weighted images and analysed using tract-based spatial statistics.ResultsPD patients showed impairments in ToM, working memory and executive functions; grey matter loss and white matter reduction compared to healthy controls. Grey matter volume decrease in the precentral and postcentral gyrus, middle and inferior frontal gyrus correlated with ToM deficit in PD. White matter in the superior longitudinal fasciculus (adjacent to the parietal lobe) and white matter adjacent to the frontal lobe correlated with ToM impairment in PD. After controlling for executive functions, the relationship between ToM deficit and white matter remained significant for white matter areas adjacent to the precuneus and the parietal lobe.ConclusionsFindings reinforce the existence of ToM impairment from the early Hoehn and Yahr stages in PD, and the findings suggest associations with white matter and grey matter volume decrease. This study contributes to better understand ToM deficit and its neural correlates in PD, which is a basic skill for development of healthy social relationships.     

Targeting miR-9 in Glioma Stem Cell-Derived Extracellular Vesicles: A Novel Diagnostic and Therapeutic Biomarker

BackgroundGlioblastoma (GBM) is a lethal brain tumor with no effective strategies in early diagnosis and treatment. This study was aimed to assess the miRNA expression profiles in EVs from CSF and tissue of glioblastoma patients to identify significantly upregulated miRNAs and investigate the underlying neoplastic mechanisms.MethodsEVs were measured by TEM and NTA assays. Differentially regulated miRNAs were measured using RNA sequencing in GBM CSF EVs and in GBM tissues compared with controls. RT-qPCR was employed to analyze miRNA and gene expression. Luciferase report assay was used to investigate gene target of miR-9. The proliferation ability was detected by EdU and CCK-8 experiment while cell migration was measured by transwell and wound healing assay.ResultsThe expression level of miR-9 was significantly higher in GBM CSF EVs and tissues than controls (p = 0.038). The area under curve for CSF EV miR-9 was 0.800 (95% CI: 0.583–1.000, p = 0.033). The expression of miR-9 was significantly higher in Glioma stem cells (GSCs) and GSC-derived EVs than in glioblastoma cells. GSC-derives EVs could promote GBM growth and migration Moreover, inhibition of miR-9 in GSCs showed the reverse anti-tumor effects through secreted EVs. MiR-9 could bind to the 3’UTR region of DACT3 and suppress its expression. The miR-9/DACT3 axis might attribute to GBM malignant phenotype.ConclusionMiR-9 in CSF EVs may act as a novel diagnostic biomarker for GBM and targeting miR-9 by GSC-derived EVs may be a specific and efficient strategy for GBM biotherapy.     

基于X线图像的膝关节周围原发性骨肿瘤辅助诊断的机器学习模型研究

摘要 目的：开发机器学习模型,并评估其在膝关节周围原发性骨肿瘤诊断方面的准确性。方法：本文将深度卷积神经网络（DCNN）这一深度学习方法应用于膝关节X线图像的影像组学分析,探讨其辅助诊断膝关节周围原发性骨肿瘤的临床价值。结果：该深度学习模型在区分正常与肿瘤影像方面展现出优异的诊断准确性,使用DCNN模型进行5轮测试的总体准确性为（99.8±0.4）%,而阳性预测值和阴性预测值分别为（100.0±0.0）%和（99.6±0.8）%,各个数据集的曲线下面积（AUC）分别为0.99、1.00、1.00、1.0和1.0,平均AUC为（0.998±0.004）;进一步使用DCNN模型进行了10轮测试显示其在区分良性与恶性骨肿瘤方面的总体准确性为（71.2±1.6）%,且达到了强阳性预测值（91.9±8.5）%,各个数据集的AUC分别为0.63、0.63、0.58、0.69、0.55、0.63、0.54、0.57、0.73、0.63,平均AUC为（0.62±0.06）。结论：本文是首个将人工智能技术应用于骨肿瘤诊断的X线图像影像组学分析方面的研究,人工智能影像组学模型能够帮助医生自动地快速筛查骨肿瘤,确定良性或恶性肿瘤时,阳性预测值较高。     

Performances de la TEP au 18F-FDG dans la caractérisation initiale de la lésion primitive du cancer du sein

