Current Issue : April-June Volume : 2024 Issue Number : 2 Articles : 5 Articles
Objective There are no specific magnetic resonance imaging (MRI) features that distinguish pilocytic astrocytoma (PA) from adamantinomatous craniopharyngioma (ACP). In this study we compared the frequency of a novel enhancement characteristic on MRI (called the cut green pepper sign) in PA and ACP. Methods Consecutive patients with PA (n = 24) and ACP (n = 36) in the suprasellar region were included in the analysis. The cut green pepper sign was evaluated on post-contrast T1WI images independently by 2 neuroradiologists who were unaware of the pathologic diagnosis. The frequency of cut green pepper sign in PA and ACP was compared with Fisher’s exact test. Results The cut green pepper sign was identified in 50% (12/24) of patients with PA, and 5.6% (2/36) with ACP. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the cut green pepper sign for diagnosing PA were 50%, 94.4%, 85.7% and 73.9%, respectively. There was a statistically significant difference in the age of patients with PA with and without the cut green pepper sign (12.3 ± 9.2 years vs. 5.5 ± 4.4 years, p = 0.035). Conclusion The novel cut green pepper sign can help distinguish suprasellar PA from ACP on MRI....
Background In recent years, there has been a growing trend towards utilizing Artificial Intelligence (AI) and machine learning techniques in medical imaging, including for the purpose of automating quality assurance. In this research, we aimed to develop and evaluate various deep learning-based approaches for automatic quality assurance of Magnetic Resonance (MR) images using the American College of Radiology (ACR) standards. Methods The study involved the development, optimization, and testing of custom convolutional neural network (CNN) models. Additionally, popular pre-trained models such as VGG16, VGG19, ResNet50, InceptionV3, EfficientNetB0, and EfficientNetB5 were trained and tested. The use of pre-trained models, particularly those trained on the ImageNet dataset, for transfer learning was also explored. Two-class classification models were employed for assessing spatial resolution and geometric distortion, while an approach classifying the image into 10 classes representing the number of visible spokes was used for the low contrast. Results Our results showed that deep learning-based methods can be effectively used for MR image quality assurance and can improve the performance of these models. The low contrast test was one of the most challenging tests within the ACR phantom. Conclusions Overall, for geometric distortion and spatial resolution, all of the deep learning models tested produced prediction accuracy of 80% or higher. The study also revealed that training the models from scratch performed slightly better compared to transfer learning. For the low contrast, our investigation emphasized the adaptability and potential of deep learning models. The custom CNN models excelled in predicting the number of visible spokes, achieving commendable accuracy, recall, precision, and F1 scores....
Ultrasound practice is a longstanding tradition for radiology departments, being part of the family of imaging techniques. Ultrasound is widely practiced by non-radiologists but becoming less popular within radiology. The position of ultrasound in radiology is reviewed, and a possible long-term solution to manage radiologist expectations is proposed. An international group of experts in the practice of ultrasound was invited to describe the current organisation of ultrasound within the radiology departments in their own countries and comment on the interaction with nonradiologists and training arrangements. Issues related to regulation, non-medical practitioners, and training principles are detailed. A consensus view was sought from the experts regarding the position of ultrasound within radiology, with the vision of the best scenario for the continuing dominance of radiologists practising ultrasound. Comments were collated from nine different countries. Variable levels of training, practice, and interaction with non-radiologist were reported, with some countries relying on non-physician input to manage the service. All experts recognised there was a diminished desire to practice ultrasound by radiologists. Models varied from practising solely ultrasound and no other imaging techniques to radiology departments being central to the practice of ultrasound by radiologists and non-radiologist, housed within radiology. The consensus view was that the model favoured in select hospitals in Germany would be the most likely setup for ultrasound radiologist to develop and maintain practice. The vision for 20 years hence is for a central ultrasound section within radiology, headed by a trained expert radiologist, with non-radiologist using the facilities. Critical relevance statement The future of ultrasound within the radiology department should encompass all ultrasound users, with radiologists expert in ultrasound, managing the ultrasound section within the radiology department. The current radiology trainees must learn of the importance of ultrasound as a component of the ‘holistic’ imaging of the patient....
