Machine learning-based systems are gaining interest in the field of medicine, mostly in medical imaging and diagnosis. In this paper, we address the problem of automatic cerebral microbleeds (CMB) detection in magnetic resonance images. It is challenging due to difficulty in distinguishing a true CMB from its mimics, however, if successfully solved, it would streamline the radiologists work. To deal with this complex three-dimensional problem, we propose a machine learning approach based on a 2D Faster RCNN network. We aimed to achieve a reliable system, i.e., with balanced sensitivity and precision. Therefore, we have researched and analysed, among others, impact of the way the training data are provided to the system, their pre-processing, the choice of model and its structure, and also the ways of regularisation. Furthermore, we also carefully analysed the network predictions and proposed an algorithm for its post-processing. The proposed approach enabled for obtaining high precision (89.74%), sensitivity (92.62%), and F1 score (90.84%). The paper presents the main challenges connected with automatic cerebral microbleeds detection, its deep analysis and developed system. The conducted research may significantly contribute to automatic medical diagnosis.
Background: The study was undertaken to compare the diagnostic performance of diffusion kurtosis imaging (DKI) with the standard monoexponential (ME) apparent diffusion coefficient (ADC) model in the detection of significant prostate cancer (PCa), using whole-mount histopathology of radical prostatectomy specimens as a reference standard. Methods: 155 patients with prostate cancer had undergone multiparametric magnetic resonance imaging (mpMRI) at 3T before prostatectomy. Quantitative diffusion parameters—the apparent diffusion coefficient corrected for non-Gaussian behavior (Dapp), kurtosis (K), ADC1200, and ADC2000 were correlated with Gleason score and compared between cancerous and benign tissue and between GS ≤ 3 + 3 and GS ≥ 3 + 4 tumors. Results: The mean values of all diffusion parameters (Dapp, K, ADC1200, ADC2000) were significantly different both between malignant and benign tissue and between GS ≤ 3 + 3 and GS ≥ 3 + 4 tumors. Although the kurtosis model was better fitted to DWI data, the diagnostic performance in receiver operating characteristic (ROC) analysis of DKI and the standard ADC model in the detection of significant PCa was similar in the peripheral zone (PZ) and in peripheral and transitional zones (TZ) together. In conclusion, our study was not able to demonstrate a clear superiority of the kurtosis model over standard ADC in the diagnosis of significant PCa in PZ and in both zones combined.
IntroductionMegaureter, described as ureter dilatation more than 7 mm in diameter, commonly associated with other anomalies, is still a diagnostic and therapeutic challenge. Magnetic resonance urography (MRU) appears as a promising method in urinary tract imaging, providing both anatomical and functional information. There are several postprocessing tools to assess renal function (including differential renal function) and severity of ureteral obstruction based on MRU. Still, the place of this method in the diagnostic algorithm of ureteropelvicalyceal dilatation with megaureter remains underestimated. Analysis of imaging findings in a group of children diagnosed with megaureter was done.Material and MethodsA retrospective analysis of magnetic resonance urography (MRU) was performed in 142 consecutive patients examined from January 2013 to September 2019. Twenty-five patients meeting the criteria of megaureter (dilatation more than 7 mm) in MRU were included in the further analysis. The MRU, ultrasound (US), and scintigraphy results were compared and analyzed together and compared with clinical data.ResultsThe sensitivity and specificity of US was comparable to the MRU in the assessment of upper urinary tract morphology (p > 0.05). In five out of 25 children, megaureter was found in each kidney; in a single case, both poles of a duplex kidney were affected. In the diagnosis of ureter ectopia, the MRU was superior to the US for which sensitivity did not exceed 16%. The US showed limited value in the diagnostics of segmental ureter dysplasia as a cause of primary megaureter when compared with MRU. Four cases were visualized in MRU studies, whereas the US examination was negative (all confirmed during surgery). There was a moderate correlation between relative renal function between fMRU and scintigraphy (t = 0.721, p = 0.477) and in the severity of obstruction assessment between both methods (r = 0.441, p < 0.05). However, in 10 kidneys with megaureter, the results in scintigraphy were inconclusive due to the signal from the megaureter imposing on the renal field.ConclusionsMRU seems to be a preferred method in the diagnostic algorithm for megaureter, providing both anatomical and functional information. MRU is superior to US and scintigraphy in diagnosing urinary tract anomalies with megaureter.