Objective¹⁸F-FDG PET/CT is for the moment not recommended for stage T of the TNM classification of breast cancer. The aim of our study was to evaluate the performance of ¹⁸F-FDG PET/CT in the initial staging of breast tumors. Tumor size, skin involvement and inflammation as well as the relationship between primary tumor maximum standardized uptake value (SUVmax) and histopathological grade (SBR), molecular tumor subtypes (luminal A and B, Her2 enriched, triple negative), estrogen receptors (ER), progesterone receptors (PR) and focality were evaluated.MethodsHistological reports of patients operated for breast cancer, without neoadjuvant chemotherapy, were compared to preoperative ¹⁸F-FDG PET/CT.ResultsSeventy-four patients who underwent surgery in 2016 were included. ¹⁸F-FDG PET/CT was able to visualize primary tumors in 91% and to correctly classify the T stage of the TNM classification in 81% of the cases, to detect multifocality in 73% and cutaneous and inflammatory breast cancers in 100%. The uptake intensity of ¹⁸F-FDG (SUVmax) was significantly correlated with histo-prognostic factors such as SBR grade (P = 0.02), lack of expression of estrogen receptors (ER) (P = 0.01) and progesterone (PR) (P = 0.02), positive HER2 status (P = 0.01) or triple negative subtype tumors (P = 0.02).Conclusion¹⁸F-FDG PET/CT provides relevant elements for local assessment, in particular, tumor focality and inflammatory character in addition to ensuring the regional and extension assessment.     

Gaussian Dropout Based Stacked Ensemble CNN for Classification of Breast Tumor in Ultrasound Images

ObjectiveBreast cancer and breast tumors have been considered to be the most pervasive form of cancer in medical practice. Breast tumors are life-threatening to women, and their early detection could save lives with the proper treatment. Physical methods for detection of Breast Cancer are time-consuming and often prone to a misdiagnosis at classifying tumors. Recent trends in radiological imaging have significantly improved the efficiency and veracity of breast tumor classification. Artificial intelligence techniques could be used as an automated detection and classification system.Materials and methodsIn this research, we propose a novel configuration of a Stacking Ensemble with custom Convolutional Neural Network architectures to classify breast tumors from ultrasound images into ‘Normal’, ‘Benign’, and ‘Malignant’ categories.ResultsAfter thorough experimentation, our ensemble has performed with an accuracy, f1-score, precision, and recall of 92.15%, 92.21%, 92.26%, 92.17% respectively.ConclusionThe presented ensemble leverages three Stacked Feature Extractors coupled with a characteristic meta-learner to provide an overall balanced classification performance, with better accuracy and lower false positives. The architecture works in association with gaussian dropout layers to improve the computation and an alternative pooling scheme to retain essential features.     

A practical methodology to improve the dosimetric accuracy of MR-based radiotherapy simulation for brain tumors

PurposeTo investigate the dosimetric accuracy of synthetic computed tomography (sCT) images generated by a clinically-ready voxel-based MRI simulation package, and to develop a simple and feasible method to improve the accuracy.Methods20 patients with brain tumor were selected to undergo CT and MRI simulation. sCT images were generated by a clinical MRI simulation package. The discrepancy between planning CT and sCT in CT number and body contour were evaluated. To resolve the discrepancies, an sCT specific CT-relative electron density (RED) calibration curve was used, and a layer of pseudo-skin was created on the sCT. The dosimetric impact of these discrepancies, and the improvement brought about by the modifications, were evaluated by a planning study. Volumetric modulated arc therapy (VMAT) treatment plans for each patient were created and optimized on the planning CT, which were then transferred to the original sCT and the modified-sCT for dose re-calculation. Dosimetric comparisons and gamma analysis between the calculated doses in different images were performed.ResultsThe average gamma passing rate with 1%/1 mm criteria was only 70.8% for the comparison of dose distribution between planning CT and original sCT. The mean dose difference between the planning CT and the original sCT were −1.2% for PTV D₉₅ and −1.7% for PTV D_max, while the mean dose difference was within 0.7 Gy for all relevant OARs. After applying the modifications on the sCT, the average gamma passing rate was increased to 92.2%. Mean dose difference in PTV D₉₅ and D_max were reduced to −0.1% and −0.3% respectively. The mean dose difference was within 0.2 Gy for all OAR structures and no statistically significant difference were found.ConclusionsThe modified-sCT demonstrated improved dosimetric agreement with the planning CT. These results indicated the overall dosimetric accuracy and practicality of this improved MR-based treatment planning method.