Background The purpose of this study is to investigate the use of radiomics and deep features obtained from multiparametric magnetic resonance imaging (mpMRI) for grading prostate cancer. We propose a novel approach called multi-flavored feature extraction or tensor, which combines four mpMRI images using eight different fusion techniques to create 52 images or datasets for each patient. We evaluate the effectiveness of this approach in grading prostate cancer and compare it to traditional methods. Methods We used the PROSTATEx-2 dataset consisting of 111 patients’ images from T2W-transverse, T2W-sagittal, DWI, and ADC images. We used eight fusion techniques to merge T2W, DWI, and ADC images, namely Laplacian Pyramid, Ratio of the low-pass pyramid, Discrete Wavelet Transform, Dual-Tree Complex Wavelet Transform, Curvelet Transform, Wavelet Fusion, Weighted Fusion, and Principal Component Analysis. Prostate cancer images were manually segmented, and radiomics features were extracted using the Pyradiomics library in Python. We also used an Autoencoder for deep feature extraction. We used five different feature sets to train the classifiers: all radiomics features, all deep features, radiomics features linked with PCA, deep features linked with PCA, and a combination of radiomics and deep features. We processed the data, including balancing, standardization, PCA, correlation, and Least Absolute Shrinkage and Selection Operator (LASSO) regression. Finally, we used nine classifiers to classify different Gleason grades. Results Our results show that the SVM classifier with deep features linked with PCA achieved the most promising results, with an AUC of 0.94 and a balanced accuracy of 0.79. Logistic regression performed best when using only the deep features, with an AUC of 0.93 and balanced accuracy of 0.76. Gaussian Naive Bayes had lower performance compared to other classifiers, while KNN achieved high performance using deep features linked with PCA. Random Forest performed well with the combination of deep features and radiomics features, achieving an AUC of 0.94 and balanced accuracy of 0.76. The Voting classifiers showed higher performance when using only the deep features, with Voting 2 achieving the highest performance, with an AUC of 0.95 and balanced accuracy of 0.78. Conclusion Our study concludes that the proposed multi-flavored feature extraction or tensor approach using radiomics and deep features can be an effective method for grading prostate cancer. Our findings suggest that deep features may be more effective than radiomics features alone in accurately classifying prostate cancer....
Background Ovarian cancer is a common cancer among women globally, and the assessment of lymph node metastasis plays a crucial role in the treatment of this malignancy. The primary objective of our study was to identify the risk factors associated with lymph node metastasis in patients with ovarian cancer and develop a predictive model to aid in the selection of the appropriate surgical procedure and treatment strategy. Methods We conducted a retrospective analysis of data from patients with ovarian cancer across three different medical centers between April 2014 and August 2022. Logistic regression analysis was employed to establish a prediction model for lymph node metastasis in patients with ovarian cancer. We evaluated the performance of the model using receiver operating characteristic (ROC) curves, calibration plots, and decision analysis curves. Results Our analysis revealed that among the 368 patients in the training set, 101 patients (27.4%) had undergone lymph node metastasis. Maximum tumor diameter, multifocal tumor, and Ki67 level were identified as independent risk factors for lymph node metastasis. The area under the curve (AUC) of the ROC curve in the training set was 0.837 (95% confidence interval [CI]: 0.792–0.881); in the validation set this value was 0.814 (95% CI: 0.744–0.884). Calibration plots and decision analysis curves revealed good calibration and clinical application value. Conclusions We successfully developed a model for predicting lymph node metastasis in patients with ovarian cancer, based on ultrasound examination results and clinical data. Our model accurately identified patients at high risk of lymph node metastasis and may guide the selection of appropriate treatment strategies. This model has the potential to significantly enhance the precision and efficacy of clinical management in patients with ovarian cancer....
Loading....