(1) Background: Multiple sclerosis (MS) is an auto-immune, chronic, neuroinflammatory, demyelinating disease that affects mainly young patients. This progressive inflammatory process causes the chronic loss of brain tissue and results in a deterioration in quality of life. To monitor neuroinflammatory process activity and predict the further development of disease, it is necessary to find a suitable biomarker that could easily be used. In this research, we verify the usability of choroid plexus (CP) volume, a new MS biomarker, in the monitoring of the progression of multiple sclerosis disease. (2) Methods: A single-center, prospective study with three groups of patients was conducted based on the following groups: MS patients who received experimental cellular therapy (Treg), treatment-naïve MS patients and healthy controls. (3) Results: This study concludes that there is a correlation between the CPV/TIV (choroid plexus/total intracranial volume) ratio and the progress of multiple sclerosis disease—patients with MS (MS + Treg) had larger volumes of choroid plexuses. CPV/TIV ratios in MS groups were constantly and significantly growing. In the Treg group, patients with relapses had larger plexuses in comparison to the group with no relapses of MS. A similar correlation was observed for the GD+ group (patients with postcontrast enhancing plaques) compared against the non-GD group (patients without postcontrast enhancing plaques). (4) Conclusion: Choroid plexus volume, due to its immunological function, correlates with the inflammatory process in the central nervous system. We consider it to become a valuable radiological biomarker of MS activity.
Introduction: Despite the significant increase in use of magnetic resonance imaging (MRI) in children, there is still a lack of normal reference values of renal size in this method and reference values are being interpolated from the ultrasound (US) studies. The study provides comparative analysis of agreement in renal length and volume measurements between MRI and ultrasound.Materials and Methods: Ninety-three children with a mean age of 8.0 ± 6.0 years, who had undergone both renal US and MRI exams, were included in the study. Participants were divided into three subgroups; each kidney was considered separately.Group 1 included 106 kidneys without any anomalies. Group 2 comprised 48 kidneys with a dilated collecting system. Group 3 included 32 kidneys with a duplicated collecting system. Measurements were taken in three dimensions, and renal volume was calculated from the ellipsoid formula.Results: We found no significant difference between US and MRI measurements in Group 1 and Group 2. In Group 3, the difference between measurements in both imaging methods was significant. The mean difference varied from 0.05% in Group 1, 2.95% in Group 2, to 4.99% in Group 3.Conclusion: The US and MRI are comparable methods in renal size measurements. The interpolation of sonographic renal length and volume reference values to the MRI in the pediatric population is justified, as there is a strong agreement between both methods. Both methods can be used interchangeably for following up of the renal size changes in the pediatric population.
The assessment of body composition in pediatric population is essential for proper nutritional support during hospitalization. However, currently available methods have limitations. This study aims to propose a novel approach for nutrition status assessment and introduce magnetic resonance imaging (MRI)-derived subcutaneous and visceral fat normative reference values. A total of 262 healthy subjects aged from 6 to 18 years underwent MRI examinations and anthropometric measurements. MRI images at the second lumbar vertebrae were used by two radiologists to perform the semi-automatic tissue segmentation. Based on obtained adipose tissue surface areas and body mass index (BMI) scores sex-specific standard percentile curves (3rd, 10th, 25th, 50th, 75th, 90th, 97th) and z-scores were constructed using LMS method. Additionally, 85th and 95th centiles of subcutaneous and visceral adipose tissue were proposed as equivalents of overweight and obesity. Bland-Altman plots revealed an excellent intra-observer reproducibility and inter-observer agreement. In conclusion, our findings demonstrate highly reproducible method and suggest that MRI-derived reference values can be implemented in clinical practice.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.